Prepare PyPI release

My reading of regularizers is that they cannot be reused, but it doesn't actually fail in any way and seems like it results in only regularizing the last layer. Having an exception prevent this would probably improve the ergonomics.

.gitignore

@@ -9,3 +9,11 @@ keras/datasets/temp/*
 docs/site/*
 docs/theme/*
 tags
+Keras.egg-info
+
+# test-related
+.coverage
+.cache
+
+# developer environments
+.idea

.travis.yml

@@ -1,9 +1,20 @@
 sudo: required
 dist: trusty
 language: python
-python:
-  - "2.7"
-  - "3.4"
+matrix:
+    include:
+        - python: 3.4
+          env: KERAS_BACKEND=theano
+        - python: 3.4
+          env: KERAS_BACKEND=tensorflow
+        - python: 2.7
+          env: KERAS_BACKEND=theano
+        - python: 2.7
+          env: KERAS_BACKEND=tensorflow
+        - python: 2.7
+          env: KERAS_BACKEND=theano TEST_MODE=INTEGRATION_TESTS
+        - python: 2.7
+          env: KERAS_BACKEND=theano TEST_MODE=PEP8
 install:
   # code below is taken from http://conda.pydata.org/docs/travis.html
   # We do this conditionally because it saves us some downloading if the
@@ -23,20 +34,40 @@ install:
 
   - conda create -q -n test-environment python=$TRAVIS_PYTHON_VERSION numpy scipy matplotlib pandas pytest h5py
   - source activate test-environment
-  - pip install pytest-cov python-coveralls
+  - pip install pytest-cov python-coveralls pytest-xdist coverage==3.7.1 #we need this version of coverage for coveralls.io to work
+  - pip install pep8 pytest-pep8
   - pip install git+git://github.com/Theano/Theano.git
+
+  # install PIL for preprocessing tests
+  - if [[ "$TRAVIS_PYTHON_VERSION" == "2.7" ]]; then
+      conda install pil;
+    elif [[ "$TRAVIS_PYTHON_VERSION" == "3.4" ]]; then
+      conda install Pillow;
+    fi
+
   - python setup.py install
 
   # install TensorFlow
   - if [[ "$TRAVIS_PYTHON_VERSION" == "2.7" ]]; then
-      pip install https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.5.0-cp27-none-linux_x86_64.whl;
+      pip install https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.9.0-cp27-none-linux_x86_64.whl;
+    elif [[ "$TRAVIS_PYTHON_VERSION" == "3.4" ]]; then
+      pip install https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.9.0-cp34-cp34m-linux_x86_64.whl;
     fi
 # command to run tests
 script:
-  - PYTHONPATH=$PWD:$PYTHONPATH py.test -v --cov-report term-missing --cov keras tests/
-  - if [[ "$TRAVIS_PYTHON_VERSION" == "2.7" ]]; then
-      sed -i -e 's/theano/tensorflow/g' ~/.keras/keras.json;
-      PYTHONPATH=$PWD:$PYTHONPATH py.test -v --cov-report term-missing --cov keras tests/;
+  # run keras backend init to initialize backend config
+  - python -c "import keras.backend"
+  # create dataset directory to avoid concurrent directory creation at runtime
+  - mkdir ~/.keras/datasets
+  # set up keras backend
+  - sed -i -e 's/"backend":[[:space:]]*"[^"]*/"backend":\ "'$KERAS_BACKEND'/g' ~/.keras/keras.json;
+  - echo -e "Running tests with the following config:\n$(cat ~/.keras/keras.json)"
+  - if [[ "$TEST_MODE" == "INTEGRATION_TESTS" ]]; then
+       PYTHONPATH=$PWD:$PYTHONPATH py.test tests/integration_tests;
+    elif [[ "$TEST_MODE" == "PEP8" ]]; then
+       PYTHONPATH=$PWD:$PYTHONPATH py.test --pep8 -m pep8 -n0;
+    else
+       PYTHONPATH=$PWD:$PYTHONPATH py.test tests/ --ignore=tests/integration_tests;
     fi
 after_success:
   - coveralls

CONTRIBUTING.md

@@ -0,0 +1,65 @@
+# On Github Issues and Pull Requests
+
+Found a bug? Have a new feature to suggest? Want to contribute changes to the codebase? Make sure to read this first.
+
+## Bug reporting
+
+Your code doesn't work, and you have determined that the issue lies with Keras? Follow these steps to report a bug.
+
+1. Your bug may already be fixed. Make sure to update to the current Keras master branch, as well as the latest Theano/TensorFlow master branch.
+To easily update Theano: `pip install git+git://github.com/Theano/Theano.git --upgrade`
+
+2. Search for similar issues. Make sure to delete `is:open` on the issue search to find solved tickets as well. It's possible somebody has encountered this bug already. Also remember to check out Keras' [FAQ](http://keras.io/faq/). Still having a problem? Open an issue on Github to let us know.
+
+3. Make sure you provide us with useful information about your configuration: what OS are you using? What Keras backend are you using? Are you running on GPU? If so, what is your version of Cuda, of cuDNN? What is your GPU?
+
+4. Provide us with a script to reproduce the issue. This script should be runnable as-is and should not require external data download (use randomly generated data if you need to run a model on some test data). We recommend that you use Github Gists to post your code. Any issue that cannot be reproduced is likely to be closed.
+
+5. If possible, take a stab at fixing the bug yourself --if you can!
+
+The more information you provide, the easier it is for us to validate that there is a bug and the faster we'll be able to take action. If you want your issue to be resolved quickly, following the steps above is crucial.
+
+
+## Requesting a Feature
+
+You can also use Github issues to request features you would like to see in Keras, or changes in the Keras API.
+
+1. Provide a clear and detailed explanation of the feature you want and why it's important to add. Keep in mind that we want features that will be useful to the majority of our users and not just a small subset. If you're just targeting a minority of users, consider writing an add-on library for Keras. It is crucial for Keras to avoid bloating the API and codebase.
+
+2. Provide code snippets demonstrating the API you have in mind and illustrating the use cases of your feature. Of course, you don't need to write any real code at this point!
+
+3. After discussing the feature you may choose to attempt a Pull Request. If you're at all able, start writing some code. We always have more work to do than time to do it. If you can write some code then that will speed the process along.
+
+## Pull Requests
+
+We love pull requests. Here's a quick guide:
+
+1. If your PR introduces a change in functionality, make sure you start by opening an issue to discuss whether the change should be made, and how to handle it. This will save you from having your PR closed down the road! Of course, if your PR is a simple bug fix, you don't need to do that.
+
+2. Write the code. This is the hard part!
+
+3. Make sure any new function or class you introduce has proper docstrings. Make sure any code you touch still has up-to-date docstrings and documentation.
+
+4. Write tests. Your code should have full unit test coverage. If you want to see your PR merged promptly, this is crucial.
+
+5. Run our test suite locally. It's easy: from the Keras folder, simply run: `py.test tests/`.
+  - You will need to install `pytest`, `coveralls`, `pytest-cov`, `pytest-xdist`: `pip install pytest pytest-cov python-coveralls pytest-xdist pep8 pytest-pep8`
+
+6. Make sure all tests are passing:
+  - with the Theano backend, on Python 2.7 and Python 3.5
+  - with the TensorFlow backend, on Python 2.7
+
+7. We use PEP8 syntax conventions, but we aren't dogmatic when it comes to line length. Make sure your lines stay reasonably sized, though. To make your life easier, we recommend running a PEP8 linter:
+  - Install PEP8 packages: `pip install pep8 pytest-pep8 autopep8`
+  - Run a standalone PEP8 check: `py.test --pep8 -m pep8`
+  - You can automatically fix some PEP8 error by running: `autopep8 -i --select <errors> <FILENAME>` for example: `autopep8 -i --select E128 tests/keras/backend/test_backends.py`
+
+8. When committing, use appropriate, descriptive commit messages. Make sure that your branch history is not a string of "bug fix", "fix", "oops", etc. When submitting your PR, squash your commits into a single commit with an appropriate commit message, to make sure the project history stays clean and readable. See ['rebase and squash'](http://rebaseandsqua.sh/) for technical help on how to squash your commits.
+
+9. Update the documentation. If introducing new functionality, make sure you include code snippets demonstrating the usage of your new feature.
+
+10. Submit your PR. If your changes have been approved in a previous discussion, and if you have complete (and passing) unit tests, your PR is likely to be merged promptly. Otherwise, well...
+
+## Adding new examples
+
+Even if you don't contribute to the Keras source code, if you have an application of Keras that is concise and powerful, please consider adding it to our collection of examples. [Existing examples](https://github.com/fchollet/keras/tree/master/examples) show idiomatic Keras code: make sure to keep your own script in the same spirit.

ISSUE_TEMPLATE.md

@@ -0,0 +1,9 @@
+Please make sure that the boxes below are checked before you submit your issue. Thank you!
+
+- [ ] Check that you are up-to-date with the master branch of Keras. You can update with:
+pip install git+git://github.com/fchollet/keras.git --upgrade --no-deps
+
+- [ ] If running on Theano, check that you are up-to-date with the master branch of Theano. You can update with:
+pip install git+git://github.com/Theano/Theano.git --upgrade --no-deps
+
+- [ ] Provide a link to a GitHub Gist of a Python script that can reproduce your issue (or just copy the script here if it is short).

README.md

@@ -1,20 +1,25 @@
-# Keras: Deep Learning library for Theano and TensorFlow
+# Keras: Deep Learning library for TensorFlow and Theano
+
+[![Build Status](https://travis-ci.org/fchollet/keras.svg?branch=master)](https://travis-ci.org/fchollet/keras)
+[![PyPI version](https://badge.fury.io/py/keras.svg)](https://badge.fury.io/py/keras)
+[![license](https://img.shields.io/github/license/mashape/apistatus.svg?maxAge=2592000)](https://github.com/fchollet/keras/blob/master/LICENSE)
+[![Join the chat at https://gitter.im/Keras-io/Lobby](https://badges.gitter.im/Keras-io/Lobby.svg)](https://gitter.im/Keras-io/Lobby)
+
 
 ## You have just found Keras.
 
-Keras is a minimalist, highly modular neural networks library, written in Python and capable of running either on top of either [TensorFlow](https://github.com/tensorflow/tensorflow) or [Theano](https://github.com/Theano/Theano). It was developed with a focus on enabling fast experimentation. Being able to go from idea to result with the least possible delay is key to doing good research.
+Keras is a minimalist, highly modular neural networks library, written in Python and capable of running on top of either [TensorFlow](https://github.com/tensorflow/tensorflow) or [Theano](https://github.com/Theano/Theano). It was developed with a focus on enabling fast experimentation. *Being able to go from idea to result with the least possible delay is key to doing good research.*
 
 Use Keras if you need a deep learning library that:
-- allows for easy and fast prototyping (through total modularity, minimalism, and extensibility).
-- supports both convolutional networks and recurrent networks, as well as combinations of the two.
-- supports arbitrary connectivity schemes (including multi-input and multi-output training).
-- runs seamlessly on CPU and GPU.
+
+- Allows for easy and fast prototyping (through total modularity, minimalism, and extensibility).
+- Supports both convolutional networks and recurrent networks, as well as combinations of the two.
+- Supports arbitrary connectivity schemes (including multi-input and multi-output training).
+- Runs seamlessly on CPU and GPU.
 
 Read the documentation at [Keras.io](http://keras.io).
 
-Keras is compatible with:
-    - __Python 2.7-3.5__ with the Theano backend
-    - __Python 2.7__ with the TensorFlow backend
+Keras is compatible with: __Python 2.7-3.5__.
 
 
 ------------------
@@ -36,9 +41,9 @@ Keras is compatible with:
 
 ## Getting started: 30 seconds to Keras
 
-The core datastructure of Keras is a __model__, a way to organize layers. There are two types of models: [`Sequential`](/models/#sequential) and [`Graph`](/models/#graph).
+The core data structure of Keras is a __model__, a way to organize layers. The main type of model is the [`Sequential`](http://keras.io/getting-started/sequential-model-guide) model, a linear stack of layers. For more complex architectures, you should use the [Keras functional API](http://keras.io/getting-started/functional-api-guide).
 
-Here's the `Sequential` model (a linear pile of layers):
+Here's the `Sequential` model:
 
 ```python
 from keras.models import Sequential
@@ -49,17 +54,17 @@ model = Sequential()
 Stacking layers is as easy as `.add()`:
 
 ```python
-from keras.layers.core import Dense, Activation
+from keras.layers import Dense, Activation
 
-model.add(Dense(output_dim=64, input_dim=100, init="glorot_uniform"))
+model.add(Dense(output_dim=64, input_dim=100))
 model.add(Activation("relu"))
-model.add(Dense(output_dim=10, init="glorot_uniform"))
+model.add(Dense(output_dim=10))
 model.add(Activation("softmax"))
 ```
 
 Once your model looks good, configure its learning process with `.compile()`:
 ```python
-model.compile(loss='categorical_crossentropy', optimizer='sgd')
+model.compile(loss='categorical_crossentropy', optimizer='sgd', metrics=['accuracy'])
 ```
 
 If you need to, you can further configure your optimizer. A core principle of Keras is to make things reasonably simple, while allowing the user to be fully in control when they need to (the ultimate control being the easy extensibility of the source code).
@@ -80,7 +85,7 @@ model.train_on_batch(X_batch, Y_batch)
 
 Evaluate your performance in one line:
 ```python
-objective_score = model.evaluate(X_test, Y_test, batch_size=32)
+loss_and_metrics = model.evaluate(X_test, Y_test, batch_size=32)
 ```
 
 Or generate predictions on new data:
@@ -89,11 +94,14 @@ classes = model.predict_classes(X_test, batch_size=32)
 proba = model.predict_proba(X_test, batch_size=32)
 ```
 
-Building a network of LSTMs, a deep CNN, a Neural Turing Machine, a word2vec embedder or any other model is just as fast. The ideas behind deep learning are simple, so why should their implementation be painful?
+Building a question answering system, an image classification model, a Neural Turing Machine, a word2vec embedder or any other model is just as fast. The ideas behind deep learning are simple, so why should their implementation be painful?
 
-Have a look at these [starter examples](http://keras.io/examples/).
+For a more in-depth tutorial about Keras, you can check out:
 
-In the [examples folder](https://github.com/fchollet/keras/tree/master/examples) of the repo, you will find more advanced models: question-answering with memory networks, text generation with stacked LSTMs, neural turing machines, etc.
+- [Getting started with the Sequential model](http://keras.io/getting-started/sequential-model-guide)
+- [Getting started with the functional API](http://keras.io/getting-started/functional-api-guide)
+
+In the [examples folder](https://github.com/fchollet/keras/tree/master/examples) of the repository, you will find more advanced models: question-answering with memory networks, text generation with stacked LSTMs, etc.
 
 
 ------------------
@@ -108,35 +116,33 @@ Keras uses the following dependencies:
 - HDF5 and h5py (optional, required if you use model saving/loading functions)
 - Optional but recommended if you use CNNs: cuDNN.
 
-When using the Theano backend:
-- Theano
-    - [See installation instructions](http://deeplearning.net/software/theano/install.html#install).
 
-**Note**: You should use the latest version of Theano, not the PyPI version. Install it with:
-```
-sudo pip install git+git://github.com/Theano/Theano.git
-```
+*When using the TensorFlow backend:*
 
-When using the TensorFlow backend:
 - TensorFlow
     - [See installation instructions](https://github.com/tensorflow/tensorflow#download-and-setup).
 
-To install, `cd` to the Keras folder and run the install command:
-```
+*When using the Theano backend:*
+
+- Theano
+    - [See installation instructions](http://deeplearning.net/software/theano/install.html#install).
+
+To install Keras, `cd` to the Keras folder and run the install command:
+```sh
 sudo python setup.py install
 ```
 
 You can also install Keras from PyPI:
-```
+```sh
 sudo pip install keras
 ```
 
 ------------------
 
 
-## Switching from Theano to TensorFlow
+## Switching from TensorFlow to Theano
 
-By default, Keras will use Theano as its tensor manipulation library. [Follow these instructions](http://keras.io/backend/) to configure the Keras backend.
+By default, Keras will use TensorFlow as its tensor manipulation library. [Follow these instructions](http://keras.io/backend/) to configure the Keras backend.
 
 ------------------
 
@@ -145,20 +151,7 @@ By default, Keras will use Theano as its tensor manipulation library. [Follow th
 
 You can ask questions and join the development discussion on the [Keras Google group](https://groups.google.com/forum/#!forum/keras-users).
 
-------------------
-
-
-## Contribution Guidelines
-
-Keras welcomes all contributions from the community. 
-
-- Keep a pragmatic mindset and avoid bloat. Only add to the source if that is the only path forward.
-- New features should be documented. Make sure you update the documentation along with your Pull Request.
-- Any new function or class should have a proper docstring.
-- The documentation for every new feature should include a usage example in the form of a code snippet. 
-- All changes should be tested. Make sure any new feature you add has a corresponding unit test.
-- Please no Pull Requests about coding style.
-- Even if you don't contribute to the Keras source code, if you have an application of Keras that is concise and powerful, please consider adding it to our collection of [examples](https://github.com/fchollet/keras/tree/master/examples).
+You can also post bug reports and feature requests in [Github issues](https://github.com/fchollet/keras/issues). Make sure to read [our guidelines](https://github.com/fchollet/keras/blob/master/CONTRIBUTING.md) first.
 
 
 ------------------
@@ -172,4 +165,4 @@ Keras was initially developed as part of the research effort of project ONEIROS
 
 >_"Oneiroi are beyond our unravelling --who can be sure what tale they tell? Not all that men look for comes to pass. Two gates there are that give passage to fleeting Oneiroi; one is made of horn, one of ivory. The Oneiroi that pass through sawn ivory are deceitful, bearing a message that will not be fulfilled; those that come out through polished horn have truth behind them, to be accomplished for men who see them."_ Homer, Odyssey 19. 562 ff (Shewring translation).
 
-------------------
+------------------

docker/Dockerfile

@@ -0,0 +1,46 @@
+FROM nvidia/cuda:7.5-cudnn5-devel
+
+ENV CONDA_DIR /opt/conda
+ENV PATH $CONDA_DIR/bin:$PATH
+
+RUN mkdir -p $CONDA_DIR && \
+    echo export PATH=$CONDA_DIR/bin:'$PATH' > /etc/profile.d/conda.sh && \
+    apt-get update && \
+    apt-get install -y wget git libhdf5-dev g++ graphviz && \
+    wget --quiet https://repo.continuum.io/miniconda/Miniconda3-3.9.1-Linux-x86_64.sh && \
+    echo "6c6b44acdd0bc4229377ee10d52c8ac6160c336d9cdd669db7371aa9344e1ac3 *Miniconda3-3.9.1-Linux-x86_64.sh" | sha256sum -c - && \
+    /bin/bash /Miniconda3-3.9.1-Linux-x86_64.sh -f -b -p $CONDA_DIR && \
+    rm Miniconda3-3.9.1-Linux-x86_64.sh
+
+ENV NB_USER keras
+ENV NB_UID 1000
+
+RUN useradd -m -s /bin/bash -N -u $NB_UID $NB_USER && \
+    mkdir -p $CONDA_DIR && \
+    chown keras $CONDA_DIR -R && \
+    mkdir -p /src && \
+    chown keras /src
+
+USER keras
+
+# Python
+ARG python_version=3.5.1
+ARG tensorflow_version=0.9.0rc0-cp35-cp35m
+RUN conda install -y python=${python_version} && \
+    pip install https://storage.googleapis.com/tensorflow/linux/gpu/tensorflow-${tensorflow_version}-linux_x86_64.whl && \
+    pip install git+git://github.com/Theano/Theano.git && \
+    pip install ipdb pytest pytest-cov python-coveralls coverage==3.7.1 pytest-xdist pep8 pytest-pep8 pydot_ng && \
+    conda install Pillow scikit-learn notebook pandas matplotlib nose pyyaml six h5py && \
+    pip install git+git://github.com/fchollet/keras.git && \
+    conda clean -yt
+
+ADD theanorc /home/keras/.theanorc
+
+ENV PYTHONPATH='/src/:$PYTHONPATH'
+
+WORKDIR /src
+
+EXPOSE 8888
+
+CMD jupyter notebook --port=8888 --ip=0.0.0.0
+

docker/Makefile

@@ -0,0 +1,26 @@
+help:
+	@cat Makefile
+
+DATA?="${HOME}/Data"
+GPU?=0
+DOCKER_FILE=Dockerfile
+DOCKER=GPU=$(GPU) nvidia-docker
+BACKEND=tensorflow
+TEST=tests/
+SRC=$(shell dirname `pwd`)
+
+build:
+	docker build -t keras --build-arg python_version=3.5 -f $(DOCKER_FILE) .
+
+bash: build
+	$(DOCKER) run -it -v $(SRC):/src -v $(DATA):/data --env KERAS_BACKEND=$(BACKEND) keras bash
+
+ipython: build
+	$(DOCKER) run -it -v $(SRC):/src -v $(DATA):/data --env KERAS_BACKEND=$(BACKEND) keras ipython
+
+notebook: build
+	$(DOCKER) run -it -v $(SRC):/src -v $(DATA):/data --net=host --env KERAS_BACKEND=$(BACKEND) keras
+
+test: build
+	$(DOCKER) run -it -v $(SRC):/src -v $(DATA):/data --env KERAS_BACKEND=$(BACKEND) keras py.test $(TEST)
+

docker/README.md

@@ -0,0 +1,58 @@
+# Using Keras via Docker
+
+This directory contains `Dockerfile` to make it easy to get up and running with
+Keras via [Docker](http://www.docker.com/).
+
+## Installing Docker
+
+General installation instructions are
+[on the Docker site](https://docs.docker.com/installation/), but we give some
+quick links here:
+
+* [OSX](https://docs.docker.com/installation/mac/): [docker toolbox](https://www.docker.com/toolbox)
+* [ubuntu](https://docs.docker.com/installation/ubuntulinux/)
+
+## Running the container
+
+We are using `Makefile` to simplify docker commands within make commands.
+
+Build the container and start a jupyter notebook
+
+    $ make notebook
+
+Build the container and start an iPython shell
+
+    $ make ipython
+
+Build the container and start a bash
+
+    $ make bash
+
+For GPU support install NVidia drivers (ideally latest) and
+[nvidia-docker](https://github.com/NVIDIA/nvidia-docker). Run using
+
+    $ make notebook GPU=0 # or [ipython, bash]
+
+Switch between Theano and TensorFlow
+
+    $ make notebook BACKEND=theano
+    $ make notebook BACKEND=tensorflow
+
+Mount a volume for external data sets
+
+    $ make DATA=~/mydata
+
+Prints all make tasks
+
+    $ make help
+
+You can change Theano parameters by editing `/docker/theanorc`.
+
+
+Note: If you would have a problem running nvidia-docker you may try the old way
+we have used. But it is not recommended. If you find a bug in the nvidia-docker report
+it there please and try using the nvidia-docker as described above.
+
+    $ export CUDA_SO=$(\ls /usr/lib/x86_64-linux-gnu/libcuda.* | xargs -I{} echo '-v {}:{}')
+    $ export DEVICES=$(\ls /dev/nvidia* | xargs -I{} echo '--device {}:{}')
+    $ docker run -it -p 8888:8888 $CUDA_SO $DEVICES gcr.io/tensorflow/tensorflow:latest-gpu

docker/theanorc

@@ -0,0 +1,5 @@
+[global]
+floatX = float32
+optimizer=None
+device = gpu
+

docs/README.md

@@ -5,5 +5,8 @@ Our documentation uses extended Markdown, as implemented by [MkDocs](http://mkdo
 
 ## Building the documentation
 
-- install MkDocs: `sudo pip install mkdocs`
-- `cd` to the `docs/` folder and run: `mkdocs serve`
+- install MkDocs: `pip install mkdocs`
+- `cd` to the `docs/` folder and run:
+    - `python autogen.py`
+    - `mkdocs serve`    # Starts a local webserver:  [localhost:8000](localhost:8000)
+    - `mkdocs build`    # Builds a static site in "site" directory

docs/autogen.py

@@ -0,0 +1,448 @@
+# -*- coding: utf-8 -*-
+'''
+General documentation architecture:
+
+Home
+Index
+
+- Getting started
+    Getting started with the sequential model
+    Getting started with the functional api
+    Examples
+    FAQ
+    Installation guide
+
+- Models
+    About Keras models
+        explain when one should use Sequential or functional API
+        explain compilation step
+        explain weight saving, weight loading
+        explain serialization, deserialization
+    Sequential
+    Model (functional API)
+
+- Layers
+    About Keras layers
+        explain common layer functions: get_weights, set_weights, get_config
+        explain input_shape
+        explain usage on non-Keras tensors
+    Core layers
+    Convolutional
+    Recurrent
+    Embeddings
+    Normalization
+    Advanced activations
+    Noise
+
+- Preprocessing
+    Image preprocessing
+    Text preprocessing
+    Sequence preprocessing
+
+Objectives
+Optimizers
+Activations
+Callbacks
+Datasets
+Backend
+Initializations
+Regularizers
+Constraints
+Visualization
+Scikit-learn API
+
+'''
+from __future__ import print_function
+from __future__ import unicode_literals
+
+import re
+import inspect
+import os
+import shutil
+import sys
+if sys.version[0] == '2':
+    reload(sys)
+    sys.setdefaultencoding('utf8')
+
+from keras.layers import convolutional
+from keras.layers import pooling
+from keras.layers import local
+from keras.layers import recurrent
+from keras.layers import core
+from keras.layers import noise
+from keras.layers import normalization
+from keras.layers import advanced_activations
+from keras.layers import embeddings
+from keras.layers import wrappers
+from keras import optimizers
+from keras import callbacks
+from keras import models
+from keras.engine import topology
+from keras import objectives
+from keras import backend
+from keras import constraints
+from keras import activations
+from keras import regularizers
+
+
+EXCLUDE = {
+    'Optimizer',
+    'Wrapper',
+    'get_session',
+    'set_session',
+    'CallbackList',
+}
+
+PAGES = [
+    {
+        'page': 'models/sequential.md',
+        'functions': [
+            models.Sequential.compile,
+            models.Sequential.fit,
+            models.Sequential.evaluate,
+            models.Sequential.predict,
+            models.Sequential.predict_classes,
+            models.Sequential.predict_proba,
+            models.Sequential.train_on_batch,
+            models.Sequential.test_on_batch,
+            models.Sequential.predict_on_batch,
+            models.Sequential.fit_generator,
+            models.Sequential.evaluate_generator,
+            models.Sequential.predict_generator,
+        ],
+    },
+    {
+        'page': 'models/model.md',
+        'functions': [
+            models.Model.compile,
+            models.Model.fit,
+            models.Model.evaluate,
+            models.Model.predict,
+            models.Model.train_on_batch,
+            models.Model.test_on_batch,
+            models.Model.predict_on_batch,
+            models.Model.fit_generator,
+            models.Model.evaluate_generator,
+            models.Model.predict_generator,
+            models.Model.get_layer,
+        ]
+    },
+    {
+        'page': 'layers/core.md',
+        'classes': [
+            core.Dense,
+            core.Activation,
+            core.Dropout,
+            core.SpatialDropout2D,
+            core.SpatialDropout3D,
+            core.Flatten,
+            core.Reshape,
+            core.Permute,
+            core.RepeatVector,
+            topology.Merge,
+            core.Lambda,
+            core.ActivityRegularization,
+            core.Masking,
+            core.Highway,
+            core.MaxoutDense,
+            core.TimeDistributedDense,
+        ],
+    },
+    {
+        'page': 'layers/convolutional.md',
+        'classes': [
+            convolutional.Convolution1D,
+            convolutional.AtrousConvolution1D,
+            convolutional.Convolution2D,
+            convolutional.AtrousConvolution2D,
+            convolutional.SeparableConvolution2D,
+            convolutional.Deconvolution2D,
+            convolutional.Convolution3D,
+            convolutional.UpSampling1D,
+            convolutional.UpSampling2D,
+            convolutional.UpSampling3D,
+            convolutional.ZeroPadding1D,
+            convolutional.ZeroPadding2D,
+            convolutional.ZeroPadding3D,
+        ],
+    },
+    {
+        'page': 'layers/pooling.md',
+        'classes': [
+            pooling.MaxPooling1D,
+            pooling.MaxPooling2D,
+            pooling.MaxPooling3D,
+            pooling.AveragePooling1D,
+            pooling.AveragePooling2D,
+            pooling.AveragePooling3D,
+            pooling.GlobalMaxPooling1D,
+            pooling.GlobalAveragePooling1D,
+            pooling.GlobalMaxPooling2D,
+            pooling.GlobalAveragePooling2D,
+        ],
+    },
+    {
+        'page': 'layers/local.md',
+        'classes': [
+            local.LocallyConnected1D,
+            local.LocallyConnected2D,
+        ],
+    },
+    {
+        'page': 'layers/recurrent.md',
+        'classes': [
+            recurrent.Recurrent,
+            recurrent.SimpleRNN,
+            recurrent.GRU,
+            recurrent.LSTM,
+        ],
+    },
+    {
+        'page': 'layers/embeddings.md',
+        'classes': [
+            embeddings.Embedding,
+        ],
+    },
+    {
+        'page': 'layers/normalization.md',
+        'classes': [
+            normalization.BatchNormalization,
+        ],
+    },
+    {
+        'page': 'layers/advanced-activations.md',
+        'all_module_classes': [advanced_activations],
+    },
+    {
+        'page': 'layers/noise.md',
+        'all_module_classes': [noise],
+    },
+    {
+        'page': 'layers/wrappers.md',
+        'all_module_classes': [wrappers],
+    },
+
+    {
+        'page': 'optimizers.md',
+        'all_module_classes': [optimizers],
+    },
+    {
+        'page': 'callbacks.md',
+        'all_module_classes': [callbacks],
+    },
+    {
+        'page': 'backend.md',
+        'all_module_functions': [backend],
+    },
+]
+
+ROOT = 'http://keras.io/'
+
+
+def get_earliest_class_that_defined_member(member, cls):
+    ancestors = get_classes_ancestors([cls])
+    result = None
+    for ancestor in ancestors:
+        if member in dir(ancestor):
+            result = ancestor
+    if not result:
+        return cls
+    return result
+
+
+def get_classes_ancestors(classes):
+    ancestors = []
+    for cls in classes:
+        ancestors += cls.__bases__
+    filtered_ancestors = []
+    for ancestor in ancestors:
+        if ancestor.__name__ in ['object']:
+            continue
+        filtered_ancestors.append(ancestor)
+    if filtered_ancestors:
+        return filtered_ancestors + get_classes_ancestors(filtered_ancestors)
+    else:
+        return filtered_ancestors
+
+
+def get_function_signature(function, method=True):
+    signature = inspect.getargspec(function)
+    defaults = signature.defaults
+    if method:
+        args = signature.args[1:]
+    else:
+        args = signature.args
+    if defaults:
+        kwargs = zip(args[-len(defaults):], defaults)
+        args = args[:-len(defaults)]
+    else:
+        kwargs = []
+    st = '%s.%s(' % (function.__module__, function.__name__)
+    for a in args:
+        st += str(a) + ', '
+    for a, v in kwargs:
+        if type(v) == str:
+            v = '\'' + v + '\''
+        st += str(a) + '=' + str(v) + ', '
+    if kwargs or args:
+        return st[:-2] + ')'
+    else:
+        return st + ')'
+
+
+def get_class_signature(cls):
+    try:
+        class_signature = get_function_signature(cls.__init__)
+        class_signature = class_signature.replace('__init__', cls.__name__)
+    except:
+        # in case the class inherits from object and does not
+        # define __init__
+        class_signature = cls.__module__ + '.' + cls.__name__ + '()'
+    return class_signature
+
+
+def class_to_docs_link(cls):
+    module_name = cls.__module__
+    assert module_name[:6] == 'keras.'
+    module_name = module_name[6:]
+    link = ROOT + module_name.replace('.', '/') + '#' + cls.__name__.lower()
+    return link
+
+
+def class_to_source_link(cls):
+    module_name = cls.__module__
+    assert module_name[:6] == 'keras.'
+    path = module_name.replace('.', '/')
+    path += '.py'
+    line = inspect.getsourcelines(cls)[-1]
+    link = 'https://github.com/fchollet/keras/blob/master/' + path + '#L' + str(line)
+    return '[[source]](' + link + ')'
+
+
+def code_snippet(snippet):
+    result = '```python\n'
+    result += snippet + '\n'
+    result += '```\n'
+    return result
+
+
+def process_class_docstring(docstring):
+    docstring = re.sub(r'\n    # (.*)\n',
+                       r'\n    __\1__\n\n',
+                       docstring)
+
+    docstring = re.sub(r'    ([^\s\\]+):(.*)\n',
+                       r'    - __\1__:\2\n',
+                       docstring)
+
+    docstring = docstring.replace('    ' * 5, '\t\t')
+    docstring = docstring.replace('    ' * 3, '\t')
+    docstring = docstring.replace('    ', '')
+    return docstring
+
+
+def process_function_docstring(docstring):
+    docstring = re.sub(r'\n    # (.*)\n',
+                       r'\n    __\1__\n\n',
+                       docstring)
+    docstring = re.sub(r'\n        # (.*)\n',
+                       r'\n        __\1__\n\n',
+                       docstring)
+
+    docstring = re.sub(r'    ([^\s\\]+):(.*)\n',
+                       r'    - __\1__:\2\n',
+                       docstring)
+
+    docstring = docstring.replace('    ' * 6, '\t\t')
+    docstring = docstring.replace('    ' * 4, '\t')
+    docstring = docstring.replace('    ', '')
+    return docstring
+
+print('Cleaning up existing sources directory.')
+if os.path.exists('sources'):
+    shutil.rmtree('sources')
+
+print('Populating sources directory with templates.')
+for subdir, dirs, fnames in os.walk('templates'):
+    for fname in fnames:
+        new_subdir = subdir.replace('templates', 'sources')
+        if not os.path.exists(new_subdir):
+            os.makedirs(new_subdir)
+        if fname[-3:] == '.md':
+            fpath = os.path.join(subdir, fname)
+            new_fpath = fpath.replace('templates', 'sources')
+            shutil.copy(fpath, new_fpath)
+
+print('Starting autogeneration.')
+for page_data in PAGES:
+    blocks = []
+    classes = page_data.get('classes', [])
+    for module in page_data.get('all_module_classes', []):
+        module_classes = []
+        for name in dir(module):
+            if name[0] == '_' or name in EXCLUDE:
+                continue
+            module_member = getattr(module, name)
+            if inspect.isclass(module_member):
+                cls = module_member
+                if cls.__module__ == module.__name__:
+                    if cls not in module_classes:
+                        module_classes.append(cls)
+        module_classes.sort(key=lambda x: id(x))
+        classes += module_classes
+
+    for cls in classes:
+        subblocks = []
+        signature = get_class_signature(cls)
+        subblocks.append('<span style="float:right;">' + class_to_source_link(cls) + '</span>')
+        subblocks.append('### ' + cls.__name__ + '\n')
+        subblocks.append(code_snippet(signature))
+        docstring = cls.__doc__
+        if docstring:
+            subblocks.append(process_class_docstring(docstring))
+        blocks.append('\n'.join(subblocks))
+
+    functions = page_data.get('functions', [])
+    for module in page_data.get('all_module_functions', []):
+        module_functions = []
+        for name in dir(module):
+            if name[0] == '_' or name in EXCLUDE:
+                continue
+            module_member = getattr(module, name)
+            if inspect.isfunction(module_member):
+                function = module_member
+                if module.__name__ in function.__module__:
+                    if function not in module_functions:
+                        module_functions.append(function)
+        module_functions.sort(key=lambda x: id(x))
+        functions += module_functions
+
+    for function in functions:
+        subblocks = []
+        signature = get_function_signature(function, method=False)
+        signature = signature.replace(function.__module__ + '.', '')
+        subblocks.append('### ' + function.__name__ + '\n')
+        subblocks.append(code_snippet(signature))
+        docstring = function.__doc__
+        if docstring:
+            subblocks.append(process_function_docstring(docstring))
+            blocks.append('\n\n'.join(subblocks))
+
+    mkdown = '\n----\n\n'.join(blocks)
+    # save module page.
+    # Either insert content into existing page,
+    # or create page otherwise
+    page_name = page_data['page']
+    path = os.path.join('sources', page_name)
+    if os.path.exists(path):
+        template = open(path).read()
+        assert '{{autogenerated}}' in template, ('Template found for ' + path +
+                                                 ' but missing {{autogenerated}} tag.')
+        mkdown = template.replace('{{autogenerated}}', mkdown)
+        print('...inserting autogenerated content into template:', path)
+    else:
+        print('...creating new page with autogenerated content:', path)
+    subdir = os.path.dirname(path)
+    if not os.path.exists(subdir):
+        os.makedirs(subdir)
+    open(path, 'w').write(mkdown)

docs/mkdocs.yml

@@ -3,8 +3,8 @@ theme: readthedocs
 docs_dir: sources
 repo_url: http://github.com/fchollet/keras
 site_url: http://keras.io/
-#theme_dir: theme
-site_description: Documentation for fast and lightweight Keras Deep Learning library.
+# theme_dir: theme
+site_description: 'Documentation for Keras, the Python Deep Learning library.'
 
 dev_addr: '0.0.0.0:8000'
 google_analytics: ['UA-61785484-1', 'keras.io']
@@ -12,30 +12,44 @@ google_analytics: ['UA-61785484-1', 'keras.io']
 
 pages:
 - Home: index.md
-- Index: documentation.md
-- Examples: examples.md
-- FAQ: faq.md
-- Optimizers: optimizers.md
-- Objectives: objectives.md
-- Models: models.md
-- Activations: activations.md
-- Initializations: initializations.md
-- Regularizers: regularizers.md
-- Constraints: constraints.md
-- Callbacks: callbacks.md
-- Datasets: datasets.md
-- Visualization: visualization.md
+- Getting started:
+  - Guide to the Sequential model: getting-started/sequential-model-guide.md
+  - Guide to the Functional API: getting-started/functional-api-guide.md
+  - FAQ: getting-started/faq.md
+- Models:
+  - About Keras models: models/about-keras-models.md
+  - Sequential: models/sequential.md
+  - Model (functional API): models/model.md
 - Layers:
+  - About Keras layers: layers/about-keras-layers.md
   - Core Layers: layers/core.md
   - Convolutional Layers: layers/convolutional.md
+  - Pooling Layers: layers/pooling.md
+  - Locally-connected Layers: layers/local.md
   - Recurrent Layers: layers/recurrent.md
-  - Advanced Activations Layers: layers/advanced_activations.md
-  - Normalization Layers: layers/normalization.md
   - Embedding Layers: layers/embeddings.md
+  - Advanced Activations Layers: layers/advanced-activations.md
+  - Normalization Layers: layers/normalization.md
   - Noise layers: layers/noise.md
-  - Containers: layers/containers.md
+  - Layer wrappers: layers/wrappers.md
+  - Writing your own Keras layers: layers/writing-your-own-keras-layers.md
 - Preprocessing:
   - Sequence Preprocessing: preprocessing/sequence.md
   - Text Preprocessing: preprocessing/text.md
   - Image Preprocessing: preprocessing/image.md
+- Objectives: objectives.md
+- Optimizers: optimizers.md
+- Activations: activations.md
+- Callbacks: callbacks.md
+- Datasets: datasets.md
+- Applications: applications.md
+- Backend: backend.md
+- Initializations: initializations.md
+- Regularizers: regularizers.md
+- Constraints: constraints.md
+- Visualization: visualization.md
+- Scikit-learn API: scikit-learn-api.md
+
+
+
 

docs/sources/documentation.md

@@ -1,40 +0,0 @@
-# Keras Documentation Index
-
-## Introduction
-
-- [Home](index.md)
-- [Index](documentation.md)
-- [Examples](examples.md)
-- [FAQ](faq.md)
-- [Backend](backend.md)
-
----
-
-## Base functionality
-
-- [Optimizers](optimizers.md)
-- [Objectives](objectives.md)
-- [Models](models.md)
-- [Activations](activations.md)
-- [Initializations](initializations.md)
-- [Regularizers](regularizers.md)
-- [Constraints](constraints.md)
-- [Callbacks](callbacks.md)
-- [Datasets](datasets.md)
-
----
-
-## Layers
-- [Core](layers/core.md)
-- [Convolutional](layers/convolutional.md)
-- [Recurrent](layers/recurrent.md)
-- [Advanced Activations](layers/advanced_activations.md)
-- [Normalization](layers/normalization.md)
-- [Embeddings](layers/embeddings.md)
-
----
-
-## Preprocessing
-- [Sequence](preprocessing/sequence.md)
-- [Text](preprocessing/text.md)
-- [Image](preprocessing/image.md)

docs/sources/examples.md

@@ -1,176 +0,0 @@
-
-Here are a few examples to get you started!
-
-### Multilayer Perceptron (MLP):
-
-```python
-from keras.models import Sequential
-from keras.layers.core import Dense, Dropout, Activation
-from keras.optimizers import SGD
-
-model = Sequential()
-# Dense(64) is a fully-connected layer with 64 hidden units.
-# in the first layer, you must specify the expected input data shape:
-# here, 20-dimensional vectors.
-model.add(Dense(64, input_dim=20, init='uniform'))
-model.add(Activation('tanh'))
-model.add(Dropout(0.5))
-model.add(Dense(64, init='uniform'))
-model.add(Activation('tanh'))
-model.add(Dropout(0.5))
-model.add(Dense(2, init='uniform'))
-model.add(Activation('softmax'))
-
-sgd = SGD(lr=0.1, decay=1e-6, momentum=0.9, nesterov=True)
-model.compile(loss='mean_squared_error', optimizer=sgd)
-
-model.fit(X_train, y_train, nb_epoch=20, batch_size=16)
-score = model.evaluate(X_test, y_test, batch_size=16)
-```
-
-### Alternative implementation of MLP:
-
-```python
-model = Sequential()
-model.add(Dense(64, input_dim=20, init='uniform', activation='tanh'))
-model.add(Dropout(0.5))
-model.add(Dense(64, init='uniform', activation='tanh'))
-model.add(Dropout(0.5))
-model.add(Dense(2, init='uniform', activation='softmax'))
-
-sgd = SGD(lr=0.1, decay=1e-6, momentum=0.9, nesterov=True)
-model.compile(loss='mean_squared_error', optimizer=sgd)
-```
-
-### VGG-like convnet:
-
-```python
-from keras.models import Sequential
-from keras.layers.core import Dense, Dropout, Activation, Flatten
-from keras.layers.convolutional import Convolution2D, MaxPooling2D
-from keras.optimizers import SGD
-
-model = Sequential()
-# input: 100x100 images with 3 channels -> (3, 100, 100) tensors.
-# this applies 32 convolution filters of size 3x3 each.
-model.add(Convolution2D(32, 3, 3, border_mode='full', input_shape=(3, 100, 100)))
-model.add(Activation('relu'))
-model.add(Convolution2D(32, 3, 3))
-model.add(Activation('relu'))
-model.add(MaxPooling2D(pool_size=(2, 2)))
-model.add(Dropout(0.25))
-
-model.add(Convolution2D(64, 3, 3, border_mode='valid'))
-model.add(Activation('relu'))
-model.add(Convolution2D(64, 3, 3))
-model.add(Activation('relu'))
-model.add(MaxPooling2D(pool_size=(2, 2)))
-model.add(Dropout(0.25))
-
-model.add(Flatten())
-# Note: Keras does automatic shape inference.
-model.add(Dense(256))
-model.add(Activation('relu'))
-model.add(Dropout(0.5))
-
-model.add(Dense(10))
-model.add(Activation('softmax'))
-
-sgd = SGD(lr=0.1, decay=1e-6, momentum=0.9, nesterov=True)
-model.compile(loss='categorical_crossentropy', optimizer=sgd)
-
-model.fit(X_train, Y_train, batch_size=32, nb_epoch=1)
-
-```
-
-### Sequence classification with LSTM:
-
-```python
-from keras.models import Sequential
-from keras.layers.core import Dense, Dropout, Activation
-from keras.layers.embeddings import Embedding
-from keras.layers.recurrent import LSTM
-
-model = Sequential()
-model.add(Embedding(max_features, 256, input_length=maxlen))
-model.add(LSTM(output_dim=128, activation='sigmoid', inner_activation='hard_sigmoid'))
-model.add(Dropout(0.5))
-model.add(Dense(1))
-model.add(Activation('sigmoid'))
-
-model.compile(loss='binary_crossentropy', optimizer='rmsprop')
-
-model.fit(X_train, Y_train, batch_size=16, nb_epoch=10)
-score = model.evaluate(X_test, Y_test, batch_size=16)
-```
-
-### Architecture for learning image captions with a convnet and a Gated Recurrent Unit:
-(word-level embedding, caption of maximum length 16 words).
-
-Note that getting this to work well will require using a bigger convnet, initialized with pre-trained weights.
-
-```python
-max_caption_len = 16
-vocab_size = 10000
-
-# first, let's define an image model that
-# will encode pictures into 128-dimensional vectors.
-# it should be initialized with pre-trained weights.
-image_model = Sequential()
-image_model.add(Convolution2D(32, 3, 3, border_mode='full', input_shape=(3, 100, 100)))
-image_model.add(Activation('relu'))
-image_model.add(Convolution2D(32, 3, 3))
-image_model.add(Activation('relu'))
-image_model.add(MaxPooling2D(pool_size=(2, 2)))
-
-image_model.add(Convolution2D(64, 3, 3, border_mode='full'))
-image_model.add(Activation('relu'))
-image_model.add(Convolution2D(64, 3, 3))
-image_model.add(Activation('relu'))
-image_model.add(MaxPooling2D(pool_size=(2, 2)))
-
-image_model.add(Flatten())
-image_model.add(Dense(128))
-
-# let's load the weights from a save file.
-image_model.load_weights('weight_file.h5')
-
-# next, let's define a RNN model that encodes sequences of words
-# into sequences of 128-dimensional word vectors.
-language_model = Sequential()
-language_model.add(Embedding(vocab_size, 256, input_length=max_caption_len))
-language_model.add(GRU(output_dim=128, return_sequences=True))
-language_model.add(TimeDistributedDense(128))
-
-# let's repeat the image vector to turn it into a sequence.
-image_model.add(RepeatVector(max_caption_len))
-
-# the output of both models will be tensors of shape (samples, max_caption_len, 128).
-# let's concatenate these 2 vector sequences.
-model = Merge([image_model, language_model], mode='concat', concat_axis=-1)
-# let's encode this vector sequence into a single vector
-model.add(GRU(256, 256, return_sequences=False))
-# which will be used to compute a probability
-# distribution over what the next word in the caption should be!
-model.add(Dense(vocab_size))
-model.add(Activation('softmax'))
-
-model.compile(loss='categorical_crossentropy', optimizer='rmsprop')
-
-# "images" is a numpy float array of shape (nb_samples, nb_channels=3, width, height).
-# "captions" is a numpy integer array of shape (nb_samples, max_caption_len)
-# containing word index sequences representing partial captions.
-# "next_words" is a numpy float array of shape (nb_samples, vocab_size)
-# containing a categorical encoding (0s and 1s) of the next word in the corresponding
-# partial caption.
-model.fit([images, partial_captions], next_words, batch_size=16, nb_epoch=100)
-```
-
-In the examples folder, you will find example models for real datasets:
-- CIFAR10 small images classification: Convolutional Neural Network (CNN) with realtime data augmentation
-- IMDB movie review sentiment classification: LSTM over sequences of words
-- Reuters newswires topic classification: Multilayer Perceptron (MLP)
-- MNIST handwritten digits classification: MLP & CNN
-- Character-level text generation with LSTM
-
-...and more.

docs/sources/faq.md

@@ -1,179 +0,0 @@
-# Keras FAQ: Frequently Asked Keras Questions
-
-[How can I run Keras on GPU?](#how-can-i-run-keras-on-gpu)
-
-[How can I save a Keras model?](#how-can-i-save-a-keras-model)
-
-[Why is the training loss much higher than the testing loss?](#why-is-the-training-loss-much-higher-than-the-testing-loss)
-
-[How can I visualize the output of an intermediate layer?](#how-can-i-visualize-the-output-of-an-intermediate-layer)
-
-[Isn't there a bug with Merge or Graph related to input concatenation?](#isnt-there-a-bug-with-merge-or-graph-related-to-input-concatenation)
-
-[How can I use Keras with datasets that don't fit in memory?](#how-can-i-use-keras-with-datasets-that-dont-fit-in-memory)
-
-[How can I interrupt training when the validation loss isn't decreasing anymore?](#how-can-i-interrupt-training-when-the-validation-loss-isnt-decreasing-anymore)
-
-[How is the validation split computed?](#how-is-the-validation-split-computed)
-
-[Is the data shuffled during training?](#is-the-data-shuffled-during-training)
-
-[How can I record the training / validation loss / accuracy at each epoch?](#how-can-i-record-the-training-validation-loss-accuracy-at-each-epoch)
-
----
-
-### How can I run Keras on GPU?
-
-Method 1: use Theano flags.
-```bash
-THEANO_FLAGS=device=gpu,floatX=float32 python my_keras_script.py
-```
-
-The name 'gpu' might have to be changed depending on your device's identifier (e.g. `gpu0`, `gpu1`, etc).
-
-Method 2: set up your `.theanorc`: [Instructions](http://deeplearning.net/software/theano/library/config.html)
-
-Method 3: manually set `theano.config.device`, `theano.config.floatX` at the beginning of your code:
-```python
-import theano
-theano.config.device = 'gpu'
-theano.config.floatX = 'float32'
-```
-
----
-
-### How can I save a Keras model?
-
-*It is not recommended to use pickle or cPickle to save a Keras model.*
-
-If you only need to save the architecture of a model, and not its weights, you can do:
-
-```python
-# save as JSON
-json_string = model.to_json()
-
-# save as YAML
-yaml_string = model.to_yaml()
-```
-
-You can then build a fresh model from this data:
-
-```python
-# model reconstruction from JSON:
-from keras.models import model_from_json
-model = model_from_json(json_string)
-
-# model reconstruction from YAML
-model = model_from_yaml(yaml_string)
-```
-
-If you need to save the weights of a model, you can do so in HDF5:
-```python
-model.save_weights('my_model_weights.h5')
-```
-
-Assuming you have code for instantiating your model, you can then load the weights you saved into a model with the same architecture:
-
-```python
-model.load_weights('my_model_weights.h5')
-```
-
-This leads us to a way to save and reconstruct models from only serialized data:
-```python
-json_string = model.to_json()
-open('my_model_architecture.json', 'w').write(json_string)
-model.save_weights('my_model_weights.h5')
-
-# elsewhere...
-model = model_from_json(open('my_model_architecture.json').read())
-model.load_weights('my_model_weights.h5')
-```
-
----
-
-### Why is the training loss much higher than the testing loss?
-
-A Keras model has two modes: training and testing. Regularization mechanisms, such as Dropout and L1/L2 weight regularization, are turned off at testing time.
-
-Besides, the training loss is the average of the losses over each batch of training data. Because your model is changing over time, the loss over the first batches of an epoch is generally higher than over the last batches. On the other hand, the testing loss for an epoch is computed using the model as it is at the end of the epoch, resulting in a lower loss.
-
----
-
-### How can I visualize the output of an intermediate layer?
-
-You can build a Theano function that will return the output of a certain layer given a certain input, for example:
-
-```python
-# with a Sequential model
-get_3rd_layer_output = theano.function([model.layers[0].input], 
-                                       model.layers[3].get_output(train=False))
-layer_output = get_3rd_layer_output(X)
-
-# with a Graph model
-get_conv_layer_output = theano.function([model.inputs[i].input for i in model.input_order],
-                                        model.outputs['conv'].get_output(train=False),
-                                        on_unused_input='ignore')
-conv_output = get_conv_output(input_data_dict)
-```
-
----
-
-### Isn't there a bug with Merge or Graph related to input concatenation?
-
-Yes, there was a known bug with tensor concatenation in Thenao that was fixed early 2015. 
-Please upgrade to the latest version of Theano:
-
-```bash
-sudo pip install git+git://github.com/Theano/Theano.git
-```
-
----
-
-### How can I use Keras with datasets that don't fit in memory?
-
-You can do batch training using `model.train_on_batch(X, y)` and `model.test_on_batch(X, y)`. See the [models documentation](models.md).
-
-You can also see batch training in action in our [CIFAR10 example](https://github.com/fchollet/keras/blob/master/examples/cifar10_cnn.py).
-
----
-
-### How can I interrupt training when the validation loss isn't decreasing anymore?
-
-You can use an `EarlyStopping` callback:
-
-```python
-from keras.callbacks import EarlyStopping
-early_stopping = EarlyStopping(monitor='val_loss', patience=2)
-model.fit(X, y, validation_split=0.2, callbacks=[early_stopping])
-```
-
-Find out more in the [callbacks documentation](callbacks.md).
-
----
-
-### How is the validation split computed?
-
-If you set the `validation_split` arugment in `model.fit` to e.g. 0.1, then the validation data used will be the *last 10%* of the data. If you set it to 0.25, it will be the last 25% of the data, etc.
-
-
----
-
-### Is the data shuffled during training?
-
-Yes, if the `shuffle` argument in `model.fit` is set to `True` (which is the default), the training data will be randomly shuffled at each epoch.
-
-Validation data isn't shuffled.
-
----
-
-
-### How can I record the training / validation loss / accuracy at each epoch?
-
-The `model.fit` method returns an `History` callback, which has a `history` attribute containing the lists of successive losses / accuracies.
-
-```python
-hist = model.fit(X, y, validation_split=0.2)
-print(hist.history)
-```
-
----

docs/sources/initializations.md

@@ -1,23 +0,0 @@
-
-## Usage of initializations
-
-Initializations define the probability distribution used to set the initial random weights of Keras layers.
-
-The keyword arguments used for passing initializations to layers will depend on the layer. Usually it is simply `init`:
-
-```python
-model.add(Dense(64, init='uniform'))
-```
-
-## Available initializations
-
-- __uniform__
-- __lecun_uniform__: Uniform initialization scaled by the square root of the number of inputs (LeCun 98).
-- __normal__
-- __identity__: Use with square 2D layers (`shape[0] == shape[1]`).
-- __orthogonal__: Use with square 2D layers (`shape[0] == shape[1]`).
-- __zero__
-- __glorot_normal__: Gaussian initialization scaled by fan_in + fan_out (Glorot 2010)
-- __glorot_uniform__
-- __he_normal__: Gaussian initialization scaled by fan_in (He et al., 2014)
-- __he_uniform__

docs/sources/layers/advanced_activations.md

@@ -1,106 +0,0 @@
-
-## LeakyReLU
-
-```python
-keras.layers.advanced_activations.LeakyReLU(alpha=0.3)
-```
-
-Special version of a Rectified Linear Unit that allows a small gradient when the unit is not active (`f(x) = alpha*x for x < 0`).
-
-
-- __Input shape__: Arbitrary. Use the keyword argument `input_shape` (tuple of integers, does not include the samples axis) when using this layer as the first layer in a model.
-
-- __Output shape__: Same as input.
-
-- __Arguments__:
-    - __alpha__: float >= 0. Negative slope coefficient.
-
----
-
-## PReLU
-
-```python
-keras.layers.advanced_activations.PReLU()
-```
-
-Parametrized linear unit. Similar to a LeakyReLU, where each input unit has its alpha coefficient, and where these coefficients are learned during training.
-
-
-- __Input shape__: Arbitrary. Use the keyword argument `input_shape` (tuple of integers, does not include the samples axis) when using this layer as the first layer in a model.
-
-- __Output shape__: Same as input.
-
-- __References__:
-    - [Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification](http://arxiv.org/pdf/1502.01852v1.pdf)
-
----
-
-## ELU
-
-```python
-keras.layers.advanced_activations.ELU()
-```
-
-Exponential linear unit. Negative values pushes mean unit activations closer to zero, with the advantage of having a noise-robust deactivation state.
-
-
-- __Input shape__: Arbitrary. Use the keyword argument `input_shape` (tuple of integers, does not include the samples axis) when using this layer as the first layer in a model.
-
-- __Output shape__: Same as input.
-
-- __References__:
-    - [Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)](http://arxiv.org/pdf/1511.07289v1.pdf)
-
----
-
-## ParametricSoftplus
-
-```python
-keras.layers.advanced_activations.ParametricSoftplus()
-```
-
-Parametric Softplus of the form: (`f(x) = alpha * (1 + exp(beta * x))`). This is essentially a smooth version of ReLU where the parameters control the sharpness of the rectification. The parameters are initialized to more closely approximate a ReLU than the standard `softplus`: `alpha` initialized to `0.2` and `beta`  initialized to `5.0`. The parameters are fit separately for each hidden unit.
-
-- __Input shape__: Arbitrary. Use the keyword argument `input_shape=...` when using this layer as the first layer in a model.
-
-- __Output shape__: Same as input.
-
-- __References__:
-    - [Inferring Nonlinear Neuronal Computation Based on Physiologically Plausible Inputs](http://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1003143)
-
-## Thresholded Linear
-
-```python
-keras.layers.advanced_activations.ThresholdedLinear(theta)
-```
-
-Parametrized linear unit. provides a threshold near zero where values are zeroed.
-
-
-- __Input shape__: Arbitrary. Use the keyword argument `input_shape` (tuple of integers, does not include the samples axis) when using this layer as the first layer in a model.
-
-- __Output shape__: Same as input.
-
-- __Arguments__:
-    - __theta__: float >= 0. Threshold location of activation
-
-- __References__:
-    - [Zero-Bias Autoencoders and the Benefits of Co-Adapting Features](http://arxiv.org/pdf/1402.3337.pdf)
-
-## Thresholded ReLu
-
-```python
-keras.layers.advanced_activations.ThresholdedReLu(theta)
-```
-
-Parametrized rectified linear unit. provides a threshold near zero where values are zeroed.
-
-- __Input shape__: Arbitrary. Use the keyword argument `input_shape=...` when using this layer as the first layer in a model.
-
-- __Output shape__: Same as input.
-
-- __Arguments__:
-    - __theta__: float >= 0. Threshold location of activation
-
-- __References__:
-    - [Zero-Bias Autoencoders and the Benefits of Co-Adapting Features](http://arxiv.org/pdf/1402.3337.pdf)

docs/sources/layers/containers.md

@@ -1,21 +0,0 @@
-Containers are ensembles of layers that can be interacted with through the same API as `Layer` objects.
-
-## Sequential
-
-```python
-keras.layers.containers.Sequential(layers=[])
-```
-
-The Sequential container is a linear stack of layers. Apart from the `add` methods and the `layers` constructor argument, the API is identical to that of the `Layer` class.
-
-This class is also the basis for the `keras.models.Sequential` architecture.
-
-The `layers` constructor argument is a list of Layer instances.
-
-__Methods__:
-
-```python
-add(layer)
-```
-
-Add a new layer to the stack.

docs/sources/layers/convolutional.md

@@ -1,202 +0,0 @@
-
-## Convolution1D
-
-```python
-keras.layers.convolutional.Convolution1D(nb_filter, filter_length, 
-        init='uniform',
-        activation='linear',
-        weights=None, 
-        border_mode='valid',
-        subsample_length=1, 
-        W_regularizer=None, b_regularizer=None,
-        W_constraint=None, b_constraint=None,
-        input_dim=None, input_length=None)
-```
-
-Convolution operator for filtering neighborhoods of one-dimensional inputs. When using this layer as the first layer in a model, either provide the keyword argument `input_dim` (int, e.g. 128 for sequences of 128-dimensional vectors), or `input_shape` (tuple of integers, e.g. (10, 128) for sequences of 10 vectors of 128-dimensional vectors).
-
-- __Input shape__: 3D tensor with shape: `(samples, steps, input_dim)`.
-
-- __Output shape__: 3D tensor with shape: `(samples, new_steps, nb_filter)`. `steps` value might have changed due to padding.
-
-- __Arguments__:
-    - __nb_filter__: Number of convolution kernels to use (dimensionality of the output).
-    - __filter_length__: The extension (spatial or temporal) of each filter.
-    - __init__: name of initialization function for the weights of the layer (see: [initializations](../initializations.md)), or alternatively, Theano function to use for weights initialization. This parameter is only relevant if you don't pass a `weights` argument.
-    - __activation__: name of activation function to use (see: [activations](../activations.md)), or alternatively, elementwise Theano function. If you don't specify anything, no activation is applied (ie. "linear" activation: a(x) = x).
-    - __weights__: list of numpy arrays to set as initial weights.
-    - __border_mode__: 'valid' or 'same'.
-    - __subsample_length__: factor by which to subsample output.
-    - __W_regularizer__: instance of [WeightRegularizer](../regularizers.md) (eg. L1 or L2 regularization), applied to the main weights matrix.
-    - __b_regularizer__: instance of [WeightRegularizer](../regularizers.md), applied to the bias.
-    - __activity_regularizer__: instance of [ActivityRegularizer](../regularizers.md), applied to the network output.
-    - __W_constraint__: instance of the [constraints](../constraints.md) module (eg. maxnorm, nonneg), applied to the main weights matrix.
-    - __b_constraint__: instance of the [constraints](../constraints.md) module, applied to the bias.
-    - __input_dim__: Number of channels/dimensions in the input. Either this argument or the keyword argument `input_shape` must be provided when using this layer as the first layer in a model.
-    - __input_length__: Length of input sequences, when it is constant. This argument is required if you are going to connect `Flatten` then `Dense` layers upstream (without it, the shape of the dense outputs cannot be computed).
-
----
-
-## Convolution2D
-
-```python
-keras.layers.convolutional.Convolution2D(nb_filter, nb_row, nb_col, 
-        init='glorot_uniform',
-        activation='linear',
-        weights=None, 
-        border_mode='valid',
-        subsample=(1, 1),
-        W_regularizer=None, b_regularizer=None,
-        W_constraint=None,
-        dim_ordering='th')
-```
-
-Convolution operator for filtering windows of two-dimensional inputs. When using this layer as the first layer in a model, provide the keyword argument `input_shape` (tuple of integers, does not include the sample axis), e.g. `input_shape=(3, 128, 128)` for 128x128 RGB pictures.
-
-- __Input shape__: 4D tensor with shape: `(samples, channels, rows, cols)` if dim_ordering='th'
-or 4D tensor with shape: `(samples, rows, cols, channels)` if dim_ordering='tf'.
-
-- __Output shape__: 4D tensor with shape: `(samples, nb_filter, nb_row, nb_col)` if dim_ordering='th'
-or 4D tensor with shape: `(samples, nb_row, nb_col, nb_filter)` if dim_ordering='tf'.
-
-
-- __Arguments__:
-
-    - __nb_filter__: Number of convolution filters to use.
-    - __nb_row__: Number of rows in the convolution kernel.
-    - __nb_col__: Number of columns in the convolution kernel.
-    - __init__: name of initialization function for the weights of the layer (see: [initializations](../initializations.md)), or alternatively, Theano function to use for weights initialization. This parameter is only relevant if you don't pass a `weights` argument.
-    - __activation__: name of activation function to use (see: [activations](../activations.md)), or alternatively, elementwise Theano function. If you don't specify anything, no activation is applied (ie. "linear" activation: a(x) = x).
-    - __weights__: list of numpy arrays to set as initial weights.
-    - __border_mode__: 'valid' or 'same'.
-    - __subsample__: tuple of length 2. Factor by which to subsample output. Also called strides elsewhere.
-    - __W_regularizer__: instance of [WeightRegularizer](../regularizers.md) (eg. L1 or L2 regularization), applied to the main weights matrix.
-    - __b_regularizer__: instance of [WeightRegularizer](../regularizers.md), applied to the bias.
-    - __activity_regularizer__: instance of [ActivityRegularizer](../regularizers.md), applied to the network output.
-    - __W_constraint__: instance of the [constraints](../constraints.md) module (eg. maxnorm, nonneg), applied to the main weights matrix.
-    - __b_constraint__: instance of the [constraints](../constraints.md) module, applied to the bias.
-    - __dim_ordering__: 'th' or 'tf'. In 'th' mode, the channels dimension (the depth) is at index 1, in 'tf' mode is it at index 3.
-
-
----
-
-## MaxPooling1D
-
-```python
-keras.layers.convolutional.MaxPooling1D(pool_length=2, stride=None, border_mode='valid')
-```
-
-Max pooling operation for temporal data.
-
-- __Input shape__: 3D tensor with shape: `(samples, steps, features)`.
-
-- __Output shape__: 3D tensor with shape: `(samples, downsampled_steps, features)`.
-
-- __Arguments__:
-
-    - __pool_length__: factor by which to downscale. 2 will halve the input.
-    - __stride__: integer or None. Stride value.
-    - __border_mode__: 'valid' or 'same'. **Note:** 'same' will only work with TensorFlow for the time being.
-
----
-
-## MaxPooling2D
-
-```python
-keras.layers.convolutional.MaxPooling2D(pool_size=(2, 2), border_mode='valid', dim_ordering='th')
-```
-
-Max pooling operation for spatial data.
-
-- __Input shape__: 4D tensor with shape: `(samples, channels, rows, cols)` if dim_ordering='th'
-or 4D tensor with shape: `(samples, rows, cols, channels)` if dim_ordering='tf'.
-
-- __Output shape__: 4D tensor with shape: `(nb_samples, channels, pooled_rows, pooled_cols)` if dim_ordering='th'
-or 4D tensor with shape: `(samples, pooled_rows, pooled_cols, channels)` if dim_ordering='tf'.
-
-- __Arguments__:
-
-    - __pool_size__: tuple of 2 integers, factors by which to downscale (vertical, horizontal). (2, 2) will halve the image in each dimension.
-    - __strides__: tuple of 2 integers, or None. Strides values.
-    - __border_mode__: 'valid' or 'same'. **Note:** 'same' will only work with TensorFlow for the time being.
-    - __dim_ordering__: 'th' or 'tf'. In 'th' mode, the channels dimension (the depth) is at index 1, in 'tf' mode is it at index 3.
-
-
----
-
-## UpSampling1D
-
-```python
-keras.layers.convolutional.UpSampling1D(length=2)
-```
-
-Repeats each temporal step `length` times along the time axis.
-
-- __Input shape__: 3D tensor with shape: `(samples, steps, features)`.
-
-- __Output shape__: 3D tensor with shape: `(samples, upsampled_steps, features)`.
-
-- __Arguments__:
-    - __length__: integer. Upsampling factor.
-
----
-
-
-## UpSampling2D
-
-```python
-keras.layers.convolutional.UpSampling2D(size=(2, 2), dim_ordering='th')
-```
-
-Repeats the rows and columns of the data by size[0] and size[1] respectively.
-
-- __Input shape__: 4D tensor with shape: `(samples, channels, rows, cols)` if dim_ordering='th'
-or 4D tensor with shape: `(samples, rows, cols, channels)` if dim_ordering='tf'.
-
-- __Output shape__: 4D tensor with shape: `(samples, channels, upsampled_rows, upsampled_cols)` if dim_ordering='th'
-or 4D tensor with shape: `(samples, upsampled_rows, upsampled_cols, channels)` if dim_ordering='tf'.
-
-- __Arguments__:
-    - __size__: tuple of 2 integers. The upsampling factors for rows and columns.
-    - __dim_ordering__: 'th' or 'tf'. In 'th' mode, the channels dimension (the depth) is at index 1, in 'tf' mode is it at index 3.
-
----
-
-
-## ZeroPadding1D
-
-```python
-keras.layers.convolutional.ZeroPaddding1D(padding=1)
-```
-
-Pads the input with zeros left and right along the time axis.
-
-- __Input shape__: 3D tensor with shape: `(nb_samples, steps, dim)`.
-
-- __Output shape__: 3D tensor with shape: `(nb_samples, padded_steps, dim)`.
-
-- __Arguments__:
-    - __padding__: integer, the size of the padding.
-
----
-
-
-## ZeroPadding2D
-
-```python
-keras.layers.convolutional.ZeroPaddding2D(padding=(1, 1), dim_ordering='th')
-```
-
-Pads the rows and columns of the input with zeros, left and right.
-
-- __Input shape__: 4D tensor with shape: `(samples, channels, rows, cols)` if dim_ordering='th'
-or 4D tensor with shape: `(samples, rows, cols, channels)` if dim_ordering='tf'.
-
-- __Output shape__: 4D tensor with shape: `(samples, channels, padded_rows, padded_cols)` if dim_ordering='th'
-or 4D tensor with shape: `(samples, padded_rows, padded_cols, channels)` if dim_ordering='tf'.
-
-- __Arguments__:
-    - __padding__: tuple of 2 integers, the size of the padding for rows and columns respectively.
-    - __dim_ordering__: 'th' or 'tf'. In 'th' mode, the channels dimension (the depth) is at index 1, in 'tf' mode is it at index 3.
-
----

docs/sources/layers/core.md

@@ -1,477 +0,0 @@
-## Base class
-
-```python
-keras.layers.core.Layer()
-```
-
-__Methods__:
-
-```python
-set_previous(previous_layer)
-```
-
-Connect the input of the current layer to the output of the argument layer.
-
-- __Return__: None.
-
-- __Arguments__: 
-    - __previous_layer__: Layer object.
-
-
-
-```python
-get_output(train)
-```
-
-Get the output of the layer.
-
-- __Return__: Theano tensor.
-
-- __Arguments__: 
-    - __train__: Boolean. Specifies whether output is computed in training mode or in testing mode, which can change the logic, for instance in there are any `Dropout` layers in the network. 
-
-
-
-```python
-get_input(train)
-```
-
-Get the input of the layer.
-
-- __Return__: Theano tensor.
-
-- __Arguments__: 
-    - __train__: Boolean. Specifies whether output is computed in training mode or in testing mode, which can change the logic, for instance in there are any `Dropout` layers in the network. 
-
-
-
-```python
-get_weights()
-```
-
-Get the weights of the parameters of the layer.
-
-- __Return__: List of numpy arrays (one per layer parameter). 
-
-
-
-```python
-set_weights(weights)
-```
-
-Set the weights of the parameters of the layer.
-
-- __Arguments__: 
-    - __weights__: List of numpy arrays (one per layer parameter). Should be in the same order as what `get_weights(self)` returns.
-
-
-```python
-get_config()
-```
-
-- __Return__: Configuration dictionary describing the layer.
-
-
----
-
-## Dense
-```python
-keras.layers.core.Dense(output_dim,
-                        init='glorot_uniform',
-                        activation='linear',
-                        weights=None,
-                        W_regularizer=None, b_regularizer=None, activity_regularizer=None,
-                        W_constraint=None, b_constraint=None,
-                        input_dim=None)
-```
-
-Standard 1D fully-connect layer. 
-
-- __Input shape__: 2D tensor with shape: `(nb_samples, input_dim)`.
-
-- __Output shape__: 2D tensor with shape: `(nb_samples, output_dim)`.
-
-- __Arguments__:
-
-    - __output_dim__: int >= 0. 
-    - __init__: name of initialization function for the weights of the layer (see: [initializations](../initializations.md)), or alternatively, Theano function to use for weights initialization. This parameter is only relevant if you don't pass a `weights` argument.
-    - __activation__: name of activation function to use (see: [activations](../activations.md)), or alternatively, elementwise Theano function. If you don't specify anything, no activation is applied (ie. "linear" activation: a(x) = x).
-    - __weights__: list of numpy arrays to set as initial weights. The list should have 1 element, of shape `(input_dim, output_dim)`.
-    - __W_regularizer__: instance of [WeightRegularizer](../regularizers.md) (eg. L1 or L2 regularization), applied to the main weights matrix.
-    - __b_regularizer__: instance of [WeightRegularizer](../regularizers.md), applied to the bias.
-    - __activity_regularizer__: instance of [ActivityRegularizer](../regularizers.md), applied to the network output.
-    - __W_constraint__: instance of the [constraints](../constraints.md) module (eg. maxnorm, nonneg), applied to the main weights matrix.
-    - __b_constraint__: instance of the [constraints](../constraints.md) module, applied to the bias.
-    - __input_dim__: dimensionality of the input (integer). This argument (or alternatively, the keyword argument `input_shape`) is required when using this layer as the first layer in a model. 
-
----
-
-## TimeDistributedDense
-```python
-keras.layers.core.TimeDistributedDense(output_dim,
-                                       init='glorot_uniform',
-                                       activation='linear',
-                                       weights=None
-                                       W_regularizer=None, b_regularizer=None, activity_regularizer=None,
-                                       W_constraint=None, b_constraint=None,
-                                       input_dim=None, input_length=None)
-```
-
-Fully-connected layer distributed over the time dimension. Useful after a recurrent network set to `return_sequences=True`.
-
-- __Input shape__: 3D tensor with shape: `(nb_samples, timesteps, input_dim)`.
-
-- __Arguments__:
-    - __output_dim__: int >= 0. 
-    - __init__: name of initialization function for the weights of the layer (see: [initializations](../initializations.md)), or alternatively, Theano function to use for weights initialization. This parameter is only relevant if you don't pass a `weights` argument.
-    - __activation__: name of activation function to use (see: [activations](../activations.md)), or alternatively, elementwise Theano function. If you don't specify anything, no activation is applied (ie. "linear" activation: a(x) = x).
-    - __weights__: list of numpy arrays to set as initial weights. The list should have 1 element, of shape `(input_dim, output_dim)`.
-    - __W_regularizer__: instance of [WeightRegularizer](../regularizers.md) (eg. L1 or L2 regularization), applied to the main weights matrix.
-    - __b_regularizer__: instance of [WeightRegularizer](../regularizers.md), applied to the bias.
-    - __activity_regularizer__: instance of [ActivityRegularizer](../regularizers.md), applied to the network output.
-    - __W_constraint__: instance of the [constraints](../constraints.md) module (eg. maxnorm, nonneg), applied to the main weights matrix.
-    - __b_constraint__: instance of the [constraints](../constraints.md) module, applied to the bias.
-    - __input_dim__: dimensionality of the input (integer). This argument (or alternatively, the keyword argument `input_shape`) is required when using this layer as the first layer in a model.
-    - __input_length__: Length of input sequences, when it is constant. This argument is required if you are going to connect `Flatten` then `Dense` layers upstream (without it, the shape of the dense outputs cannot be computed).
-
-- __Example__:
-```python
-# input shape: (nb_samples, timesteps, 10)
-model.add(LSTM(5, return_sequences=True, input_dim=10)) # output shape: (nb_samples, timesteps, 5)
-model.add(TimeDistributedDense(15)) # output shape: (nb_samples, timesteps, 15)
-```
-
-
----
-
-## AutoEncoder
-```python
-keras.layers.core.AutoEncoder(encoder, decoder, output_reconstruction=True, weights=None):
-```
-
-A customizable autoencoder model. If `output_reconstruction = True` then dim(input) = dim(output) else dim(output) = dim(hidden)
-
-
-- __Input shape__: The layer shape is defined by the encoder definitions
-
-- __Output shape__: The layer shape is defined by the decoder definitions
-
-- __Arguments__:
-
-    - __encoder__: A [layer](./) or [layer container](./containers.md).
-
-    - __decoder__: A [layer](./) or [layer container](./containers.md).
-    
-    - __output_reconstruction__: If this is False, then when .predict() is called, the output is the deepest hidden layer's activation. Otherwise, the output of the final decoder layer is presented. Be sure your validation data conforms to this logic if you decide to use any.
-    
-    - __weights__: list of numpy arrays to set as initial weights. The list should have 1 element, of shape `(input_dim, output_dim)`.
-
-- __Example__:
-```python
-from keras.layers import containers
-
-# input shape: (nb_samples, 32)
-encoder = containers.Sequential([Dense(16, input_dim=32), Dense(8)])
-decoder = containers.Sequential([Dense(16, input_dim=8), Dense(32)])
-
-autoencoder = Sequential()
-autoencoder.add(AutoEncoder(encoder=encoder, decoder=decoder, output_reconstruction=False))
-```
-
-
----
-
-## Activation
-```python
-keras.layers.core.Activation(activation)
-```
-Apply an activation function to the input. 
-
-
-- __Input shape__: Arbitrary. Use the keyword argument `input_shape` (tuple of integers, does not include the samples axis) when using this layer as the first layer in a model.
-
-- __Output shape__: Same as input.
-
-- __Arguments__:
-
-    - __activation__: name of activation function to use (see: [activations](../activations.md)), or alternatively, elementwise Theano function.
-
-
----
-
-## Dropout
-```python
-keras.layers.core.Dropout(p)
-```
-Apply dropout to the input. Dropout consists in randomly setting a fraction `p` of input units to 0 at each update during training time, which helps prevent overfitting. Reference: [Dropout: A Simple Way to Prevent Neural Networks from Overfitting](http://www.cs.toronto.edu/~rsalakhu/papers/srivastava14a.pdf)
-
-
-- __Input shape__: Arbitrary. Use the keyword argument `input_shape` (tuple of integers, does not include the samples axis) when using this layer as the first layer in a model.
-
-- __Output shape__: Same as input.
-
-- __Arguments__:
-
-    - __p__: float (0 <= p < 1). Fraction of the input that gets dropped out at training time. 
-
----
-
-
-## Reshape
-```python
-keras.layers.core.Reshape(dims)
-```
-
-Reshape the input to a new shape containing the same number of units. 
-
-
-- __Input shape__: Arbitrary. Use the keyword argument `input_shape` (tuple of integers, does not include the samples axis) when using this layer as the first layer in a model.
-
-- __Output shape__: `(nb_samples, dims)`.
-
-- __Arguments__:
-
-    - dims: tuple of integers. Dimensions of the new shape.
-
-- __Example__:
-```python
-# input shape: (nb_samples, 10)
-model.add(Dense(100, input_dim=10)) # output shape: (nb_samples, 100)
-model.add(Reshape(dims=(10, 10)))  # output shape: (nb_samples, 10, 10)
-```
-
----
-
-## Flatten
-```python
-keras.layers.core.Flatten()
-```
-
-Convert a nD input to 1D. 
-
-- __Input shape__: Arbitrary. Use the keyword argument `input_shape` (tuple of integers, does not include the samples axis) when using this layer as the first layer in a model.
-
-- __Output shape__: `(nb_samples, nb_input_units)`.
-
----
-
-## RepeatVector
-```python
-keras.layers.core.RepeatVector(n)
-```
-
-Repeat the 1D input n times. Dimensions of input are assumed to be `(nb_samples, dim)`. Output will have the shape `(nb_samples, n, dim)`.
-
-Note that the output is still a single tensor; `RepeatVector` does not split the data flow.
-
-- __Input shape__: Arbitrary. Use the keyword argument `input_shape` (tuple of integers, does not include the samples axis) when using this layer as the first layer in a model.
-
-- __Output shape__: `(nb_samples, n, input_dims)`.
-
-- __Arguments__:
-    - __n__: int. 
-
----
-
-## Permute
-```python
-keras.layers.core.Permute(dims)
-```
-Permute the dimensions of the input data according to the given tuple. Sometimes useful for connecting RNNs and convnets together.
-
-- __Input shape__: Arbitrary. Use the keyword argument `input_shape` (tuple of integers, does not include the samples axis) when using this layer as the first layer in a model.
-
-- __Output shape__: Same as the input shape, but with the dimensions re-ordered according to the ordering specified by the tuple.
-
-- __Argument__: tuple specifying the permutation scheme (e.g. `(2, 1)` permutes the first and second dimension of the input).
-
-- __Example__:
-```python
-# input shape: (nb_samples, 10)
-model.add(Dense(50, input_dim=10)) # output shape: (nb_samples, 50)
-model.add(Reshape(dims=(10, 5))) # output shape: (nb_samples, 10, 5)
-model.add(Permute(dims=(2, 1))) #output shape: (nb_samples, 5, 10)
-```
-
----
-
-## ActivityRegularization
-```python
-keras.layers.core.ActivityRegularization(l1=0., l2=0.)
-```
-
-Leaves the input unchanged, but adds a term to the loss function based on the input activity. L1 and L2 regularization supported.
-
-This layer can be used, for instance, to induce activation sparsity in the previous layer.
-
----
-
-## MaxoutDense
-```python
-keras.layers.core.MaxoutDense(output_dim, nb_feature=4,
-        init='glorot_uniform',
-        weights=None,
-        W_regularizer=None, b_regularizer=None, activity_regularizer=None,
-        W_constraint=None, b_constraint=None,
-        input_dim=None)
-```
-
-A dense maxout layer. A `MaxoutDense` layer takes the element-wise maximum of `nb_feature` `Dense(input_dim, output_dim)` linear layers. This allows the layer to learn a convex, piecewise linear activation function over the inputs. See [this paper](http://arxiv.org/pdf/1302.4389.pdf) for more details. Note that this is a *linear* layer -- if you wish to apply activation function (you shouldn't need to -- they are universal function approximators), an `Activation` layer must be added after.
-
-- __Input shape__: 2D tensor with shape: `(nb_samples, input_dim)`.
-
-- __Output shape__: 2D tensor with shape: `(nb_samples, output_dim)`.
-
-- __Arguments__:
-
-    - __output_dim__: int >= 0. 
-    - __nb_feature__: int >= 0. the number of features to create for the maxout. This is equivalent to the number of piecewise elements to be allowed for the activation function. 
-    - __init__: name of initialization function for the weights of the layer (see: [initializations](../initializations.md)), or alternatively, Theano function to use for weights initialization. This parameter is only relevant if you don't pass a `weights` argument.
-    - __weights__: list of numpy arrays to set as initial weights. The list should have 1 element, of shape `(input_dim, output_dim)`.
-    - __W_regularizer__: instance of [WeightRegularizer](../regularizers.md) (eg. L1 or L2 regularization), applied to the main weights matrix.
-    - __b_regularizer__: instance of [WeightRegularizer](../regularizers.md), applied to the bias.
-    - __activity_regularizer__: instance of [ActivityRegularizer](../regularizers.md), applied to the network output.
-    - __W_constraint__: instance of the [constraints](../constraints.md) module (eg. maxnorm, nonneg), applied to the main weights matrix.
-    - __b_constraint__: instance of the [constraints](../constraints.md) module, applied to the bias.
-    - __input_dim__: dimensionality of the input (integer). This argument (or alternatively, the keyword argument `input_shape`) is required when using this layer as the first layer in a model.
-
-```python
-# input shape: (nb_samples, 10)
-model.add(Dense(100, input_dim=10)) # output shape: (nb_samples, 100)
-model.add(MaxoutDense(50, nb_feature=10)) # output shape: (nb_samples, 50)
-```
-
-## Merge
-```python
-keras.layers.core.Merge(layers, mode='sum', concat_axis=-1, dot_axes=-1)
-```
-
-Merge the output of a list of layers (or containers) into a single tensor.
-
-- __Arguments__:
-    - __layers__: List of layers or [containers](/layers/containers/).
-    - __mode__: String, one of `{'sum', 'mul', 'concat', 'ave', 'dot'}`. `sum`, `mul` and `ave` will simply sum/multiply/average the outputs of the layers (therefore all layers should have an output with the same shape). `concat` will concatenate the outputs along the dimension specified by `concate_axis` (therefore all layers should have an output that only differ along this dimension). `dot` will dot tensor contraction on the axes specified by `dot_axes` (see [the Numpy documentation](http://docs.scipy.org/doc/numpy-1.10.1/reference/generated/numpy.tensordot.html) for more details).
-    - __concat_axis__: axis to use in `concat` mode.
-    - __dot_axes__: axis or axes to use in `dot` mode (see [the Numpy documentation](http://docs.scipy.org/doc/numpy-1.10.1/reference/generated/numpy.tensordot.html) for more details).
-
-
-- __Notes__:
-    - `dot` mode only works with Theano for the time being.
-
-- __Example__:
-
-```python
-left = Sequential()
-left.add(Dense(50, input_shape=(784,)))
-left.add(Activation('relu'))
-
-right = Sequential()
-right.add(Dense(50, input_shape=(784,)))
-right.add(Activation('relu'))
-
-model = Sequential()
-model.add(Merge([left, right], mode='sum'))
-
-model.add(Dense(10))
-model.add(Activation('softmax'))
-
-model.compile(loss='categorical_crossentropy', optimizer='rmsprop')
-
-model.fit([X_train, X_train], Y_train, batch_size=128, nb_epoch=20, validation_data=([X_test, X_test], Y_test))
-```
-
-## Masking
-```python
-keras.layers.core.Masking(mask_value=0.)
-```
-
-Create a mask for the input data by using `mask_value` as the sentinel value which should be masked out.
-Given an input of dimensions `(nb_samples, timesteps, input_dim)`, return the input untouched as output, and supply a mask of shape `(nb_samples, timesteps)` where all timesteps which had *all* their values equal to `mask_value` are masked out.
-
-- __Input shape__: 3D tensor with shape: `(nb_samples, timesteps, features)`.
-
-- __Output shape__: 3D tensor with shape: `(nb_samples, timesteps, features)`.
-
-- __Notes__: Masking only works in Theano for the time being.
-
-## Lambda
-```python
-keras.layers.core.Lambda(function, output_shape=None)
-```
-
-Used for evaluating an arbitrary Theano expression on the output of the previous layer.
-
-- __Input shape__: Arbitrary. Use the keyword argument input_shape (tuple of integers, does not include the samples axis) when using this layer as the first layer in a model.
-
-- __Output shape__: Specified by the `output_shape` argument.
-
-- __Arguments__:
-
-    - __function__: The expression to be evaluated. Takes one argument: the output of the previous layer.
-    - __output_shape__: Shape of the tensor returned by `function`. Should be a shape tuple (not including the samples dimension) or a function of the full input shape tuple (including samples dimension).
-
-- __Example__:
-
-```python
-# custom softmax function
-def sharp_softmax(X, beta=1.5):
-    return theano.tensor.nnet.softmax(X * beta)
-
-def output_shape(input_shape):
-    # here input_shape includes the samples dimension
-    return input_shape  # shape is unchanged
-
-model = Sequential()
-model.add(Dense(input_dim=10, output_dim=10))
-model.add(Lambda(sharp_softmax, output_shape))
-model.add(Dense(1))
-model.add(Activation('sigmoid'))
-```
-
-
-## LambdaMerge
-```python
-keras.layers.core.LambdaMerge(layers, function, output_shape=None)
-```
-
-Merge the output of a list of layers (or containers) into a single tensor, using an arbitrary Theano expression.
-
-- __Arguments__:
-    - __layers__: List of layers or [containers](/layers/containers/).
-    - __function__: The expression to be evaluated. Takes one argument: the list of input tensors.
-    - __output_shape__: Shape of the tensor returned by `function`. Should be a shape tuple (not including samples dimension) or a function of the list of input shape tuples (including samples dimension).
-
-- __Example__:
-
-```python
-# root mean square function
-def rms(inputs):
-    # inputs is a list of tensors
-    s = inputs[0] ** 2
-    for i in range(1, len(inputs)):
-        s += inputs[i] ** 2
-    s /= len(inputs)
-    s = theano.tensor.sqrt(s)
-    # return a single tensor
-    return s
-
-def output_shape(input_shapes):
-    # return the shape of the first tensor
-    return input_shapes[0]
-
-left = Sequential()
-left.add(Dense(input_dim=10, output_dim=10))
-left.add(Activation('sigmoid'))
-
-right = Sequential()
-right.add(Dense(input_dim=10, output_dim=10))
-right.add(Activation('sigmoid'))
-
-model = Sequential()
-model.add(LambdaMerge([left, right], rms, output_shape))
-
-model.add(Dense(1))
-model.add(Activation('sigmoid'))
-```
-
----

docs/sources/layers/embeddings.md

@@ -1,31 +0,0 @@
-
-## Embedding
-
-```python
-keras.layers.embeddings.Embedding(input_dim, output_dim,
-                                  init='uniform',
-                                  weights=None,
-                                  W_regularizer=None, W_constraint=None,
-                                  mask_zero=False,
-                                  input_length=None)
-```
-
-Turn positive integers (indexes) into denses vectors of fixed size,
-eg. `[[4], [20]] -> [[0.25, 0.1], [0.6, -0.2]]`
-
-- __Input shape__: 2D tensor with shape: `(nb_samples, sequence_length)`.
-
-- __Output shape__: 3D tensor with shape: `(nb_samples, sequence_length, output_dim)`.
-
-- __Arguments__:
-
-    - __input_dim__: int >= 0. Size of the vocabulary, ie. 1+maximum integer index occurring in the input data.
-    - __output_dim__: int >= 0. Dimension of the dense embedding.
-    - __init__: name of initialization function for the weights of the layer (see: [initializations](../initializations.md)), or alternatively, Theano function to use for weights initialization. This parameter is only relevant if you don't pass a `weights` argument.
-    - __weights__: list of numpy arrays to set as initial weights. The list should have 1 element, of shape `(input_dim, output_dim)`.
-    - __W_regularizer__: instance of the [regularizers](../regularizers.md) module (eg. L1 or L2 regularization), applied to the embedding matrix.
-    - __W_constraint__: instance of the [constraints](../constraints.md) module (eg. maxnorm, nonneg), applied to the embedding matrix.
-	- __mask_zero__: Whether or not the input value 0 is a special "padding" value that should be masked out. This is useful for [recurrent layers](recurrent.md) which may take variable length input. If this is `True` then all subsequent layers in the model need to support masking or an exception will be raised.
-    - __input_length__: Length of input sequences, when it is constant. This argument is required if you are going to connect `Flatten` then `Dense` layers upstream (without it, the shape of the dense outputs cannot be computed).
-
----

docs/sources/layers/noise.md

@@ -1,37 +0,0 @@
-
-
-## GaussianNoise
-```python
-keras.layers.noise.GaussianNoise(sigma)
-```
-Apply to the input an additive zero-centred gaussian noise with standard deviation `sigma`. This is useful to mitigate overfitting (you could see it as a kind of random data augmentation). Gaussian Noise (GS) is a natural choice as corruption process for real valued inputs.
-
-Only active at training time.
-
-- __Input shape__: Arbitrary. Use the keyword argument `input_shape` (tuple of integers, does not include the samples axis) when using this layer as the first layer in a model.
-
-- __Output shape__: Same as input.
-
-- __Arguments__:
-
-    - __sigma__: float, standard deviation of the noise distribution.
-
----
-
-## GaussianDropout
-```python
-keras.layers.noise.GaussianDropout(p)
-```
-Apply to the input an multiplicative one-centred gaussian noise with standard deviation `sqrt(p/(1-p))`. p refers to drop probability to match Dropout layer syntax. 
-
-Only active at training time.
-
-- __Input shape__: Arbitrary. Use the keyword argument `input_shape` (tuple of integers, does not include the samples axis) when using this layer as the first layer in a model.
-
-- __Output shape__: Same as input.
-
-- __Arguments__:
-
-    - __p__: float, drop probability as with Dropout.
-
-

docs/sources/layers/normalization.md

@@ -1,19 +0,0 @@
-
-## BatchNormalization
-
-```python
-keras.layers.normalization.BatchNormalization(epsilon=1e-6, weights=None)
-```
-
-Normalize the activations of the previous layer at each batch.
-
-- __Input shape__: Arbitrary. Use the keyword argument `input_shape` (tuple of integers, does not include the samples axis) when using this layer as the first layer in a model.
-
-- __Output shape__: Same as input.
-
-- __Arguments__: 
-    - __epsilon__: small float > 0. Fuzz parameter.
-    - __weights__: Initialization weights. List of 2 numpy arrays, with shapes: `[(input_shape,), (input_shape,)]`
-
-- __References__:
-    - [Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift](http://arxiv.org/pdf/1502.03167v3.pdf)

docs/sources/layers/recurrent.md

@@ -1,130 +0,0 @@
-
-## SimpleRNN
-
-```python
-keras.layers.recurrent.SimpleRNN(output_dim,
-        init='glorot_uniform', inner_init='orthogonal',
-        activation='sigmoid',
-        weights=None,
-        return_sequences=False,
-        go_backwards=False,
-        stateful=False,
-        input_dim=None, input_length=None)
-```
-Fully connected RNN where output is to fed back to input. 
-
-- __Input shape__: 3D tensor with shape: `(nb_samples, timesteps, input_dim)`.
-
-- __Output shape__: 
-    - if `return_sequences`: 3D tensor with shape: `(nb_samples, timesteps, output_dim)`.
-    - else: 2D tensor with shape: `(nb_samples, output_dim)`.
-
-- __Masking__: This layer supports masking for input data with a variable number of timesteps To introduce masks to your data, use an [Embedding](embeddings.md) layer with the `mask_zero` parameter set to `True`. **Note:** for the time being, masking in only supported with Theano.
-
-- __Notes__: When using the TensorFlow backend, the number of timesteps used must be fixed. Make sure to pass an `input_length` int argument or a complete `input_shape` tuple argument.
-
-
-- __Arguments__:
-    - __output_dim__: dimension of the internal projections and the final output.
-    - __init__: weight initialization function. Can be the name of an existing function (str), or a Theano function (see: [initializations](../initializations.md)).
-    - __activation__: activation function. Can be the name of an existing function (str), or a Theano function (see: [activations](../activations.md)).
-    - __weights__: list of numpy arrays to set as initial weights. The list should have 3 elements, of shapes: `[(input_dim, output_dim), (output_dim, output_dim), (output_dim,)]`.
-    - __return_sequences__: Boolean. Whether to return the last output in the output sequence, or the full sequence.
-    - __go_backwards__: Boolean (default False). If True, rocess the input sequence backwards.
-    - __stateful__: Boolean (default False). If True, the last state for each sample at index i in a batch will be used as initial state for the sample of index i in the following batch.
-    - __input_dim__: dimensionality of the input (integer). This argument (or alternatively, the keyword argument `input_shape`) is required when using this layer as the first layer in a model.
-    - __input_length__: Length of input sequences, when it is constant. This argument is required if you are going to connect `Flatten` then `Dense` layers upstream (without it, the shape of the dense outputs cannot be computed).
-
-
----
-
-## GRU
-
-```python
-keras.layers.recurrent.GRU(output_dim,
-        init='glorot_uniform', inner_init='orthogonal',
-        activation='sigmoid', inner_activation='hard_sigmoid',
-        return_sequences=False,
-        go_backwards=False,
-        stateful=False,
-        input_dim=None, input_length=None)
-```
-
-Gated Recurrent Unit - Cho et al. 2014.
-
-- __Input shape__: 3D tensor with shape: `(nb_samples, timesteps, input_dim)`.
-
-- __Output shape__:
-    - if `return_sequences`: 3D tensor with shape: `(nb_samples, timesteps, output_dim)`.
-    - else: 2D tensor with shape: `(nb_samples, output_dim)`.
-
-- __Masking__: This layer supports masking for input data with a variable number of timesteps To introduce masks to your data, use an [Embedding](embeddings.md) layer with the `mask_zero` parameter set to true. **Note:** for the time being, masking in only supported with Theano.
-
-- __Notes__: When using the TensorFlow backend, the number of timesteps used must be fixed. Make sure to pass an `input_length` int argument or a complete `input_shape` tuple argument.
-
-- __Arguments__:
-    - __output_dim__: dimension of the internal projections and the final output.
-    - __init__: weight initialization function for the output cell. Can be the name of an existing function (str), or a Theano function (see: [initializations](../initializations.md)).
-    - __inner_init__: weight initialization function for the inner cells.
-    - __activation__: activation function for the output. Can be the name of an existing function (str), or a Theano function (see: [activations](../activations.md)).
-    - __inner_activation__: activation function for the inner cells.
-    - __weights__: list of numpy arrays to set as initial weights. The list should have 9 elements.
-    - __return_sequences__: Boolean. Whether to return the last output in the output sequence, or the full sequence.
-    - __go_backwards__: Boolean (default False). If True, rocess the input sequence backwards.
-    - __stateful__: Boolean (default False). If True, the last state for each sample at index i in a batch will be used as initial state for the sample of index i in the following batch.
-    - __input_dim__: dimensionality of the input (integer). This argument (or alternatively, the keyword argument `input_shape`) is required when using this layer as the first layer in a model.
-    - __input_length__: Length of input sequences, when it is constant. This argument is required if you are going to connect `Flatten` then `Dense` layers upstream (without it, the shape of the dense outputs cannot be computed).
-
-
-- __References__: 
-    - [On the Properties of Neural Machine Translation: Encoder–Decoder Approaches](http://www.aclweb.org/anthology/W14-4012)
-    - [Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling](http://arxiv.org/pdf/1412.3555v1.pdf)
-
----
-
-## LSTM
-
-```python
-keras.layers.recurrent.LSTM(output_dim,
-        init='glorot_uniform', inner_init='orthogonal', forget_bias_init='one',
-        activation='tanh', inner_activation='hard_sigmoid',
-        weights=None,
-        return_sequences=False,
-        go_backwards=False,
-        stateful=False,
-        input_dim=None, input_length=None)
-```
-
-Long-Short Term Memory unit - Hochreiter 1997.
-
-- __Input shape__: 3D tensor with shape: `(nb_samples, timesteps, input_dim)`.
-
-- __Output shape__:
-    - if `return_sequences`: 3D tensor with shape: `(nb_samples, timesteps, output_dim)`.
-    - else: 2D tensor with shape: `(nb_samples, output_dim)`.
-
-- __Masking__: This layer supports masking for input data with a variable number of timesteps To introduce masks to your data, use an [Embedding](embeddings.md) layer with the `mask_zero` parameter set to true. **Note:** for the time being, masking in only supported with Theano.
-
-- __Notes__: When using the TensorFlow backend, the number of timesteps used must be fixed. Make sure to pass an `input_length` int argument or a complete `input_shape` tuple argument.
-
-- __Arguments__:
-    - __output_dim__: dimension of the internal projections and the final output.
-    - __init__: weight initialization function for the output cell. Can be the name of an existing function (str), or a Theano function (see: [initializations](../initializations.md)).
-    - __inner_init__: weight initialization function for the inner cells.
-    - __forget_bias_init__: initialization function for the bias of the forget gate. [Jozefowicz et al.](http://www.jmlr.org/proceedings/papers/v37/jozefowicz15.pdf) recommend initializing with ones.
-    - __activation__: activation function for the output. Can be the name of an existing function (str), or a Theano function (see: [activations](../activations.md)).
-    - __inner_activation__: activation function for the inner cells.
-    - __weights__: list of numpy arrays to set as initial weights. The list should have 12 elements.
-    - __return_sequences__: Boolean. Whether to return the last output in the output sequence, or the full sequence.
-    - __go_backwards__: Boolean (default False). If True, rocess the input sequence backwards.
-    - __stateful__: Boolean (default False). If True, the last state for each sample at index i in a batch will be used as initial state for the sample of index i in the following batch.
-    - __input_dim__: dimensionality of the input (integer). This argument (or alternatively, the keyword argument `input_shape`) is required when using this layer as the first layer in a model.
-    - __input_length__: Length of input sequences, when it is constant. This argument is required if you are going to connect `Flatten` then `Dense` layers upstream (without it, the shape of the dense outputs cannot be computed).
-
-
-- __References__: 
-    - [Long short-term memory](http://deeplearning.cs.cmu.edu/pdfs/Hochreiter97_lstm.pdf) (original 1997 paper)
-    - [Learning to forget: Continual prediction with LSTM](http://www.mitpressjournals.org/doi/pdf/10.1162/089976600300015015)
-    - [Supervised sequence labelling with recurrent neural networks](http://www.cs.toronto.edu/~graves/preprint.pdf)
-
----

docs/sources/models.md

@@ -1,211 +0,0 @@
-## Sequential
-
-Linear stack of layers.
-
-```python
-model = keras.models.Sequential()
-```
-- __Methods__:
-    - __add__(layer): Add a layer to the model.
-    - __compile__(optimizer, loss, class_mode="categorical"):
-        - __Arguments__:
-            - __optimizer__: str (name of optimizer) or optimizer object. See [optimizers](optimizers.md).
-            - __loss__: str (name of objective function) or objective function. See [objectives](objectives.md).
-            - __class_mode__: one of "categorical", "binary". This is only used for computing classification accuracy or using the predict_classes method.
-    - __fit__(X, y, batch_size=128, nb_epoch=100, verbose=1, validation_split=0., validation_data=None, shuffle=True, show_accuracy=False, callbacks=[], class_weight=None, sample_weight=None): Train a model for a fixed number of epochs.
-        - __Return__: a history object. It `history` attribute is a record of training loss values at successive epochs, as well as validation loss values (if applicable).
-        - __Arguments__:
-            - __X__: data.
-            - __y__: labels.
-            - __batch_size__: int. Number of samples per gradient update.
-            - __nb_epoch__: int.
-            - __verbose__: 0 for no logging to stdout, 1 for progress bar logging, 2 for one log line per epoch.
-            - __callbacks__: `keras.callbacks.Callback` list. List of callbacks to apply during training. See [callbacks](callbacks.md).
-            - __validation_split__: float (0. < x < 1). Fraction of the data to use as held-out validation data.
-            - __validation_data__: tuple (X, y) to be used as held-out validation data. Will override validation_split.
-            - __shuffle__: boolean or str (for 'batch'). Whether to shuffle the samples at each epoch. 'batch' is a special option for dealing with the limitations of HDF5 data; it shuffles in batch-sized chunks.
-            - __show_accuracy__: boolean. Whether to display class accuracy in the logs to stdout at each epoch.
-            - __class_weight__: dictionary mapping classes to a weight value, used for scaling the loss function (during training only).
-            - __sample_weight__: list or numpy array with 1:1 mapping to the training samples, used for scaling the loss function (during training only). For time-distributed data, there is one weight per sample *per timestep*, i.e. if your output data is shaped `(nb_samples, timesteps, output_dim)`, your mask should be of shape `(nb_samples, timesteps, 1)`. This allows you to mask out or reweight individual output timesteps, which is useful in sequence to sequence learning.
-    - __evaluate__(X, y, batch_size=128, show_accuracy=False, verbose=1, sample_weight=None): Show performance of the model over some validation data.
-        - __Return__: The loss score over the data, or a `(loss, accuracy)` tuple if `show_accuracy=True`.
-        - __Arguments__: Same meaning as fit method above. verbose is used as a binary flag (progress bar or nothing).
-    - __predict__(X, batch_size=128, verbose=1):
-        - __Return__: An array of predictions for some test data.
-        - __Arguments__: Same meaning as fit method above.
-    - __predict_classes__(X, batch_size=128, verbose=1): Return an array of class predictions for some test data.
-        - __Return__: An array of labels for some test data.
-        - __Arguments__: Same meaning as fit method above. verbose is used as a binary flag (progress bar or nothing).
-    - __train_on_batch__(X, y, accuracy=False, class_weight=None, sample_weight=None): Single gradient update on one batch.
-        - __Return__: loss over the data, or tuple `(loss, accuracy)` if `accuracy=True`.
-    - __test_on_batch__(X, y, accuracy=False, sample_weight=None): Single performance evaluation on one batch.
-        - __Return__: loss over the data, or tuple `(loss, accuracy)` if `accuracy=True`.
-    - __save_weights__(fname, overwrite=False): Store the weights of all layers to a HDF5 file. If overwrite==False and the file already exists, an exception will be thrown.
-    - __load_weights__(fname): Sets the weights of a model, based to weights stored by __save_weights__. You can only __load_weights__ on a savefile from a model with an identical architecture. __load_weights__ can be called either before or after the __compile__ step.
-
-__Examples__:
-
-```python
-from keras.models import Sequential
-from keras.layers.core import Dense, Dropout, Activation
-from keras.optimizers import SGD
-
-model = Sequential()
-model.add(Dense(2, init='uniform', input_dim=64))
-model.add(Activation('softmax'))
-
-model.compile(loss='mse', optimizer='sgd')
-
-'''
-Demonstration of verbose modes 1 and 2
-'''
-model.fit(X_train, y_train, nb_epoch=3, batch_size=16, verbose=1)
-# outputs
-'''
-Train on 37800 samples, validate on 4200 samples
-Epoch 0
-37800/37800 [==============================] - 7s - loss: 0.0385
-Epoch 1
-37800/37800 [==============================] - 8s - loss: 0.0140
-Epoch 2
-10960/37800 [=======>......................] - ETA: 4s - loss: 0.0109
-'''
-
-model.fit(X_train, y_train, nb_epoch=3, batch_size=16, verbose=2)
-# outputs
-'''
-Train on 37800 samples, validate on 4200 samples
-Epoch 0
-loss: 0.0190
-Epoch 1
-loss: 0.0146
-Epoch 2
-loss: 0.0049
-'''
-
-'''
-Demonstration of show_accuracy
-'''
-model.fit(X_train, y_train, nb_epoch=3, batch_size=16, verbose=2, show_accuracy=True)
-# outputs
-'''
-Train on 37800 samples, validate on 4200 samples
-Epoch 0
-loss: 0.0190 - acc.: 0.8750
-Epoch 1
-loss: 0.0146 - acc.: 0.8750
-Epoch 2
-loss: 0.0049 - acc.: 1.0000
-'''
-
-'''
-Demonstration of validation_split
-'''
-model.fit(X_train, y_train, nb_epoch=3, batch_size=16, validation_split=0.1, show_accuracy=True, verbose=1)
-# outputs
-'''
-Train on 37800 samples, validate on 4200 samples
-Epoch 0
-37800/37800 [==============================] - 7s - loss: 0.0385 - acc.: 0.7258 - val. loss: 0.0160 - val. acc.: 0.9136
-Epoch 1
-37800/37800 [==============================] - 8s - loss: 0.0140 - acc.: 0.9265 - val. loss: 0.0109 - val. acc.: 0.9383
-Epoch 2
-10960/37800 [=======>......................] - ETA: 4s - loss: 0.0109 - acc.: 0.9420
-'''
-```
-
----
-
-## Graph
-
-Arbitrary connection graph. It can have any number of inputs and outputs, with each output trained with its own loss function. The quantity being optimized by a Graph model is the sum of all loss functions over the different outputs.
-
-```python
-model = keras.models.Graph()
-```
-- __Methods__:
-    - __add_input__(name, input_shape, dtype='float'): Add an input with shape dimensionality `ndim`. 
-        - __Arguments__:
-            - __input_shape__: Integer tuple, shape of the expected input (not including the samples axis). E.g. (10,) for 10-dimensional vectors, (None, 128) for sequences (of variable length) of 128-dimensional vectors, (3, 32, 32) for 32x32 images with RGB channels.
-            - __dtype__: `float` or `int`. Type of the expected input data.
-    - __add_output__(name, input=None, inputs=[], merge_mode='concat'): Add an output connect to `input` or `inputs`.
-        - __Arguments__:
-            - __name__: str. unique identifier of the output.
-            - __input__: str name of the node that the output is connected to. Only specify *one* of either `input` or `inputs`.
-            - __inputs__: list of str names of the node that the output is connected to.
-            - __merge_mode__: "sum" or "concat". Only applicable if `inputs` list is specified. Merge mode for the different inputs.
-    - __add_node__(layer, name, input=None, inputs=[], merge_mode='concat'): Add an output connect to `input` or `inputs`.
-        - __Arguments__:
-            - __layer__: Layer instance.
-            - __name__: str. unique identifier of the node.
-            - __input__: str name of the node/input that the node is connected to. Only specify *one* of either `input` or `inputs`.
-            - __inputs__: list of str names of the node that the node is connected to.
-            - __merge_mode__: "sum" or "concat". Only applicable if `inputs` list is specified. Merge mode for the different inputs.
-    - __add_shared_node__(layer, name, inputs=[], merge_mode=None, outputs=[]): Add a shared node connected to `inputs`. A shared node is a layer that will be applied separately to every incoming input, and that uses only one set of weights. The merging operation occurs on the outputs of the layer. 
-        - __Arguments__:
-            - __layer__: Layer instance.
-            - __name__: str. unique identifier of the node.
-            - __inputs__: list of str names of the node that the node is connected to.
-            - __merge_mode__: Merge mode for the different inputs.
-            - __outputs__: Optional. List of names for outputs, when merge_mode = None.
-    - __compile__(optimizer, loss):
-        - __Arguments__:
-            - __optimizer__: str (name of optimizer) or optimizer object. See [optimizers](optimizers.md).
-            - __loss__: dictionary mapping the name(s) of the output(s) to a loss function (string name of objective function or objective function. See [objectives](objectives.md)).
-    - __fit__(data, batch_size=128, nb_epoch=100, verbose=1, validation_split=0., validation_data=None, shuffle=True, callbacks=[]): Train a model for a fixed number of epochs.
-        - __Return__: a history object. It `history` attribute is a record of training loss values at successive epochs, as well as validation loss values (if applicable).
-        - __Arguments__:
-            - __data__:dictionary mapping input names out outputs names to appropriate numpy arrays. All arrays should contain the same number of samples.
-            - __batch_size__: int. Number of samples per gradient update.
-            - __nb_epoch__: int.
-            - __verbose__: 0 for no logging to stdout, 1 for progress bar logging, 2 for one log line per epoch.
-            - __callbacks__: `keras.callbacks.Callback` list. List of callbacks to apply during training. See [callbacks](callbacks.md).
-            - __validation_split__: float (0. < x < 1). Fraction of the data to use as held-out validation data.
-            - __validation_data__: tuple (X, y) to be used as held-out validation data. Will override validation_split.
-            - __shuffle__: boolean. Whether to shuffle the samples at each epoch.
-    - __evaluate__(data, batch_size=128, verbose=1): Show performance of the model over some validation data.
-        - __Return__: The loss score over the data.
-        - __Arguments__: Same meaning as fit method above. verbose is used as a binary flag (progress bar or nothing).
-    - __predict__(data, batch_size=128, verbose=1):
-        - __Return__: A dictionary mapping output names to arrays of predictions over the data.
-        - __Arguments__: Same meaning as fit method above. Only inputs need to be specified in `data`.
-    - __train_on_batch__(data): Single gradient update on one batch.
-        - __Return__: loss over the data.
-    - __test_on_batch__(data): Single performance evaluation on one batch.
-        - __Return__: loss over the data.
-    - __save_weights__(fname, overwrite=False): Store the weights of all layers to a HDF5 file. If `overwrite==False` and the file already exists, an exception will be thrown.
-    - __load_weights__(fname): Sets the weights of a model, based to weights stored by __save_weights__. You can only __load_weights__ on a savefile from a model with an identical architecture. __load_weights__ can be called either before or after the __compile__ step.
-
-
-__Examples__:
-
-```python
-# graph model with one input and two outputs
-graph = Graph()
-graph.add_input(name='input', input_shape=(32,))
-graph.add_node(Dense(16), name='dense1', input='input')
-graph.add_node(Dense(4), name='dense2', input='input')
-graph.add_node(Dense(4), name='dense3', input='dense1')
-graph.add_output(name='output1', input='dense2')
-graph.add_output(name='output2', input='dense3')
-
-graph.compile('rmsprop', {'output1':'mse', 'output2':'mse'})
-history = graph.fit({'input':X_train, 'output1':y_train, 'output2':y2_train}, nb_epoch=10)
-
-```
-
-```python
-# graph model with two inputs and one output
-graph = Graph()
-graph.add_input(name='input1', input_shape=(32,))
-graph.add_input(name='input2', input_shape=(32,))
-graph.add_node(Dense(16), name='dense1', input='input1')
-graph.add_node(Dense(4), name='dense2', input='input2')
-graph.add_node(Dense(4), name='dense3', input='dense1')
-graph.add_output(name='output', inputs=['dense2', 'dense3'], merge_mode='sum')
-graph.compile('rmsprop', {'output':'mse'})
-
-history = graph.fit({'input1':X_train, 'input2':X2_train, 'output':y_train}, nb_epoch=10)
-predictions = graph.predict({'input1':X_test, 'input2':X2_test}) # {'output':...}
-
-```

docs/sources/optimizers.md

@@ -1,117 +0,0 @@
-
-## Usage of optimizers
-
-An optimizer is one of the two arguments required for compiling a Keras model:
-
-```python
-model = Sequential()
-model.add(Dense(64, init='uniform', input_dim=10))
-model.add(Activation('tanh'))
-model.add(Activation('softmax'))
-
-sgd = SGD(lr=0.1, decay=1e-6, momentum=0.9, nesterov=True)
-model.compile(loss='mean_squared_error', optimizer=sgd)
-```
-
-You can either instantiate an optimizer before passing it to `model.compile()` , as in the above example, or you can call it by its name. In the latter case, the default parameters for the optimizer will be used.
-
-```python
-# pass optimizer by name: default parameters will be used
-model.compile(loss='mean_squared_error', optimizer='sgd')
-```
-
----
-
-## Base class
-
-```python
-keras.optimizers.Optimizer(**kwargs)
-```
-
-All optimizers descended from this class support the following keyword argument:
-
-- __clipnorm__: float >= 0.
-
-Note: this is base class for building optimizers, not an actual optimizer that can be used for training models.
-
----
-
-##  SGD
-
-```python
-keras.optimizers.SGD(lr=0.01, momentum=0., decay=0., nesterov=False)
-``` 
-
-__Arguments__:
-
-- __lr__: float >= 0. Learning rate.
-- __momentum__: float >= 0. Parameter updates momentum.
-- __decay__: float >= 0. Learning rate decay over each update.
-- __nesterov__: boolean. Whether to apply Nesterov momentum.
-
----
-
-##  Adagrad
-
-```python
-keras.optimizers.Adagrad(lr=0.01, epsilon=1e-6)
-```
-
-It is recommended to leave the parameters of this optimizer at their default values.
-
-__Arguments__:
-
-- __lr__: float >= 0. Learning rate. 
-- __epsilon__: float >= 0. 
-
----
-
-##  Adadelta
-
-```python
-keras.optimizers.Adadelta(lr=1.0, rho=0.95, epsilon=1e-6)
-```
-
-It is recommended to leave the parameters of this optimizer at their default values.
-
-__Arguments__:
-
-- __lr__: float >= 0. Learning rate. It is recommended to leave it at the default value.
-- __rho__: float >= 0. 
-- __epsilon__: float >= 0. Fuzz factor.
-
-For more info, see *"Adadelta: an adaptive learning rate method"* by Matthew Zeiler.
-
----
-
-##  RMSprop 
-
-```python
-keras.optimizers.RMSprop(lr=0.001, rho=0.9, epsilon=1e-6)
-```
-
-It is recommended to leave the parameters of this optimizer at their default values.
-
-__Arguments__:
-
-- __lr__: float >= 0. Learning rate. 
-- __rho__: float >= 0.
-- __epsilon__: float >= 0. Fuzz factor.
-
----
-
-## Adam
-
-```python
-keras.optimizers.Adam(lr=0.001, beta_1=0.9, beta_2=0.999, epsilon=1e-8)
-```
-
-Adam optimizer, proposed by Kingma and Lei Ba in [Adam: A Method For Stochastic Optimization](http://arxiv.org/pdf/1412.6980v8.pdf). Default parameters are those suggested in the paper.
-
-__Arguments__:
-
-- __lr__: float >= 0. Learning rate.
-- __beta_1__, __beta_2__: floats, 0 < beta < 1. Generally close to 1.
-- __epsilon__: float >= 0. Fuzz factor.
-
----

docs/sources/preprocessing/image.md

@@ -1,70 +0,0 @@
-
-## ImageDataGenerator
-
-```python
-keras.preprocessing.image.ImageDataGenerator(featurewise_center=True,
-    samplewise_center=False,
-    featurewise_std_normalization=True,
-    samplewise_std_normalization=False,
-    zca_whitening=False,
-    rotation_range=0.,
-    width_shift_range=0.,
-    height_shift_range=0.,
-    horizontal_flip=False,
-    vertical_flip=False)
-```
-
-Generate batches of tensor image data with real-time data augmentation.
-
-- __Arguments__:
-    - __featurewise_center__: Boolean. Set input mean to 0 over the dataset.
-    - __samplewise_center__: Boolean. Set each sample mean to 0.
-    - __featurewise_std_normalization__: Boolean. Divide inputs by std of the dataset.
-    - __samplewise_std_normalization__: Boolean. Divide each input by its std.
-    - __zca_whitening__: Boolean. Apply ZCA whitening.
-    - __rotation_range__: Int. Degree range for random rotations.
-    - __width_shift_range__: Float (fraction of total width). Range for random horizontal shifts.
-    - __height_shift_range__: Float (fraction of total height). Range for random vertical shifts.
-    - __horizontal_flip__: Boolean. Randomly flip inputs horizontally.
-    - __vertical_flip__: Boolean. Randomly flip inputs vertically.
-
-- __Methods__:
-    - __fit(X)__: Required if featurewise_center or featurewise_std_normalization or zca_whitening. Compute necessary quantities on some sample data.
-        - __Arguments__: 
-            - __X__: sample data.
-            - __augment__: Boolean (default: False). Whether to fit on randomly augmented samples.
-            - __rounds__: int (default: 1). If augment, how many augmentation passes over the data to use.
-    - __flow(X, y)__:
-        - __Arguments__: 
-            - __X__: data.
-            - __y__: labels.
-            - __batch_size__: int (default: 32).
-            - __shuffle__: boolean (defaut: False).
-            - __save_to_dir__: None or str. This allows you to optimally specify a directory to which to save the augmented pictures being generated (useful for visualizing what you are doing). 
-            - __save_prefix__: str. Prefix to use for filenames of saved pictures.
-            - __save_format__: one of "png", jpeg". 
-
-- __Example__:
-```python
-(X_train, y_train), (X_test, y_test) = cifar10.load_data(test_split=0.1)
-Y_train = np_utils.to_categorical(y_train, nb_classes)
-Y_test = np_utils.to_categorical(y_test, nb_classes)
-
-datagen = ImageDataGenerator(
-    featurewise_center=True,
-    featurewise_std_normalization=True,
-    rotation_range=20,
-    width_shift_range=0.2,
-    height_shift_range=0.2,
-    horizontal_flip=True)
-
-# compute quantities required for featurewise normalization 
-# (std, mean, and principal components if ZCA whitening is applied)
-datagen.fit(X_train)
-
-for e in range(nb_epoch):
-    print 'Epoch', e
-    # batch train with realtime data augmentation
-    for X_batch, Y_batch in datagen.flow(X_train, Y_train):
-        loss = model.train(X_batch, Y_batch)
-```

docs/templates/activations.md

@@ -14,11 +14,13 @@ is equivalent to:
 model.add(Dense(64, activation='tanh'))
 ```
 
-You can also pass an element-wise Theano function as an activation:
+You can also pass an element-wise Theano/TensorFlow function as an activation:
 
 ```python
+from keras import backend as K
+
 def tanh(x):
-    return theano.tensor.tanh(x)
+    return K.tanh(x)
 
 model.add(Dense(64, activation=tanh))
 model.add(Activation(tanh))
@@ -28,6 +30,7 @@ model.add(Activation(tanh))
 
 - __softmax__: Softmax applied across inputs last dimension. Expects shape either `(nb_samples, nb_timesteps, nb_dims)` or `(nb_samples, nb_dims)`.
 - __softplus__
+- __softsign__
 - __relu__
 - __tanh__
 - __sigmoid__
@@ -36,4 +39,4 @@ model.add(Activation(tanh))
 
 ## On Advanced Activations
 
-Activations that are more complex than a simple Theano function (eg. learnable activations, configurable activations, etc.) are available as [Advanced Activation layers](layers/advanced_activations.md), and can be found in the module `keras.layers.advanced_activations`. These include PReLU and LeakyReLU.
+Activations that are more complex than a simple Theano/TensorFlow function (eg. learnable activations, configurable activations, etc.) are available as [Advanced Activation layers](layers/advanced-activations.md), and can be found in the module `keras.layers.advanced_activations`. These include PReLU and LeakyReLU.

docs/templates/applications.md

@@ -0,0 +1,260 @@
+# Applications
+
+Keras Applications are deep learning models that are made available alongside pre-trained weights.
+These models can be used for prediction, feature extraction, and fine-tuning.
+
+Weights are downloaded automatically when instantiating a model. They are stored at `~/.keras/models/`.
+
+## Available models
+
+Models for image classification with weights trained on ImageNet:
+
+- [VGG16](#vgg16)
+- [VGG19](#vgg19)
+- [ResNet50](#resnet50)
+- [InceptionV3](#inceptionv3)
+
+All of these architectures are compatible with both TensorFlow and Theano, and upon instantiation the models will be built according to the image dimension ordering set in your Keras configuration file at `~/.keras/keras.json`. For instance, if you have set `image_dim_ordering=tf`, then any model loaded from this repository will get built according to the TensorFlow dimension ordering convention, "Width-Height-Depth".
+
+-----
+
+## Examples
+
+### Classify ImageNet classes with ResNet50
+
+```python
+from keras.applications.resnet50 import ResNet50
+from keras.preprocessing import image
+from keras.applications.resnet50 import preprocess_input, decode_predictions
+
+model = ResNet50(weights='imagenet')
+
+img_path = 'elephant.jpg'
+img = image.load_img(img_path, target_size=(224, 224))
+x = image.img_to_array(img)
+x = np.expand_dims(x, axis=0)
+x = preprocess_input(x)
+
+preds = model.predict(x)
+print('Predicted:', decode_predictions(preds))
+# print: [[u'n02504458', u'African_elephant']]
+```
+
+### Extract features with VGG16
+
+```python
+from keras.applications.vgg16 import VGG16
+from keras.preprocessing import image
+from keras.applications.vgg16 import preprocess_input
+
+model = VGG16(weights='imagenet', include_top=False)
+
+img_path = 'elephant.jpg'
+img = image.load_img(img_path, target_size=(224, 224))
+x = image.img_to_array(img)
+x = np.expand_dims(x, axis=0)
+x = preprocess_input(x)
+
+features = model.predict(x)
+```
+
+### Extract features from an arbitrary intermediate layer with VGG19
+
+```python
+from keras.applications.vgg19 import VGG19
+from keras.preprocessing import image
+from keras.applications.vgg19 import preprocess_input
+from keras.models import Model
+
+base_model = VGG19(weights='imagenet')
+model = Model(input=base_model.input, output=base_model.get_layer('block4_pool').output)
+
+img_path = 'elephant.jpg'
+img = image.load_img(img_path, target_size=(224, 224))
+x = image.img_to_array(img)
+x = np.expand_dims(x, axis=0)
+x = preprocess_input(x)
+
+block4_pool_features = model.predict(x)
+```
+
+### Fine-tune InceptionV3 on a new set of classes
+
+```python
+from keras.applications.inception_v3 import InceptionV3
+from keras.preprocessing import image
+from keras.models import Model
+from keras.layers import Dense, GlobalAveragePooling2D
+from keras import backend as K
+
+# create the base pre-trained model
+base_model = InceptionV3(weights='imagenet', include_top=False)
+
+# add a global spatial average pooling layer
+x = base_model.output
+x = GlobalAveragePooling2D()(x)
+# let's add a fully-connected layer
+x = Dense(1024, activation='relu')(x)
+# and a logistic layer -- let's say we have 200 classes
+predictions = Dense(200, activation='softmax')(x)
+
+# this is the model we will train
+model = Model(input=base_model.input, output=predictions)
+
+# first: train only the top layers (which were randomly initialized)
+# i.e. freeze all convolutional InceptionV3 layers
+for layer in base_model.layers:
+    layer.trainable = False
+
+# compile the model (should be done *after* setting layers to non-trainable)
+model.compile(optimizer='rmsprop', loss='categorical_crossentropy')
+
+# train the model on the new data for a few epochs
+model.fit_generator(...)
+
+# at this point, the top layers are well trained and we can start fine-tuning
+# convolutional layers from inception V3. We will freeze the bottom N layers
+# and train the remaining top layers.
+
+# let's visualize layer names and layer indices to see how many layers
+# we should freeze:
+for i, layer in enumerate(base_model.layers):
+   print(i, layer.name)
+
+# we chose to train the top 2 inception blocks, i.e. we will freeze
+# the first 172 layers and unfreeze the rest:
+for layer in model.layers[:172]:
+   layer.trainable = False
+for layer in model.layers[172:]:
+   layer.trainable = True
+
+# we need to recompile the model for these modifications to take effect
+# we use SGD with a low learning rate
+from keras.optimizers import SGD
+model.compile(optimizer=SGD(lr=0.0001, momentum=0.9), loss='categorical_crossentropy')
+
+# we train our model again (this time fine-tuning the top 2 inception blocks
+# alongside the top Dense layers
+model.fit_generator(...)
+```
+
+
+### Build InceptionV3 over a custom input tensor
+
+```python
+from keras.applications.inception_v3 import InceptionV3
+from keras.layers import Input
+
+# this could also be the output a different Keras model or layer
+input_tensor = Input(shape=(224, 224, 3))  # this assumes K.image_dim_ordering() == 'tf'
+
+model = InceptionV3(input_tensor=input_tensor, weights='imagenet', include_top=True)
+```
+
+-----
+
+## VGG16
+
+```python
+keras.applications.vgg16.VGG16(include_top=True, weights='imagenet', input_tensor=None)
+```
+
+### Arguments
+
+- include_top: whether to include the 3 fully-connected layers at the top of the network.
+- weights: one of `None` (random initialization) or "imagenet" (pre-training on ImageNet).
+- input_tensor: optional Keras tensor (i.e. output of `layers.Input()`) to use as image input for the model.
+
+### Returns
+
+A Keras model instance.
+
+### References
+
+- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://arxiv.org/abs/1409.1556): please cite this paper if you use the VGG models in your work.
+
+### License
+
+These weights are ported from the ones [released by VGG at Oxford](http://www.robots.ox.ac.uk/~vgg/research/very_deep/) under the [Creative Commons Attribution License](https://creativecommons.org/licenses/by/4.0/).
+
+-----
+
+## VGG19
+
+
+```python
+keras.applications.vgg19.VGG19(include_top=True, weights='imagenet', input_tensor=None)
+```
+
+### Arguments
+
+- include_top: whether to include the 3 fully-connected layers at the top of the network.
+- weights: one of `None` (random initialization) or "imagenet" (pre-training on ImageNet).
+- input_tensor: optional Keras tensor (i.e. output of `layers.Input()`) to use as image input for the model.
+
+### Returns
+
+A Keras model instance.
+
+
+### References
+
+- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://arxiv.org/abs/1409.1556)
+
+### License
+
+These weights are ported from the ones [released by VGG at Oxford](http://www.robots.ox.ac.uk/~vgg/research/very_deep/) under the [Creative Commons Attribution License](https://creativecommons.org/licenses/by/4.0/).
+
+-----
+
+## ResNet50
+
+
+```python
+keras.applications.resnet50.ResNet50(include_top=True, weights='imagenet', input_tensor=None)
+```
+
+### Arguments
+
+- include_top: whether to include the 3 fully-connected layers at the top of the network.
+- weights: one of `None` (random initialization) or "imagenet" (pre-training on ImageNet).
+- input_tensor: optional Keras tensor (i.e. output of `layers.Input()`) to use as image input for the model.
+
+### Returns
+
+A Keras model instance.
+
+### References
+
+- [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385)
+
+### License
+
+These weights are ported from the ones [released by Kaiming He](https://github.com/KaimingHe/deep-residual-networks) under the [MIT license](https://github.com/KaimingHe/deep-residual-networks/blob/master/LICENSE).
+
+-----
+
+## InceptionV3
+
+
+```python
+keras.applications.inception_v3.InceptionV3(include_top=True, weights='imagenet', input_tensor=None)
+```
+
+### Arguments
+
+- include_top: whether to include the 3 fully-connected layers at the top of the network.
+- weights: one of `None` (random initialization) or "imagenet" (pre-training on ImageNet).
+- input_tensor: optional Keras tensor (i.e. output of `layers.Input()`) to use as image input for the model.
+
+### Returns
+
+A Keras model instance.
+
+### References
+
+- [Rethinking the Inception Architecture for Computer Vision](http://arxiv.org/abs/1512.00567)
+
+### License
+
+These weights are trained by ourselves and are released under the MIT license.

docs/templates/backend.md

@@ -4,10 +4,14 @@
 
 Keras is a model-level library, providing high-level building blocks for developing deep learning models. It does not handle itself low-level operations such as tensor products, convolutions and so on. Instead, it relies on a specialized, well-optimized tensor manipulation library to do so, serving as the "backend engine" of Keras. Rather than picking one single tensor library and making the implementation of Keras tied to that library, Keras handles the problem in a modular way, and several different backend engines can be plugged seamlessly into Keras.
 
-At this time, Keras has two backend implementations available: the **Theano** backend and the **TensorFlow** backend.
+At this time, Keras has two backend implementations available: the **TensorFlow** backend and the **Theano** backend.
 
-- [Theano](http://deeplearning.net/software/theano/) is an open-source symbolic tensor manipulation framework developed by LISA/MILA Lab at Université de Montréal.
 - [TensorFlow](http://www.tensorflow.org/) is an open-source symbolic tensor manipulation framework developed by Google, Inc.
+- [Theano](http://deeplearning.net/software/theano/) is an open-source symbolic tensor manipulation framework developed by LISA/MILA Lab at Université de Montréal.
+
+In the future, we are likely to add more backend options. If you are interested in developing a new backend, get in touch!
+
+----
 
 ## Switching from one backend to another
 
@@ -17,12 +21,29 @@ If you have run Keras at least once, you will find the Keras configuration file
 
 If it isn't there, you can create it.
 
-It probably looks like this:
+The default configuration file looks like this:
 
-`{"epsilon": 1e-07, "floatx": "float32", "backend": "theano"}`
+```
+{
+    "image_dim_ordering": "tf",
+    "epsilon": 1e-07,
+    "floatx": "float32",
+    "backend": "tensorflow"
+}
+```
 
 Simply change the field `backend` to either `"theano"` or `"tensorflow"`, and Keras will use the new configuration next time you run any Keras code.
 
+You can also define the environment variable ``KERAS_BACKEND`` and this will
+override what is defined in your config file :
+
+```bash
+KERAS_BACKEND=tensorflow python -c "from keras import backend"
+Using TensorFlow backend.
+```
+
+----
+
 ## Using the abstract Keras backend to write new code
 
 If you want the Keras modules you write to be compatible with both Theano and TensorFlow, you have to write them via the abstract Keras backend API. Here's an intro.
@@ -32,7 +53,7 @@ You can import the backend module via:
 from keras import backend as K
 ```
 
-This instantiates an input placeholder. It's equivalent to `tf.placeholder()` or `T.matrix()`, `T.tensor3()`, etc.
+The code below instantiates an input placeholder. It's equivalent to `tf.placeholder()` or `T.matrix()`, `T.tensor3()`, etc.
 
 ```python
 input = K.placeholder(shape=(2, 4, 5))
@@ -42,16 +63,16 @@ input = K.placeholder(shape=(None, 4, 5))
 input = K.placeholder(ndim=3)
 ```
 
-This instantiates a shared variable. It's equivalent to `tf.variable()` or `theano.shared()`.
+The code below instantiates a shared variable. It's equivalent to `tf.variable()` or `theano.shared()`.
 
 ```python
 val = np.random.random((3, 4, 5))
 var = K.variable(value=val)
 
 # all-zeros variable:
-var = K.ones(shape=(3, 4, 5))
-# all-ones:
 var = K.zeros(shape=(3, 4, 5))
+# all-ones:
+var = K.ones(shape=(3, 4, 5))
 ```
 
 Most tensor operations you will need can be done as you would in TensorFlow or Theano:
@@ -65,8 +86,12 @@ a = concatenate([b, c], axis=-1)
 # etc...
 ```
 
-For more information, see the code at `keras/backend/theano_backend.py` and `keras/backend/tensorflow_backend.py`.
-
+----
+
+## Backend functions
+
+
+{{autogenerated}}
 
 
 

docs/templates/callbacks.md

@@ -4,51 +4,12 @@ A callback is a set of functions to be applied at given stages of the training p
 
 ---
 
-## Base class
-
-```python
-keras.callbacks.Callback()
-```
-- __Properties__:
-    - __params__: dict. Training parameters (eg. verbosity, batch size, number of epochs...).
-    - __model__: `keras.models.Model`. Reference of the model being trained.
-- __Methods__:
-    - __on_train_begin__(logs={}): Method called at the beginning of training.
-    - __on_train_end__(logs={}): Method called at the end of training.
-    - __on_epoch_begin__(epoch, logs={}): Method called at the beginning of epoch `epoch`.
-    - __on_epoch_end__(epoch, logs={}): Method called at the end of epoch `epoch`.
-    - __on_batch_begin__(batch, logs={}): Method called at the beginning of batch `batch`.
-    - __on_batch_end__(batch, logs={}): Method called at the end of batch `batch`.
-
-The `logs` dictionary will contain keys for quantities relevant to the current batch or epoch. Currently, the `.fit()` method of the `Sequential` model class will include the following quantities in the `logs` that it passes to its callbacks:
-- __on_epoch_end__: logs optionally include `val_loss` (if validation is enabled in `fit`), and `val_accuracy` (if validation and accuracy monitoring are enabled).
-- __on_batch_begin__: logs include `size`, the number of samples in the current batch.
-- __on_batch_end__: logs include `loss`, and optionally `accuracy` (if accuracy monitoring is enabled).
-
----
-
-## Available callbacks
-
-```python
-keras.callbacks.ModelCheckpoint(filepath, verbose=0, save_best_only=False)
-```
-
-Save the model after every epoch. If `save_best_only=True`, the latest best model according to the validation loss will not be overwritten.
-`filepath` can contain named formatting options, which will be filled the value of `epoch` and keys in `logs` (passed in `on_epoch_end`).
-
-For example: if `filepath` is `weights.{epoch:02d}-{val_loss:.2f}.hdf5`, then multiple files will be save with the epoch number and the validation loss.
-
-
-```python
-keras.callbacks.EarlyStopping(monitor='val_loss', patience=0, verbose=0)
-```
-
-Stop training after no improvement of the metric `monitor` is seen for `patience` epochs.
+{{autogenerated}}
 
 ---
 
 
-## Create a callback
+# Create a callback
 
 You can create a custom callback by extending the base class `keras.callbacks.Callback`. A callback has access to its associated model through the class property `self.model`.
 

docs/templates/datasets.md

@@ -2,13 +2,13 @@
 
 ## CIFAR10 small image classification
 
-`keras.datasets.cifar10`
-
 Dataset of 50,000 32x32 color training images, labeled over 10 categories, and 10,000 test images.
 
 ### Usage:
 
 ```python
+from keras.datasets import cifar10
+
 (X_train, y_train), (X_test, y_test) = cifar10.load_data()
 ```
 
@@ -21,13 +21,13 @@ Dataset of 50,000 32x32 color training images, labeled over 10 categories, and 1
 
 ## CIFAR100 small image classification
 
-`keras.datasets.cifar100`
-
 Dataset of 50,000 32x32 color training images, labeled over 100 categories, and 10,000 test images.
 
 ### Usage:
 
 ```python
+from keras.datasets import cifar100
+
 (X_train, y_train), (X_test, y_test) = cifar100.load_data(label_mode='fine')
 ```
 
@@ -44,8 +44,6 @@ Dataset of 50,000 32x32 color training images, labeled over 100 categories, and
 
 ## IMDB Movie reviews sentiment classification
 
-`keras.datasets.imdb`
-
 Dataset of 25,000 movies reviews from IMDB, labeled by sentiment (positive/negative). Reviews have been preprocessed, and each review is encoded as a [sequence](preprocessing/sequence.md) of word indexes (integers). For convenience, words are indexed by overall frequency in the dataset, so that for instance the integer "3" encodes the 3rd most frequent word in the data. This allows for quick filtering operations such as: "only consider the top 10,000 most common words, but eliminate the top 20 most common words".
 
 As a convention, "0" does not stand for a specific word, but instead is used to encode any unknown word.
@@ -53,8 +51,16 @@ As a convention, "0" does not stand for a specific word, but instead is used to
 ### Usage:
 
 ```python
-(X_train, y_train), (X_test, y_test) = imdb.load_data(path="imdb.pkl", \
-nb_words=None, skip_top=0, maxlen=None, test_split=0.1, seed=113)
+from keras.datasets import imdb
+
+(X_train, y_train), (X_test, y_test) = imdb.load_data(path="imdb_full.pkl",
+                                                      nb_words=None,
+                                                      skip_top=0,
+                                                      maxlen=None,
+                                                      seed=113,
+                                                      start_char=1,
+                                                      oov_char=2,
+                                                      index_from=3)
 ```
 - __Return:__
     - 2 tuples:
@@ -67,25 +73,38 @@ nb_words=None, skip_top=0, maxlen=None, test_split=0.1, seed=113)
     - __nb_words__: integer or None. Top most frequent words to consider. Any less frequent word will appear as 0 in the sequence data.
     - __skip_top__: integer. Top most frequent words to ignore (they will appear as 0s in the sequence data).
     - __maxlen__: int. Maximum sequence length. Any longer sequence will be truncated.
-    - __test_split__: float. Fraction of the dataset to be used as test data.
     - __seed__: int. Seed for reproducible data shuffling.
+    - __start_char__: char. The start of a sequence will be marked with this character.
+        Set to 1 because 0 is usually the padding character.
+    - __oov_char__: char. words that were cut out because of the `nb_words`
+        or `skip_top` limit will be replaced with this character.
+    - __index_from__: int. Index actual words with this index and higher.
 
 ---
 
 ## Reuters newswire topics classification
 
-`keras.datasets.reuters`
-
 Dataset of 11,228 newswires from Reuters, labeled over 46 topics. As with the IMDB dataset, each wire is encoded as a sequence of word indexes (same conventions).
 
 ### Usage:
 
 ```python
-(X_train, y_train), (X_test, y_test) = reuters.load_data(path="reuters.pkl", \
-nb_words=None, skip_top=0, maxlen=None, test_split=0.1, seed=113)
+from keras.datasets import reuters
+
+(X_train, y_train), (X_test, y_test) = reuters.load_data(path="reuters.pkl",
+                                                         nb_words=None,
+                                                         skip_top=0,
+                                                         maxlen=None,
+                                                         test_split=0.2,
+                                                         seed=113,
+                                                         start_char=1,
+                                                         oov_char=2,
+                                                         index_from=3)
 ```
 
-The specifications are the same as that of the IMDB dataset.
+The specifications are the same as that of the IMDB dataset, with the addition of:
+
+    - __test_split__: float. Fraction of the dataset to be used as test data.
 
 This dataset also makes available the word index used for encoding the sequences:
 
@@ -101,13 +120,13 @@ word_index = reuters.get_word_index(path="reuters_word_index.pkl")
     
 ## MNIST database of handwritten digits
 
-`keras.datasets.mnist`
-
 Dataset of 60,000 28x28 grayscale images of the 10 digits, along with a test set of 10,000 images.
 
 ### Usage:
 
 ```python
+from keras.datasets import mnist
+
 (X_train, y_train), (X_test, y_test) = mnist.load_data()
 ```
 

docs/templates/getting-started/faq.md

@@ -0,0 +1,385 @@
+# Keras FAQ: Frequently Asked Keras Questions
+
+- [How should I cite Keras?](#how-should-i-cite-keras)
+- [How can I run Keras on GPU?](#how-can-i-run-keras-on-gpu)
+- [How can I save a Keras model?](#how-can-i-save-a-keras-model)
+- [Why is the training loss much higher than the testing loss?](#why-is-the-training-loss-much-higher-than-the-testing-loss)
+- [How can I visualize the output of an intermediate layer?](#how-can-i-visualize-the-output-of-an-intermediate-layer)
+- [How can I use Keras with datasets that don't fit in memory?](#how-can-i-use-keras-with-datasets-that-dont-fit-in-memory)
+- [How can I interrupt training when the validation loss isn't decreasing anymore?](#how-can-i-interrupt-training-when-the-validation-loss-isnt-decreasing-anymore)
+- [How is the validation split computed?](#how-is-the-validation-split-computed)
+- [Is the data shuffled during training?](#is-the-data-shuffled-during-training)
+- [How can I record the training / validation loss / accuracy at each epoch?](#how-can-i-record-the-training-validation-loss-accuracy-at-each-epoch)
+- [How can I "freeze" layers?](#how-can-i-freeze-keras-layers)
+- [How can I use stateful RNNs?](#how-can-i-use-stateful-rnns)
+- [How can I remove a layer from a Sequential model?](#how-can-i-remove-a-layer-from-a-sequential-model)
+- [How can I use pre-trained models in Keras?](#how-can-i-use-pre-trained-models-in-keras)
+
+---
+
+### How should I cite Keras?
+
+Please cite Keras in your publications if it helps your research. Here is an example BibTeX entry:
+
+```
+@misc{chollet2015keras,
+  title={Keras},
+  author={Chollet, Fran\c{c}ois},
+  year={2015},
+  publisher={GitHub},
+  howpublished={\url{https://github.com/fchollet/keras}},
+}
+```
+
+### How can I run Keras on GPU?
+
+If you are running on the TensorFlow backend, your code will automatically run on GPU if any available GPU is detected.
+If you are running on the Theano backend, you can use one of the following methods:
+
+Method 1: use Theano flags.
+```bash
+THEANO_FLAGS=device=gpu,floatX=float32 python my_keras_script.py
+```
+
+The name 'gpu' might have to be changed depending on your device's identifier (e.g. `gpu0`, `gpu1`, etc).
+
+Method 2: set up your `.theanorc`: [Instructions](http://deeplearning.net/software/theano/library/config.html)
+
+Method 3: manually set `theano.config.device`, `theano.config.floatX` at the beginning of your code:
+```python
+import theano
+theano.config.device = 'gpu'
+theano.config.floatX = 'float32'
+```
+
+---
+
+### How can I save a Keras model?
+
+*It is not recommended to use pickle or cPickle to save a Keras model.*
+
+You can use `model.save(filepath)` to save a Keras model into a single HDF5 file which will contain:
+
+- the architecture of the model, allowing to re-create the model
+- the weights of the model
+- the training configuration (loss, optimizer)
+- the state of the optimizer, allowing to resume training exactly where you left off.
+
+You can then use `keras.models.load_model(filepath)` to reinstantiate your model.
+`load_model` will also take care of compiling the model using the saved training configuration
+(unless the model was never compiled in the first place).
+
+Example:
+
+```python
+from keras.models import load_model
+
+model.save('my_model.h5')  # creates a HDF5 file 'my_model.h5'
+del model  # deletes the existing model
+
+# returns a compiled model
+# identical to the previous one
+model = load_model('my_model.h5')
+```
+
+If you only need to save the **architecture of a model**, and not its weights or its training configuration, you can do:
+
+```python
+# save as JSON
+json_string = model.to_json()
+
+# save as YAML
+yaml_string = model.to_yaml()
+```
+
+The generated JSON / YAML files are human-readable and can be manually edited if needed.
+
+You can then build a fresh model from this data:
+
+```python
+# model reconstruction from JSON:
+from keras.models import model_from_json
+model = model_from_json(json_string)
+
+# model reconstruction from YAML
+model = model_from_yaml(yaml_string)
+```
+
+If you need to save the **weights of a model**, you can do so in HDF5 with the code below.
+
+Note that you will first need to install HDF5 and the Python library h5py, which do not come bundled with Keras.
+
+```python
+model.save_weights('my_model_weights.h5')
+```
+
+Assuming you have code for instantiating your model, you can then load the weights you saved into a model with the *same* architecture:
+
+```python
+model.load_weights('my_model_weights.h5')
+```
+
+If you need to load weights into a *different* architecture (with some layers in common), for instance for fine-tuning or transfer-learning, you can load weights by *layer name*:
+
+```python
+model.load_weights('my_model_weights.h5', by_name=True)
+```
+
+For example:
+
+```python
+"""
+Assume original model looks like this:
+    model = Sequential()
+    model.add(Dense(2, input_dim=3, name="dense_1"))
+    model.add(Dense(3, name="dense_2"))
+    ...
+    model.save_weights(fname)
+"""
+
+# new model
+model = Sequential()
+model.add(Dense(2, input_dim=3, name="dense_1"))  # will be loaded
+model.add(Dense(10, name="new_dense"))  # will not be loaded
+
+# load weights from first model; will only affect the first layer, dense_1.
+model.load_weights(fname, by_name=True)
+```
+
+---
+
+### Why is the training loss much higher than the testing loss?
+
+A Keras model has two modes: training and testing. Regularization mechanisms, such as Dropout and L1/L2 weight regularization, are turned off at testing time.
+
+Besides, the training loss is the average of the losses over each batch of training data. Because your model is changing over time, the loss over the first batches of an epoch is generally higher than over the last batches. On the other hand, the testing loss for an epoch is computed using the model as it is at the end of the epoch, resulting in a lower loss.
+
+---
+
+### How can I visualize the output of an intermediate layer?
+
+You can build a Keras function that will return the output of a certain layer given a certain input, for example:
+
+```python
+from keras import backend as K
+
+# with a Sequential model
+get_3rd_layer_output = K.function([model.layers[0].input],
+                                  [model.layers[3].output])
+layer_output = get_3rd_layer_output([X])[0]
+```
+
+Similarly, you could build a Theano and TensorFlow function directly.
+
+Note that if your model has a different behavior in training and testing phase (e.g. if it uses `Dropout`, `BatchNormalization`, etc.), you will need
+to pass the learning phase flag to your function:
+
+```python
+get_3rd_layer_output = K.function([model.layers[0].input, K.learning_phase()],
+                                  [model.layers[3].output])
+
+# output in test mode = 0
+layer_output = get_3rd_layer_output([X, 0])[0]
+
+# output in train mode = 1
+layer_output = get_3rd_layer_output([X, 1])[0]
+```
+
+Another more flexible way of getting output from intermediate layers is to use the [functional API](/getting-started/functional-api-guide). For example, if you have created an autoencoder for MNIST:
+
+```python
+inputs = Input(shape=(784,))
+encoded = Dense(32, activation='relu')(inputs)
+decoded = Dense(784)(encoded)
+model = Model(input=inputs, output=decoded)
+```
+
+After compiling and training the model, you can get the output of the data from the encoder like this:
+
+```python
+encoder = Model(input=inputs, output=encoded)
+X_encoded = encoder.predict(X)
+```
+
+---
+
+### How can I use Keras with datasets that don't fit in memory?
+
+You can do batch training using `model.train_on_batch(X, y)` and `model.test_on_batch(X, y)`. See the [models documentation](/models/sequential).
+
+Alternatively, you can write a generator that yields batches of training data and use the method `model.fit_generator(data_generator, samples_per_epoch, nb_epoch)`.
+
+You can see batch training in action in our [CIFAR10 example](https://github.com/fchollet/keras/blob/master/examples/cifar10_cnn.py).
+
+---
+
+### How can I interrupt training when the validation loss isn't decreasing anymore?
+
+You can use an `EarlyStopping` callback:
+
+```python
+from keras.callbacks import EarlyStopping
+early_stopping = EarlyStopping(monitor='val_loss', patience=2)
+model.fit(X, y, validation_split=0.2, callbacks=[early_stopping])
+```
+
+Find out more in the [callbacks documentation](/callbacks).
+
+---
+
+### How is the validation split computed?
+
+If you set the `validation_split` argument in `model.fit` to e.g. 0.1, then the validation data used will be the *last 10%* of the data. If you set it to 0.25, it will be the last 25% of the data, etc.
+
+
+---
+
+### Is the data shuffled during training?
+
+Yes, if the `shuffle` argument in `model.fit` is set to `True` (which is the default), the training data will be randomly shuffled at each epoch.
+
+Validation data is never shuffled.
+
+---
+
+
+### How can I record the training / validation loss / accuracy at each epoch?
+
+The `model.fit` method returns an `History` callback, which has a `history` attribute containing the lists of successive losses and other metrics.
+
+```python
+hist = model.fit(X, y, validation_split=0.2)
+print(hist.history)
+```
+
+---
+
+### How can I "freeze" Keras layers?
+
+To "freeze" a layer means to exclude it from training, i.e. its weights will never be updated. This is useful in the context of fine-tuning a model, or using fixed embeddings for a text input.
+
+You can pass a `trainable` argument (boolean) to a layer constructor to set a layer to be non-trainable:
+
+```python
+frozen_layer = Dense(32, trainable=False)
+```
+
+Additionally, you can set the `trainable` property of a layer to `True` or `False` after instantiation. For this to take effect, you will need to call `compile()` on your model after modifying the `trainable` property. Here's an example:
+
+```python
+x = Input(shape=(32,))
+layer = Dense(32)
+layer.trainable = False
+y = layer(x)
+
+frozen_model = Model(x, y)
+# in the model below, the weights of `layer` will not be updated during training
+frozen_model.compile(optimizer='rmsprop', loss='mse')
+
+layer.trainable = True
+trainable_model = Model(x, y)
+# with this model the weights of the layer will be updated during training
+# (which will also affect the above model since it uses the same layer instance)
+trainable_model.compile(optimizer='rmsprop', loss='mse')
+
+frozen_model.fit(data, labels)  # this does NOT update the weights of `layer`
+trainable_model.fit(data, labels)  # this updates the weights of `layer`
+```
+
+---
+
+### How can I use stateful RNNs?
+
+Making a RNN stateful means that the states for the samples of each batch will be reused as initial states for the samples in the next batch.
+
+When using stateful RNNs, it is therefore assumed that:
+
+- all batches have the same number of samples
+- If `X1` and `X2` are successive batches of samples, then `X2[i]` is the follow-up sequence to `X1[i]`, for every `i`.
+
+To use statefulness in RNNs, you need to:
+
+- explicitly specify the batch size you are using, by passing a `batch_input_shape` argument to the first layer in your model. It should be a tuple of integers, e.g. `(32, 10, 16)` for a 32-samples batch of sequences of 10 timesteps with 16 features per timestep.
+- set `stateful=True` in your RNN layer(s).
+
+To reset the states accumulated:
+
+- use `model.reset_states()` to reset the states of all layers in the model
+- use `layer.reset_states()` to reset the states of a specific stateful RNN layer
+
+Example:
+
+```python
+
+X  # this is our input data, of shape (32, 21, 16)
+# we will feed it to our model in sequences of length 10
+
+model = Sequential()
+model.add(LSTM(32, batch_input_shape=(32, 10, 16), stateful=True))
+model.add(Dense(16, activation='softmax'))
+
+model.compile(optimizer='rmsprop', loss='categorical_crossentropy')
+
+# we train the network to predict the 11th timestep given the first 10:
+model.train_on_batch(X[:, :10, :], np.reshape(X[:, 10, :], (32, 16)))
+
+# the state of the network has changed. We can feed the follow-up sequences:
+model.train_on_batch(X[:, 10:20, :], np.reshape(X[:, 20, :], (32, 16)))
+
+# let's reset the states of the LSTM layer:
+model.reset_states()
+
+# another way to do it in this case:
+model.layers[0].reset_states()
+```
+
+Notes that the methods `predict`, `fit`, `train_on_batch`, `predict_classes`, etc. will *all* update the states of the stateful layers in a model. This allows you to do not only stateful training, but also stateful prediction.
+
+---
+
+### How can I remove a layer from a Sequential model?
+
+You can remove the last added layer in a Sequential model by calling `.pop()`:
+
+```python
+model = Sequential()
+model.add(Dense(32, activation='relu', input_dim=784))
+model.add(Dense(32, activation='relu'))
+
+print(len(model.layers))  # "2"
+
+model.pop()
+print(len(model.layers))  # "1"
+```
+
+---
+
+### How can I use pre-trained models in Keras?
+
+Code and pre-trained weights are available for the following image classification models:
+
+- VGG16
+- VGG19
+- ResNet50
+- Inception v3
+
+They can be imported from the module `keras.applications`:
+
+```python
+from keras.applications.vgg16 import VGG16
+from keras.applications.vgg19 import VGG19
+from keras.applications.resnet50 import ResNet50
+from keras.applications.inception_v3 import InceptionV3
+
+model = VGG16(weights='imagenet', include_top=True)
+```
+
+For a few simple usage examples, see [the documentation for the Applications module](/applications).
+
+For a detailed example of how to use such a pre-trained model for feature extraction or for fine-tuning, see [this blog post](http://blog.keras.io/building-powerful-image-classification-models-using-very-little-data.html).
+
+The VGG16 model is also the basis for several Keras example scripts:
+
+- [Style transfer](https://github.com/fchollet/keras/blob/master/examples/neural_style_transfer.py)
+- [Feature visualization](https://github.com/fchollet/keras/blob/master/examples/conv_filter_visualization.py)
+- [Deep dream](https://github.com/fchollet/keras/blob/master/examples/deep_dream.py)

docs/templates/getting-started/functional-api-guide.md

@@ -0,0 +1,421 @@
+# Getting started with the Keras functional API
+
+The Keras functional API is the way to go for defining complex models, such as multi-output models, directed acyclic graphs, or models with shared layers.
+
+This guide assumes that you are already familiar with the `Sequential` model.
+
+Let's start with something simple.
+
+-----
+
+## First example: fully connected network
+
+The `Sequential` model is probably a better choice to implement such a network, but it helps to start with something really simple.
+
+- A layer instance is callable (on a tensor), and it returns a tensor
+- Input tensor(s) and output tensor(s) can then be used to define a `Model`
+- Such a model can be trained just like Keras `Sequential` models.
+
+```python
+from keras.layers import Input, Dense
+from keras.models import Model
+
+# this returns a tensor
+inputs = Input(shape=(784,))
+
+# a layer instance is callable on a tensor, and returns a tensor
+x = Dense(64, activation='relu')(inputs)
+x = Dense(64, activation='relu')(x)
+predictions = Dense(10, activation='softmax')(x)
+
+# this creates a model that includes
+# the Input layer and three Dense layers
+model = Model(input=inputs, output=predictions)
+model.compile(optimizer='rmsprop',
+              loss='categorical_crossentropy',
+              metrics=['accuracy'])
+model.fit(data, labels)  # starts training
+```
+
+-----
+
+## All models are callable, just like layers
+
+With the functional API, it is easy to re-use trained models: you can treat any model as if it were a layer, by calling it on a tensor. Note that by calling a model you aren't just re-using the *architecture* of the model, you are also re-using its weights.
+
+```python
+x = Input(shape=(784,))
+# this works, and returns the 10-way softmax we defined above.
+y = model(x)
+```
+
+This can allow, for instance, to quickly create models that can process *sequences* of inputs. You could turn an image classification model into a video classification model, in just one line.
+
+```python
+from keras.layers import TimeDistributed
+
+# input tensor for sequences of 20 timesteps,
+# each containing a 784-dimensional vector
+input_sequences = Input(shape=(20, 784))
+
+# this applies our previous model to every timestep in the input sequences.
+# the output of the previous model was a 10-way softmax,
+# so the output of the layer below will be a sequence of 20 vectors of size 10.
+processed_sequences = TimeDistributed(model)(input_sequences)
+```
+
+-----
+
+## Multi-input and multi-output models
+
+Here's a good use case for the functional API: models with multiple inputs and outputs. The functional API makes it easy to manipulate a large number of intertwined datastreams.
+
+Let's consider the following model. We seek to predict how many retweets and likes a news headline will receive on Twitter. The main input to the model will be the headline itself, as a sequence of words, but to spice things up, our model will also have an auxiliary input, receiving extra data such as the time of day when the headline was posted, etc.
+The model will also be supervised via two loss functions. Using the main loss function earlier in a model is a good regularization mechanism for deep models.
+
+Here's what our model looks like:
+
+<img src="https://s3.amazonaws.com/keras.io/img/multi-input-multi-output-graph.png" alt="multi-input-multi-output-graph" style="width: 400px;"/>
+
+Let's implement it with the functional API.
+
+The main input will receive the headline, as a sequence of integers (each integer encodes a word).
+The integers will be between 1 and 10,000 (a vocabulary of 10,000 words) and the sequences will be 100 words long.
+
+```python
+from keras.layers import Input, Embedding, LSTM, Dense, merge
+from keras.models import Model
+
+# headline input: meant to receive sequences of 100 integers, between 1 and 10000.
+# note that we can name any layer by passing it a "name" argument.
+main_input = Input(shape=(100,), dtype='int32', name='main_input')
+
+# this embedding layer will encode the input sequence
+# into a sequence of dense 512-dimensional vectors.
+x = Embedding(output_dim=512, input_dim=10000, input_length=100)(main_input)
+
+# a LSTM will transform the vector sequence into a single vector,
+# containing information about the entire sequence
+lstm_out = LSTM(32)(x)
+```
+
+Here we insert the auxiliary loss, allowing the LSTM and Embedding layer to be trained smoothly even though the main loss will be much higher in the model.
+
+```python
+auxiliary_loss = Dense(1, activation='sigmoid', name='aux_output')(lstm_out)
+```
+
+At this point, we feed into the model our auxiliary input data by concatenating it with the LSTM output:
+
+```python
+auxiliary_input = Input(shape=(5,), name='aux_input')
+x = merge([lstm_out, auxiliary_input], mode='concat')
+
+# we stack a deep fully-connected network on top
+x = Dense(64, activation='relu')(x)
+x = Dense(64, activation='relu')(x)
+x = Dense(64, activation='relu')(x)
+
+# and finally we add the main logistic regression layer
+main_loss = Dense(1, activation='sigmoid', name='main_output')(x)
+```
+
+This defines a model with two inputs and two outputs:
+
+```python
+model = Model(input=[main_input, auxiliary_input], output=[main_loss, auxiliary_loss])
+```
+
+We compile the model and assign a weight of 0.2 to the auxiliary loss.
+To specify different `loss_weights` or `loss` for each different output, you can use a list or a dictionary.
+Here we pass a single loss as the `loss` argument, so the same loss will be used on all outputs.
+
+```python
+model.compile(optimizer='rmsprop', loss='binary_crossentropy',
+              loss_weights=[1., 0.2])
+```
+
+We can train the model by passing it lists of input arrays and target arrays:
+
+```python
+model.fit([headline_data, additional_data], [labels, labels],
+          nb_epoch=50, batch_size=32)
+```
+
+Since our inputs and outputs are named (we passed them a "name" argument),
+We could also have compiled the model via:
+
+```python
+model.compile(optimizer='rmsprop',
+              loss={'main_output': 'binary_crossentropy', 'aux_output': 'binary_crossentropy'},
+              loss_weights={'main_output': 1., 'aux_output': 0.2})
+
+# and trained it via:
+model.fit({'main_input': headline_data, 'aux_input': additional_data},
+          {'main_output': labels, 'aux_output': labels},
+          nb_epoch=50, batch_size=32)
+```
+
+-----
+
+## Shared layers
+
+Another good use for the functional API are models that use shared layers. Let's take a look at shared layers.
+
+Let's consider a dataset of tweets. We want to build a model that can tell whether two tweets are from the same person or not (this can allow us to compare users by the similarity of their tweets, for instance).
+
+One way to achieve this is to build a model that encodes two tweets into two vectors, concatenates the vectors and adds a logistic regression of top, outputting a probability that the two tweets share the same author. The model would then be trained on positive tweet pairs and negative tweet pairs.
+
+Because the problem is symmetric, the mechanism that encodes the first tweet should be reused (weights and all) to encode the second tweet. Here we use a shared LSTM layer to encode the tweets.
+
+Let's build this with the functional API. We will take as input for a tweet a binary matrix of shape `(140, 256)`, i.e. a sequence of 140 vectors of size 256, where each dimension in the 256-dimensional vector encodes the presence/absence of a character (out of an alphabet of 256 frequent characters).
+
+```python
+from keras.layers import Input, LSTM, Dense, merge
+from keras.models import Model
+
+tweet_a = Input(shape=(140, 256))
+tweet_b = Input(shape=(140, 256))
+```
+
+To share a layer across different inputs, simply instantiate the layer once, then call it on as many inputs as you want:
+
+```python
+# this layer can take as input a matrix
+# and will return a vector of size 64
+shared_lstm = LSTM(64)
+
+# when we reuse the same layer instance
+# multiple times, the weights of the layer
+# are also being reused
+# (it is effectively *the same* layer)
+encoded_a = shared_lstm(tweet_a)
+encoded_b = shared_lstm(tweet_b)
+
+# we can then concatenate the two vectors:
+merged_vector = merge([encoded_a, encoded_b], mode='concat', concat_axis=-1)
+
+# and add a logistic regression on top
+predictions = Dense(1, activation='sigmoid')(merged_vector)
+
+# we define a trainable model linking the
+# tweet inputs to the predictions
+model = Model(input=[tweet_a, tweet_b], output=predictions)
+
+model.compile(optimizer='rmsprop',
+              loss='binary_crossentropy',
+              metrics=['accuracy'])
+model.fit([data_a, data_b], labels, nb_epoch=10)
+```
+
+Let's pause to take a look at how to read the shared layer's output or output shape.
+
+-----
+
+## The concept of layer "node"
+
+Whenever you are calling a layer on some input, you are creating a new tensor (the output of the layer), and you are adding a "node" to the layer, linking the input tensor to the output tensor. When you are calling the same layer multiple times, that layer owns multiple nodes indexed as 0, 1, 2...
+
+In previous versions of Keras, you could obtain the output tensor of a layer instance via `layer.get_output()`, or its output shape via `layer.output_shape`. You still can (except `get_output()` has been replaced by the property `output`). But what if a layer is connected to multiple inputs?
+
+As long as a layer is only connected to one input, there is no confusion, and `.output` will return the one output of the layer:
+
+```python
+a = Input(shape=(140, 256))
+
+lstm = LSTM(32)
+encoded_a = lstm(a)
+
+assert lstm.output == encoded_a
+```
+
+Not so if the layer has multiple inputs:
+```python
+a = Input(shape=(140, 256))
+b = Input(shape=(140, 256))
+
+lstm = LSTM(32)
+encoded_a = lstm(a)
+encoded_b = lstm(b)
+
+lstm.output
+```
+```
+>> AssertionError: Layer lstm_1 has multiple inbound nodes,
+hence the notion of "layer output" is ill-defined.
+Use `get_output_at(node_index)` instead.
+```
+
+Okay then. The following works:
+
+```python
+assert lstm.get_output_at(0) == encoded_a
+assert lstm.get_output_at(1) == encoded_b
+```
+
+Simple enough, right?
+
+The same is true for the properties `input_shape` and `output_shape`: as long as the layer has only one node, or as long as all nodes have the same input/output shape, then the notion of "layer output/input shape" is well defined, and that one shape will be returned by `layer.output_shape`/`layer.input_shape`. But if, for instance, you apply a same `Convolution2D` layer to an input of shape `(3, 32, 32)`, and then to an input of shape `(3, 64, 64)`, the layer will have multiple input/output shapes, and you will have to fetch them by specifying the index of the node they belong to:
+
+```python
+a = Input(shape=(3, 32, 32))
+b = Input(shape=(3, 64, 64))
+
+conv = Convolution2D(16, 3, 3, border_mode='same')
+conved_a = conv(a)
+
+# only one input so far, the following will work:
+assert conv.input_shape == (None, 3, 32, 32)
+
+conved_b = conv(b)
+# now the `.input_shape` property wouldn't work, but this does:
+assert conv.get_input_shape_at(0) == (None, 3, 32, 32)
+assert conv.get_input_shape_at(1) == (None, 3, 64, 64)
+```
+
+-----
+
+## More examples
+
+Code examples are still the best way to get started, so here are a few more.
+
+### Inception module
+
+For more information about the Inception architecture, see [Going Deeper with Convolutions](http://arxiv.org/abs/1409.4842).
+
+```python
+from keras.layers import merge, Convolution2D, MaxPooling2D, Input
+
+input_img = Input(shape=(3, 256, 256))
+
+tower_1 = Convolution2D(64, 1, 1, border_mode='same', activation='relu')(input_img)
+tower_1 = Convolution2D(64, 3, 3, border_mode='same', activation='relu')(tower_1)
+
+tower_2 = Convolution2D(64, 1, 1, border_mode='same', activation='relu')(input_img)
+tower_2 = Convolution2D(64, 5, 5, border_mode='same', activation='relu')(tower_2)
+
+tower_3 = MaxPooling2D((3, 3), strides=(1, 1), border_mode='same')(input_img)
+tower_3 = Convolution2D(64, 1, 1, border_mode='same', activation='relu')(tower_3)
+
+output = merge([tower_1, tower_2, tower_3], mode='concat', concat_axis=1)
+```
+
+### Residual connection on a convolution layer
+
+For more information about residual networks, see [Deep Residual Learning for Image Recognition](http://arxiv.org/abs/1512.03385).
+
+```python
+from keras.layers import merge, Convolution2D, Input
+
+# input tensor for a 3-channel 256x256 image
+x = Input(shape=(3, 256, 256))
+# 3x3 conv with 3 output channels (same as input channels)
+y = Convolution2D(3, 3, 3, border_mode='same')(x)
+# this returns x + y.
+z = merge([x, y], mode='sum')
+```
+
+### Shared vision model
+
+This model re-uses the same image-processing module on two inputs, to classify whether two MNIST digits are the same digit or different digits.
+
+```python
+from keras.layers import merge, Convolution2D, MaxPooling2D, Input, Dense, Flatten
+from keras.models import Model
+
+# first, define the vision modules
+digit_input = Input(shape=(1, 27, 27))
+x = Convolution2D(64, 3, 3)(digit_input)
+x = Convolution2D(64, 3, 3)(x)
+x = MaxPooling2D((2, 2))(x)
+out = Flatten()(x)
+
+vision_model = Model(digit_input, out)
+
+# then define the tell-digits-apart model
+digit_a = Input(shape=(1, 27, 27))
+digit_b = Input(shape=(1, 27, 27))
+
+# the vision model will be shared, weights and all
+out_a = vision_model(digit_a)
+out_b = vision_model(digit_b)
+
+concatenated = merge([out_a, out_b], mode='concat')
+out = Dense(1, activation='sigmoid')(concatenated)
+
+classification_model = Model([digit_a, digit_b], out)
+```
+
+### Visual question answering model
+
+This model can select the correct one-word answer when asked a natural-language question about a picture.
+
+It works by encoding the question into a vector, encoding the image into a vector, concatenating the two, and training on top a logistic regression over some vocabulary of potential answers.
+
+```python
+from keras.layers import Convolution2D, MaxPooling2D, Flatten
+from keras.layers import Input, LSTM, Embedding, Dense, merge
+from keras.models import Model, Sequential
+
+# first, let's define a vision model using a Sequential model.
+# this model will encode an image into a vector.
+vision_model = Sequential()
+vision_model.add(Convolution2D(64, 3, 3, activation='relu', border_mode='same', input_shape=(3, 224, 224)))
+vision_model.add(Convolution2D(64, 3, 3, activation='relu'))
+vision_model.add(MaxPooling2D((2, 2)))
+vision_model.add(Convolution2D(128, 3, 3, activation='relu', border_mode='same'))
+vision_model.add(Convolution2D(128, 3, 3, activation='relu'))
+vision_model.add(MaxPooling2D((2, 2)))
+vision_model.add(Convolution2D(256, 3, 3, activation='relu', border_mode='same'))
+vision_model.add(Convolution2D(256, 3, 3, activation='relu'))
+vision_model.add(Convolution2D(256, 3, 3, activation='relu'))
+vision_model.add(MaxPooling2D((2, 2)))
+vision_model.add(Flatten())
+
+# now let's get a tensor with the output of our vision model:
+image_input = Input(shape=(3, 224, 224))
+encoded_image = vision_model(image_input)
+
+# next, let's define a language model to encode the question into a vector.
+# each question will be at most 100 word long,
+# and we will index words as integers from 1 to 9999.
+question_input = Input(shape=(100,), dtype='int32')
+embedded_question = Embedding(input_dim=10000, output_dim=256, input_length=100)(question_input)
+encoded_question = LSTM(256)(embedded_question)
+
+# let's concatenate the question vector and the image vector:
+merged = merge([encoded_question, encoded_image], mode='concat')
+
+# and let's train a logistic regression over 1000 words on top:
+output = Dense(1000, activation='softmax')(merged)
+
+# this is our final model:
+vqa_model = Model(input=[image_input, question_input], output=output)
+
+# the next stage would be training this model on actual data.
+```
+
+### Video question answering model
+
+Now that we have trained our image QA model, we can quickly turn it into a video QA model. With appropriate training, you will be able to show it a short video (e.g. 100-frame human action) and ask a natural language question about the video (e.g. "what sport is the boy playing?" -> "football").
+
+```python
+from keras.layers import TimeDistributed
+
+video_input = Input(shape=(100, 3, 224, 224))
+# this is our video encoded via the previously trained vision_model (weights are reused)
+encoded_frame_sequence = TimeDistributed(vision_model)(video_input)  # the output will be a sequence of vectors
+encoded_video = LSTM(256)(encoded_frame_sequence)  # the output will be a vector
+
+# this is a model-level representation of the question encoder, reusing the same weights as before:
+question_encoder = Model(input=question_input, output=encoded_question)
+
+# let's use it to encode the question:
+video_question_input = Input(shape=(100,), dtype='int32')
+encoded_video_question = question_encoder(video_question_input)
+
+# and this is our video question answering model:
+merged = merge([encoded_video, encoded_video_question], mode='concat')
+output = Dense(1000, activation='softmax')(merged)
+video_qa_model = Model(input=[video_input, video_question_input], output=output)
+```

docs/templates/getting-started/sequential-model-guide.md

@@ -0,0 +1,549 @@
+# Getting started with the Keras Sequential model
+
+The `Sequential` model is a linear stack of layers.
+
+You can create a `Sequential` model by passing a list of layer instances to the constructor:
+
+```python
+from keras.models import Sequential
+from keras.layers import Dense, Activation
+
+model = Sequential([
+    Dense(32, input_dim=784),
+    Activation('relu'),
+    Dense(10),
+    Activation('softmax'),
+])
+```
+
+You can also simply add layers via the `.add()` method:
+
+```python
+model = Sequential()
+model.add(Dense(32, input_dim=784))
+model.add(Activation('relu'))
+```
+
+----
+
+## Specifying the input shape
+
+The model needs to know what input shape it should expect. For this reason, the first layer in a `Sequential` model (and only the first, because following layers can do automatic shape inference) needs to receive information about its input shape. There are several possible ways to do this:
+
+- pass an `input_shape` argument to the first layer. This is a shape tuple (a tuple of integers or `None` entries, where `None` indicates that any positive integer may be expected). In `input_shape`, the batch dimension is not included.
+- pass instead a `batch_input_shape` argument, where the batch dimension is included. This is useful for specifying a fixed batch size (e.g. with stateful RNNs).
+- some 2D layers, such as `Dense`, support the specification of their input shape via the argument `input_dim`, and some 3D temporal layers support the arguments `input_dim` and `input_length`.
+
+As such, the following three snippets are strictly equivalent:
+```python
+model = Sequential()
+model.add(Dense(32, input_shape=(784,)))
+```
+```python
+model = Sequential()
+model.add(Dense(32, batch_input_shape=(None, 784)))
+# note that batch dimension is "None" here,
+# so the model will be able to process batches of any size.
+```
+```python
+model = Sequential()
+model.add(Dense(32, input_dim=784))
+```
+
+And so are the following three snippets:
+```python
+model = Sequential()
+model.add(LSTM(32, input_shape=(10, 64)))
+```
+```python
+model = Sequential()
+model.add(LSTM(32, batch_input_shape=(None, 10, 64)))
+```
+```python
+model = Sequential()
+model.add(LSTM(32, input_length=10, input_dim=64))
+```
+
+----
+
+## The Merge layer
+
+Multiple `Sequential` instances can be merged into a single output via a `Merge` layer. The output is a layer that can be added as first layer in a new `Sequential` model. For instance, here's a model with two separate input branches getting merged:
+
+```python
+from keras.layers import Merge
+
+left_branch = Sequential()
+left_branch.add(Dense(32, input_dim=784))
+
+right_branch = Sequential()
+right_branch.add(Dense(32, input_dim=784))
+
+merged = Merge([left_branch, right_branch], mode='concat')
+
+final_model = Sequential()
+final_model.add(merged)
+final_model.add(Dense(10, activation='softmax'))
+```
+
+<img src="https://s3.amazonaws.com/keras.io/img/two_branches_sequential_model.png" alt="two branch Sequential" style="width: 400px;"/>
+
+Such a two-branch model can then be trained via e.g.:
+
+```python
+final_model.compile(optimizer='rmsprop', loss='categorical_crossentropy')
+final_model.fit([input_data_1, input_data_2], targets)  # we pass one data array per model input
+```
+
+The `Merge` layer supports a number of pre-defined modes:
+
+- `sum` (default): element-wise sum
+- `concat`: tensor concatenation. You can specify the concatenation axis via the argument `concat_axis`.
+- `mul`: element-wise multiplication
+- `ave`: tensor average
+- `dot`: dot product. You can specify which axes to reduce along via the argument `dot_axes`.
+- `cos`: cosine proximity between vectors in 2D tensors.
+
+You can also pass a function as the `mode` argument, allowing for arbitrary transformations:
+
+```python
+merged = Merge([left_branch, right_branch], mode=lambda x: x[0] - x[1])
+```
+
+Now you know enough to be able to define *almost* any model with Keras. For complex models that cannot be expressed via `Sequential` and `Merge`, you can use [the functional API](/getting-started/functional-api-guide).
+
+
+----
+
+## Compilation
+
+Before training a model, you need to configure the learning process, which is done via the `compile` method. It receives three arguments:
+
+- an optimizer. This could be the string identifier of an existing optimizer (such as `rmsprop` or `adagrad`), or an instance of the `Optimizer` class. See: [optimizers](/optimizers).
+- a loss function. This is the objective that the model will try to minimize. It can be the string identifier of an existing loss function (such as `categorical_crossentropy` or `mse`), or it can be an objective function. See: [objectives](/objectives).
+- a list of metrics. For any classification problem you will want to set this to `metrics=['accuracy']`. A metric could be the string identifier of an existing metric (only `accuracy` is supported at this point), or a custom metric function.
+
+```python
+# for a multi-class classification problem
+model.compile(optimizer='rmsprop',
+              loss='categorical_crossentropy',
+              metrics=['accuracy'])
+
+# for a binary classification problem
+model.compile(optimizer='rmsprop',
+              loss='binary_crossentropy',
+              metrics=['accuracy'])
+
+# for a mean squared error regression problem
+model.compile(optimizer='rmsprop',
+              loss='mse')
+```
+
+----
+
+## Training
+
+Keras models are trained on Numpy arrays of input data and labels. For training a model, you will typically use the `fit` function. [Read its documentation here](/models/sequential). 
+
+```python
+# for a single-input model with 2 classes (binary):
+
+model = Sequential()
+model.add(Dense(1, input_dim=784, activation='sigmoid'))
+model.compile(optimizer='rmsprop',
+              loss='binary_crossentropy',
+              metrics=['accuracy'])
+
+# generate dummy data
+import numpy as np
+data = np.random.random((1000, 784))
+labels = np.random.randint(2, size=(1000, 1))
+
+# train the model, iterating on the data in batches
+# of 32 samples
+model.fit(data, labels, nb_epoch=10, batch_size=32)
+```
+```python
+# for a multi-input model with 10 classes:
+
+left_branch = Sequential()
+left_branch.add(Dense(32, input_dim=784))
+
+right_branch = Sequential()
+right_branch.add(Dense(32, input_dim=784))
+
+merged = Merge([left_branch, right_branch], mode='concat')
+
+model = Sequential()
+model.add(merged)
+model.add(Dense(10, activation='softmax'))
+
+model.compile(optimizer='rmsprop',
+              loss='categorical_crossentropy',
+              metrics=['accuracy'])
+
+# generate dummy data
+import numpy as np
+from keras.utils.np_utils import to_categorical
+data_1 = np.random.random((1000, 784))
+data_2 = np.random.random((1000, 784))
+
+# these are integers between 0 and 9
+labels = np.random.randint(10, size=(1000, 1))
+# we convert the labels to a binary matrix of size (1000, 10)
+# for use with categorical_crossentropy
+labels = to_categorical(labels, 10)
+
+# train the model
+# note that we are passing a list of Numpy arrays as training data
+# since the model has 2 inputs
+model.fit([data_1, data_2], labels, nb_epoch=10, batch_size=32)
+```
+
+----
+
+
+## Examples
+
+Here are a few examples to get you started!
+
+In the examples folder, you will also find example models for real datasets:
+
+- CIFAR10 small images classification: Convolutional Neural Network (CNN) with realtime data augmentation
+- IMDB movie review sentiment classification: LSTM over sequences of words
+- Reuters newswires topic classification: Multilayer Perceptron (MLP)
+- MNIST handwritten digits classification: MLP & CNN
+- Character-level text generation with LSTM
+
+...and more.
+
+
+### Multilayer Perceptron (MLP) for multi-class softmax classification:
+
+```python
+from keras.models import Sequential
+from keras.layers import Dense, Dropout, Activation
+from keras.optimizers import SGD
+
+model = Sequential()
+# Dense(64) is a fully-connected layer with 64 hidden units.
+# in the first layer, you must specify the expected input data shape:
+# here, 20-dimensional vectors.
+model.add(Dense(64, input_dim=20, init='uniform'))
+model.add(Activation('tanh'))
+model.add(Dropout(0.5))
+model.add(Dense(64, init='uniform'))
+model.add(Activation('tanh'))
+model.add(Dropout(0.5))
+model.add(Dense(10, init='uniform'))
+model.add(Activation('softmax'))
+
+sgd = SGD(lr=0.1, decay=1e-6, momentum=0.9, nesterov=True)
+model.compile(loss='categorical_crossentropy',
+              optimizer=sgd,
+              metrics=['accuracy'])
+
+model.fit(X_train, y_train,
+          nb_epoch=20,
+          batch_size=16)
+score = model.evaluate(X_test, y_test, batch_size=16)
+```
+
+
+### Alternative implementation of a similar MLP:
+
+```python
+model = Sequential()
+model.add(Dense(64, input_dim=20, activation='relu'))
+model.add(Dropout(0.5))
+model.add(Dense(64, activation='relu'))
+model.add(Dropout(0.5))
+model.add(Dense(10, activation='softmax'))
+
+model.compile(loss='categorical_crossentropy',
+              optimizer='adadelta',
+              metrics=['accuracy'])
+```
+
+
+### MLP for binary classification:
+```python
+model = Sequential()
+model.add(Dense(64, input_dim=20, init='uniform', activation='relu'))
+model.add(Dropout(0.5))
+model.add(Dense(64, activation='relu'))
+model.add(Dropout(0.5))
+model.add(Dense(1, activation='sigmoid'))
+
+model.compile(loss='binary_crossentropy',
+              optimizer='rmsprop',
+              metrics=['accuracy'])
+```
+
+
+### VGG-like convnet:
+
+```python
+from keras.models import Sequential
+from keras.layers import Dense, Dropout, Activation, Flatten
+from keras.layers import Convolution2D, MaxPooling2D
+from keras.optimizers import SGD
+
+model = Sequential()
+# input: 100x100 images with 3 channels -> (3, 100, 100) tensors.
+# this applies 32 convolution filters of size 3x3 each.
+model.add(Convolution2D(32, 3, 3, border_mode='valid', input_shape=(3, 100, 100)))
+model.add(Activation('relu'))
+model.add(Convolution2D(32, 3, 3))
+model.add(Activation('relu'))
+model.add(MaxPooling2D(pool_size=(2, 2)))
+model.add(Dropout(0.25))
+
+model.add(Convolution2D(64, 3, 3, border_mode='valid'))
+model.add(Activation('relu'))
+model.add(Convolution2D(64, 3, 3))
+model.add(Activation('relu'))
+model.add(MaxPooling2D(pool_size=(2, 2)))
+model.add(Dropout(0.25))
+
+model.add(Flatten())
+# Note: Keras does automatic shape inference.
+model.add(Dense(256))
+model.add(Activation('relu'))
+model.add(Dropout(0.5))
+
+model.add(Dense(10))
+model.add(Activation('softmax'))
+
+sgd = SGD(lr=0.1, decay=1e-6, momentum=0.9, nesterov=True)
+model.compile(loss='categorical_crossentropy', optimizer=sgd)
+
+model.fit(X_train, Y_train, batch_size=32, nb_epoch=1)
+```
+
+
+### Sequence classification with LSTM:
+
+```python
+from keras.models import Sequential
+from keras.layers import Dense, Dropout, Activation
+from keras.layers import Embedding
+from keras.layers import LSTM
+
+model = Sequential()
+model.add(Embedding(max_features, 256, input_length=maxlen))
+model.add(LSTM(output_dim=128, activation='sigmoid', inner_activation='hard_sigmoid'))
+model.add(Dropout(0.5))
+model.add(Dense(1))
+model.add(Activation('sigmoid'))
+
+model.compile(loss='binary_crossentropy',
+              optimizer='rmsprop',
+              metrics=['accuracy'])
+
+model.fit(X_train, Y_train, batch_size=16, nb_epoch=10)
+score = model.evaluate(X_test, Y_test, batch_size=16)
+```
+
+### Architecture for learning image captions with a convnet and a Gated Recurrent Unit:
+(word-level embedding, caption of maximum length 16 words).
+
+Note that getting this to work well will require using a bigger convnet, initialized with pre-trained weights.
+
+```python
+max_caption_len = 16
+vocab_size = 10000
+
+# first, let's define an image model that
+# will encode pictures into 128-dimensional vectors.
+# it should be initialized with pre-trained weights.
+image_model = Sequential()
+image_model.add(Convolution2D(32, 3, 3, border_mode='valid', input_shape=(3, 100, 100)))
+image_model.add(Activation('relu'))
+image_model.add(Convolution2D(32, 3, 3))
+image_model.add(Activation('relu'))
+image_model.add(MaxPooling2D(pool_size=(2, 2)))
+
+image_model.add(Convolution2D(64, 3, 3, border_mode='valid'))
+image_model.add(Activation('relu'))
+image_model.add(Convolution2D(64, 3, 3))
+image_model.add(Activation('relu'))
+image_model.add(MaxPooling2D(pool_size=(2, 2)))
+
+image_model.add(Flatten())
+image_model.add(Dense(128))
+
+# let's load the weights from a save file.
+image_model.load_weights('weight_file.h5')
+
+# next, let's define a RNN model that encodes sequences of words
+# into sequences of 128-dimensional word vectors.
+language_model = Sequential()
+language_model.add(Embedding(vocab_size, 256, input_length=max_caption_len))
+language_model.add(GRU(output_dim=128, return_sequences=True))
+language_model.add(TimeDistributed(Dense(128)))
+
+# let's repeat the image vector to turn it into a sequence.
+image_model.add(RepeatVector(max_caption_len))
+
+# the output of both models will be tensors of shape (samples, max_caption_len, 128).
+# let's concatenate these 2 vector sequences.
+model = Sequential()
+model.add(Merge([image_model, language_model], mode='concat', concat_axis=-1))
+# let's encode this vector sequence into a single vector
+model.add(GRU(256, return_sequences=False))
+# which will be used to compute a probability
+# distribution over what the next word in the caption should be!
+model.add(Dense(vocab_size))
+model.add(Activation('softmax'))
+
+model.compile(loss='categorical_crossentropy', optimizer='rmsprop')
+
+# "images" is a numpy float array of shape (nb_samples, nb_channels=3, width, height).
+# "captions" is a numpy integer array of shape (nb_samples, max_caption_len)
+# containing word index sequences representing partial captions.
+# "next_words" is a numpy float array of shape (nb_samples, vocab_size)
+# containing a categorical encoding (0s and 1s) of the next word in the corresponding
+# partial caption.
+model.fit([images, partial_captions], next_words, batch_size=16, nb_epoch=100)
+```
+
+
+### Stacked LSTM for sequence classification
+
+In this model, we stack 3 LSTM layers on top of each other,
+making the model capable of learning higher-level temporal representations.
+
+The first two LSTMs return their full output sequences, but the last one only returns
+the last step in its output sequence, thus dropping the temporal dimension
+(i.e. converting the input sequence into a single vector).
+
+<img src="https://keras.io/img/regular_stacked_lstm.png" alt="stacked LSTM" style="width: 300px;"/>
+
+```python
+from keras.models import Sequential
+from keras.layers import LSTM, Dense
+import numpy as np
+
+data_dim = 16
+timesteps = 8
+nb_classes = 10
+
+# expected input data shape: (batch_size, timesteps, data_dim)
+model = Sequential()
+model.add(LSTM(32, return_sequences=True,
+               input_shape=(timesteps, data_dim)))  # returns a sequence of vectors of dimension 32
+model.add(LSTM(32, return_sequences=True))  # returns a sequence of vectors of dimension 32
+model.add(LSTM(32))  # return a single vector of dimension 32
+model.add(Dense(10, activation='softmax'))
+
+model.compile(loss='categorical_crossentropy',
+              optimizer='rmsprop',
+              metrics=['accuracy'])
+
+# generate dummy training data
+x_train = np.random.random((1000, timesteps, data_dim))
+y_train = np.random.random((1000, nb_classes))
+
+# generate dummy validation data
+x_val = np.random.random((100, timesteps, data_dim))
+y_val = np.random.random((100, nb_classes))
+
+model.fit(x_train, y_train,
+          batch_size=64, nb_epoch=5,
+          validation_data=(x_val, y_val))
+```
+
+
+### Same stacked LSTM model, rendered "stateful"
+
+A stateful recurrent model is one for which the internal states (memories) obtained after processing a batch
+of samples are reused as initial states for the samples of the next batch. This allows to process longer sequences
+while keeping computational complexity manageable.
+
+[You can read more about stateful RNNs in the FAQ.](/faq/#how-can-i-use-stateful-rnns)
+
+```python
+from keras.models import Sequential
+from keras.layers import LSTM, Dense
+import numpy as np
+
+data_dim = 16
+timesteps = 8
+nb_classes = 10
+batch_size = 32
+
+# expected input batch shape: (batch_size, timesteps, data_dim)
+# note that we have to provide the full batch_input_shape since the network is stateful.
+# the sample of index i in batch k is the follow-up for the sample i in batch k-1.
+model = Sequential()
+model.add(LSTM(32, return_sequences=True, stateful=True,
+               batch_input_shape=(batch_size, timesteps, data_dim)))
+model.add(LSTM(32, return_sequences=True, stateful=True))
+model.add(LSTM(32, stateful=True))
+model.add(Dense(10, activation='softmax'))
+
+model.compile(loss='categorical_crossentropy',
+              optimizer='rmsprop',
+              metrics=['accuracy'])
+
+# generate dummy training data
+x_train = np.random.random((batch_size * 10, timesteps, data_dim))
+y_train = np.random.random((batch_size * 10, nb_classes))
+
+# generate dummy validation data
+x_val = np.random.random((batch_size * 3, timesteps, data_dim))
+y_val = np.random.random((batch_size * 3, nb_classes))
+
+model.fit(x_train, y_train,
+          batch_size=batch_size, nb_epoch=5,
+          validation_data=(x_val, y_val))
+```
+
+
+### Two merged LSTM encoders for classification over two parallel sequences
+
+In this model, two input sequences are encoded into vectors by two separate LSTM modules.
+
+These two vectors are then concatenated, and a fully connected network is trained on top of the concatenated representations.
+
+<img src="https://keras.io/img/dual_lstm.png" alt="Dual LSTM" style="width: 600px;"/>
+
+```python
+from keras.models import Sequential
+from keras.layers import Merge, LSTM, Dense
+import numpy as np
+
+data_dim = 16
+timesteps = 8
+nb_classes = 10
+
+encoder_a = Sequential()
+encoder_a.add(LSTM(32, input_shape=(timesteps, data_dim)))
+
+encoder_b = Sequential()
+encoder_b.add(LSTM(32, input_shape=(timesteps, data_dim)))
+
+decoder = Sequential()
+decoder.add(Merge([encoder_a, encoder_b], mode='concat'))
+decoder.add(Dense(32, activation='relu'))
+decoder.add(Dense(nb_classes, activation='softmax'))
+
+decoder.compile(loss='categorical_crossentropy',
+                optimizer='rmsprop',
+                metrics=['accuracy'])
+
+# generate dummy training data
+x_train_a = np.random.random((1000, timesteps, data_dim))
+x_train_b = np.random.random((1000, timesteps, data_dim))
+y_train = np.random.random((1000, nb_classes))
+
+# generate dummy validation data
+x_val_a = np.random.random((100, timesteps, data_dim))
+x_val_b = np.random.random((100, timesteps, data_dim))
+y_val = np.random.random((100, nb_classes))
+
+decoder.fit([x_train_a, x_train_b], y_train,
+            batch_size=64, nb_epoch=5,
+            validation_data=([x_val_a, x_val_b], y_val))
+```

docs/templates/index.md

@@ -2,19 +2,18 @@
 
 ## You have just found Keras.
 
-Keras is a minimalist, highly modular neural networks library, written in Python and capable of running either on top of either [TensorFlow](https://github.com/tensorflow/tensorflow) or [Theano](https://github.com/Theano/Theano). It was developed with a focus on enabling fast experimentation. Being able to go from idea to result with the least possible delay is key to doing good research.
+Keras is a minimalist, highly modular neural networks library, written in Python and capable of running on top of either [TensorFlow](https://github.com/tensorflow/tensorflow) or [Theano](https://github.com/Theano/Theano). It was developed with a focus on enabling fast experimentation. *Being able to go from idea to result with the least possible delay is key to doing good research.*
 
 Use Keras if you need a deep learning library that:
-- allows for easy and fast prototyping (through total modularity, minimalism, and extensibility).
-- supports both convolutional networks and recurrent networks, as well as combinations of the two.
-- supports arbitrary connectivity schemes (including multi-input and multi-output training).
-- runs seamlessly on CPU and GPU.
+
+- Allows for easy and fast prototyping (through total modularity, minimalism, and extensibility).
+- Supports both convolutional networks and recurrent networks, as well as combinations of the two.
+- Supports arbitrary connectivity schemes (including multi-input and multi-output training).
+- Runs seamlessly on CPU and GPU.
 
 Read the documentation at [Keras.io](http://keras.io).
 
-Keras is compatible with:
-    - __Python 2.7-3.5__ with the Theano backend
-    - __Python 2.7__ with the TensorFlow backend
+Keras is compatible with: __Python 2.7-3.5__.
 
 
 ------------------
@@ -36,9 +35,9 @@ Keras is compatible with:
 
 ## Getting started: 30 seconds to Keras
 
-The core datastructure of Keras is a __model__, a way to organize layers. There are two types of models: [`Sequential`](/models/#sequential) and [`Graph`](/models/#graph).
+The core data structure of Keras is a __model__, a way to organize layers. The main type of model is the [`Sequential`](http://keras.io/getting-started/sequential-model-guide) model, a linear stack of layers. For more complex architectures, you should use the [Keras functional API](http://keras.io/getting-started/functional-api-guide).
 
-Here's the `Sequential` model (a linear pile of layers):
+Here's the `Sequential` model:
 
 ```python
 from keras.models import Sequential
@@ -49,17 +48,17 @@ model = Sequential()
 Stacking layers is as easy as `.add()`:
 
 ```python
-from keras.layers.core import Dense, Activation
+from keras.layers import Dense, Activation
 
-model.add(Dense(output_dim=64, input_dim=100, init="glorot_uniform"))
+model.add(Dense(output_dim=64, input_dim=100))
 model.add(Activation("relu"))
-model.add(Dense(output_dim=10, init="glorot_uniform"))
+model.add(Dense(output_dim=10))
 model.add(Activation("softmax"))
 ```
 
 Once your model looks good, configure its learning process with `.compile()`:
 ```python
-model.compile(loss='categorical_crossentropy', optimizer='sgd')
+model.compile(loss='categorical_crossentropy', optimizer='sgd', metrics=['accuracy'])
 ```
 
 If you need to, you can further configure your optimizer. A core principle of Keras is to make things reasonably simple, while allowing the user to be fully in control when they need to (the ultimate control being the easy extensibility of the source code).
@@ -80,7 +79,7 @@ model.train_on_batch(X_batch, Y_batch)
 
 Evaluate your performance in one line:
 ```python
-objective_score = model.evaluate(X_test, Y_test, batch_size=32)
+loss_and_metrics = model.evaluate(X_test, Y_test, batch_size=32)
 ```
 
 Or generate predictions on new data:
@@ -89,11 +88,14 @@ classes = model.predict_classes(X_test, batch_size=32)
 proba = model.predict_proba(X_test, batch_size=32)
 ```
 
-Building a network of LSTMs, a deep CNN, a Neural Turing Machine, a word2vec embedder or any other model is just as fast. The ideas behind deep learning are simple, so why should their implementation be painful?
+Building a question answering system, an image classification model, a Neural Turing Machine, a word2vec embedder or any other model is just as fast. The ideas behind deep learning are simple, so why should their implementation be painful?
 
-Have a look at these [starter examples](http://keras.io/examples/).
+For a more in-depth tutorial about Keras, you can check out:
 
-In the [examples folder](https://github.com/fchollet/keras/tree/master/examples) of the repo, you will find more advanced models: question-answering with memory networks, text generation with stacked LSTMs, neural turing machines, etc.
+- [Getting started with the Sequential model](http://keras.io/getting-started/sequential-model-guide)
+- [Getting started with the functional API](http://keras.io/getting-started/functional-api-guide)
+
+In the [examples folder](https://github.com/fchollet/keras/tree/master/examples) of the repository, you will find more advanced models: question-answering with memory networks, text generation with stacked LSTMs, etc.
 
 
 ------------------
@@ -108,35 +110,33 @@ Keras uses the following dependencies:
 - HDF5 and h5py (optional, required if you use model saving/loading functions)
 - Optional but recommended if you use CNNs: cuDNN.
 
-When using the Theano backend:
-- Theano
-    - [See installation instructions](http://deeplearning.net/software/theano/install.html#install).
 
-**Note**: You should use the latest version of Theano, not the PyPI version. Install it with:
-```
-sudo pip install git+git://github.com/Theano/Theano.git
-```
+*When using the TensorFlow backend:*
 
-When using the TensorFlow backend:
 - TensorFlow
     - [See installation instructions](https://github.com/tensorflow/tensorflow#download-and-setup).
 
-To install, `cd` to the Keras folder and run the install command:
-```
+*When using the Theano backend:*
+
+- Theano
+    - [See installation instructions](http://deeplearning.net/software/theano/install.html#install).
+
+To install Keras, `cd` to the Keras folder and run the install command:
+```sh
 sudo python setup.py install
 ```
 
 You can also install Keras from PyPI:
-```
+```sh
 sudo pip install keras
 ```
 
 ------------------
 
 
-## Switching from Theano to TensorFlow
+## Switching from TensorFlow to Theano
 
-By default, Keras will use Theano as its tensor manipulation library. [Follow these instructions](http://keras.io/backend/) to configure the Keras backend.
+By default, Keras will use TensorFlow as its tensor manipulation library. [Follow these instructions](http://keras.io/backend/) to configure the Keras backend.
 
 ------------------
 
@@ -145,20 +145,7 @@ By default, Keras will use Theano as its tensor manipulation library. [Follow th
 
 You can ask questions and join the development discussion on the [Keras Google group](https://groups.google.com/forum/#!forum/keras-users).
 
-------------------
-
-
-## Contribution Guidelines
-
-Keras welcomes all contributions from the community. 
-
-- Keep a pragmatic mindset and avoid bloat. Only add to the source if that is the only path forward.
-- New features should be documented. Make sure you update the documentation along with your Pull Request.
-- Any new function or class should have a proper docstring.
-- The documentation for every new feature should include a usage example in the form of a code snippet. 
-- All changes should be tested. Make sure any new feature you add has a corresponding unit test.
-- Please no Pull Requests about coding style.
-- Even if you don't contribute to the Keras source code, if you have an application of Keras that is concise and powerful, please consider adding it to our collection of [examples](https://github.com/fchollet/keras/tree/master/examples).
+You can also post bug reports and feature requests in [Github issues](https://github.com/fchollet/keras/issues). Make sure to read [our guidelines](https://github.com/fchollet/keras/blob/master/CONTRIBUTING.md) first.
 
 
 ------------------
@@ -172,4 +159,4 @@ Keras was initially developed as part of the research effort of project ONEIROS
 
 >_"Oneiroi are beyond our unravelling --who can be sure what tale they tell? Not all that men look for comes to pass. Two gates there are that give passage to fleeting Oneiroi; one is made of horn, one of ivory. The Oneiroi that pass through sawn ivory are deceitful, bearing a message that will not be fulfilled; those that come out through polished horn have truth behind them, to be accomplished for men who see them."_ Homer, Odyssey 19. 562 ff (Shewring translation).
 
-------------------
+------------------

docs/templates/initializations.md

@@ -0,0 +1,50 @@
+
+## Usage of initializations
+
+Initializations define the way to set the initial random weights of Keras layers.
+
+The keyword arguments used for passing initializations to layers will depend on the layer. Usually it is simply `init`:
+
+```python
+model.add(Dense(64, init='uniform'))
+```
+
+## Available initializations
+
+- __uniform__
+- __lecun_uniform__: Uniform initialization scaled by the square root of the number of inputs (LeCun 98).
+- __normal__
+- __identity__: Use with square 2D layers (`shape[0] == shape[1]`).
+- __orthogonal__: Use with square 2D layers (`shape[0] == shape[1]`).
+- __zero__
+- __glorot_normal__: Gaussian initialization scaled by fan_in + fan_out (Glorot 2010)
+- __glorot_uniform__
+- __he_normal__: Gaussian initialization scaled by fan_in (He et al., 2014)
+- __he_uniform__
+
+
+An initialization may be passed as a string (must match one of the available initializations above), or as a callable.
+If a callable, then it must take two arguments: `shape` (shape of the variable to initialize) and `name` (name of the variable),
+and it must return a variable (e.g. output of `K.variable()`):
+
+```python
+from keras import backend as K
+import numpy as np
+
+def my_init(shape, name=None):
+    value = np.random.random(shape)
+    return K.variable(value, name=name)
+
+model.add(Dense(64, init=my_init))
+```
+
+You could also use functions from `keras.initializations` in this way:
+
+```python
+from keras import initializations
+
+def my_init(shape, name=None):
+    return initializations.normal(shape, scale=0.01, name=name)
+
+model.add(Dense(64, init=my_init))
+```

docs/templates/layers/about-keras-layers.md

@@ -0,0 +1,27 @@
+# About Keras layers
+
+All Keras layers have a number of methods in common:
+
+- `layer.get_weights()`: returns the weights of the layer as a list of Numpy arrays.
+- `layer.set_weights(weights)`: sets the weights of the layer from a list of Numpy arrays (with the same shapes as the output of `get_weights`).
+- `layer.get_config()`: returns a dictionary containing the configuration of the layer. The layer can be reinstantiated from its config via:
+```python
+from keras.utils.layer_utils import layer_from_config
+
+config = layer.get_config()
+layer = layer_from_config(config)
+```
+
+If a layer has a single node (i.e. if it isn't a shared layer), you can get its input tensor, output tensor, input shape and output shape via:
+
+- `layer.input`
+- `layer.output`
+- `layer.input_shape`
+- `layer.output_shape`
+
+If the layer has multiple nodes (see: [the concept of layer node and shared layers](/getting-started/functional-api-guide/#the-concept-of-layer-node)), you can use the following methods:
+
+- `layer.get_input_at(node_index)`
+- `layer.get_output_at(node_index)`
+- `layer.get_input_shape_at(node_index)`
+- `layer.get_output_shape_at(node_index)`

docs/templates/layers/writing-your-own-keras-layers.md

@@ -0,0 +1,34 @@
+# Writing your own Keras layers
+
+For simple, stateless custom operations, you are probably better off using `layers.core.Lambda` layers. But for any custom operation that has trainable weights, you should implement your own layer.
+
+Here is the skeleton of a Keras layer. There are only three methods you need to implement:
+
+- `build(input_shape)`: this is where you will define your weights. Trainable weights should be added to the list `self.trainable_weights`. Other attributes of note are: `self.non_trainable_weights` (list) and `self.updates` (list of update tuples (tensor, new_tensor)). For an example of how to use `non_trainable_weights` and `updates`, see the code for the `BatchNormalization` layer.
+- `call(x)`: this is where the layer's logic lives. Unless you want your layer to support masking, you only have to care about the first argument passed to `call`: the input tensor.
+- `get_output_shape_for(input_shape)`: in case your layer modifies the shape of its input, you should specify here the shape transformation logic. This allows Keras to do automatic shape inference.
+
+```python
+from keras import backend as K
+from keras.engine.topology import Layer
+import numpy as np
+
+class MyLayer(Layer):
+    def __init__(self, output_dim, **kwargs):
+        self.output_dim = output_dim
+        super(MyLayer, self).__init__(**kwargs)
+
+    def build(self, input_shape):
+        input_dim = input_shape[1]
+        initial_weight_value = np.random.random((input_dim, output_dim))
+        self.W = K.variable(initial_weight_value)
+        self.trainable_weights = [self.W]
+
+    def call(self, x, mask=None):
+        return K.dot(x, self.W)
+
+    def get_output_shape_for(self, input_shape):
+        return (input_shape[0], self.output_dim)
+```
+
+The existing Keras layers provide ample examples of how to implement almost anything. Never hesitate to read the source code!

docs/templates/models/about-keras-models.md

@@ -0,0 +1,33 @@
+# About Keras models
+
+There are two types of models available in Keras: [the Sequential model](/models/sequential) and [the Model class used with functional API](/models/model).
+
+These models have a number of methods in common:
+
+- `model.summary()`: prints a summary representation of your model.
+- `model.get_config()`: returns a dictionary containing the configuration of the model. The model can be reinstantiated from its config via:
+```python
+config = model.get_config()
+model = Model.from_config(config)
+# or, for Sequential:
+model = Sequential.from_config(config)
+```
+
+- `model.get_weights()`: returns a list of all weight tensors in the model, as Numpy arrays.
+- `model.set_weights(weights)`: sets the values of the weights of the model, from a list of Numpy arrays. The arrays in the list should have the same shape as those returned by `get_weights()`.
+- `model.to_json()`: returns a representation of the model as a JSON string. Note that the representation does not include the weights, only the architecture. You can reinstantiate the same model (with reinitialized weights) from the JSON string via:
+```python
+from models import model_from_json
+
+json_string = model.to_json()
+model = model_from_json(json_string)
+```
+- `model.to_yaml()`: returns a representation of the model as a YAML string. Note that the representation does not include the weights, only the architecture. You can reinstantiate the same model (with reinitialized weights) from the YAML string via:
+```python
+from models import model_from_yaml
+
+yaml_string = model.to_yaml()
+model = model_from_yaml(yaml_string)
+```
+- `model.save_weights(filepath)`: saves the weights of the model as a HDF5 file.
+- `model.load_weights(filepath, by_name=False)`: loads the weights of the model from a HDF5 file (created by `save_weights`). By default, the architecture is expected to be unchanged. To load weights into a different architecture (with some layers in common), use `by_name=True` to load only those layers with the same name.

docs/templates/models/model.md

@@ -0,0 +1,32 @@
+# Model class API
+
+In the functional API, given an input tensor and output tensor, you can instantiate a `Model` via:
+
+```python
+from keras.models import Model
+from keras.layers import Input, Dense
+
+a = Input(shape=(32,))
+b = Dense(32)(a)
+model = Model(input=a, output=b)
+```
+
+This model will include all layers required in the computation of `b` given `a`.
+
+In the case of multi-input or multi-output models, you can use lists as well:
+
+```python
+model = Model(input=[a1, a2], output=[b1, b3, b3])
+```
+
+For a detailed introduction of what `Model` can do, read [this guide to the Keras functional API](/getting-started/functional-api-guide).
+
+## Useful attributes of Model
+
+- `model.layers` is a flattened list of the layers comprising the model graph.
+- `model.inputs` is the list of input tensors.
+- `model.outputs` is the list of output tensors.
+
+## Methods
+
+{{autogenerated}}

docs/templates/models/sequential.md

@@ -0,0 +1,14 @@
+# The Sequential model API
+
+To get started, read [this guide to the Keras Sequential model](/getting-started/sequential-model-guide).
+
+## Useful attributes of Model
+
+- `model.layers` is a list of the layers added to the model.
+
+
+----
+
+## Sequential model methods
+
+{{autogenerated}}

docs/templates/objectives.md

@@ -7,10 +7,10 @@ An objective function (or loss function, or optimization score function) is one
 model.compile(loss='mean_squared_error', optimizer='sgd')
 ```
 
-You can either pass the name of an existing objective, or pass a Theano symbolic function that returns a scalar for each data-point and takes the following two arguments:
+You can either pass the name of an existing objective, or pass a Theano/TensorFlow symbolic function that returns a scalar for each data-point and takes the following two arguments:
 
-- __y_true__: True labels. Theano tensor.
-- __y_pred__: Predictions. Theano tensor of the same shape as y_true.
+- __y_true__: True labels. Theano/TensorFlow tensor.
+- __y_pred__: Predictions. Theano/TensorFlow tensor of the same shape as y_true.
 
 The actual optimized objective is the mean of the output array across all datapoints.
 
@@ -19,7 +19,6 @@ For a few examples of such functions, check out the [objectives source](https://
 ## Available objectives
 
 - __mean_squared_error__ / __mse__
-- __root_mean_squared_error__ / __rmse__
 - __mean_absolute_error__ / __mae__
 - __mean_absolute_percentage_error__ / __mape__
 - __mean_squared_logarithmic_error__ / __msle__
@@ -27,3 +26,7 @@ For a few examples of such functions, check out the [objectives source](https://
 - __hinge__
 - __binary_crossentropy__: Also known as logloss. 
 - __categorical_crossentropy__: Also known as multiclass logloss. __Note__: using this objective requires that your labels are binary arrays of shape `(nb_samples, nb_classes)`.
+- __sparse_categorical_crossentropy__: As above but accepts sparse labels. __Note__: this objective still requires that your labels have the same number of dimensions as your outputs; you may need to add a length-1 dimension to the shape of your labels, e.g with `np.expand_dims(y, -1)`.
+- __kullback_leibler_divergence__ / __kld__: Information gain from a predicted probability distribution Q to a true probability distribution P. Gives a measure of difference between both distributions.
+- __poisson__: Mean of `(predictions - targets * log(predictions))`
+- __cosine_proximity__: The opposite (negative) of the mean cosine proximity between predictions and targets.

docs/templates/optimizers.md

@@ -0,0 +1,44 @@
+
+## Usage of optimizers
+
+An optimizer is one of the two arguments required for compiling a Keras model:
+
+```python
+model = Sequential()
+model.add(Dense(64, init='uniform', input_dim=10))
+model.add(Activation('tanh'))
+model.add(Activation('softmax'))
+
+sgd = SGD(lr=0.01, decay=1e-6, momentum=0.9, nesterov=True)
+model.compile(loss='mean_squared_error', optimizer=sgd)
+```
+
+You can either instantiate an optimizer before passing it to `model.compile()` , as in the above example, or you can call it by its name. In the latter case, the default parameters for the optimizer will be used.
+
+```python
+# pass optimizer by name: default parameters will be used
+model.compile(loss='mean_squared_error', optimizer='sgd')
+```
+
+---
+
+## Parameters common to all Keras optimizers
+
+The parameters `clipnorm` and `clipvalue` can be used with all optimizers to control gradient clipping:
+
+```python
+# all parameter gradients will be clipped to
+# a maximum norm of 1.
+sgd = SGD(lr=0.01, clipnorm=1.)
+```
+
+```python
+# all parameter gradients will be clipped to
+# a maximum value of 0.5 and
+# a minimum value of -0.5.
+sgd = SGD(lr=0.01, clipvalue=0.5)
+```
+
+---
+
+{{autogenerated}}

docs/templates/preprocessing/image.md

@@ -0,0 +1,153 @@
+
+## ImageDataGenerator
+
+```python
+keras.preprocessing.image.ImageDataGenerator(featurewise_center=False,
+    samplewise_center=False,
+    featurewise_std_normalization=False,
+    samplewise_std_normalization=False,
+    zca_whitening=False,
+    rotation_range=0.,
+    width_shift_range=0.,
+    height_shift_range=0.,
+    shear_range=0.,
+    zoom_range=0.,
+    channel_shift_range=0.,
+    fill_mode='nearest',
+    cval=0.,
+    horizontal_flip=False,
+    vertical_flip=False,
+    rescale=None,
+    dim_ordering=K.image_dim_ordering())
+```
+
+Generate batches of tensor image data with real-time data augmentation. The data will be looped over (in batches) indefinitely.
+
+- __Arguments__:
+    - __featurewise_center__: Boolean. Set input mean to 0 over the dataset.
+    - __samplewise_center__: Boolean. Set each sample mean to 0.
+    - __featurewise_std_normalization__: Boolean. Divide inputs by std of the dataset.
+    - __samplewise_std_normalization__: Boolean. Divide each input by its std.
+    - __zca_whitening__: Boolean. Apply ZCA whitening.
+    - __rotation_range__: Int. Degree range for random rotations.
+    - __width_shift_range__: Float (fraction of total width). Range for random horizontal shifts.
+    - __height_shift_range__: Float (fraction of total height). Range for random vertical shifts.
+    - __shear_range__: Float. Shear Intensity (Shear angle in counter-clockwise direction as radians)
+    - __zoom_range__: Float or [lower, upper]. Range for random zoom. If a float, `[lower, upper] = [1-zoom_range, 1+zoom_range]`.
+    - __channel_shift_range__: Float. Range for random channel shifts.
+    - __fill_mode__: One of {"constant", "nearest", "reflect" or "wrap"}.  Points outside the boundaries of the input are filled according to the given mode.
+    - __cval__: Float or Int. Value used for points outside the boundaries when `fill_mode = "constant"`.
+    - __horizontal_flip__: Boolean. Randomly flip inputs horizontally.
+    - __vertical_flip__: Boolean. Randomly flip inputs vertically.
+    - __rescale__: rescaling factor. Defaults to None. If None or 0, no rescaling is applied,
+            otherwise we multiply the data by the value provided (before applying
+            any other transformation).
+    - __dim_ordering__: One of {"th", "tf"}.
+        "tf" mode means that the images should have shape `(samples, width, height, channels)`,
+        "th" mode means that the images should have shape `(samples, channels, width, height)`.
+        It defaults to the `image_dim_ordering` value found in your
+        Keras config file at `~/.keras/keras.json`.
+        If you never set it, then it will be "th".
+
+- __Methods__:
+    - __fit(X)__: Compute the internal data stats related to the data-dependent transformations, based on an array of sample data.
+        Only required if featurewise_center or featurewise_std_normalization or zca_whitening.
+        - __Arguments__:
+            - __X__: sample data.
+            - __augment__: Boolean (default: False). Whether to fit on randomly augmented samples.
+            - __rounds__: int (default: 1). If augment, how many augmentation passes over the data to use.
+    - __flow(X, y)__: Takes numpy data & label arrays, and generates batches of augmented/normalized data. Yields batches indefinitely, in an infinite loop.
+        - __Arguments__:
+            - __X__: data.
+            - __y__: labels.
+            - __batch_size__: int (default: 32).
+            - __shuffle__: boolean (defaut: True).
+            - __save_to_dir__: None or str (default: None). This allows you to optimally specify a directory to which to save the augmented pictures being generated (useful for visualizing what you are doing).
+            - __save_prefix__: str (default: `''`). Prefix to use for filenames of saved pictures (only relevant if `save_to_dir` is set).
+            - __save_format__: one of "png", "jpeg" (only relevant if `save_to_dir` is set). Default: "jpeg".
+        - __yields__: Tuples of `(x, y)` where `x` is a numpy array of image data and `y` is a numpy array of corresponding labels.
+            The generator loops indefinitely.
+    - __flow_from_directory(directory)__: Takes the path to a directory, and generates batches of augmented/normalized data. Yields batches indefinitely, in an infinite loop.
+        - __Arguments__:
+            - __directory__: path to the target directory. It should contain one subdirectory per class,
+                and the subdirectories should contain PNG or JPG images. See [this script](https://gist.github.com/fchollet/0830affa1f7f19fd47b06d4cf89ed44d) for more details.
+            - __target_size__: tuple of integers, default: `(256, 256)`. The dimensions to which all images found will be resized.
+            - __color_mode__: one of "grayscale", "rbg". Default: "rgb". Whether the images will be converted to have 1 or 3 color channels.
+            - __classes__: optional list of class subdirectories (e.g. `['dogs', 'cats']`). Default: None. If not provided, the list of classes will be automatically inferred (and the order of the classes, which will map to the label indices, will be alphanumeric).
+            - __class_mode__: one of "categorical", "binary", "sparse" or None. Default: "categorical". Determines the type of label arrays that are returned: "categorical" will be 2D one-hot encoded labels, "binary" will be 1D binary labels, "sparse" will be 1D integer labels. If None, no labels are returned (the generator will only yield batches of image data, which is useful to use `model.predict_generator()`, `model.evaluate_generator()`, etc.).
+            - __batch_size__: size of the batches of data (default: 32).
+            - __shuffle__: whether to shuffle the data (default: True)
+            - __seed__: optional random seed for shuffling.
+            - __save_to_dir__: None or str (default: None). This allows you to optimally specify a directory to which to save the augmented pictures being generated (useful for visualizing what you are doing).
+            - __save_prefix__: str. Prefix to use for filenames of saved pictures (only relevant if `save_to_dir` is set).
+            - __save_format__: one of "png", "jpeg" (only relevant if `save_to_dir` is set). Default: "jpeg".
+
+
+- __Examples__:
+
+Example of using `.flow(X, y)`:
+
+```python
+(X_train, y_train), (X_test, y_test) = cifar10.load_data()
+Y_train = np_utils.to_categorical(y_train, nb_classes)
+Y_test = np_utils.to_categorical(y_test, nb_classes)
+
+datagen = ImageDataGenerator(
+    featurewise_center=True,
+    featurewise_std_normalization=True,
+    rotation_range=20,
+    width_shift_range=0.2,
+    height_shift_range=0.2,
+    horizontal_flip=True)
+
+# compute quantities required for featurewise normalization
+# (std, mean, and principal components if ZCA whitening is applied)
+datagen.fit(X_train)
+
+# fits the model on batches with real-time data augmentation:
+model.fit_generator(datagen.flow(X_train, Y_train, batch_size=32),
+                    samples_per_epoch=len(X_train), nb_epoch=nb_epoch)
+
+# here's a more "manual" example
+for e in range(nb_epoch):
+    print 'Epoch', e
+    batches = 0
+    for X_batch, Y_batch in datagen.flow(X_train, Y_train, batch_size=32):
+        loss = model.train(X_batch, Y_batch)
+        batches += 1
+        if batches >= len(X_train) / 32:
+            # we need to break the loop by hand because
+            # the generator loops indefinitely
+            break
+```
+
+Example of using `.flow_from_directory(directory)`:
+
+```python
+train_datagen = ImageDataGenerator(
+        rescale=1./255,
+        shear_range=0.2,
+        zoom_range=0.2,
+        horizontal_flip=True)
+
+test_datagen = ImageDataGenerator(rescale=1./255)
+
+train_generator = train_datagen.flow_from_directory(
+        'data/train',
+        target_size=(150, 150),
+        batch_size=32,
+        class_mode='binary')
+
+validation_generator = test_datagen.flow_from_directory(
+        'data/validation',
+        target_size=(150, 150),
+        batch_size=32,
+        class_mode='binary')
+
+model.fit_generator(
+        train_generator,
+        samples_per_epoch=2000,
+        nb_epoch=50,
+        validation_data=validation_generator,
+        nb_val_samples=800)
+```

docs/templates/preprocessing/sequence.md

@@ -4,26 +4,29 @@
 keras.preprocessing.sequence.pad_sequences(sequences, maxlen=None, dtype='int32')
 ```
 
-Transform a list of `nb_samples sequences` (lists of scalars) into a 2D numpy array of shape `(nb_samples, nb_timesteps)`. `nb_timesteps` is either the `maxlen` argument if provided, or the length of the longest sequence otherwise. Sequences that are shorter than `nb_timesteps` are padded with zeros at the end.
+Transform a list of `nb_samples sequences` (lists of scalars) into a 2D Numpy array of shape `(nb_samples, nb_timesteps)`. `nb_timesteps` is either the `maxlen` argument if provided, or the length of the longest sequence otherwise. Sequences that are shorter than `nb_timesteps` are padded with zeros at the end.
 
-- __Return__: 2D numpy array of shape `(nb_samples, nb_timesteps)`.
+- __Return__: 2D Numpy array of shape `(nb_samples, nb_timesteps)`.
 
 - __Arguments__:
     - __sequences__: List of lists of int or float.
     - __maxlen__: None or int. Maximum sequence length, longer sequences are truncated and shorter sequences are padded with zeros at the end.
-    - __dtype__: datatype of the numpy array returned.
+    - __dtype__: datatype of the Numpy array returned.
+    - __padding__: 'pre' or 'post', pad either before or after each sequence.
+    - __truncating__: 'pre' or 'post', remove values from sequences larger than maxlen either in the beginning or in the end of the sequence
+    - __value__: float, value to pad the sequences to the desired value.
 
 ---
 
 ## skipgrams
 
 ```python
-keras.preprocessing.sequence.skipgrams(sequence, vocabulary_size, 
-    window_size=4, negative_samples=1., shuffle=True, 
+keras.preprocessing.sequence.skipgrams(sequence, vocabulary_size,
+    window_size=4, negative_samples=1., shuffle=True,
     categorical=False, sampling_table=None)
 ```
 
-Transforms a sequence of word indexes (list of int) into couples of the form: 
+Transforms a sequence of word indexes (list of int) into couples of the form:
 
 - (word, word in the same window), with label 1 (positive samples).
 - (word, random word from the vocabulary), with label 0 (negative samples).
@@ -31,8 +34,8 @@ Transforms a sequence of word indexes (list of int) into couples of the form:
 Read more about Skipgram in this gnomic paper by Mikolov et al.: [Efficient Estimation of Word Representations in
 Vector Space](http://arxiv.org/pdf/1301.3781v3.pdf)
 
-- __Return__: tuple `(couples, labels)`. 
-    - `couples` is a list of 2-elements lists of int: `[word_index, other_word_index]`. 
+- __Return__: tuple `(couples, labels)`.
+    - `couples` is a list of 2-elements lists of int: `[word_index, other_word_index]`.
     - `labels` is a list of 0 and 1, where 1 indicates that `other_word_index` was found in the same window as `word_index`, and 0 indicates that `other_word_index` was random.
     - if categorical is set to True, the labels are categorical, ie. 1 becomes [0,1], and 0 becomes [1, 0].
 
@@ -43,7 +46,7 @@ Vector Space](http://arxiv.org/pdf/1301.3781v3.pdf)
     - __negative_samples__: float >= 0. 0 for no negative (=random) samples. 1 for same number as positive samples. etc.
     - __shuffle__: boolean. Whether to shuffle the samples.
     - __categorical__: boolean. Whether to make the returned labels categorical.
-    - __sampling_table__: numpy array of shape `(vocabulary_size,)` where `sampling_table[i]` is the probability of sampling the word with index i (assumed to be i-th most common word in the dataset).
+    - __sampling_table__: Numpy array of shape `(vocabulary_size,)` where `sampling_table[i]` is the probability of sampling the word with index i (assumed to be i-th most common word in the dataset).
 
 
 ---
@@ -56,7 +59,7 @@ keras.preprocessing.sequence.make_sampling_table(size, sampling_factor=1e-5)
 
 Used for generating the `sampling_table` argument for `skipgrams`. `sampling_table[i]` is the probability of sampling the word i-th most common word in a dataset (more common words should be sampled less frequently, for balance).
 
-- __Return__: numpy array of shape `(size,)`.
+- __Return__: Numpy array of shape `(size,)`.
 
 - __Arguments__:
     - __size__: size of the vocabulary considered.

docs/templates/scikit-learn-api.md

@@ -0,0 +1,45 @@
+# Wrappers for the Scikit-Learn API
+
+You can use `Sequential` Keras models (single-input only) as part of your Scikit-Learn workflow via the wrappers found at `keras.wrappers.scikit_learn.py`.
+
+There are two wrappers available:
+
+`keras.wrappers.scikit_learn.KerasClassifier(build_fn=None, **sk_params)`, which implements the Scikit-Learn classifier interface,
+
+`keras.wrappers.scikit_learn.KerasRegressor(build_fn=None, **sk_params)`, which implements the Scikit-Learn regressor interface.
+
+### Arguments
+
+- __build_fn__: callable function or class instance
+- __sk_params__: model parameters & fitting parameters
+
+`build_fn` should construct, compile and return a Keras model, which
+will then be used to fit/predict. One of the following
+three values could be passed to build_fn:
+
+1. A function
+2. An instance of a class that implements the __call__ method
+3. None. This means you implement a class that inherits from either
+`KerasClassifier` or `KerasRegressor`. The __call__ method of the
+present class will then be treated as the default build_fn.
+
+`sk_params` takes both model parameters and fitting parameters. Legal model
+parameters are the arguments of `build_fn`. Note that like all other
+estimators in scikit-learn, 'build_fn' should provide default values for
+its arguments, so that you could create the estimator without passing any
+values to `sk_params`.
+
+`sk_params` could also accept parameters for calling `fit`, `predict`,
+`predict_proba`, and `score` methods (e.g., `nb_epoch`, `batch_size`).
+fitting (predicting) parameters are selected in the following order:
+
+1. Values passed to the dictionary arguments of
+`fit`, `predict`, `predict_proba`, and `score` methods
+2. Values passed to `sk_params`
+3. The default values of the `keras.models.Sequential`
+`fit`, `predict`, `predict_proba` and `score` methods
+
+When using scikit-learn's `grid_search` API, legal tunable parameters are
+those you could pass to `sk_params`, including fitting parameters.
+In other words, you could use `grid_search` to search for the best
+`batch_size` or `nb_epoch` as well as the model parameters.

docs/templates/visualization.md

@@ -10,11 +10,16 @@ from keras.utils.visualize_util import plot
 plot(model, to_file='model.png')
 ```
 
+`plot` takes two optional arguments:
+
+- `show_shapes` (defaults to False) controls whether output shapes are shown in the graph.
+- `show_layer_names` (defaults to True) controls whether layer names are shown in the graph.
+
 You can also directly obtain the `pydot.Graph` object and render it yourself,
 for example to show it in an ipython notebook :
 ```python
 from IPython.display import SVG
-from keras.utils.visualize_util import to_graph
+from keras.utils.visualize_util import model_to_dot
 
-SVG(to_graph(model).create(prog='dot', format='svg'))
+SVG(model_to_dot(model).create(prog='dot', format='svg'))
 ```

examples/addition_rnn.py

@@ -1,13 +1,5 @@
 # -*- coding: utf-8 -*-
-from __future__ import print_function
-from keras.models import Sequential, slice_X
-from keras.layers.core import Activation, TimeDistributedDense, RepeatVector
-from keras.layers import recurrent
-import numpy as np
-from six.moves import range
-
-"""
-An implementation of sequence to sequence learning for performing addition
+'''An implementation of sequence to sequence learning for performing addition
 Input: "535+61"
 Output: "596"
 Padding is handled by using a repeated sentinel character (space)
@@ -32,16 +24,23 @@ Four digits inverted:
 Five digits inverted:
 + One layer LSTM (128 HN), 550k training examples = 99% train/test accuracy in 30 epochs
 
-"""
+'''
+
+from __future__ import print_function
+from keras.models import Sequential
+from keras.engine.training import slice_X
+from keras.layers import Activation, TimeDistributed, Dense, RepeatVector, recurrent
+import numpy as np
+from six.moves import range
 
 
 class CharacterTable(object):
-    """
+    '''
     Given a set of characters:
     + Encode them to a one hot integer representation
     + Decode the one hot integer representation to their character output
-    + Decode a vector of probabilties to their character output
-    """
+    + Decode a vector of probabilities to their character output
+    '''
     def __init__(self, chars, maxlen):
         self.chars = sorted(set(chars))
         self.char_indices = dict((c, i) for i, c in enumerate(self.chars))
@@ -140,17 +139,20 @@ for _ in range(LAYERS):
     model.add(RNN(HIDDEN_SIZE, return_sequences=True))
 
 # For each of step of the output sequence, decide which character should be chosen
-model.add(TimeDistributedDense(len(chars)))
+model.add(TimeDistributed(Dense(len(chars))))
 model.add(Activation('softmax'))
 
-model.compile(loss='categorical_crossentropy', optimizer='adam')
+model.compile(loss='categorical_crossentropy',
+              optimizer='adam',
+              metrics=['accuracy'])
 
 # Train the model each generation and show predictions against the validation dataset
 for iteration in range(1, 200):
     print()
     print('-' * 50)
     print('Iteration', iteration)
-    model.fit(X_train, y_train, batch_size=BATCH_SIZE, nb_epoch=1, validation_data=(X_val, y_val), show_accuracy=True)
+    model.fit(X_train, y_train, batch_size=BATCH_SIZE, nb_epoch=1,
+              validation_data=(X_val, y_val))
     ###
     # Select 10 samples from the validation set at random so we can visualize errors
     for i in range(10):

examples/antirectifier.py

@@ -0,0 +1,104 @@
+'''The example demonstrates how to write custom layers for Keras.
+
+We build a custom activation layer called 'Antirectifier',
+which modifies the shape of the tensor that passes through it.
+We need to specify two methods: `get_output_shape_for` and `call`.
+
+Note that the same result can also be achieved via a Lambda layer.
+
+Because our custom layer is written with primitives from the Keras
+backend (`K`), our code can run both on TensorFlow and Theano.
+'''
+
+from __future__ import print_function
+from keras.models import Sequential
+from keras.layers import Dense, Dropout, Layer, Activation
+from keras.datasets import mnist
+from keras import backend as K
+from keras.utils import np_utils
+
+
+class Antirectifier(Layer):
+    '''This is the combination of a sample-wise
+    L2 normalization with the concatenation of the
+    positive part of the input with the negative part
+    of the input. The result is a tensor of samples that are
+    twice as large as the input samples.
+
+    It can be used in place of a ReLU.
+
+    # Input shape
+        2D tensor of shape (samples, n)
+
+    # Output shape
+        2D tensor of shape (samples, 2*n)
+
+    # Theoretical justification
+        When applying ReLU, assuming that the distribution
+        of the previous output is approximately centered around 0.,
+        you are discarding half of your input. This is inefficient.
+
+        Antirectifier allows to return all-positive outputs like ReLU,
+        without discarding any data.
+
+        Tests on MNIST show that Antirectifier allows to train networks
+        with twice less parameters yet with comparable
+        classification accuracy as an equivalent ReLU-based network.
+    '''
+    def get_output_shape_for(self, input_shape):
+        shape = list(input_shape)
+        assert len(shape) == 2  # only valid for 2D tensors
+        shape[-1] *= 2
+        return tuple(shape)
+
+    def call(self, x, mask=None):
+        x -= K.mean(x, axis=1, keepdims=True)
+        x = K.l2_normalize(x, axis=1)
+        pos = K.relu(x)
+        neg = K.relu(-x)
+        return K.concatenate([pos, neg], axis=1)
+
+# global parameters
+batch_size = 128
+nb_classes = 10
+nb_epoch = 40
+
+# the data, shuffled and split between train and test sets
+(X_train, y_train), (X_test, y_test) = mnist.load_data()
+
+X_train = X_train.reshape(60000, 784)
+X_test = X_test.reshape(10000, 784)
+X_train = X_train.astype('float32')
+X_test = X_test.astype('float32')
+X_train /= 255
+X_test /= 255
+print(X_train.shape[0], 'train samples')
+print(X_test.shape[0], 'test samples')
+
+# convert class vectors to binary class matrices
+Y_train = np_utils.to_categorical(y_train, nb_classes)
+Y_test = np_utils.to_categorical(y_test, nb_classes)
+
+# build the model
+model = Sequential()
+model.add(Dense(256, input_shape=(784,)))
+model.add(Antirectifier())
+model.add(Dropout(0.1))
+model.add(Dense(256))
+model.add(Antirectifier())
+model.add(Dropout(0.1))
+model.add(Dense(10))
+model.add(Activation('softmax'))
+
+# compile the model
+model.compile(loss='categorical_crossentropy',
+              optimizer='rmsprop',
+              metrics=['accuracy'])
+
+# train the model
+model.fit(X_train, Y_train,
+          batch_size=batch_size, nb_epoch=nb_epoch,
+          verbose=1, validation_data=(X_test, Y_test))
+
+# next, compare with an equivalent network
+# with2x bigger Dense layers and ReLU

examples/babi_memnn.py

@@ -1,30 +1,29 @@
-from __future__ import print_function
-from keras.models import Sequential
-from keras.layers.embeddings import Embedding
-from keras.layers.core import Activation, Dense, Merge, Permute, Dropout
-from keras.layers.recurrent import LSTM
-from keras.datasets.data_utils import get_file
-from keras.preprocessing.sequence import pad_sequences
-from functools import reduce
-import tarfile
-import numpy as np
-import re
-
-"""
-Train a memory network on the bAbI dataset.
+'''Trains a memory network on the bAbI dataset.
 
 References:
 - Jason Weston, Antoine Bordes, Sumit Chopra, Tomas Mikolov, Alexander M. Rush,
   "Towards AI-Complete Question Answering: A Set of Prerequisite Toy Tasks",
-  http://arxiv.org/abs/1503.08895
+  http://arxiv.org/abs/1502.05698
 
 - Sainbayar Sukhbaatar, Arthur Szlam, Jason Weston, Rob Fergus,
   "End-To-End Memory Networks",
   http://arxiv.org/abs/1503.08895
 
-Reaches 93% accuracy on task 'single_supporting_fact_10k' after 70 epochs.
+Reaches 98.6% accuracy on task 'single_supporting_fact_10k' after 120 epochs.
 Time per epoch: 3s on CPU (core i7).
-"""
+'''
+
+from __future__ import print_function
+from keras.models import Sequential
+from keras.layers.embeddings import Embedding
+from keras.layers import Activation, Dense, Merge, Permute, Dropout
+from keras.layers import LSTM
+from keras.utils.data_utils import get_file
+from keras.preprocessing.sequence import pad_sequences
+from functools import reduce
+import tarfile
+import numpy as np
+import re
 
 
 def tokenize(sent):
@@ -95,8 +94,13 @@ def vectorize_stories(data, word_idx, story_maxlen, query_maxlen):
             pad_sequences(Xq, maxlen=query_maxlen), np.array(Y))
 
 
-path = get_file('babi-tasks-v1-2.tar.gz',
-                origin='http://www.thespermwhale.com/jaseweston/babi/tasks_1-20_v1-2.tar.gz')
+try:
+    path = get_file('babi-tasks-v1-2.tar.gz', origin='https://s3.amazonaws.com/text-datasets/babi_tasks_1-20_v1-2.tar.gz')
+except:
+    print('Error downloading dataset, please download it manually:\n'
+          '$ wget http://www.thespermwhale.com/jaseweston/babi/tasks_1-20_v1-2.tar.gz\n'
+          '$ mv tasks_1-20_v1-2.tar.gz ~/.keras/datasets/babi-tasks-v1-2.tar.gz')
+    raise
 tar = tarfile.open(path)
 
 challenges = {
@@ -154,25 +158,29 @@ input_encoder_m = Sequential()
 input_encoder_m.add(Embedding(input_dim=vocab_size,
                               output_dim=64,
                               input_length=story_maxlen))
+input_encoder_m.add(Dropout(0.3))
 # output: (samples, story_maxlen, embedding_dim)
 # embed the question into a sequence of vectors
 question_encoder = Sequential()
 question_encoder.add(Embedding(input_dim=vocab_size,
                                output_dim=64,
                                input_length=query_maxlen))
+question_encoder.add(Dropout(0.3))
 # output: (samples, query_maxlen, embedding_dim)
 # compute a 'match' between input sequence elements (which are vectors)
 # and the question vector sequence
 match = Sequential()
 match.add(Merge([input_encoder_m, question_encoder],
                 mode='dot',
-                dot_axes=[(2,), (2,)]))
+                dot_axes=[2, 2]))
+match.add(Activation('softmax'))
 # output: (samples, story_maxlen, query_maxlen)
 # embed the input into a single vector with size = story_maxlen:
 input_encoder_c = Sequential()
 input_encoder_c.add(Embedding(input_dim=vocab_size,
                               output_dim=query_maxlen,
                               input_length=story_maxlen))
+input_encoder_c.add(Dropout(0.3))
 # output: (samples, story_maxlen, query_maxlen)
 # sum the match vector with the input vector:
 response = Sequential()
@@ -186,18 +194,17 @@ answer = Sequential()
 answer.add(Merge([response, question_encoder], mode='concat', concat_axis=-1))
 # the original paper uses a matrix multiplication for this reduction step.
 # we choose to use a RNN instead.
-answer.add(LSTM(64))
+answer.add(LSTM(32))
 # one regularization layer -- more would probably be needed.
-answer.add(Dropout(0.25))
+answer.add(Dropout(0.3))
 answer.add(Dense(vocab_size))
 # we output a probability distribution over the vocabulary
 answer.add(Activation('softmax'))
 
-answer.compile(optimizer='rmsprop', loss='categorical_crossentropy')
+answer.compile(optimizer='rmsprop', loss='categorical_crossentropy',
+               metrics=['accuracy'])
 # Note: you could use a Graph model to avoid repeat the input twice
 answer.fit([inputs_train, queries_train, inputs_train], answers_train,
            batch_size=32,
-           nb_epoch=70,
-           show_accuracy=True,
+           nb_epoch=120,
            validation_data=([inputs_test, queries_test, inputs_test], answers_test))
-

examples/babi_rnn.py

@@ -1,21 +1,4 @@
-from __future__ import absolute_import
-from __future__ import print_function
-from functools import reduce
-import re
-import tarfile
-
-import numpy as np
-np.random.seed(1337)  # for reproducibility
-
-from keras.datasets.data_utils import get_file
-from keras.layers.embeddings import Embedding
-from keras.layers.core import Dense, Merge
-from keras.layers import recurrent
-from keras.models import Sequential
-from keras.preprocessing.sequence import pad_sequences
-
-'''
-Trains two recurrent neural networks based upon a story and a question.
+'''Trains two recurrent neural networks based upon a story and a question.
 The resulting merged vector is then queried to answer a range of bAbI tasks.
 
 The results are comparable to those for an LSTM model provided in Weston et al.:
@@ -24,8 +7,8 @@ http://arxiv.org/abs/1502.05698
 
 Task Number                  | FB LSTM Baseline | Keras QA
 ---                          | ---              | ---
-QA1 - Single Supporting Fact | 50               | 52.1
-QA2 - Two Supporting Facts   | 20               | 37.0
+QA1 - Single Supporting Fact | 50               | 100.0
+QA2 - Two Supporting Facts   | 20               | 50.0
 QA3 - Three Supporting Facts | 20               | 20.5
 QA4 - Two Arg. Relations     | 61               | 62.9
 QA5 - Three Arg. Relations   | 70               | 61.9
@@ -51,8 +34,8 @@ https://research.facebook.com/researchers/1543934539189348
 Notes:
 
 - With default word, sentence, and query vector sizes, the GRU model achieves:
-  - 52.1% test accuracy on QA1 in 20 epochs (2 seconds per epoch on CPU)
-  - 37.0% test accuracy on QA2 in 20 epochs (16 seconds per epoch on CPU)
+  - 100% test accuracy on QA1 in 20 epochs (2 seconds per epoch on CPU)
+  - 50% test accuracy on QA2 in 20 epochs (16 seconds per epoch on CPU)
 In comparison, the Facebook paper achieves 50% and 20% for the LSTM baseline.
 
 - The task does not traditionally parse the question separately. This likely
@@ -73,6 +56,21 @@ noise to find the relevant statements, improving performance substantially.
 This becomes especially obvious on QA2 and QA3, both far longer than QA1.
 '''
 
+from __future__ import print_function
+from functools import reduce
+import re
+import tarfile
+
+import numpy as np
+np.random.seed(1337)  # for reproducibility
+
+from keras.utils.data_utils import get_file
+from keras.layers.embeddings import Embedding
+from keras.layers import Dense, Merge, Dropout, RepeatVector
+from keras.layers import recurrent
+from keras.models import Sequential
+from keras.preprocessing.sequence import pad_sequences
+
 
 def tokenize(sent):
     '''Return the tokens of a sentence including punctuation.
@@ -140,15 +138,21 @@ def vectorize_stories(data, word_idx, story_maxlen, query_maxlen):
         Y.append(y)
     return pad_sequences(X, maxlen=story_maxlen), pad_sequences(Xq, maxlen=query_maxlen), np.array(Y)
 
-RNN = recurrent.GRU
+RNN = recurrent.LSTM
 EMBED_HIDDEN_SIZE = 50
 SENT_HIDDEN_SIZE = 100
 QUERY_HIDDEN_SIZE = 100
 BATCH_SIZE = 32
-EPOCHS = 20
+EPOCHS = 40
 print('RNN / Embed / Sent / Query = {}, {}, {}, {}'.format(RNN, EMBED_HIDDEN_SIZE, SENT_HIDDEN_SIZE, QUERY_HIDDEN_SIZE))
 
-path = get_file('babi-tasks-v1-2.tar.gz', origin='http://www.thespermwhale.com/jaseweston/babi/tasks_1-20_v1-2.tar.gz')
+try:
+    path = get_file('babi-tasks-v1-2.tar.gz', origin='https://s3.amazonaws.com/text-datasets/babi_tasks_1-20_v1-2.tar.gz')
+except:
+    print('Error downloading dataset, please download it manually:\n'
+          '$ wget http://www.thespermwhale.com/jaseweston/babi/tasks_1-20_v1-2.tar.gz\n'
+          '$ mv tasks_1-20_v1-2.tar.gz ~/.keras/datasets/babi-tasks-v1-2.tar.gz')
+    raise
 tar = tarfile.open(path)
 # Default QA1 with 1000 samples
 # challenge = 'tasks_1-20_v1-2/en/qa1_single-supporting-fact_{}.txt'
@@ -180,20 +184,28 @@ print('story_maxlen, query_maxlen = {}, {}'.format(story_maxlen, query_maxlen))
 print('Build model...')
 
 sentrnn = Sequential()
-sentrnn.add(Embedding(vocab_size, EMBED_HIDDEN_SIZE, mask_zero=True))
-sentrnn.add(RNN(SENT_HIDDEN_SIZE, return_sequences=False))
+sentrnn.add(Embedding(vocab_size, EMBED_HIDDEN_SIZE,
+                      input_length=story_maxlen))
+sentrnn.add(Dropout(0.3))
 
 qrnn = Sequential()
-qrnn.add(Embedding(vocab_size, EMBED_HIDDEN_SIZE))
-qrnn.add(RNN(QUERY_HIDDEN_SIZE, return_sequences=False))
+qrnn.add(Embedding(vocab_size, EMBED_HIDDEN_SIZE,
+                   input_length=query_maxlen))
+qrnn.add(Dropout(0.3))
+qrnn.add(RNN(EMBED_HIDDEN_SIZE, return_sequences=False))
+qrnn.add(RepeatVector(story_maxlen))
 
 model = Sequential()
-model.add(Merge([sentrnn, qrnn], mode='concat'))
+model.add(Merge([sentrnn, qrnn], mode='sum'))
+model.add(RNN(EMBED_HIDDEN_SIZE, return_sequences=False))
+model.add(Dropout(0.3))
 model.add(Dense(vocab_size, activation='softmax'))
 
-model.compile(optimizer='adam', loss='categorical_crossentropy', class_mode='categorical')
+model.compile(optimizer='adam',
+              loss='categorical_crossentropy',
+              metrics=['accuracy'])
 
 print('Training')
-model.fit([X, Xq], Y, batch_size=BATCH_SIZE, nb_epoch=EPOCHS, validation_split=0.05, show_accuracy=True)
-loss, acc = model.evaluate([tX, tXq], tY, batch_size=BATCH_SIZE, show_accuracy=True)
+model.fit([X, Xq], Y, batch_size=BATCH_SIZE, nb_epoch=EPOCHS, validation_split=0.05)
+loss, acc = model.evaluate([tX, tXq], tY, batch_size=BATCH_SIZE)
 print('Test loss / test accuracy = {:.4f} / {:.4f}'.format(loss, acc))

examples/cifar10_cnn.py

@@ -1,27 +1,24 @@
-from __future__ import absolute_import
+'''Train a simple deep CNN on the CIFAR10 small images dataset.
+
+GPU run command:
+    THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 python cifar10_cnn.py
+
+It gets down to 0.65 test logloss in 25 epochs, and down to 0.55 after 50 epochs.
+(it's still underfitting at that point, though).
+
+Note: the data was pickled with Python 2, and some encoding issues might prevent you
+from loading it in Python 3. You might have to load it in Python 2,
+save it in a different format, load it in Python 3 and repickle it.
+'''
+
 from __future__ import print_function
 from keras.datasets import cifar10
 from keras.preprocessing.image import ImageDataGenerator
 from keras.models import Sequential
-from keras.layers.core import Dense, Dropout, Activation, Flatten
-from keras.layers.convolutional import Convolution2D, MaxPooling2D
-from keras.optimizers import SGD, Adadelta, Adagrad
-from keras.utils import np_utils, generic_utils
-from six.moves import range
-
-'''
-    Train a (fairly simple) deep CNN on the CIFAR10 small images dataset.
-
-    GPU run command:
-        THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 python cifar10_cnn.py
-
-    It gets down to 0.65 test logloss in 25 epochs, and down to 0.55 after 50 epochs.
-    (it's still underfitting at that point, though).
-
-    Note: the data was pickled with Python 2, and some encoding issues might prevent you
-    from loading it in Python 3. You might have to load it in Python 2,
-    save it in a different format, load it in Python 3 and repickle it.
-'''
+from keras.layers import Dense, Dropout, Activation, Flatten
+from keras.layers import Convolution2D, MaxPooling2D
+from keras.optimizers import SGD
+from keras.utils import np_utils
 
 batch_size = 32
 nb_classes = 10
@@ -46,7 +43,7 @@ Y_test = np_utils.to_categorical(y_test, nb_classes)
 model = Sequential()
 
 model.add(Convolution2D(32, 3, 3, border_mode='same',
-                        input_shape=(img_channels, img_rows, img_cols)))
+                        input_shape=X_train.shape[1:]))
 model.add(Activation('relu'))
 model.add(Convolution2D(32, 3, 3))
 model.add(Activation('relu'))
@@ -69,32 +66,35 @@ model.add(Activation('softmax'))
 
 # let's train the model using SGD + momentum (how original).
 sgd = SGD(lr=0.01, decay=1e-6, momentum=0.9, nesterov=True)
-model.compile(loss='categorical_crossentropy', optimizer=sgd)
+model.compile(loss='categorical_crossentropy',
+              optimizer=sgd,
+              metrics=['accuracy'])
 
-X_train = X_train.astype("float32")
-X_test = X_test.astype("float32")
+X_train = X_train.astype('float32')
+X_test = X_test.astype('float32')
 X_train /= 255
 X_test /= 255
 
 if not data_augmentation:
-    print("Not using data augmentation or normalization")
-    model.fit(X_train, Y_train, batch_size=batch_size, nb_epoch=nb_epoch)
-    score = model.evaluate(X_test, Y_test, batch_size=batch_size)
-    print('Test score:', score)
-
+    print('Not using data augmentation.')
+    model.fit(X_train, Y_train,
+              batch_size=batch_size,
+              nb_epoch=nb_epoch,
+              validation_data=(X_test, Y_test),
+              shuffle=True)
 else:
-    print("Using real time data augmentation")
+    print('Using real-time data augmentation.')
 
     # this will do preprocessing and realtime data augmentation
     datagen = ImageDataGenerator(
-        featurewise_center=True,  # set input mean to 0 over the dataset
+        featurewise_center=False,  # set input mean to 0 over the dataset
         samplewise_center=False,  # set each sample mean to 0
-        featurewise_std_normalization=True,  # divide inputs by std of the dataset
+        featurewise_std_normalization=False,  # divide inputs by std of the dataset
         samplewise_std_normalization=False,  # divide each input by its std
         zca_whitening=False,  # apply ZCA whitening
-        rotation_range=20,  # randomly rotate images in the range (degrees, 0 to 180)
-        width_shift_range=0.2,  # randomly shift images horizontally (fraction of total width)
-        height_shift_range=0.2,  # randomly shift images vertically (fraction of total height)
+        rotation_range=0,  # randomly rotate images in the range (degrees, 0 to 180)
+        width_shift_range=0.1,  # randomly shift images horizontally (fraction of total width)
+        height_shift_range=0.1,  # randomly shift images vertically (fraction of total height)
         horizontal_flip=True,  # randomly flip images
         vertical_flip=False)  # randomly flip images
 
@@ -102,20 +102,9 @@ else:
     # (std, mean, and principal components if ZCA whitening is applied)
     datagen.fit(X_train)
 
-    for e in range(nb_epoch):
-        print('-'*40)
-        print('Epoch', e)
-        print('-'*40)
-        print("Training...")
-        # batch train with realtime data augmentation
-        progbar = generic_utils.Progbar(X_train.shape[0])
-        for X_batch, Y_batch in datagen.flow(X_train, Y_train):
-            loss = model.train_on_batch(X_batch, Y_batch)
-            progbar.add(X_batch.shape[0], values=[("train loss", loss)])
-
-        print("Testing...")
-        # test time!
-        progbar = generic_utils.Progbar(X_test.shape[0])
-        for X_batch, Y_batch in datagen.flow(X_test, Y_test):
-            score = model.test_on_batch(X_batch, Y_batch)
-            progbar.add(X_batch.shape[0], values=[("test loss", score)])
+    # fit the model on the batches generated by datagen.flow()
+    model.fit_generator(datagen.flow(X_train, Y_train,
+                        batch_size=batch_size),
+                        samples_per_epoch=X_train.shape[0],
+                        nb_epoch=nb_epoch,
+                        validation_data=(X_test, Y_test))

examples/conv_filter_visualization.py

@@ -0,0 +1,132 @@
+'''Visualization of the filters of VGG16, via gradient ascent in input space.
+
+This script can run on CPU in a few minutes (with the TensorFlow backend).
+
+Results example: http://i.imgur.com/4nj4KjN.jpg
+'''
+from __future__ import print_function
+from scipy.misc import imsave
+import numpy as np
+import time
+from keras.applications import vgg16
+from keras import backend as K
+
+# dimensions of the generated pictures for each filter.
+img_width = 128
+img_height = 128
+
+# the name of the layer we want to visualize
+# (see model definition at keras/applications/vgg16.py)
+layer_name = 'block5_conv1'
+
+# util function to convert a tensor into a valid image
+def deprocess_image(x):
+    # normalize tensor: center on 0., ensure std is 0.1
+    x -= x.mean()
+    x /= (x.std() + 1e-5)
+    x *= 0.1
+
+    # clip to [0, 1]
+    x += 0.5
+    x = np.clip(x, 0, 1)
+
+    # convert to RGB array
+    x *= 255
+    if K.image_dim_ordering() == 'th':
+        x = x.transpose((1, 2, 0))
+    x = np.clip(x, 0, 255).astype('uint8')
+    return x
+
+# build the VGG16 network with ImageNet weights
+model = vgg16.VGG16(weights='imagenet', include_top=False)
+print('Model loaded.')
+
+model.summary()
+
+# this is the placeholder for the input images
+input_img = model.input
+
+# get the symbolic outputs of each "key" layer (we gave them unique names).
+layer_dict = dict([(layer.name, layer) for layer in model.layers[1:]])
+
+
+def normalize(x):
+    # utility function to normalize a tensor by its L2 norm
+    return x / (K.sqrt(K.mean(K.square(x))) + 1e-5)
+
+
+kept_filters = []
+for filter_index in range(0, 200):
+    # we only scan through the first 200 filters,
+    # but there are actually 512 of them
+    print('Processing filter %d' % filter_index)
+    start_time = time.time()
+
+    # we build a loss function that maximizes the activation
+    # of the nth filter of the layer considered
+    layer_output = layer_dict[layer_name].output
+    if K.image_dim_ordering() == 'th':
+        loss = K.mean(layer_output[:, filter_index, :, :])
+    else:
+        loss = K.mean(layer_output[:, :, :, filter_index])
+
+    # we compute the gradient of the input picture wrt this loss
+    grads = K.gradients(loss, input_img)[0]
+
+    # normalization trick: we normalize the gradient
+    grads = normalize(grads)
+
+    # this function returns the loss and grads given the input picture
+    iterate = K.function([input_img], [loss, grads])
+
+    # step size for gradient ascent
+    step = 1.
+
+    # we start from a gray image with some random noise
+    if K.image_dim_ordering() == 'th':
+        input_img_data = np.random.random((1, 3, img_width, img_height))
+    else:
+        input_img_data = np.random.random((1, img_width, img_height, 3))
+    input_img_data = (input_img_data - 0.5) * 20 + 128
+
+    # we run gradient ascent for 20 steps
+    for i in range(20):
+        loss_value, grads_value = iterate([input_img_data])
+        input_img_data += grads_value * step
+
+        print('Current loss value:', loss_value)
+        if loss_value <= 0.:
+            # some filters get stuck to 0, we can skip them
+            break
+
+    # decode the resulting input image
+    if loss_value > 0:
+        img = deprocess_image(input_img_data[0])
+        kept_filters.append((img, loss_value))
+    end_time = time.time()
+    print('Filter %d processed in %ds' % (filter_index, end_time - start_time))
+
+# we will stich the best 64 filters on a 8 x 8 grid.
+n = 8
+
+# the filters that have the highest loss are assumed to be better-looking.
+# we will only keep the top 64 filters.
+kept_filters.sort(key=lambda x: x[1], reverse=True)
+kept_filters = kept_filters[:n * n]
+
+# build a black picture with enough space for
+# our 8 x 8 filters of size 128 x 128, with a 5px margin in between
+margin = 5
+width = n * img_width + (n - 1) * margin
+height = n * img_height + (n - 1) * margin
+stitched_filters = np.zeros((width, height, 3))
+
+# fill the picture with our saved filters
+for i in range(n):
+    for j in range(n):
+        img, loss = kept_filters[i * n + j]
+        stitched_filters[(img_width + margin) * i: (img_width + margin) * i + img_width,
+                         (img_height + margin) * j: (img_height + margin) * j + img_height, :] = img
+
+# save the result to disk
+imsave('stitched_filters_%dx%d.png' % (n, n), stitched_filters)

examples/deep_dream.py

@@ -0,0 +1,233 @@
+'''Deep Dreaming in Keras.
+
+Run the script with:
+```
+python deep_dream.py path_to_your_base_image.jpg prefix_for_results
+```
+e.g.:
+```
+python deep_dream.py img/mypic.jpg results/dream
+```
+
+It is preferable to run this script on GPU, for speed.
+If running on CPU, prefer the TensorFlow backend (much faster).
+
+Example results: http://i.imgur.com/FX6ROg9.jpg
+'''
+from __future__ import print_function
+from scipy.misc import imread, imresize, imsave
+import numpy as np
+from scipy.optimize import fmin_l_bfgs_b
+import time
+import argparse
+import h5py
+import os
+
+from keras.models import Sequential
+from keras.layers import Convolution2D, ZeroPadding2D, MaxPooling2D
+from keras import backend as K
+
+parser = argparse.ArgumentParser(description='Deep Dreams with Keras.')
+parser.add_argument('base_image_path', metavar='base', type=str,
+                    help='Path to the image to transform.')
+parser.add_argument('result_prefix', metavar='res_prefix', type=str,
+                    help='Prefix for the saved results.')
+
+args = parser.parse_args()
+base_image_path = args.base_image_path
+result_prefix = args.result_prefix
+
+# dimensions of the generated picture.
+img_width = 600
+img_height = 600
+
+# path to the model weights file.
+weights_path = 'vgg16_weights.h5'
+
+# some settings we found interesting
+saved_settings = {
+    'bad_trip': {'features': {'conv4_1': 0.05,
+                              'conv4_2': 0.01,
+                              'conv4_3': 0.01},
+                 'continuity': 0.1,
+                 'dream_l2': 0.8,
+                 'jitter': 5},
+    'dreamy': {'features': {'conv5_1': 0.05,
+                            'conv5_2': 0.02},
+               'continuity': 0.1,
+               'dream_l2': 0.02,
+               'jitter': 0},
+}
+# the settings we will use in this experiment
+settings = saved_settings['dreamy']
+
+# util function to open, resize and format pictures into appropriate tensors
+def preprocess_image(image_path):
+    img = imresize(imread(image_path), (img_width, img_height))
+    img = img.transpose((2, 0, 1)).astype('float64')
+    img = np.expand_dims(img, axis=0)
+    return img
+
+# util function to convert a tensor into a valid image
+def deprocess_image(x):
+    x = x.transpose((1, 2, 0))
+    x = np.clip(x, 0, 255).astype('uint8')
+    return x
+
+# build the VGG16 network
+model = Sequential()
+model.add(ZeroPadding2D((1, 1), batch_input_shape=(1, 3, img_width, img_height)))
+first_layer = model.layers[-1]
+# this is a placeholder tensor that will contain our generated images
+dream = first_layer.input
+
+model.add(Convolution2D(64, 3, 3, activation='relu', name='conv1_1'))
+model.add(ZeroPadding2D((1, 1)))
+model.add(Convolution2D(64, 3, 3, activation='relu', name='conv1_2'))
+model.add(MaxPooling2D((2, 2), strides=(2, 2)))
+
+model.add(ZeroPadding2D((1, 1)))
+model.add(Convolution2D(128, 3, 3, activation='relu', name='conv2_1'))
+model.add(ZeroPadding2D((1, 1)))
+model.add(Convolution2D(128, 3, 3, activation='relu', name='conv2_2'))
+model.add(MaxPooling2D((2, 2), strides=(2, 2)))
+
+model.add(ZeroPadding2D((1, 1)))
+model.add(Convolution2D(256, 3, 3, activation='relu', name='conv3_1'))
+model.add(ZeroPadding2D((1, 1)))
+model.add(Convolution2D(256, 3, 3, activation='relu', name='conv3_2'))
+model.add(ZeroPadding2D((1, 1)))
+model.add(Convolution2D(256, 3, 3, activation='relu', name='conv3_3'))
+model.add(MaxPooling2D((2, 2), strides=(2, 2)))
+
+model.add(ZeroPadding2D((1, 1)))
+model.add(Convolution2D(512, 3, 3, activation='relu', name='conv4_1'))
+model.add(ZeroPadding2D((1, 1)))
+model.add(Convolution2D(512, 3, 3, activation='relu', name='conv4_2'))
+model.add(ZeroPadding2D((1, 1)))
+model.add(Convolution2D(512, 3, 3, activation='relu', name='conv4_3'))
+model.add(MaxPooling2D((2, 2), strides=(2, 2)))
+
+model.add(ZeroPadding2D((1, 1)))
+model.add(Convolution2D(512, 3, 3, activation='relu', name='conv5_1'))
+model.add(ZeroPadding2D((1, 1)))
+model.add(Convolution2D(512, 3, 3, activation='relu', name='conv5_2'))
+model.add(ZeroPadding2D((1, 1)))
+model.add(Convolution2D(512, 3, 3, activation='relu', name='conv5_3'))
+model.add(MaxPooling2D((2, 2), strides=(2, 2)))
+
+# load the weights of the VGG16 networks
+# (trained on ImageNet, won the ILSVRC competition in 2014)
+# note: when there is a complete match between your model definition
+# and your weight savefile, you can simply call model.load_weights(filename)
+assert os.path.exists(weights_path), 'Model weights not found (see "weights_path" variable in script).'
+f = h5py.File(weights_path)
+for k in range(f.attrs['nb_layers']):
+    if k >= len(model.layers):
+        # we don't look at the last (fully-connected) layers in the savefile
+        break
+    g = f['layer_{}'.format(k)]
+    weights = [g['param_{}'.format(p)] for p in range(g.attrs['nb_params'])]
+    model.layers[k].set_weights(weights)
+f.close()
+print('Model loaded.')
+
+# get the symbolic outputs of each "key" layer (we gave them unique names).
+layer_dict = dict([(layer.name, layer) for layer in model.layers])
+
+# continuity loss util function
+def continuity_loss(x):
+    assert K.ndim(x) == 4
+    a = K.square(x[:, :, :img_width-1, :img_height-1] - x[:, :, 1:, :img_height-1])
+    b = K.square(x[:, :, :img_width-1, :img_height-1] - x[:, :, :img_width-1, 1:])
+    return K.sum(K.pow(a + b, 1.25))
+
+# define the loss
+loss = K.variable(0.)
+for layer_name in settings['features']:
+    # add the L2 norm of the features of a layer to the loss
+    assert layer_name in layer_dict.keys(), 'Layer ' + layer_name + ' not found in model.'
+    coeff = settings['features'][layer_name]
+    x = layer_dict[layer_name].output
+    shape = layer_dict[layer_name].output_shape
+    # we avoid border artifacts by only involving non-border pixels in the loss
+    loss -= coeff * K.sum(K.square(x[:, :, 2: shape[2]-2, 2: shape[3]-2])) / np.prod(shape[1:])
+
+# add continuity loss (gives image local coherence, can result in an artful blur)
+loss += settings['continuity'] * continuity_loss(dream) / (3 * img_width * img_height)
+# add image L2 norm to loss (prevents pixels from taking very high values, makes image darker)
+loss += settings['dream_l2'] * K.sum(K.square(dream)) / (3 * img_width * img_height)
+
+# feel free to further modify the loss as you see fit, to achieve new effects...
+
+# compute the gradients of the dream wrt the loss
+grads = K.gradients(loss, dream)
+
+outputs = [loss]
+if type(grads) in {list, tuple}:
+    outputs += grads
+else:
+    outputs.append(grads)
+
+f_outputs = K.function([dream], outputs)
+def eval_loss_and_grads(x):
+    x = x.reshape((1, 3, img_width, img_height))
+    outs = f_outputs([x])
+    loss_value = outs[0]
+    if len(outs[1:]) == 1:
+        grad_values = outs[1].flatten().astype('float64')
+    else:
+        grad_values = np.array(outs[1:]).flatten().astype('float64')
+    return loss_value, grad_values
+
+# this Evaluator class makes it possible
+# to compute loss and gradients in one pass
+# while retrieving them via two separate functions,
+# "loss" and "grads". This is done because scipy.optimize
+# requires separate functions for loss and gradients,
+# but computing them separately would be inefficient.
+class Evaluator(object):
+    def __init__(self):
+        self.loss_value = None
+        self.grad_values = None
+
+    def loss(self, x):
+        assert self.loss_value is None
+        loss_value, grad_values = eval_loss_and_grads(x)
+        self.loss_value = loss_value
+        self.grad_values = grad_values
+        return self.loss_value
+
+    def grads(self, x):
+        assert self.loss_value is not None
+        grad_values = np.copy(self.grad_values)
+        self.loss_value = None
+        self.grad_values = None
+        return grad_values
+
+evaluator = Evaluator()
+
+# run scipy-based optimization (L-BFGS) over the pixels of the generated image
+# so as to minimize the loss
+x = preprocess_image(base_image_path)
+for i in range(5):
+    print('Start of iteration', i)
+    start_time = time.time()
+
+    # add a random jitter to the initial image. This will be reverted at decoding time
+    random_jitter = (settings['jitter'] * 2) * (np.random.random((3, img_width, img_height)) - 0.5)
+    x += random_jitter
+
+    # run L-BFGS for 7 steps
+    x, min_val, info = fmin_l_bfgs_b(evaluator.loss, x.flatten(),
+                                     fprime=evaluator.grads, maxfun=7)
+    print('Current loss value:', min_val)
+    # decode the dream and save it
+    x = x.reshape((3, img_width, img_height))
+    x -= random_jitter
+    img = deprocess_image(x)
+    fname = result_prefix + '_at_iteration_%d.png' % i
+    imsave(fname, img)
+    end_time = time.time()
+    print('Image saved as', fname)
+    print('Iteration %d completed in %ds' % (i, end_time - start_time))

examples/image_ocr.py

@@ -0,0 +1,470 @@
+'''This example uses a convolutional stack followed by a recurrent stack
+and a CTC logloss function to perform optical character recognition
+of generated text images. I have no evidence of whether it actually
+learns general shapes of text, or just is able to recognize all
+the different fonts thrown at it...the purpose is more to demonstrate CTC
+inside of Keras.  Note that the font list may need to be updated
+for the particular OS in use.
+
+This starts off with 4 letter words. After 10 or so epochs, CTC
+learns translational invariance, so longer words and groups of words
+with spaces are gradually fed in.  This gradual increase in difficulty
+is handled using the TextImageGenerator class which is both a generator
+class for test/train data and a Keras callback class. Every 10 epochs
+the wordlist that the generator draws from increases in difficulty.
+
+The table below shows normalized edit distance values. Theano uses
+a slightly different CTC implementation, so some Theano-specific
+hyperparameter tuning would be needed to get it to match Tensorflow.
+
+            Norm. ED
+Epoch |   TF   |   TH
+------------------------
+    10   0.072    0.272
+    20   0.032    0.115
+    30   0.024    0.098
+    40   0.023    0.108
+
+This requires cairo and editdistance packages:
+pip install cairocffi
+pip install editdistance
+
+Due to the use of a dummy loss function, Theano requires the following flags:
+on_unused_input='ignore'
+
+Created by Mike Henry
+https://github.com/mbhenry/
+'''
+
+import os
+import itertools
+import re
+import datetime
+import cairocffi as cairo
+import editdistance
+import numpy as np
+from scipy import ndimage
+import pylab
+from keras import backend as K
+from keras.layers.convolutional import Convolution2D, MaxPooling2D
+from keras.layers import Input, Layer, Dense, Activation, Flatten
+from keras.layers import Reshape, Lambda, merge, Permute, TimeDistributed
+from keras.models import Model
+from keras.layers.recurrent import GRU
+from keras.optimizers import SGD
+from keras.utils import np_utils
+from keras.utils.data_utils import get_file
+from keras.preprocessing import image
+import keras.callbacks
+
+OUTPUT_DIR = "image_ocr"
+
+np.random.seed(55)
+
+
+# this creates larger "blotches" of noise which look
+# more realistic than just adding gaussian noise
+# assumes greyscale with pixels ranging from 0 to 1
+
+def speckle(img):
+    severity = np.random.uniform(0, 0.6)
+    blur = ndimage.gaussian_filter(np.random.randn(*img.shape) * severity, 1)
+    img_speck = (img + blur)
+    img_speck[img_speck > 1] = 1
+    img_speck[img_speck <= 0] = 0
+    return img_speck
+
+
+# paints the string in a random location the bounding box
+# also uses a random font, a slight random rotation,
+# and a random amount of speckle noise
+
+def paint_text(text, w, h):
+    surface = cairo.ImageSurface(cairo.FORMAT_RGB24, w, h)
+    with cairo.Context(surface) as context:
+        context.set_source_rgb(1, 1, 1)  # White
+        context.paint()
+        # this font list works in Centos 7
+        fonts = ['Century Schoolbook', 'Courier', 'STIX', 'URW Chancery L', 'FreeMono']
+        context.select_font_face(np.random.choice(fonts), cairo.FONT_SLANT_NORMAL,
+                                 np.random.choice([cairo.FONT_WEIGHT_BOLD, cairo.FONT_WEIGHT_NORMAL]))
+        context.set_font_size(40)
+        box = context.text_extents(text)
+        if box[2] > w or box[3] > h:
+            raise IOError('Could not fit string into image. Max char count is too large for given image width.')
+
+        # teach the RNN translational invariance by
+        # fitting text box randomly on canvas, with some room to rotate
+        border_w_h = (10, 16)
+        max_shift_x = w - box[2] - border_w_h[0]
+        max_shift_y = h - box[3] - border_w_h[1]
+        top_left_x = np.random.randint(0, int(max_shift_x))
+        top_left_y = np.random.randint(0, int(max_shift_y))
+
+        context.move_to(top_left_x - int(box[0]), top_left_y - int(box[1]))
+        context.set_source_rgb(0, 0, 0)
+        context.show_text(text)
+
+    buf = surface.get_data()
+    a = np.frombuffer(buf, np.uint8)
+    a.shape = (h, w, 4)
+    a = a[:, :, 0]  # grab single channel
+    a /= 255
+    a = np.expand_dims(a, 0)
+    a = speckle(a)
+    a = image.random_rotation(a, 3 * (w - top_left_x) / w + 1)
+
+    return a
+
+
+def shuffle_mats_or_lists(matrix_list, stop_ind=None):
+    ret = []
+    assert all([len(i) == len(matrix_list[0]) for i in matrix_list])
+    len_val = len(matrix_list[0])
+    if stop_ind is None:
+        stop_ind = len_val
+    assert stop_ind <= len_val
+
+    a = range(stop_ind)
+    np.random.shuffle(a)
+    a += range(stop_ind, len_val)
+    for mat in matrix_list:
+        if isinstance(mat, np.ndarray):
+            ret.append(mat[a])
+        elif isinstance(mat, list):
+            ret.append([mat[i] for i in a])
+        else:
+            raise TypeError('shuffle_mats_or_lists only supports '
+                            'numpy.array and list objects')
+    return ret
+
+
+def text_to_labels(text, num_classes):
+    ret = []
+    for char in text:
+        if char >= 'a' and char <= 'z':
+            ret.append(ord(char) - ord('a'))
+        elif char == ' ':
+            ret.append(26)
+    return ret
+
+
+# only a-z and space..probably not to difficult
+# to expand to uppercase and symbols
+
+def is_valid_str(in_str):
+    search = re.compile(r'[^a-z\ ]').search
+    return not bool(search(in_str))
+
+
+# Uses generator functions to supply train/test with
+# data. Image renderings are text are created on the fly
+# each time with random perturbations
+
+class TextImageGenerator(keras.callbacks.Callback):
+
+    def __init__(self, monogram_file, bigram_file, minibatch_size,
+                 img_w, img_h, downsample_width, val_split,
+                 absolute_max_string_len=16):
+
+        self.minibatch_size = minibatch_size
+        self.img_w = img_w
+        self.img_h = img_h
+        self.monogram_file = monogram_file
+        self.bigram_file = bigram_file
+        self.downsample_width = downsample_width
+        self.val_split = val_split
+        self.blank_label = self.get_output_size() - 1
+        self.absolute_max_string_len = absolute_max_string_len
+
+    def get_output_size(self):
+        return 28
+
+    # num_words can be independent of the epoch size due to the use of generators
+    # as max_string_len grows, num_words can grow
+    def build_word_list(self, num_words, max_string_len=None, mono_fraction=0.5):
+        assert max_string_len <= self.absolute_max_string_len
+        assert num_words % self.minibatch_size == 0
+        assert (self.val_split * num_words) % self.minibatch_size == 0
+        self.num_words = num_words
+        self.string_list = []
+        self.max_string_len = max_string_len
+        self.Y_data = np.ones([self.num_words, self.absolute_max_string_len]) * -1
+        self.X_text = []
+        self.Y_len = [0] * self.num_words
+
+        # monogram file is sorted by frequency in english speech
+        with open(self.monogram_file, 'rt') as f:
+            for line in f:
+                if len(self.string_list) == int(self.num_words * mono_fraction):
+                    break
+                word = line.rstrip()
+                if max_string_len == -1 or max_string_len is None or len(word) <= max_string_len:
+                    self.string_list.append(word)
+
+        # bigram file contains common word pairings in english speech
+        with open(self.bigram_file, 'rt') as f:
+            lines = f.readlines()
+            for line in lines:
+                if len(self.string_list) == self.num_words:
+                    break
+                columns = line.lower().split()
+                word = columns[0] + ' ' + columns[1]
+                if is_valid_str(word) and \
+                        (max_string_len == -1 or max_string_len is None or len(word) <= max_string_len):
+                    self.string_list.append(word)
+        if len(self.string_list) != self.num_words:
+            raise IOError('Could not pull enough words from supplied monogram and bigram files. ')
+
+        for i, word in enumerate(self.string_list):
+            self.Y_len[i] = len(word)
+            self.Y_data[i, 0:len(word)] = text_to_labels(word, self.get_output_size())
+            self.X_text.append(word)
+        self.Y_len = np.expand_dims(np.array(self.Y_len), 1)
+
+        self.cur_val_index = self.val_split
+        self.cur_train_index = 0
+
+    # each time an image is requested from train/val/test, a new random
+    # painting of the text is performed
+    def get_batch(self, index, size, train):
+        if K.image_dim_ordering() == 'th':
+            X_data = np.ones([size, 1, self.img_h, self.img_w])
+        else:
+            X_data = np.ones([size, self.img_h, self.img_w, 1])
+        labels = np.ones([size, self.absolute_max_string_len])
+        input_length = np.zeros([size, 1])
+        label_length = np.zeros([size, 1])
+        source_str = []
+
+        for i in range(0, size):
+            # Mix in some blank inputs.  This seems to be important for
+            # achieving translational invariance
+            if train and i > size - 4:
+                if K.image_dim_ordering() == 'th':
+                    X_data[i, 0, :, :] = paint_text('', self.img_w, self.img_h)
+                else:
+                    X_data[i, :, :, 0] = paint_text('', self.img_w, self.img_h)
+                labels[i, 0] = self.blank_label
+                input_length[i] = self.downsample_width
+                label_length[i] = 1
+                source_str.append('')
+            else:
+                if K.image_dim_ordering() == 'th':
+                    X_data[i, 0, :, :] = paint_text(self.X_text[index + i], self.img_w, self.img_h)
+                else:
+                    X_data[i, :, :, 0] = paint_text(self.X_text[index + i], self.img_w, self.img_h)
+                labels[i, :] = self.Y_data[index + i]
+                input_length[i] = self.downsample_width
+                label_length[i] = self.Y_len[index + i]
+                source_str.append(self.X_text[index + i])
+
+        inputs = {'the_input': X_data,
+                  'the_labels': labels,
+                  'input_length': input_length,
+                  'label_length': label_length,
+                  'source_str': source_str  # used for visualization only
+                  }
+        outputs = {'ctc': np.zeros([size])}  # dummy data for dummy loss function
+        return (inputs, outputs)
+
+    def next_train(self):
+        while 1:
+            ret = self.get_batch(self.cur_train_index, self.minibatch_size, train=True)
+            self.cur_train_index += self.minibatch_size
+            if self.cur_train_index >= self.val_split:
+                self.cur_train_index = self.cur_train_index % 32
+                (self.X_text, self.Y_data, self.Y_len) = shuffle_mats_or_lists(
+                    [self.X_text, self.Y_data, self.Y_len], self.val_split)
+            yield ret
+
+    def next_val(self):
+        while 1:
+            ret = self.get_batch(self.cur_val_index, self.minibatch_size, train=False)
+            self.cur_val_index += self.minibatch_size
+            if self.cur_val_index >= self.num_words:
+                self.cur_val_index = self.val_split + self.cur_val_index % 32
+            yield ret
+
+    def on_train_begin(self, logs={}):
+        # translational invariance seems to be the hardest thing
+        # for the RNN to learn, so start with <= 4 letter words.
+        self.build_word_list(16000, 4, 1)
+
+    def on_epoch_begin(self, epoch, logs={}):
+        # After 10 epochs, translational invariance should be learned
+        # so start feeding longer words and eventually multiple words with spaces
+        if epoch == 10:
+            self.build_word_list(32000, 8, 1)
+        if epoch == 20:
+            self.build_word_list(32000, 8, 0.6)
+        if epoch == 30:
+            self.build_word_list(64000, 12, 0.5)
+
+
+# the actual loss calc occurs here despite it not being
+# an internal Keras loss function
+
+def ctc_lambda_func(args):
+    y_pred, labels, input_length, label_length = args
+    # the 2 is critical here since the first couple outputs of the RNN
+    # tend to be garbage:
+    y_pred = y_pred[:, 2:, :]
+    return K.ctc_batch_cost(labels, y_pred, input_length, label_length)
+
+
+# For a real OCR application, this should be beam search with a dictionary
+# and language model.  For this example, best path is sufficient.
+
+def decode_batch(test_func, word_batch):
+    out = test_func([word_batch])[0]
+    ret = []
+    for j in range(out.shape[0]):
+        out_best = list(np.argmax(out[j, 2:], 1))
+        out_best = [k for k, g in itertools.groupby(out_best)]
+        # 26 is space, 27 is CTC blank char
+        outstr = ''
+        for c in out_best:
+            if c >= 0 and c < 26:
+                outstr += chr(c + ord('a'))
+            elif c == 26:
+                outstr += ' '
+        ret.append(outstr)
+    return ret
+
+
+class VizCallback(keras.callbacks.Callback):
+
+    def __init__(self, test_func, text_img_gen, num_display_words=6):
+        self.test_func = test_func
+        self.output_dir = os.path.join(
+            OUTPUT_DIR, datetime.datetime.now().strftime('%A, %d. %B %Y %I.%M%p'))
+        self.text_img_gen = text_img_gen
+        self.num_display_words = num_display_words
+        os.makedirs(self.output_dir)
+
+    def show_edit_distance(self, num):
+        num_left = num
+        mean_norm_ed = 0.0
+        mean_ed = 0.0
+        while num_left > 0:
+            word_batch = next(self.text_img_gen)[0]
+            num_proc = min(word_batch['the_input'].shape[0], num_left)
+            decoded_res = decode_batch(self.test_func, word_batch['the_input'][0:num_proc])
+            for j in range(0, num_proc):
+                edit_dist = editdistance.eval(decoded_res[j], word_batch['source_str'][j])
+                mean_ed += float(edit_dist)
+                mean_norm_ed += float(edit_dist) / len(word_batch['source_str'][j])
+            num_left -= num_proc
+        mean_norm_ed = mean_norm_ed / num
+        mean_ed = mean_ed / num
+        print('\nOut of %d samples:  Mean edit distance: %.3f Mean normalized edit distance: %0.3f'
+              % (num, mean_ed, mean_norm_ed))
+
+    def on_epoch_end(self, epoch, logs={}):
+        self.model.save_weights(os.path.join(self.output_dir, 'weights%02d.h5' % epoch))
+        self.show_edit_distance(256)
+        word_batch = next(self.text_img_gen)[0]
+        res = decode_batch(self.test_func, word_batch['the_input'][0:self.num_display_words])
+
+        for i in range(self.num_display_words):
+            pylab.subplot(self.num_display_words, 1, i + 1)
+            if K.image_dim_ordering() == 'th':
+                the_input = word_batch['the_input'][i, 0, :, :]
+            else:
+                the_input = word_batch['the_input'][i, :, :, 0]
+            pylab.imshow(the_input, cmap='Greys_r')
+            pylab.xlabel('Truth = \'%s\' Decoded = \'%s\'' % (word_batch['source_str'][i], res[i]))
+        fig = pylab.gcf()
+        fig.set_size_inches(10, 12)
+        pylab.savefig(os.path.join(self.output_dir, 'e%02d.png' % epoch))
+        pylab.close()
+
+# Input Parameters
+img_h = 64
+img_w = 512
+nb_epoch = 50
+minibatch_size = 32
+words_per_epoch = 16000
+val_split = 0.2
+val_words = int(words_per_epoch * (val_split))
+
+# Network parameters
+conv_num_filters = 16
+filter_size = 3
+pool_size_1 = 4
+pool_size_2 = 2
+time_dense_size = 32
+rnn_size = 512
+time_steps = img_w / (pool_size_1 * pool_size_2)
+
+if K.image_dim_ordering() == 'th':
+    input_shape = (1, img_h, img_w)
+else:
+    input_shape = (img_h, img_w, 1)
+
+fdir = os.path.dirname(get_file('wordlists.tgz',
+                                origin='http://www.isosemi.com/datasets/wordlists.tgz', untar=True))
+
+img_gen = TextImageGenerator(monogram_file=os.path.join(fdir, 'wordlist_mono_clean.txt'),
+                             bigram_file=os.path.join(fdir, 'wordlist_bi_clean.txt'),
+                             minibatch_size=32,
+                             img_w=img_w,
+                             img_h=img_h,
+                             downsample_width=img_w / (pool_size_1 * pool_size_2) - 2,
+                             val_split=words_per_epoch - val_words)
+
+act = 'relu'
+input_data = Input(name='the_input', shape=input_shape, dtype='float32')
+inner = Convolution2D(conv_num_filters, filter_size, filter_size, border_mode='same',
+                      activation=act, name='conv1')(input_data)
+inner = MaxPooling2D(pool_size=(pool_size_1, pool_size_1), name='max1')(inner)
+inner = Convolution2D(conv_num_filters, filter_size, filter_size, border_mode='same',
+                      activation=act, name='conv2')(inner)
+inner = MaxPooling2D(pool_size=(pool_size_2, pool_size_2), name='max2')(inner)
+
+conv_to_rnn_dims = ((img_h / (pool_size_1 * pool_size_2)) * conv_num_filters, img_w / (pool_size_1 * pool_size_2))
+inner = Reshape(target_shape=conv_to_rnn_dims, name='reshape')(inner)
+inner = Permute(dims=(2, 1), name='permute')(inner)
+
+# cuts down input size going into RNN:
+inner = TimeDistributed(Dense(time_dense_size, activation=act, name='dense1'))(inner)
+
+# Two layers of bidirecitonal GRUs
+# GRU seems to work as well, if not better than LSTM:
+gru_1 = GRU(rnn_size, return_sequences=True, name='gru1')(inner)
+gru_1b = GRU(rnn_size, return_sequences=True, go_backwards=True, name='gru1_b')(inner)
+gru1_merged = merge([gru_1, gru_1b], mode='sum')
+gru_2 = GRU(rnn_size, return_sequences=True, name='gru2')(gru1_merged)
+gru_2b = GRU(rnn_size, return_sequences=True, go_backwards=True)(gru1_merged)
+
+# transforms RNN output to character activations:
+inner = TimeDistributed(Dense(img_gen.get_output_size(), name='dense2'))(merge([gru_2, gru_2b], mode='concat'))
+y_pred = Activation('softmax', name='softmax')(inner)
+Model(input=[input_data], output=y_pred).summary()
+
+labels = Input(name='the_labels', shape=[img_gen.absolute_max_string_len], dtype='float32')
+input_length = Input(name='input_length', shape=[1], dtype='int64')
+label_length = Input(name='label_length', shape=[1], dtype='int64')
+# Keras doesn't currently support loss funcs with extra parameters
+# so CTC loss is implemented in a lambda layer
+loss_out = Lambda(ctc_lambda_func, output_shape=(1,), name="ctc")([y_pred, labels, input_length, label_length])
+
+lr = 0.03
+# clipnorm seems to speeds up convergence
+clipnorm = 5
+sgd = SGD(lr=lr, decay=3e-7, momentum=0.9, nesterov=True, clipnorm=clipnorm)
+
+model = Model(input=[input_data, labels, input_length, label_length], output=[loss_out])
+
+# the loss calc occurs elsewhere, so use a dummy lambda func for the loss
+model.compile(loss={'ctc': lambda y_true, y_pred: y_pred}, optimizer=sgd)
+
+# captures output of softmax so we can decode the output during visualization
+test_func = K.function([input_data], [y_pred])
+
+viz_cb = VizCallback(test_func, img_gen.next_val())
+
+model.fit_generator(generator=img_gen.next_train(), samples_per_epoch=(words_per_epoch - val_words),
+                    nb_epoch=nb_epoch, validation_data=img_gen.next_val(), nb_val_samples=val_words,
+                    callbacks=[viz_cb, img_gen])

examples/imdb_bidirectional_lstm.py

@@ -1,33 +1,25 @@
-from __future__ import absolute_import
+'''Train a Bidirectional LSTM on the IMDB sentiment classification task.
+
+Output after 4 epochs on CPU: ~0.8146
+Time per epoch on CPU (Core i7): ~150s.
+'''
+
 from __future__ import print_function
 import numpy as np
 np.random.seed(1337)  # for reproducibility
 
 from keras.preprocessing import sequence
-from keras.utils.np_utils import accuracy
-from keras.models import Graph
-from keras.layers.core import Dense, Dropout
-from keras.layers.embeddings import Embedding
-from keras.layers.recurrent import LSTM
+from keras.models import Sequential
+from keras.layers import Dense, Dropout, Embedding, LSTM, Input, Bidirectional
 from keras.datasets import imdb
 
-'''
-    Train a Bidirectional LSTM on the IMDB sentiment classification task.
-
-    GPU command:
-        THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 python imdb_bidirectional_lstm.py
-
-    Output after 4 epochs on CPU: ~0.8146
-    Time per epoch on CPU (Core i7): ~150s.
-'''
 
 max_features = 20000
 maxlen = 100  # cut texts after this number of words (among top max_features most common words)
 batch_size = 32
 
-print("Loading data...")
-(X_train, y_train), (X_test, y_test) = imdb.load_data(nb_words=max_features,
-                                                      test_split=0.2)
+print('Loading data...')
+(X_train, y_train), (X_test, y_test) = imdb.load_data(nb_words=max_features)
 print(len(X_train), 'train sequences')
 print(len(X_test), 'test sequences')
 
@@ -39,25 +31,17 @@ print('X_test shape:', X_test.shape)
 y_train = np.array(y_train)
 y_test = np.array(y_test)
 
-print('Build model...')
-model = Graph()
-model.add_input(name='input', input_shape=(maxlen,), dtype=int)
-model.add_node(Embedding(max_features, 128, input_length=maxlen),
-               name='embedding', input='input')
-model.add_node(LSTM(64), name='forward', input='embedding')
-model.add_node(LSTM(64, go_backwards=True), name='backward', input='embedding')
-model.add_node(Dropout(0.5), name='dropout', inputs=['forward', 'backward'])
-model.add_node(Dense(1, activation='sigmoid'), name='sigmoid', input='dropout')
-model.add_output(name='output', input='sigmoid')
+model = Sequential()
+model.add(Embedding(max_features, 128, input_length=maxlen))
+model.add(Bidirectional(LSTM(64)))
+model.add(Dropout(0.5))
+model.add(Dense(1, activation='sigmoid'))
 
 # try using different optimizers and different optimizer configs
-model.compile('adam', {'output': 'binary_crossentropy'})
+model.compile('adam', 'binary_crossentropy', metrics=['accuracy'])
 
-print("Train...")
-model.fit({'input': X_train, 'output': y_train},
+print('Train...')
+model.fit(X_train, y_train,
           batch_size=batch_size,
-          nb_epoch=4)
-acc = accuracy(y_test,
-               np.round(np.array(model.predict({'input': X_test},
-                                               batch_size=batch_size)['output'])))
-print('Test accuracy:', acc)
+          nb_epoch=4,
+          validation_data=[X_test, y_test])

examples/imdb_cnn.py

@@ -1,41 +1,40 @@
-from __future__ import absolute_import
+'''This example demonstrates the use of Convolution1D for text classification.
+
+Gets to 0.89 test accuracy after 2 epochs.
+90s/epoch on Intel i5 2.4Ghz CPU.
+10s/epoch on Tesla K40 GPU.
+
+'''
+
 from __future__ import print_function
 import numpy as np
 np.random.seed(1337)  # for reproducibility
 
 from keras.preprocessing import sequence
 from keras.models import Sequential
-from keras.layers.core import Dense, Dropout, Activation, Flatten
-from keras.layers.embeddings import Embedding
-from keras.layers.convolutional import Convolution1D, MaxPooling1D
+from keras.layers import Dense, Dropout, Activation, Flatten
+from keras.layers import Embedding
+from keras.layers import Convolution1D, MaxPooling1D
 from keras.datasets import imdb
+from keras import backend as K
 
-'''
-    This example demonstrates the use of Convolution1D
-    for text classification.
-
-    Run on GPU: THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 python imdb_cnn.py
-
-    Get to 0.835 test accuracy after 2 epochs. 100s/epoch on K520 GPU.
-'''
 
 # set parameters:
 max_features = 5000
-maxlen = 100
+maxlen = 400
 batch_size = 32
-embedding_dims = 100
+embedding_dims = 50
 nb_filter = 250
 filter_length = 3
 hidden_dims = 250
 nb_epoch = 2
 
-print("Loading data...")
-(X_train, y_train), (X_test, y_test) = imdb.load_data(nb_words=max_features,
-                                                      test_split=0.2)
+print('Loading data...')
+(X_train, y_train), (X_test, y_test) = imdb.load_data(nb_words=max_features)
 print(len(X_train), 'train sequences')
 print(len(X_test), 'test sequences')
 
-print("Pad sequences (samples x time)")
+print('Pad sequences (samples x time)')
 X_train = sequence.pad_sequences(X_train, maxlen=maxlen)
 X_test = sequence.pad_sequences(X_test, maxlen=maxlen)
 print('X_train shape:', X_train.shape)
@@ -46,18 +45,20 @@ model = Sequential()
 
 # we start off with an efficient embedding layer which maps
 # our vocab indices into embedding_dims dimensions
-model.add(Embedding(max_features, embedding_dims, input_length=maxlen))
-model.add(Dropout(0.25))
+model.add(Embedding(max_features,
+                    embedding_dims,
+                    input_length=maxlen,
+                    dropout=0.2))
 
 # we add a Convolution1D, which will learn nb_filter
 # word group filters of size filter_length:
 model.add(Convolution1D(nb_filter=nb_filter,
                         filter_length=filter_length,
-                        border_mode="valid",
-                        activation="relu",
+                        border_mode='valid',
+                        activation='relu',
                         subsample_length=1))
-# we use standard max pooling (halving the output of the previous layer):
-model.add(MaxPooling1D(pool_length=2))
+# we use max pooling:
+model.add(MaxPooling1D(pool_length=model.output_shape[1]))
 
 # We flatten the output of the conv layer,
 # so that we can add a vanilla dense layer:
@@ -65,7 +66,7 @@ model.add(Flatten())
 
 # We add a vanilla hidden layer:
 model.add(Dense(hidden_dims))
-model.add(Dropout(0.25))
+model.add(Dropout(0.2))
 model.add(Activation('relu'))
 
 # We project onto a single unit output layer, and squash it with a sigmoid:
@@ -73,8 +74,9 @@ model.add(Dense(1))
 model.add(Activation('sigmoid'))
 
 model.compile(loss='binary_crossentropy',
-              optimizer='rmsprop',
-              class_mode="binary")
-model.fit(X_train, y_train, batch_size=batch_size,
-          nb_epoch=nb_epoch, show_accuracy=True,
+              optimizer='adam',
+              metrics=['accuracy'])
+model.fit(X_train, y_train,
+          batch_size=batch_size,
+          nb_epoch=nb_epoch,
           validation_data=(X_test, y_test))

examples/imdb_cnn_lstm.py

@@ -1,26 +1,20 @@
-from __future__ import absolute_import
+'''Train a recurrent convolutional network on the IMDB sentiment
+classification task.
+
+Gets to 0.8498 test accuracy after 2 epochs. 41s/epoch on K520 GPU.
+'''
 from __future__ import print_function
 import numpy as np
 np.random.seed(1337)  # for reproducibility
 
 from keras.preprocessing import sequence
-from keras.optimizers import SGD, RMSprop, Adagrad
-from keras.utils import np_utils
 from keras.models import Sequential
-from keras.layers.core import Dense, Dropout, Activation
-from keras.layers.embeddings import Embedding
-from keras.layers.recurrent import LSTM, GRU, SimpleRNN
-from keras.layers.convolutional import Convolution1D, MaxPooling1D
+from keras.layers import Dense, Dropout, Activation
+from keras.layers import Embedding
+from keras.layers import LSTM, GRU, SimpleRNN
+from keras.layers import Convolution1D, MaxPooling1D
 from keras.datasets import imdb
 
-'''
-    Train a recurrent convolutional network on the IMDB sentiment classification task.
-
-    GPU command:
-        THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 python imdb_lstm.py
-
-    Get to 0.8498 test accuracy after 2 epochs. 41s/epoch on K520 GPU.
-'''
 
 # Embedding
 max_features = 20000
@@ -28,9 +22,9 @@ maxlen = 100
 embedding_size = 128
 
 # Convolution
-filter_length = 3
+filter_length = 5
 nb_filter = 64
-pool_length = 2
+pool_length = 4
 
 # LSTM
 lstm_output_size = 70
@@ -45,12 +39,12 @@ batch_size is highly sensitive.
 Only 2 epochs are needed as the dataset is very small.
 '''
 
-print("Loading data...")
-(X_train, y_train), (X_test, y_test) = imdb.load_data(nb_words=max_features, test_split=0.2)
+print('Loading data...')
+(X_train, y_train), (X_test, y_test) = imdb.load_data(nb_words=max_features)
 print(len(X_train), 'train sequences')
 print(len(X_test), 'test sequences')
 
-print("Pad sequences (samples x time)")
+print('Pad sequences (samples x time)')
 X_train = sequence.pad_sequences(X_train, maxlen=maxlen)
 X_test = sequence.pad_sequences(X_test, maxlen=maxlen)
 print('X_train shape:', X_train.shape)
@@ -63,8 +57,8 @@ model.add(Embedding(max_features, embedding_size, input_length=maxlen))
 model.add(Dropout(0.25))
 model.add(Convolution1D(nb_filter=nb_filter,
                         filter_length=filter_length,
-                        border_mode="valid",
-                        activation="relu",
+                        border_mode='valid',
+                        activation='relu',
                         subsample_length=1))
 model.add(MaxPooling1D(pool_length=pool_length))
 model.add(LSTM(lstm_output_size))
@@ -73,12 +67,11 @@ model.add(Activation('sigmoid'))
 
 model.compile(loss='binary_crossentropy',
               optimizer='adam',
-              class_mode="binary")
+              metrics=['accuracy'])
 
-print("Train...")
+print('Train...')
 model.fit(X_train, y_train, batch_size=batch_size, nb_epoch=nb_epoch,
-          validation_data=(X_test, y_test), show_accuracy=True)
-score, acc = model.evaluate(X_test, y_test, batch_size=batch_size,
-                            show_accuracy=True)
+          validation_data=(X_test, y_test))
+score, acc = model.evaluate(X_test, y_test, batch_size=batch_size)
 print('Test score:', score)
 print('Test accuracy:', acc)

examples/imdb_fasttext.py

@@ -0,0 +1,139 @@
+'''This example demonstrates the use of fasttext for text classification
+
+Based on Joulin et al's paper:
+
+Bags of Tricks for Efficient Text Classification
+https://arxiv.org/abs/1607.01759
+
+Results on IMDB datasets with uni and bi-gram embeddings:
+    Uni-gram: 0.8813 test accuracy after 5 epochs. 15s/epoch on i7 cpu.
+    Bi-gram : 0.9056 test accuracy after 5 epochs. 5s/epoch on GTX 1080 gpu.
+'''
+
+from __future__ import print_function
+import numpy as np
+np.random.seed(1337)  # for reproducibility
+
+from keras.preprocessing import sequence
+from keras.models import Sequential
+from keras.layers import Dense, Flatten
+from keras.layers import Embedding
+from keras.layers import AveragePooling1D
+from keras.datasets import imdb
+
+
+def create_ngram_set(input_list, ngram_value=2):
+    """
+    Extract a set of n-grams from a list of integers.
+
+    >>> create_ngram_set([1, 4, 9, 4, 1, 4], ngram_value=2)
+    {(4, 9), (4, 1), (1, 4), (9, 4)}
+
+    >>> create_ngram_set([1, 4, 9, 4, 1, 4], ngram_value=3)
+    [(1, 4, 9), (4, 9, 4), (9, 4, 1), (4, 1, 4)]
+    """
+    return set(zip(*[input_list[i:] for i in range(ngram_value)]))
+
+
+def add_ngram(sequences, token_indice, ngram_range=2):
+    """
+    Augment the input list of list (sequences) by appending n-grams values.
+
+    Example: adding bi-gram
+    >>> sequences = [[1, 3, 4, 5], [1, 3, 7, 9, 2]]
+    >>> token_indice = {(1, 3): 1337, (9, 2): 42, (4, 5): 2017}
+    >>> add_ngram(sequences, token_indice, ngram_range=2)
+    [[1, 3, 4, 5, 1337, 2017], [1, 3, 7, 9, 2, 1337, 42]]
+
+    Example: adding tri-gram
+    >>> sequences = [[1, 3, 4, 5], [1, 3, 7, 9, 2]]
+    >>> token_indice = {(1, 3): 1337, (9, 2): 42, (4, 5): 2017, (7, 9, 2): 2018}
+    >>> add_ngram(sequences, token_indice, ngram_range=3)
+    [[1, 3, 4, 5, 1337], [1, 3, 7, 9, 2, 1337, 2018]]
+    """
+    new_sequences = []
+    for input_list in sequences:
+        new_list = input_list[:]
+        for i in range(len(new_list)-ngram_range+1):
+            for ngram_value in range(2, ngram_range+1):
+                ngram = tuple(new_list[i:i+ngram_value])
+                if ngram in token_indice:
+                    new_list.append(token_indice[ngram])
+        new_sequences.append(new_list)
+
+    return new_sequences
+
+# Set parameters:
+# ngram_range = 2 will add bi-grams features
+ngram_range = 1
+max_features = 20000
+maxlen = 400
+batch_size = 32
+embedding_dims = 50
+nb_epoch = 5
+
+print('Loading data...')
+(X_train, y_train), (X_test, y_test) = imdb.load_data(nb_words=max_features)
+print(len(X_train), 'train sequences')
+print(len(X_test), 'test sequences')
+print('Average train sequence length: {}'.format(np.mean(list(map(len, X_train)), dtype=int)))
+print('Average test sequence length: {}'.format(np.mean(list(map(len, X_test)), dtype=int)))
+
+if ngram_range > 1:
+    print('Adding {}-gram features'.format(ngram_range))
+    # Create set of unique n-gram from the training set.
+    ngram_set = set()
+    for input_list in X_train:
+        for i in range(2, ngram_range+1):
+            set_of_ngram = create_ngram_set(input_list, ngram_value=i)
+            ngram_set.update(set_of_ngram)
+
+    # Dictionary mapping n-gram token to a unique integer.
+    # Integer values are greater than max_features in order
+    # to avoid collision with existing features.
+    start_index = max_features + 1
+    token_indice = {v: k+start_index for k, v in enumerate(ngram_set)}
+    indice_token = {token_indice[k]: k for k in token_indice}
+
+    # max_features is the highest integer that could be found in the dataset.
+    max_features = np.max(list(indice_token.keys())) + 1
+
+    # Augmenting X_train and X_test with n-grams features
+    X_train = add_ngram(X_train, token_indice, ngram_range)
+    X_test = add_ngram(X_test, token_indice, ngram_range)
+    print('Average train sequence length: {}'.format(np.mean(list(map(len, X_train)), dtype=int)))
+    print('Average test sequence length: {}'.format(np.mean(list(map(len, X_test)), dtype=int)))
+
+print('Pad sequences (samples x time)')
+X_train = sequence.pad_sequences(X_train, maxlen=maxlen)
+X_test = sequence.pad_sequences(X_test, maxlen=maxlen)
+print('X_train shape:', X_train.shape)
+print('X_test shape:', X_test.shape)
+
+print('Build model...')
+model = Sequential()
+
+# we start off with an efficient embedding layer which maps
+# our vocab indices into embedding_dims dimensions
+model.add(Embedding(max_features,
+                    embedding_dims,
+                    input_length=maxlen))
+
+# we add a AveragePooling1D, which will average the embeddings
+# of all words in the document
+model.add(AveragePooling1D(pool_length=model.output_shape[1]))
+
+# We flatten the output of the AveragePooling1D layer
+model.add(Flatten())
+
+# We project onto a single unit output layer, and squash it with a sigmoid:
+model.add(Dense(1, activation='sigmoid'))
+
+model.compile(loss='binary_crossentropy',
+              optimizer='adam',
+              metrics=['accuracy'])
+
+model.fit(X_train, y_train,
+          batch_size=batch_size,
+          nb_epoch=nb_epoch,
+          validation_data=(X_test, y_test))

examples/imdb_lstm.py

@@ -1,4 +1,15 @@
-from __future__ import absolute_import
+'''Trains a LSTM on the IMDB sentiment classification task.
+The dataset is actually too small for LSTM to be of any advantage
+compared to simpler, much faster methods such as TF-IDF + LogReg.
+Notes:
+
+- RNNs are tricky. Choice of batch size is important,
+choice of loss and optimizer is critical, etc.
+Some configurations won't converge.
+
+- LSTM loss decrease patterns during training can be quite different
+from what you see with CNNs/MLPs/etc.
+'''
 from __future__ import print_function
 import numpy as np
 np.random.seed(1337)  # for reproducibility
@@ -6,41 +17,20 @@ np.random.seed(1337)  # for reproducibility
 from keras.preprocessing import sequence
 from keras.utils import np_utils
 from keras.models import Sequential
-from keras.layers.core import Dense, Dropout, Activation
-from keras.layers.embeddings import Embedding
-from keras.layers.recurrent import LSTM
+from keras.layers import Dense, Dropout, Activation, Embedding
+from keras.layers import LSTM, SimpleRNN, GRU
 from keras.datasets import imdb
 
-'''
-    Train a LSTM on the IMDB sentiment classification task.
-
-    The dataset is actually too small for LSTM to be of any advantage
-    compared to simpler, much faster methods such as TF-IDF+LogReg.
-
-    Notes:
-
-    - RNNs are tricky. Choice of batch size is important,
-    choice of loss and optimizer is critical, etc.
-    Some configurations won't converge.
-
-    - LSTM loss decrease patterns during training can be quite different
-    from what you see with CNNs/MLPs/etc.
-
-    GPU command:
-        THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 python imdb_lstm.py
-'''
-
 max_features = 20000
-maxlen = 100  # cut texts after this number of words (among top max_features most common words)
+maxlen = 80  # cut texts after this number of words (among top max_features most common words)
 batch_size = 32
 
-print("Loading data...")
-(X_train, y_train), (X_test, y_test) = imdb.load_data(nb_words=max_features,
-                                                      test_split=0.2)
+print('Loading data...')
+(X_train, y_train), (X_test, y_test) = imdb.load_data(nb_words=max_features)
 print(len(X_train), 'train sequences')
 print(len(X_test), 'test sequences')
 
-print("Pad sequences (samples x time)")
+print('Pad sequences (samples x time)')
 X_train = sequence.pad_sequences(X_train, maxlen=maxlen)
 X_test = sequence.pad_sequences(X_test, maxlen=maxlen)
 print('X_train shape:', X_train.shape)
@@ -48,22 +38,20 @@ print('X_test shape:', X_test.shape)
 
 print('Build model...')
 model = Sequential()
-model.add(Embedding(max_features, 128, input_length=maxlen))
-model.add(LSTM(128))  # try using a GRU instead, for fun
-model.add(Dropout(0.5))
+model.add(Embedding(max_features, 128, dropout=0.2))
+model.add(LSTM(128, dropout_W=0.2, dropout_U=0.2))  # try using a GRU instead, for fun
 model.add(Dense(1))
 model.add(Activation('sigmoid'))
 
 # try using different optimizers and different optimizer configs
 model.compile(loss='binary_crossentropy',
               optimizer='adam',
-              class_mode="binary")
+              metrics=['accuracy'])
 
-print("Train...")
-model.fit(X_train, y_train, batch_size=batch_size, nb_epoch=3,
-          validation_data=(X_test, y_test), show_accuracy=True)
+print('Train...')
+model.fit(X_train, y_train, batch_size=batch_size, nb_epoch=15,
+          validation_data=(X_test, y_test))
 score, acc = model.evaluate(X_test, y_test,
-                            batch_size=batch_size,
-                            show_accuracy=True)
+                            batch_size=batch_size)
 print('Test score:', score)
 print('Test accuracy:', acc)

examples/kaggle_otto_nn.py

@@ -1,123 +0,0 @@
-from __future__ import absolute_import
-from __future__ import print_function
-
-import numpy as np
-import pandas as pd
-np.random.seed(1337)  # for reproducibility
-
-from keras.models import Sequential
-from keras.layers.core import Dense, Dropout, Activation
-from keras.layers.normalization import BatchNormalization
-from keras.layers.advanced_activations import PReLU
-from keras.utils import np_utils, generic_utils
-
-from sklearn.preprocessing import LabelEncoder
-from sklearn.preprocessing import StandardScaler
-
-'''
-    This demonstrates how to reach a score of 0.4890 (local validation)
-    on the Kaggle Otto challenge, with a deep net using Keras.
-
-    Compatible Python 2.7-3.4. Requires Scikit-Learn and Pandas.
-
-    Recommended to run on GPU:
-        Command: THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 python kaggle_otto_nn.py
-        On EC2 g2.2xlarge instance: 19s/epoch. 6-7 minutes total training time.
-
-    Best validation score at epoch 21: 0.4881
-
-    Try it at home:
-        - with/without BatchNormalization (BatchNormalization helps!)
-        - with ReLU or with PReLU (PReLU helps!)
-        - with smaller layers, largers layers
-        - with more layers, less layers
-        - with different optimizers (SGD+momentum+decay is probably better than Adam!)
-
-    Get the data from Kaggle: https://www.kaggle.com/c/otto-group-product-classification-challenge/data
-'''
-
-
-def load_data(path, train=True):
-    df = pd.read_csv(path)
-    X = df.values.copy()
-    if train:
-        np.random.shuffle(X)  # https://youtu.be/uyUXoap67N8
-        X, labels = X[:, 1:-1].astype(np.float32), X[:, -1]
-        return X, labels
-    else:
-        X, ids = X[:, 1:].astype(np.float32), X[:, 0].astype(str)
-        return X, ids
-
-
-def preprocess_data(X, scaler=None):
-    if not scaler:
-        scaler = StandardScaler()
-        scaler.fit(X)
-    X = scaler.transform(X)
-    return X, scaler
-
-
-def preprocess_labels(labels, encoder=None, categorical=True):
-    if not encoder:
-        encoder = LabelEncoder()
-        encoder.fit(labels)
-    y = encoder.transform(labels).astype(np.int32)
-    if categorical:
-        y = np_utils.to_categorical(y)
-    return y, encoder
-
-
-def make_submission(y_prob, ids, encoder, fname):
-    with open(fname, 'w') as f:
-        f.write('id,')
-        f.write(','.join([str(i) for i in encoder.classes_]))
-        f.write('\n')
-        for i, probs in zip(ids, y_prob):
-            probas = ','.join([i] + [str(p) for p in probs.tolist()])
-            f.write(probas)
-            f.write('\n')
-    print("Wrote submission to file {}.".format(fname))
-
-print("Loading data...")
-X, labels = load_data('train.csv', train=True)
-X, scaler = preprocess_data(X)
-y, encoder = preprocess_labels(labels)
-
-X_test, ids = load_data('test.csv', train=False)
-X_test, _ = preprocess_data(X_test, scaler)
-
-nb_classes = y.shape[1]
-print(nb_classes, 'classes')
-
-dims = X.shape[1]
-print(dims, 'dims')
-
-print("Building model...")
-
-model = Sequential()
-model.add(Dense(512, input_shape=(dims,)))
-model.add(PReLU())
-model.add(BatchNormalization())
-model.add(Dropout(0.5))
-
-model.add(Dense(512))
-model.add(PReLU())
-model.add(BatchNormalization())
-model.add(Dropout(0.5))
-
-model.add(Dense(512))
-model.add(PReLU())
-model.add(BatchNormalization())
-model.add(Dropout(0.5))
-
-model.add(Dense(nb_classes))
-model.add(Activation('softmax'))
-
-model.compile(loss='categorical_crossentropy', optimizer="adam")
-
-print("Training model...")
-model.fit(X, y, nb_epoch=20, batch_size=128, validation_split=0.15)
-
-print("Generating submission...")
-proba = model.predict_proba(X_test)
-make_submission(proba, ids, encoder, fname='keras-otto.csv')

examples/lstm_benchmark.py

@@ -0,0 +1,83 @@
+'''Compare LSTM implementations on the IMDB sentiment classification task.
+
+consume_less='cpu' preprocesses input to the LSTM which typically results in
+faster computations at the expense of increased peak memory usage as the
+preprocessed input must be kept in memory.
+
+consume_less='mem' does away with the preprocessing, meaning that it might take
+a little longer, but should require less peak memory.
+
+consume_less='gpu' concatenates the input, output and forget gate's weights
+into one, large matrix, resulting in faster computation time as the GPU can
+utilize more cores, at the expense of reduced regularization because the same
+dropout is shared across the gates.
+
+Note that the relative performance of the different `consume_less` modes
+can vary depending on your device, your model and the size of your data.
+'''
+
+import time
+import numpy as np
+import matplotlib.pyplot as plt
+
+from keras.preprocessing import sequence
+from keras.models import Sequential
+from keras.layers import Embedding, Dense, LSTM
+from keras.datasets import imdb
+
+max_features = 20000
+max_length = 80
+embedding_dim = 256
+batch_size = 128
+epochs = 10
+modes = ['cpu', 'mem', 'gpu']
+
+print('Loading data...')
+(X_train, y_train), (X_test, y_test) = imdb.load_data(nb_words=max_features)
+X_train = sequence.pad_sequences(X_train, max_length)
+X_test = sequence.pad_sequences(X_test, max_length)
+
+# Compile and train different models while meauring performance.
+results = []
+for mode in modes:
+    print('Testing mode: consume_less="{}"'.format(mode))
+
+    model = Sequential()
+    model.add(Embedding(max_features, embedding_dim, input_length=max_length, dropout=0.2))
+    model.add(LSTM(embedding_dim, dropout_W=0.2, dropout_U=0.2, consume_less=mode))
+    model.add(Dense(1, activation='sigmoid'))
+    model.compile(loss='binary_crossentropy',
+                  optimizer='adam',
+                  metrics=['accuracy'])
+
+    start_time = time.time()
+    history = model.fit(X_train, y_train,
+                        batch_size=batch_size,
+                        nb_epoch=epochs,
+                        validation_data=(X_test, y_test))
+    average_time_per_epoch = (time.time() - start_time) / epochs
+
+    results.append((history, average_time_per_epoch))
+
+# Compare models' accuracy, loss and elapsed time per epoch.
+plt.style.use('ggplot')
+ax1 = plt.subplot2grid((2, 2), (0, 0))
+ax1.set_title('Accuracy')
+ax1.set_ylabel('Validation Accuracy')
+ax1.set_xlabel('Epochs')
+ax2 = plt.subplot2grid((2, 2), (1, 0))
+ax2.set_title('Loss')
+ax2.set_ylabel('Validation Loss')
+ax2.set_xlabel('Epochs')
+ax3 = plt.subplot2grid((2, 2), (0, 1), rowspan=2)
+ax3.set_title('Time')
+ax3.set_ylabel('Seconds')
+for mode, result in zip(modes, results):
+    ax1.plot(result[0].epoch, result[0].history['val_acc'], label=mode)
+    ax2.plot(result[0].epoch, result[0].history['val_loss'], label=mode)
+ax1.legend()
+ax2.legend()
+ax3.bar(np.arange(len(results)), [x[1] for x in results],
+        tick_label=modes, align='center')
+plt.tight_layout()
+plt.show()

examples/lstm_text_generation.py

@@ -1,36 +1,36 @@
+'''Example script to generate text from Nietzsche's writings.
+
+At least 20 epochs are required before the generated text
+starts sounding coherent.
+
+It is recommended to run this script on GPU, as recurrent
+networks are quite computationally intensive.
+
+If you try this script on new data, make sure your corpus
+has at least ~100k characters. ~1M is better.
+'''
+
 from __future__ import print_function
 from keras.models import Sequential
-from keras.layers.core import Dense, Activation, Dropout
-from keras.layers.recurrent import LSTM
-from keras.datasets.data_utils import get_file
+from keras.layers import Dense, Activation, Dropout
+from keras.layers import LSTM
+from keras.optimizers import RMSprop
+from keras.utils.data_utils import get_file
 import numpy as np
 import random
 import sys
 
-'''
-    Example script to generate text from Nietzsche's writings.
-
-    At least 20 epochs are required before the generated text
-    starts sounding coherent.
-
-    It is recommended to run this script on GPU, as recurrent
-    networks are quite computationally intensive.
-
-    If you try this script on new data, make sure your corpus
-    has at least ~100k characters. ~1M is better.
-'''
-
 path = get_file('nietzsche.txt', origin="https://s3.amazonaws.com/text-datasets/nietzsche.txt")
 text = open(path).read().lower()
 print('corpus length:', len(text))
 
-chars = set(text)
+chars = sorted(list(set(text)))
 print('total chars:', len(chars))
 char_indices = dict((c, i) for i, c in enumerate(chars))
 indices_char = dict((i, c) for i, c in enumerate(chars))
 
 # cut the text in semi-redundant sequences of maxlen characters
-maxlen = 20
+maxlen = 40
 step = 3
 sentences = []
 next_chars = []
@@ -48,24 +48,25 @@ for i, sentence in enumerate(sentences):
     y[i, char_indices[next_chars[i]]] = 1
 
 
-# build the model: 2 stacked LSTM
+# build the model: a single LSTM
 print('Build model...')
 model = Sequential()
-model.add(LSTM(512, return_sequences=True, input_shape=(maxlen, len(chars))))
-model.add(Dropout(0.2))
-model.add(LSTM(512, return_sequences=False))
-model.add(Dropout(0.2))
+model.add(LSTM(128, input_shape=(maxlen, len(chars))))
 model.add(Dense(len(chars)))
 model.add(Activation('softmax'))
 
-model.compile(loss='categorical_crossentropy', optimizer='rmsprop')
+optimizer = RMSprop(lr=0.01)
+model.compile(loss='categorical_crossentropy', optimizer=optimizer)
 
 
-def sample(a, temperature=1.0):
+def sample(preds, temperature=1.0):
     # helper function to sample an index from a probability array
-    a = np.log(a) / temperature
-    a = np.exp(a) / np.sum(np.exp(a))
-    return np.argmax(np.random.multinomial(1, a, 1))
+    preds = np.asarray(preds).astype('float64')
+    preds = np.log(preds) / temperature
+    exp_preds = np.exp(preds)
+    preds = exp_preds / np.sum(exp_preds)
+    probas = np.random.multinomial(1, preds, 1)
+    return np.argmax(probas)
 
 # train the model, output generated text after each iteration
 for iteration in range(1, 60):
@@ -86,7 +87,7 @@ for iteration in range(1, 60):
         print('----- Generating with seed: "' + sentence + '"')
         sys.stdout.write(generated)
 
-        for iteration in range(400):
+        for i in range(400):
             x = np.zeros((1, maxlen, len(chars)))
             for t, char in enumerate(sentence):
                 x[0, t, char_indices[char]] = 1.

examples/mnist_cnn.py

@@ -1,22 +1,20 @@
-from __future__ import absolute_import
+'''Trains a simple convnet on the MNIST dataset.
+
+Gets to 99.25% test accuracy after 12 epochs
+(there is still a lot of margin for parameter tuning).
+16 seconds per epoch on a GRID K520 GPU.
+'''
+
 from __future__ import print_function
 import numpy as np
 np.random.seed(1337)  # for reproducibility
 
 from keras.datasets import mnist
 from keras.models import Sequential
-from keras.layers.core import Dense, Dropout, Activation, Flatten
-from keras.layers.convolutional import Convolution2D, MaxPooling2D
+from keras.layers import Dense, Dropout, Activation, Flatten
+from keras.layers import Convolution2D, MaxPooling2D
 from keras.utils import np_utils
-
-'''
-    Train a simple convnet on the MNIST dataset.
-
-    Run on GPU: THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 python mnist_cnn.py
-
-    Get to 99.25% test accuracy after 12 epochs (there is still a lot of margin for parameter tuning).
-    16 seconds per epoch on a GRID K520 GPU.
-'''
+from keras import backend as K
 
 batch_size = 128
 nb_classes = 10
@@ -27,17 +25,24 @@ img_rows, img_cols = 28, 28
 # number of convolutional filters to use
 nb_filters = 32
 # size of pooling area for max pooling
-nb_pool = 2
+pool_size = (2, 2)
 # convolution kernel size
-nb_conv = 3
+kernel_size = (3, 3)
 
-# the data, shuffled and split between tran and test sets
+# the data, shuffled and split between train and test sets
 (X_train, y_train), (X_test, y_test) = mnist.load_data()
 
-X_train = X_train.reshape(X_train.shape[0], 1, img_rows, img_cols)
-X_test = X_test.reshape(X_test.shape[0], 1, img_rows, img_cols)
-X_train = X_train.astype("float32")
-X_test = X_test.astype("float32")
+if K.image_dim_ordering() == 'th':
+    X_train = X_train.reshape(X_train.shape[0], 1, img_rows, img_cols)
+    X_test = X_test.reshape(X_test.shape[0], 1, img_rows, img_cols)
+    input_shape = (1, img_rows, img_cols)
+else:
+    X_train = X_train.reshape(X_train.shape[0], img_rows, img_cols, 1)
+    X_test = X_test.reshape(X_test.shape[0], img_rows, img_cols, 1)
+    input_shape = (img_rows, img_cols, 1)
+
+X_train = X_train.astype('float32')
+X_test = X_test.astype('float32')
 X_train /= 255
 X_test /= 255
 print('X_train shape:', X_train.shape)
@@ -50,13 +55,13 @@ Y_test = np_utils.to_categorical(y_test, nb_classes)
 
 model = Sequential()
 
-model.add(Convolution2D(nb_filters, nb_conv, nb_conv,
-                        border_mode='same',
-                        input_shape=(1, img_rows, img_cols)))
+model.add(Convolution2D(nb_filters, kernel_size[0], kernel_size[1],
+                        border_mode='valid',
+                        input_shape=input_shape))
 model.add(Activation('relu'))
-model.add(Convolution2D(nb_filters, nb_conv, nb_conv))
+model.add(Convolution2D(nb_filters, kernel_size[0], kernel_size[1]))
 model.add(Activation('relu'))
-model.add(MaxPooling2D(pool_size=(nb_pool, nb_pool)))
+model.add(MaxPooling2D(pool_size=pool_size))
 model.add(Dropout(0.25))
 
 model.add(Flatten())
@@ -66,9 +71,12 @@ model.add(Dropout(0.5))
 model.add(Dense(nb_classes))
 model.add(Activation('softmax'))
 
-model.compile(loss='categorical_crossentropy', optimizer='adadelta')
+model.compile(loss='categorical_crossentropy',
+              optimizer='adadelta',
+              metrics=['accuracy'])
 
-model.fit(X_train, Y_train, batch_size=batch_size, nb_epoch=nb_epoch, show_accuracy=True, verbose=1, validation_data=(X_test, Y_test))
-score = model.evaluate(X_test, Y_test, show_accuracy=True, verbose=0)
+model.fit(X_train, Y_train, batch_size=batch_size, nb_epoch=nb_epoch,
+          verbose=1, validation_data=(X_test, Y_test))
+score = model.evaluate(X_test, Y_test, verbose=0)
 print('Test score:', score[0])
 print('Test accuracy:', score[1])

examples/mnist_hierarchical_rnn.py

@@ -0,0 +1,87 @@
+"""This is an example of using Hierarchical RNN (HRNN) to classify MNIST digits.
+
+HRNNs can learn across multiple levels of temporal hiearchy over a complex sequence.
+Usually, the first recurrent layer of an HRNN encodes a sentence (e.g. of word vectors)
+into a  sentence vector. The second recurrent layer then encodes a sequence of
+such vectors (encoded by the first layer) into a document vector. This
+document vector is considered to preserve both the word-level and
+sentence-level structure of the context.
+
+# References
+    - [A Hierarchical Neural Autoencoder for Paragraphs and Documents](https://web.stanford.edu/~jurafsky/pubs/P15-1107.pdf)
+        Encodes paragraphs and documents with HRNN.
+        Results have shown that HRNN outperforms standard
+        RNNs and may play some role in more sophisticated generation tasks like
+        summarization or question answering.
+    - [Hierarchical recurrent neural network for skeleton based action recognition](http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7298714)
+        Achieved state-of-the-art results on skeleton based action recognition with 3 levels
+        of bidirectional HRNN combined with fully connected layers.
+
+In the below MNIST example the first LSTM layer first encodes every
+column of pixels of shape (28, 1) to a column vector of shape (128,). The second LSTM
+layer encodes then these 28 column vectors of shape (28, 128) to a image vector
+representing the whole image. A final Dense layer is added for prediction.
+
+After 5 epochs: train acc: 0.9858, val acc: 0.9864
+"""
+from __future__ import print_function
+
+from keras.datasets import mnist
+from keras.models import Sequential, Model
+from keras.layers import Input, Dense, TimeDistributed
+from keras.layers import LSTM
+from keras.utils import np_utils
+
+# Training parameters.
+batch_size = 32
+nb_classes = 10
+nb_epochs = 5
+
+# Embedding dimensions.
+row_hidden = 128
+col_hidden = 128
+
+# The data, shuffled and split between train and test sets.
+(X_train, y_train), (X_test, y_test) = mnist.load_data()
+
+# Reshapes data to 4D for Hierarchical RNN.
+X_train = X_train.reshape(X_train.shape[0], 28, 28, 1)
+X_test = X_test.reshape(X_test.shape[0], 28, 28, 1)
+X_train = X_train.astype('float32')
+X_test = X_test.astype('float32')
+X_train /= 255
+X_test /= 255
+print('X_train shape:', X_train.shape)
+print(X_train.shape[0], 'train samples')
+print(X_test.shape[0], 'test samples')
+
+# Converts class vectors to binary class matrices.
+Y_train = np_utils.to_categorical(y_train, nb_classes)
+Y_test = np_utils.to_categorical(y_test, nb_classes)
+
+row, col, pixel = X_train.shape[1:]
+
+# 4D input.
+x = Input(shape=(row, col, pixel))
+
+# Encodes a row of pixels using TimeDistributed Wrapper.
+encoded_rows = TimeDistributed(LSTM(output_dim=row_hidden))(x)
+
+# Encodes columns of encoded rows.
+encoded_columns = LSTM(col_hidden)(encoded_rows)
+
+# Final predictions and model.
+prediction = Dense(nb_classes, activation='softmax')(encoded_columns)
+model = Model(input=x, output=prediction)
+model.compile(loss='categorical_crossentropy',
+              optimizer='rmsprop',
+              metrics=['accuracy'])
+
+# Training.
+model.fit(X_train, Y_train, batch_size=batch_size, nb_epoch=nb_epochs,
+          verbose=1, validation_data=(X_test, Y_test))
+
+# Evaluation.
+scores = model.evaluate(X_test, Y_test, verbose=0)
+print('Test loss:', scores[0])
+print('Test accuracy:', scores[1])

examples/mnist_irnn.py

@@ -1,32 +1,28 @@
-from __future__ import absolute_import
+'''This is a reproduction of the IRNN experiment
+with pixel-by-pixel sequential MNIST in
+"A Simple Way to Initialize Recurrent Networks of Rectified Linear Units"
+by Quoc V. Le, Navdeep Jaitly, Geoffrey E. Hinton
+
+arXiv:1504.00941v2 [cs.NE] 7 Apr 2015
+http://arxiv.org/pdf/1504.00941v2.pdf
+
+Optimizer is replaced with RMSprop which yields more stable and steady
+improvement.
+
+Reaches 0.93 train/test accuracy after 900 epochs
+(which roughly corresponds to 1687500 steps in the original paper.)
+'''
+
 from __future__ import print_function
-import numpy as np
-np.random.seed(1337)  # for reproducibility
 
 from keras.datasets import mnist
 from keras.models import Sequential
-from keras.layers.core import Dense, Activation
+from keras.layers import Dense, Activation
+from keras.layers import SimpleRNN
 from keras.initializations import normal, identity
-from keras.layers.recurrent import SimpleRNN, LSTM
 from keras.optimizers import RMSprop
 from keras.utils import np_utils
 
-'''
-    This is a reproduction of the IRNN experiment
-    with pixel-by-pixel sequential MNIST in
-    "A Simple Way to Initialize Recurrent Networks of Rectified Linear Units "
-    by Quoc V. Le, Navdeep Jaitly, Geoffrey E. Hinton
-
-    arXiv:1504.00941v2 [cs.NE] 7 Apr 201
-    http://arxiv.org/pdf/1504.00941v2.pdf
-
-    Optimizer is replaced with RMSprop which yields more stable and steady
-    improvement.
-
-    Reaches 0.93 train/test accuracy after 900 epochs
-    (which roughly corresponds to 1687500 steps in the original paper.)
-'''
-
 batch_size = 32
 nb_classes = 10
 nb_epochs = 200
@@ -40,8 +36,8 @@ clip_norm = 1.0
 
 X_train = X_train.reshape(X_train.shape[0], -1, 1)
 X_test = X_test.reshape(X_test.shape[0], -1, 1)
-X_train = X_train.astype("float32")
-X_test = X_test.astype("float32")
+X_train = X_train.astype('float32')
+X_test = X_test.astype('float32')
 X_train /= 255
 X_test /= 255
 print('X_train shape:', X_train.shape)
@@ -55,32 +51,20 @@ Y_test = np_utils.to_categorical(y_test, nb_classes)
 print('Evaluate IRNN...')
 model = Sequential()
 model.add(SimpleRNN(output_dim=hidden_units,
-                    init=lambda shape: normal(shape, scale=0.001),
-                    inner_init=lambda shape: identity(shape, scale=1.0),
-                    activation='relu', input_shape=X_train.shape[1:]))
+                    init=lambda shape, name: normal(shape, scale=0.001, name=name),
+                    inner_init=lambda shape, name: identity(shape, scale=1.0, name=name),
+                    activation='relu',
+                    input_shape=X_train.shape[1:]))
 model.add(Dense(nb_classes))
 model.add(Activation('softmax'))
 rmsprop = RMSprop(lr=learning_rate)
-model.compile(loss='categorical_crossentropy', optimizer=rmsprop)
+model.compile(loss='categorical_crossentropy',
+              optimizer=rmsprop,
+              metrics=['accuracy'])
 
 model.fit(X_train, Y_train, batch_size=batch_size, nb_epoch=nb_epochs,
-          show_accuracy=True, verbose=1, validation_data=(X_test, Y_test))
+          verbose=1, validation_data=(X_test, Y_test))
 
-scores = model.evaluate(X_test, Y_test, show_accuracy=True, verbose=0)
+scores = model.evaluate(X_test, Y_test, verbose=0)
 print('IRNN test score:', scores[0])
 print('IRNN test accuracy:', scores[1])
-
-print('Compare to LSTM...')
-model = Sequential()
-model.add(LSTM(hidden_units, input_shape=X_train.shape[1:]))
-model.add(Dense(nb_classes))
-model.add(Activation('softmax'))
-rmsprop = RMSprop(lr=learning_rate)
-model.compile(loss='categorical_crossentropy', optimizer=rmsprop)
-
-model.fit(X_train, Y_train, batch_size=batch_size, nb_epoch=nb_epochs,
-          show_accuracy=True, verbose=1, validation_data=(X_test, Y_test))
-
-scores = model.evaluate(X_test, Y_test, show_accuracy=True, verbose=0)
-print('LSTM test score:', scores[0])
-print('LSTM test accuracy:', scores[1])

examples/mnist_mlp.py

@@ -1,4 +1,10 @@
-from __future__ import absolute_import
+'''Trains a simple deep NN on the MNIST dataset.
+
+Gets to 98.40% test accuracy after 20 epochs
+(there is *a lot* of margin for parameter tuning).
+2 seconds per epoch on a K520 GPU.
+'''
+
 from __future__ import print_function
 import numpy as np
 np.random.seed(1337)  # for reproducibility
@@ -9,25 +15,18 @@ from keras.layers.core import Dense, Dropout, Activation
 from keras.optimizers import SGD, Adam, RMSprop
 from keras.utils import np_utils
 
-'''
-    Train a simple deep NN on the MNIST dataset.
-
-    Get to 98.40% test accuracy after 20 epochs
-    (there is *a lot* of margin for parameter tuning).
-    2 seconds per epoch on a K520 GPU.
-'''
 
 batch_size = 128
 nb_classes = 10
 nb_epoch = 20
 
-# the data, shuffled and split between tran and test sets
+# the data, shuffled and split between train and test sets
 (X_train, y_train), (X_test, y_test) = mnist.load_data()
 
 X_train = X_train.reshape(60000, 784)
 X_test = X_test.reshape(10000, 784)
-X_train = X_train.astype("float32")
-X_test = X_test.astype("float32")
+X_train = X_train.astype('float32')
+X_test = X_test.astype('float32')
 X_train /= 255
 X_test /= 255
 print(X_train.shape[0], 'train samples')
@@ -47,14 +46,15 @@ model.add(Dropout(0.2))
 model.add(Dense(10))
 model.add(Activation('softmax'))
 
-rms = RMSprop()
-model.compile(loss='categorical_crossentropy', optimizer=rms)
+model.summary()
 
-model.fit(X_train, Y_train,
-          batch_size=batch_size, nb_epoch=nb_epoch,
-          show_accuracy=True, verbose=2,
-          validation_data=(X_test, Y_test))
-score = model.evaluate(X_test, Y_test,
-                       show_accuracy=True, verbose=0)
+model.compile(loss='categorical_crossentropy',
+              optimizer=RMSprop(),
+              metrics=['accuracy'])
+
+history = model.fit(X_train, Y_train,
+                    batch_size=batch_size, nb_epoch=nb_epoch,
+                    verbose=1, validation_data=(X_test, Y_test))
+score = model.evaluate(X_test, Y_test, verbose=0)
 print('Test score:', score[0])
 print('Test accuracy:', score[1])

examples/mnist_net2net.py

@@ -0,0 +1,384 @@
+'''This is an implementation of Net2Net experiment with MNIST in
+'Net2Net: Accelerating Learning via Knowledge Transfer'
+by Tianqi Chen, Ian Goodfellow, and Jonathon Shlens
+
+arXiv:1511.05641v4 [cs.LG] 23 Apr 2016
+http://arxiv.org/abs/1511.05641
+
+Notes
+- What:
+  + Net2Net is a group of methods to transfer knowledge from a teacher neural
+    net to a student net,so that the student net can be trained faster than
+    from scratch.
+  + The paper discussed two specific methods of Net2Net, i.e. Net2WiderNet
+    and Net2DeeperNet.
+  + Net2WiderNet replaces a model with an equivalent wider model that has
+    more units in each hidden layer.
+  + Net2DeeperNet replaces a model with an equivalent deeper model.
+  + Both are based on the idea of 'function-preserving transformations of
+    neural nets'.
+- Why:
+  + Enable fast exploration of multiple neural nets in experimentation and
+    design process,by creating a series of wider and deeper models with
+    transferable knowledge.
+  + Enable 'lifelong learning system' by gradually adjusting model complexity
+    to data availability,and reusing transferable knowledge.
+
+Experiments
+- Teacher model: a basic CNN model trained on MNIST for 3 epochs.
+- Net2WiderNet exepriment:
+  + Student model has a wider Conv2D layer and a wider FC layer.
+  + Comparison of 'random-padding' vs 'net2wider' weight initialization.
+  + With both methods, student model should immediately perform as well as
+    teacher model, but 'net2wider' is slightly better.
+- Net2DeeperNet experiment:
+  + Student model has an extra Conv2D layer and an extra FC layer.
+  + Comparison of 'random-init' vs 'net2deeper' weight initialization.
+  + Starting performance of 'net2deeper' is better than 'random-init'.
+- Hyper-parameters:
+  + SGD with momentum=0.9 is used for training teacher and student models.
+  + Learning rate adjustment: it's suggested to reduce learning rate
+    to 1/10 for student model.
+  + Addition of noise in 'net2wider' is used to break weight symmetry
+    and thus enable full capacity of student models. It is optional
+    when a Dropout layer is used.
+
+Results
+- Tested with 'Theano' backend and 'th' image_dim_ordering.
+- Running on GPU GeForce GTX 980M
+- Performance Comparisons - validation loss values during first 3 epochs:
+(1) teacher_model:             0.075    0.041    0.041
+(2) wider_random_pad:          0.036    0.034    0.032
+(3) wider_net2wider:           0.032    0.030    0.030
+(4) deeper_random_init:        0.061    0.043    0.041
+(5) deeper_net2deeper:         0.032    0.031    0.029
+'''
+
+from __future__ import print_function
+import numpy as np
+np.random.seed(1337)
+
+from keras.models import Sequential
+from keras.layers import Conv2D, MaxPooling2D, Dense, Flatten
+from keras.optimizers import SGD
+from keras.utils import np_utils
+from keras.datasets import mnist
+
+input_shape = (1, 28, 28)  # image shape
+nb_class = 10  # number of class
+
+
+# load and pre-process data
+def preprocess_input(x):
+    return x.reshape((-1, ) + input_shape) / 255.
+
+
+def preprocess_output(y):
+    return np_utils.to_categorical(y)
+
+(train_x, train_y), (validation_x, validation_y) = mnist.load_data()
+train_x, validation_x = map(preprocess_input, [train_x, validation_x])
+train_y, validation_y = map(preprocess_output, [train_y, validation_y])
+print('Loading MNIST data...')
+print('train_x shape:', train_x.shape, 'train_y shape:', train_y.shape)
+print('validation_x shape:', validation_x.shape,
+      'validation_y shape', validation_y.shape)
+
+
+# knowledge transfer algorithms
+def wider2net_conv2d(teacher_w1, teacher_b1, teacher_w2, new_width, init):
+    '''Get initial weights for a wider conv2d layer with a bigger nb_filter,
+    by 'random-padding' or 'net2wider'.
+
+    # Arguments
+        teacher_w1: `weight` of conv2d layer to become wider,
+          of shape (nb_filter1, nb_channel1, kh1, kw1)
+        teacher_b1: `bias` of conv2d layer to become wider,
+          of shape (nb_filter1, )
+        teacher_w2: `weight` of next connected conv2d layer,
+          of shape (nb_filter2, nb_channel2, kh2, kw2)
+        new_width: new `nb_filter` for the wider conv2d layer
+        init: initialization algorithm for new weights,
+          either 'random-pad' or 'net2wider'
+    '''
+    assert teacher_w1.shape[0] == teacher_w2.shape[1], (
+        'successive layers from teacher model should have compatible shapes')
+    assert teacher_w1.shape[0] == teacher_b1.shape[0], (
+        'weight and bias from same layer should have compatible shapes')
+    assert new_width > teacher_w1.shape[0], (
+        'new width (nb_filter) should be bigger than the existing one')
+
+    n = new_width - teacher_w1.shape[0]
+    if init == 'random-pad':
+        new_w1 = np.random.normal(0, 0.1, size=(n, ) + teacher_w1.shape[1:])
+        new_b1 = np.ones(n) * 0.1
+        new_w2 = np.random.normal(0, 0.1, size=(
+            teacher_w2.shape[0], n) + teacher_w2.shape[2:])
+    elif init == 'net2wider':
+        index = np.random.randint(teacher_w1.shape[0], size=n)
+        factors = np.bincount(index)[index] + 1.
+        new_w1 = teacher_w1[index, :, :, :]
+        new_b1 = teacher_b1[index]
+        new_w2 = teacher_w2[:, index, :, :] / factors.reshape((1, -1, 1, 1))
+    else:
+        raise ValueError('Unsupported weight initializer: %s' % init)
+
+    student_w1 = np.concatenate((teacher_w1, new_w1), axis=0)
+    if init == 'random-pad':
+        student_w2 = np.concatenate((teacher_w2, new_w2), axis=1)
+    elif init == 'net2wider':
+        # add small noise to break symmetry, so that student model will have
+        # full capacity later
+        noise = np.random.normal(0, 5e-2 * new_w2.std(), size=new_w2.shape)
+        student_w2 = np.concatenate((teacher_w2, new_w2 + noise), axis=1)
+        student_w2[:, index, :, :] = new_w2
+    student_b1 = np.concatenate((teacher_b1, new_b1), axis=0)
+
+    return student_w1, student_b1, student_w2
+
+
+def wider2net_fc(teacher_w1, teacher_b1, teacher_w2, new_width, init):
+    '''Get initial weights for a wider fully connected (dense) layer
+       with a bigger nout, by 'random-padding' or 'net2wider'.
+
+    # Arguments
+        teacher_w1: `weight` of fc layer to become wider,
+          of shape (nin1, nout1)
+        teacher_b1: `bias` of fc layer to become wider,
+          of shape (nout1, )
+        teacher_w2: `weight` of next connected fc layer,
+          of shape (nin2, nout2)
+        new_width: new `nout` for the wider fc layer
+        init: initialization algorithm for new weights,
+          either 'random-pad' or 'net2wider'
+    '''
+    assert teacher_w1.shape[1] == teacher_w2.shape[0], (
+        'successive layers from teacher model should have compatible shapes')
+    assert teacher_w1.shape[1] == teacher_b1.shape[0], (
+        'weight and bias from same layer should have compatible shapes')
+    assert new_width > teacher_w1.shape[1], (
+        'new width (nout) should be bigger than the existing one')
+
+    n = new_width - teacher_w1.shape[1]
+    if init == 'random-pad':
+        new_w1 = np.random.normal(0, 0.1, size=(teacher_w1.shape[0], n))
+        new_b1 = np.ones(n) * 0.1
+        new_w2 = np.random.normal(0, 0.1, size=(n, teacher_w2.shape[1]))
+    elif init == 'net2wider':
+        index = np.random.randint(teacher_w1.shape[1], size=n)
+        factors = np.bincount(index)[index] + 1.
+        new_w1 = teacher_w1[:, index]
+        new_b1 = teacher_b1[index]
+        new_w2 = teacher_w2[index, :] / factors[:, np.newaxis]
+    else:
+        raise ValueError('Unsupported weight initializer: %s' % init)
+
+    student_w1 = np.concatenate((teacher_w1, new_w1), axis=1)
+    if init == 'random-pad':
+        student_w2 = np.concatenate((teacher_w2, new_w2), axis=0)
+    elif init == 'net2wider':
+        # add small noise to break symmetry, so that student model will have
+        # full capacity later
+        noise = np.random.normal(0, 5e-2 * new_w2.std(), size=new_w2.shape)
+        student_w2 = np.concatenate((teacher_w2, new_w2 + noise), axis=0)
+        student_w2[index, :] = new_w2
+    student_b1 = np.concatenate((teacher_b1, new_b1), axis=0)
+
+    return student_w1, student_b1, student_w2
+
+
+def deeper2net_conv2d(teacher_w):
+    '''Get initial weights for a deeper conv2d layer by net2deeper'.
+
+    # Arguments
+        teacher_w: `weight` of previous conv2d layer,
+          of shape (nb_filter, nb_channel, kh, kw)
+    '''
+    nb_filter, nb_channel, kh, kw = teacher_w.shape
+    student_w = np.zeros((nb_filter, nb_filter, kh, kw))
+    for i in xrange(nb_filter):
+        student_w[i, i, (kh - 1) / 2, (kw - 1) / 2] = 1.
+    student_b = np.zeros(nb_filter)
+    return student_w, student_b
+
+
+def copy_weights(teacher_model, student_model, layer_names):
+    '''Copy weights from teacher_model to student_model,
+     for layers with names listed in layer_names
+    '''
+    for name in layer_names:
+        weights = teacher_model.get_layer(name=name).get_weights()
+        student_model.get_layer(name=name).set_weights(weights)
+
+
+# methods to construct teacher_model and student_models
+def make_teacher_model(train_data, validation_data, nb_epoch=3):
+    '''Train a simple CNN as teacher model.
+    '''
+    model = Sequential()
+    model.add(Conv2D(64, 3, 3, input_shape=input_shape,
+                     border_mode='same', name='conv1'))
+    model.add(MaxPooling2D(name='pool1'))
+    model.add(Conv2D(64, 3, 3, border_mode='same', name='conv2'))
+    model.add(MaxPooling2D(name='pool2'))
+    model.add(Flatten(name='flatten'))
+    model.add(Dense(64, activation='relu', name='fc1'))
+    model.add(Dense(nb_class, activation='softmax', name='fc2'))
+    model.compile(loss='categorical_crossentropy',
+                  optimizer=SGD(lr=0.01, momentum=0.9),
+                  metrics=['accuracy'])
+
+    train_x, train_y = train_data
+    history = model.fit(train_x, train_y, nb_epoch=nb_epoch,
+                        validation_data=validation_data)
+    return model, history
+
+
+def make_wider_student_model(teacher_model, train_data,
+                             validation_data, init, nb_epoch=3):
+    '''Train a wider student model based on teacher_model,
+       with either 'random-pad' (baseline) or 'net2wider'
+    '''
+    new_conv1_width = 128
+    new_fc1_width = 128
+
+    model = Sequential()
+    # a wider conv1 compared to teacher_model
+    model.add(Conv2D(new_conv1_width, 3, 3, input_shape=input_shape,
+                     border_mode='same', name='conv1'))
+    model.add(MaxPooling2D(name='pool1'))
+    model.add(Conv2D(64, 3, 3, border_mode='same', name='conv2'))
+    model.add(MaxPooling2D(name='pool2'))
+    model.add(Flatten(name='flatten'))
+    # a wider fc1 compared to teacher model
+    model.add(Dense(new_fc1_width, activation='relu', name='fc1'))
+    model.add(Dense(nb_class, activation='softmax', name='fc2'))
+
+    # The weights for other layers need to be copied from teacher_model
+    # to student_model, except for widened layers
+    # and their immediate downstreams, which will be initialized separately.
+    # For this example there are no other layers that need to be copied.
+
+    w_conv1, b_conv1 = teacher_model.get_layer('conv1').get_weights()
+    w_conv2, b_conv2 = teacher_model.get_layer('conv2').get_weights()
+    new_w_conv1, new_b_conv1, new_w_conv2 = wider2net_conv2d(
+        w_conv1, b_conv1, w_conv2, new_conv1_width, init)
+    model.get_layer('conv1').set_weights([new_w_conv1, new_b_conv1])
+    model.get_layer('conv2').set_weights([new_w_conv2, b_conv2])
+
+    w_fc1, b_fc1 = teacher_model.get_layer('fc1').get_weights()
+    w_fc2, b_fc2 = teacher_model.get_layer('fc2').get_weights()
+    new_w_fc1, new_b_fc1, new_w_fc2 = wider2net_fc(
+        w_fc1, b_fc1, w_fc2, new_fc1_width, init)
+    model.get_layer('fc1').set_weights([new_w_fc1, new_b_fc1])
+    model.get_layer('fc2').set_weights([new_w_fc2, b_fc2])
+
+    model.compile(loss='categorical_crossentropy',
+                  optimizer=SGD(lr=0.001, momentum=0.9),
+                  metrics=['accuracy'])
+
+    train_x, train_y = train_data
+    history = model.fit(train_x, train_y, nb_epoch=nb_epoch,
+                        validation_data=validation_data)
+    return model, history
+
+
+def make_deeper_student_model(teacher_model, train_data,
+                              validation_data, init, nb_epoch=3):
+    '''Train a deeper student model based on teacher_model,
+       with either 'random-init' (baseline) or 'net2deeper'
+    '''
+    model = Sequential()
+    model.add(Conv2D(64, 3, 3, input_shape=input_shape,
+                     border_mode='same', name='conv1'))
+    model.add(MaxPooling2D(name='pool1'))
+    model.add(Conv2D(64, 3, 3, border_mode='same', name='conv2'))
+    # add another conv2d layer to make original conv2 deeper
+    if init == 'net2deeper':
+        prev_w, _ = model.get_layer('conv2').get_weights()
+        new_weights = deeper2net_conv2d(prev_w)
+        model.add(Conv2D(64, 3, 3, border_mode='same',
+                         name='conv2-deeper', weights=new_weights))
+    elif init == 'random-init':
+        model.add(Conv2D(64, 3, 3, border_mode='same', name='conv2-deeper'))
+    else:
+        raise ValueError('Unsupported weight initializer: %s' % init)
+    model.add(MaxPooling2D(name='pool2'))
+    model.add(Flatten(name='flatten'))
+    model.add(Dense(64, activation='relu', name='fc1'))
+    # add another fc layer to make original fc1 deeper
+    if init == 'net2deeper':
+        # net2deeper for fc layer with relu, is just an identity initializer
+        model.add(Dense(64, init='identity',
+                        activation='relu', name='fc1-deeper'))
+    elif init == 'random-init':
+        model.add(Dense(64, activation='relu', name='fc1-deeper'))
+    else:
+        raise ValueError('Unsupported weight initializer: %s' % init)
+    model.add(Dense(nb_class, activation='softmax', name='fc2'))
+
+    # copy weights for other layers
+    copy_weights(teacher_model, model, layer_names=[
+                 'conv1', 'conv2', 'fc1', 'fc2'])
+
+    model.compile(loss='categorical_crossentropy',
+                  optimizer=SGD(lr=0.001, momentum=0.9),
+                  metrics=['accuracy'])
+
+    train_x, train_y = train_data
+    history = model.fit(train_x, train_y, nb_epoch=nb_epoch,
+                        validation_data=validation_data)
+    return model, history
+
+
+# experiments setup
+def net2wider_experiment():
+    '''Benchmark performances of
+    (1) a teacher model,
+    (2) a wider student model with `random_pad` initializer
+    (3) a wider student model with `Net2WiderNet` initializer
+    '''
+    train_data = (train_x, train_y)
+    validation_data = (validation_x, validation_y)
+    print('\nExperiment of Net2WiderNet ...')
+    print('\nbuilding teacher model ...')
+    teacher_model, _ = make_teacher_model(train_data,
+                                          validation_data,
+                                          nb_epoch=3)
+
+    print('\nbuilding wider student model by random padding ...')
+    make_wider_student_model(teacher_model, train_data,
+                             validation_data, 'random-pad',
+                             nb_epoch=3)
+    print('\nbuilding wider student model by net2wider ...')
+    make_wider_student_model(teacher_model, train_data,
+                             validation_data, 'net2wider',
+                             nb_epoch=3)
+
+
+def net2deeper_experiment():
+    '''Benchmark performances of
+    (1) a teacher model,
+    (2) a deeper student model with `random_init` initializer
+    (3) a deeper student model with `Net2DeeperNet` initializer
+    '''
+    train_data = (train_x, train_y)
+    validation_data = (validation_x, validation_y)
+    print('\nExperiment of Net2DeeperNet ...')
+    print('\nbuilding teacher model ...')
+    teacher_model, _ = make_teacher_model(train_data,
+                                          validation_data,
+                                          nb_epoch=3)
+
+    print('\nbuilding deeper student model by random init ...')
+    make_deeper_student_model(teacher_model, train_data,
+                              validation_data, 'random-init',
+                              nb_epoch=3)
+    print('\nbuilding deeper student model by net2deeper ...')
+    make_deeper_student_model(teacher_model, train_data,
+                              validation_data, 'net2deeper',
+                              nb_epoch=3)
+
+# run the experiments
+net2wider_experiment()
+net2deeper_experiment()

examples/mnist_siamese_graph.py

@@ -0,0 +1,130 @@
+'''Train a Siamese MLP on pairs of digits from the MNIST dataset.
+
+It follows Hadsell-et-al.'06 [1] by computing the Euclidean distance on the
+output of the shared network and by optimizing the contrastive loss (see paper
+for mode details).
+
+[1] "Dimensionality Reduction by Learning an Invariant Mapping"
+    http://yann.lecun.com/exdb/publis/pdf/hadsell-chopra-lecun-06.pdf
+
+Gets to 99.5% test accuracy after 20 epochs.
+3 seconds per epoch on a Titan X GPU
+'''
+from __future__ import absolute_import
+from __future__ import print_function
+import numpy as np
+np.random.seed(1337)  # for reproducibility
+
+import random
+from keras.datasets import mnist
+from keras.models import Sequential, Model
+from keras.layers import Dense, Dropout, Input, Lambda
+from keras.optimizers import SGD, RMSprop
+from keras import backend as K
+
+
+def euclidean_distance(vects):
+    x, y = vects
+    return K.sqrt(K.sum(K.square(x - y), axis=1, keepdims=True))
+
+
+def eucl_dist_output_shape(shapes):
+    shape1, shape2 = shapes
+    return (shape1[0], 1)
+
+
+def contrastive_loss(y_true, y_pred):
+    '''Contrastive loss from Hadsell-et-al.'06
+    http://yann.lecun.com/exdb/publis/pdf/hadsell-chopra-lecun-06.pdf
+    '''
+    margin = 1
+    return K.mean(y_true * K.square(y_pred) + (1 - y_true) * K.square(K.maximum(margin - y_pred, 0)))
+
+
+def create_pairs(x, digit_indices):
+    '''Positive and negative pair creation.
+    Alternates between positive and negative pairs.
+    '''
+    pairs = []
+    labels = []
+    n = min([len(digit_indices[d]) for d in range(10)]) - 1
+    for d in range(10):
+        for i in range(n):
+            z1, z2 = digit_indices[d][i], digit_indices[d][i+1]
+            pairs += [[x[z1], x[z2]]]
+            inc = random.randrange(1, 10)
+            dn = (d + inc) % 10
+            z1, z2 = digit_indices[d][i], digit_indices[dn][i]
+            pairs += [[x[z1], x[z2]]]
+            labels += [1, 0]
+    return np.array(pairs), np.array(labels)
+
+
+def create_base_network(input_dim):
+    '''Base network to be shared (eq. to feature extraction).
+    '''
+    seq = Sequential()
+    seq.add(Dense(128, input_shape=(input_dim,), activation='relu'))
+    seq.add(Dropout(0.1))
+    seq.add(Dense(128, activation='relu'))
+    seq.add(Dropout(0.1))
+    seq.add(Dense(128, activation='relu'))
+    return seq
+
+
+def compute_accuracy(predictions, labels):
+    '''Compute classification accuracy with a fixed threshold on distances.
+    '''
+    return labels[predictions.ravel() < 0.5].mean()
+
+
+# the data, shuffled and split between train and test sets
+(X_train, y_train), (X_test, y_test) = mnist.load_data()
+X_train = X_train.reshape(60000, 784)
+X_test = X_test.reshape(10000, 784)
+X_train = X_train.astype('float32')
+X_test = X_test.astype('float32')
+X_train /= 255
+X_test /= 255
+input_dim = 784
+nb_epoch = 20
+
+# create training+test positive and negative pairs
+digit_indices = [np.where(y_train == i)[0] for i in range(10)]
+tr_pairs, tr_y = create_pairs(X_train, digit_indices)
+
+digit_indices = [np.where(y_test == i)[0] for i in range(10)]
+te_pairs, te_y = create_pairs(X_test, digit_indices)
+
+# network definition
+base_network = create_base_network(input_dim)
+
+input_a = Input(shape=(input_dim,))
+input_b = Input(shape=(input_dim,))
+
+# because we re-use the same instance `base_network`,
+# the weights of the network
+# will be shared across the two branches
+processed_a = base_network(input_a)
+processed_b = base_network(input_b)
+
+distance = Lambda(euclidean_distance, output_shape=eucl_dist_output_shape)([processed_a, processed_b])
+
+model = Model(input=[input_a, input_b], output=distance)
+
+# train
+rms = RMSprop()
+model.compile(loss=contrastive_loss, optimizer=rms)
+model.fit([tr_pairs[:, 0], tr_pairs[:, 1]], tr_y,
+          validation_data=([te_pairs[:, 0], te_pairs[:, 1]], te_y),
+          batch_size=128,
+          nb_epoch=nb_epoch)
+
+# compute final accuracy on training and test sets
+pred = model.predict([tr_pairs[:, 0], tr_pairs[:, 1]])
+tr_acc = compute_accuracy(pred, tr_y)
+pred = model.predict([te_pairs[:, 0], te_pairs[:, 1]])
+te_acc = compute_accuracy(pred, te_y)
+
+print('* Accuracy on training set: %0.2f%%' % (100 * tr_acc))
+print('* Accuracy on test set: %0.2f%%' % (100 * te_acc))

examples/mnist_sklearn_wrapper.py

@@ -0,0 +1,94 @@
+'''Example of how to use sklearn wrapper
+
+Builds simple CNN models on MNIST and uses sklearn's GridSearchCV to find best model
+'''
+
+from __future__ import print_function
+import numpy as np
+np.random.seed(1337)  # for reproducibility
+
+from keras.datasets import mnist
+from keras.models import Sequential
+from keras.layers import Dense, Dropout, Activation, Flatten
+from keras.layers import Convolution2D, MaxPooling2D
+from keras.utils import np_utils
+from keras.wrappers.scikit_learn import KerasClassifier
+from sklearn.grid_search import GridSearchCV
+
+
+nb_classes = 10
+
+# input image dimensions
+img_rows, img_cols = 28, 28
+
+# load training data and do basic data normalization
+(X_train, y_train), (X_test, y_test) = mnist.load_data()
+X_train = X_train.reshape(X_train.shape[0], 1, img_rows, img_cols)
+X_test = X_test.reshape(X_test.shape[0], 1, img_rows, img_cols)
+X_train = X_train.astype('float32')
+X_test = X_test.astype('float32')
+X_train /= 255
+X_test /= 255
+
+# convert class vectors to binary class matrices
+y_train = np_utils.to_categorical(y_train, nb_classes)
+y_test = np_utils.to_categorical(y_test, nb_classes)
+
+def make_model(dense_layer_sizes, nb_filters, nb_conv, nb_pool):
+    '''Creates model comprised of 2 convolutional layers followed by dense layers
+
+    dense_layer_sizes: List of layer sizes. This list has one number for each layer
+    nb_filters: Number of convolutional filters in each convolutional layer
+    nb_conv: Convolutional kernel size
+    nb_pool: Size of pooling area for max pooling
+    '''
+
+    model = Sequential()
+
+    model.add(Convolution2D(nb_filters, nb_conv, nb_conv,
+                            border_mode='valid',
+                            input_shape=(1, img_rows, img_cols)))
+    model.add(Activation('relu'))
+    model.add(Convolution2D(nb_filters, nb_conv, nb_conv))
+    model.add(Activation('relu'))
+    model.add(MaxPooling2D(pool_size=(nb_pool, nb_pool)))
+    model.add(Dropout(0.25))
+
+    model.add(Flatten())
+    for layer_size in dense_layer_sizes:
+        model.add(Dense(layer_size))
+    model.add(Activation('relu'))
+    model.add(Dropout(0.5))
+    model.add(Dense(nb_classes))
+    model.add(Activation('softmax'))
+
+    model.compile(loss='categorical_crossentropy',
+                  optimizer='adadelta',
+                  metrics=['accuracy'])
+
+    return model
+
+dense_size_candidates = [[32], [64], [32, 32], [64, 64]]
+my_classifier = KerasClassifier(make_model, batch_size=32)
+validator = GridSearchCV(my_classifier,
+                         param_grid={'dense_layer_sizes': dense_size_candidates,
+                                     # nb_epoch is avail for tuning even when not
+                                     # an argument to model building function
+                                     'nb_epoch': [3, 6],
+                                     'nb_filters': [8],
+                                     'nb_conv': [3],
+                                     'nb_pool': [2]},
+                         scoring='log_loss',
+                         n_jobs=1)
+validator.fit(X_train, y_train)
+
+print('The parameters of the best model are: ')
+print(validator.best_params_)
+
+# validator.best_estimator_ returns sklearn-wrapped version of best model.
+# validator.best_estimator_.model returns the (unwrapped) keras model
+best_model = validator.best_estimator_.model
+metric_names = best_model.metrics_names
+metric_values = best_model.evaluate(X_test, y_test)
+for metric, value in zip(metric_names, metric_values):
+    print(metric, ': ', value)

examples/mnist_swwae.py

@@ -0,0 +1,167 @@
+'''Trains a stacked what-where autoencoder built on residual blocks on the
+MNIST dataset.  It exemplifies two influential methods that have been developed
+in the past few years.
+
+The first is the idea of properly "unpooling." During any max pool, the
+exact location (the "where") of the maximal value in a pooled receptive field
+is lost, however it can be very useful in the overall reconstruction of an
+input image.  Therefore, if the "where" is handed from the encoder
+to the corresponding decoder layer, features being decoded can be "placed" in
+the right location, allowing for reconstructions of much higher fidelity.
+
+References:
+[1]
+"Visualizing and Understanding Convolutional Networks"
+Matthew D Zeiler, Rob Fergus
+https://arxiv.org/abs/1311.2901v3
+
+[2]
+"Stacked What-Where Auto-encoders"
+Junbo Zhao, Michael Mathieu, Ross Goroshin, Yann LeCun
+https://arxiv.org/abs/1506.02351v8
+
+The second idea exploited here is that of residual learning.  Residual blocks
+ease the training process by allowing skip connections that give the network
+the ability to be as linear (or non-linear) as the data sees fit.  This allows
+for much deep networks to be easily trained.  The residual element seems to
+be advantageous in the context of this example as it allows a nice symmetry
+between the encoder and decoder.  Normally, in the decoder, the final
+projection to the space where the image is reconstructed is linear, however
+this does not have to be the case for a residual block as the degree to which
+its output is linear or non-linear is determined by the data it is fed.
+However, in order to cap the reconstruction in this example, a hard softmax is
+applied as a bias because we know the MNIST digits are mapped to [0,1].
+
+References:
+[3]
+"Deep Residual Learning for Image Recognition"
+Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun
+https://arxiv.org/abs/1512.03385v1
+
+[4]
+"Identity Mappings in Deep Residual Networks"
+Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun
+https://arxiv.org/abs/1603.05027v3
+
+'''
+
+from __future__ import print_function
+import numpy as np
+np.random.seed(1337)  # for reproducibility
+
+from keras.datasets import mnist
+from keras.models import Model
+from keras.layers import Activation, merge
+from keras.layers import UpSampling2D, Convolution2D, MaxPooling2D
+from keras.layers import Input, BatchNormalization
+import matplotlib.pyplot as plt
+import keras.backend as K
+
+
+def convresblock(x, nfeats=8, ksize=3, nskipped=2):
+    ''' The proposed residual block from [4]'''
+    y0 = Convolution2D(nfeats, ksize, ksize, border_mode='same')(x)
+    y = y0
+    for i in range(nskipped):
+        y = BatchNormalization(mode=0, axis=1)(y)
+        y = Activation('relu')(y)
+        y = Convolution2D(nfeats, ksize, ksize, border_mode='same')(y)
+    return merge([y0, y], mode='sum')
+
+
+def getwhere(x):
+    ''' Calculate the "where" mask that contains switches indicating which
+    index contained the max value when MaxPool2D was applied.  Using the
+    gradient of the sum is a nice trick to keep everything high level.'''
+    y_prepool, y_postpool = x
+    return K.gradients(K.sum(y_postpool), y_prepool)
+
+# input image dimensions
+img_rows, img_cols = 28, 28
+
+# the data, shuffled and split between train and test sets
+(X_train, _), (X_test, _) = mnist.load_data()
+
+X_train = X_train.reshape(X_train.shape[0], 1, img_rows, img_cols)
+X_test = X_test.reshape(X_test.shape[0], 1, img_rows, img_cols)
+X_train = X_train.astype('float32')
+X_test = X_test.astype('float32')
+X_train /= 255
+X_test /= 255
+print('X_train shape:', X_train.shape)
+print(X_train.shape[0], 'train samples')
+print(X_test.shape[0], 'test samples')
+
+# The size of the kernel used for the MaxPooling2D
+pool_size = 2
+# The total number of feature maps at each layer
+nfeats = [8, 16, 32, 64, 128]
+# The sizes of the pooling kernel at each layer
+pool_sizes = np.array([1, 1, 1, 1, 1]) * pool_size
+# The convolution kernel size
+ksize = 3
+# Number of epochs to train for
+nb_epoch = 5
+# Batch size during training
+batch_size = 128
+
+if pool_size == 2:
+    # if using a 5 layer net of pool_size = 2
+    X_train = np.pad(X_train, [[0, 0], [0, 0], [2, 2], [2, 2]],
+                     mode='constant')
+    X_test = np.pad(X_test, [[0, 0], [0, 0], [2, 2], [2, 2]], mode='constant')
+    nlayers = 5
+elif pool_size == 3:
+    # if using a 3 layer net of pool_size = 3
+    X_train = X_train[:, :, :-1, :-1]
+    X_test = X_test[:, :, :-1, :-1]
+    nlayers = 3
+else:
+    import sys
+    sys.exit("Script supports pool_size of 2 and 3.")
+
+# Shape of input to train on (note that model is fully convolutional however)
+input_shape = X_train.shape[1:]
+# The final list of the size of axis=1 for all layers, including input
+nfeats_all = [input_shape[0]] + nfeats
+
+# First build the encoder, all the while keeping track of the "where" masks
+img_input = Input(shape=input_shape)
+
+# We push the "where" masks to the following list
+wheres = [None] * nlayers
+y = img_input
+for i in range(nlayers):
+    y_prepool = convresblock(y, nfeats=nfeats_all[i + 1], ksize=ksize)
+    y = MaxPooling2D(pool_size=(pool_sizes[i], pool_sizes[i]))(y_prepool)
+    wheres[i] = merge([y_prepool, y], mode=getwhere,
+                      output_shape=lambda x: x[0])
+
+# Now build the decoder, and use the stored "where" masks to place the features
+for i in range(nlayers):
+    ind = nlayers - 1 - i
+    y = UpSampling2D(size=(pool_sizes[ind], pool_sizes[ind]))(y)
+    y = merge([y, wheres[ind]], mode='mul')
+    y = convresblock(y, nfeats=nfeats_all[ind], ksize=ksize)
+
+# Use hard_simgoid to clip range of reconstruction
+y = Activation('hard_sigmoid')(y)
+
+# Define the model and it's mean square error loss, and compile it with Adam
+model = Model(img_input, y)
+model.compile('adam', 'mse')
+
+# Fit the model
+model.fit(X_train, X_train, validation_data=(X_test, X_test),
+          batch_size=batch_size, nb_epoch=nb_epoch)
+
+# Plot
+X_recon = model.predict(X_test[:25])
+X_plot = np.concatenate((X_test[:25], X_recon), axis=1)
+X_plot = X_plot.reshape((5, 10, input_shape[-2], input_shape[-1]))
+X_plot = np.vstack([np.hstack(x) for x in X_plot])
+plt.figure()
+plt.axis('off')
+plt.title('Test Samples: Originals/Reconstructions')
+plt.imshow(X_plot, interpolation='none', cmap='gray')
+plt.savefig('reconstructions.png')

examples/mnist_transfer_cnn.py

@@ -1,4 +1,16 @@
-from __future__ import absolute_import
+'''Transfer learning toy example:
+
+1- Train a simple convnet on the MNIST dataset the first 5 digits [0..4].
+2- Freeze convolutional layers and fine-tune dense layers
+   for the classification of digits [5..9].
+
+Run on GPU: THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 python mnist_transfer_cnn.py
+
+Get to 99.8% test accuracy after 5 epochs
+for the first five digits classifier
+and 99.2% for the last five digits after transfer + fine-tuning.
+'''
+
 from __future__ import print_function
 import numpy as np
 import datetime
@@ -7,22 +19,10 @@ np.random.seed(1337)  # for reproducibility
 
 from keras.datasets import mnist
 from keras.models import Sequential
-from keras.layers.core import Dense, Dropout, Activation, Flatten
-from keras.layers.convolutional import Convolution2D, MaxPooling2D
+from keras.layers import Dense, Dropout, Activation, Flatten
+from keras.layers import Convolution2D, MaxPooling2D
 from keras.utils import np_utils
-
-'''
-    Transfer learning toy example:
-        1- Train a simple convnet on the MNIST dataset the first 5 digits [0..4].
-        2- Freeze convolutional layers and fine-tune dense layers
-           for the classification of digits [5..9].
-
-    Run on GPU: THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 python mnist_cnn.py
-
-    Get to 99.8% test accuracy after 5 epochs
-    for the first five digits classifier
-    and 99.2% for the last five digits after transfer + fine-tuning.
-'''
+from keras import backend as K
 
 now = datetime.datetime.now
 
@@ -35,16 +35,21 @@ img_rows, img_cols = 28, 28
 # number of convolutional filters to use
 nb_filters = 32
 # size of pooling area for max pooling
-nb_pool = 2
+pool_size = 2
 # convolution kernel size
-nb_conv = 3
+kernel_size = 3
+
+if K.image_dim_ordering() == 'th':
+    input_shape = (1, img_rows, img_cols)
+else:
+    input_shape = (img_rows, img_cols, 1)
 
 
 def train_model(model, train, test, nb_classes):
-    X_train = train[0].reshape(train[0].shape[0], 1, img_rows, img_cols)
-    X_test = test[0].reshape(test[0].shape[0], 1, img_rows, img_cols)
-    X_train = X_train.astype("float32")
-    X_test = X_test.astype("float32")
+    X_train = train[0].reshape((train[0].shape[0],) + input_shape)
+    X_test = test[0].reshape((test[0].shape[0],) + input_shape)
+    X_train = X_train.astype('float32')
+    X_test = X_test.astype('float32')
     X_train /= 255
     X_test /= 255
     print('X_train shape:', X_train.shape)
@@ -55,15 +60,17 @@ def train_model(model, train, test, nb_classes):
     Y_train = np_utils.to_categorical(train[1], nb_classes)
     Y_test = np_utils.to_categorical(test[1], nb_classes)
 
-    model.compile(loss='categorical_crossentropy', optimizer='adadelta')
+    model.compile(loss='categorical_crossentropy',
+                  optimizer='adadelta',
+                  metrics=['accuracy'])
 
     t = now()
     model.fit(X_train, Y_train,
               batch_size=batch_size, nb_epoch=nb_epoch,
-              show_accuracy=True, verbose=1,
+              verbose=1,
               validation_data=(X_test, Y_test))
     print('Training time: %s' % (now() - t))
-    score = model.evaluate(X_test, Y_test, show_accuracy=True, verbose=0)
+    score = model.evaluate(X_test, Y_test, verbose=0)
     print('Test score:', score[0])
     print('Test accuracy:', score[1])
 
@@ -84,13 +91,13 @@ y_test_gte5 = y_test[y_test >= 5] - 5
 
 # define two groups of layers: feature (convolutions) and classification (dense)
 feature_layers = [
-    Convolution2D(nb_filters, nb_conv, nb_conv,
+    Convolution2D(nb_filters, kernel_size, kernel_size,
                   border_mode='valid',
-                  input_shape=(1, img_rows, img_cols)),
+                  input_shape=input_shape),
     Activation('relu'),
-    Convolution2D(nb_filters, nb_conv, nb_conv),
+    Convolution2D(nb_filters, kernel_size, kernel_size),
     Activation('relu'),
-    MaxPooling2D(pool_size=(nb_pool, nb_pool)),
+    MaxPooling2D(pool_size=(pool_size, pool_size)),
     Dropout(0.25),
     Flatten(),
 ]
@@ -103,9 +110,7 @@ classification_layers = [
 ]
 
 # create complete model
-model = Sequential()
-for l in feature_layers + classification_layers:
-    model.add(l)
+model = Sequential(feature_layers + classification_layers)
 
 # train model for 5-digit classification [0..4]
 train_model(model,

examples/neural_doodle.py

@@ -0,0 +1,364 @@
+'''Neural doodle with Keras
+
+Script Usage:
+    # Arguments:
+    ```
+    --nlabels:              # of regions (colors) in mask images
+    --style-image:          image to learn style from
+    --style-mask:           semantic labels for style image
+    --target-mask:          semantic labels for target image (your doodle)
+    --content-image:        optional image to learn content from
+    --target-image-prefix:  path prefix for generated target images
+    ```
+
+    # Example 1: doodle using a style image, style mask
+    and target mask.
+    ```
+    python neural_doodle.py --nlabels 4 --style-image Monet/style.png \
+    --style-mask Monet/style_mask.png --target-mask Monet/target_mask.png \
+    --target-image-prefix generated/monet
+    ```
+
+    # Example 2: doodle using a style image, style mask,
+    target mask and an optional content image.
+    ```
+    python neural_doodle.py --nlabels 4 --style-image Renoir/style.png \
+    --style-mask Renoir/style_mask.png --target-mask Renoir/target_mask.png \
+    --content-image Renoir/creek.jpg \
+    --target-image-prefix generated/renoir
+    ```
+
+References:
+[Dmitry Ulyanov's blog on fast-neural-doodle](http://dmitryulyanov.github.io/feed-forward-neural-doodle/)
+[Torch code for fast-neural-doodle](https://github.com/DmitryUlyanov/fast-neural-doodle)
+[Torch code for online-neural-doodle](https://github.com/DmitryUlyanov/online-neural-doodle)
+[Paper Texture Networks: Feed-forward Synthesis of Textures and Stylized Images](http://arxiv.org/abs/1603.03417)
+[Discussion on parameter tuning](https://github.com/fchollet/keras/issues/3705)
+
+Resources:
+Example images can be downloaded from
+https://github.com/DmitryUlyanov/fast-neural-doodle/tree/master/data
+'''
+from __future__ import print_function
+import time
+import argparse
+import numpy as np
+from scipy.optimize import fmin_l_bfgs_b
+from scipy.misc import imread, imsave
+
+from keras import backend as K
+from keras.layers import Input, Convolution2D, MaxPooling2D, AveragePooling2D
+from keras.models import Model
+from keras.preprocessing.image import load_img, img_to_array
+from keras.applications import vgg19
+
+# Command line arguments
+parser = argparse.ArgumentParser(description='Keras neural doodle example')
+parser.add_argument('--nlabels', type=int,
+                    help='number of semantic labels'
+                    ' (regions in differnet colors)'
+                    ' in style_mask/target_mask')
+parser.add_argument('--style-image', type=str,
+                    help='path to image to learn style from')
+parser.add_argument('--style-mask', type=str,
+                    help='path to semantic mask of style image')
+parser.add_argument('--target-mask', type=str,
+                    help='path to semantic mask of target image')
+parser.add_argument('--content-image', type=str, default=None,
+                    help='path to optional content image')
+parser.add_argument('--target-image-prefix', type=str,
+                    help='path prefix for generated results')
+args = parser.parse_args()
+
+style_img_path = args.style_image
+style_mask_path = args.style_mask
+target_mask_path = args.target_mask
+content_img_path = args.content_image
+target_img_prefix = args.target_image_prefix
+use_content_img = content_img_path is not None
+
+nb_labels = args.nlabels
+nb_colors = 3  # RGB
+# determine image sizes based on target_mask
+ref_img = imread(target_mask_path)
+img_nrows, img_ncols = ref_img.shape[:2]
+
+total_variation_weight = 50.
+style_weight = 1.
+content_weight = 0.1 if use_content_img else 0
+
+content_feature_layers = ['block5_conv2']
+# To get better generation qualities, use more conv layers for style features
+style_feature_layers = ['block1_conv1', 'block2_conv1', 'block3_conv1',
+                        'block4_conv1', 'block5_conv1']
+
+
+# helper functions for reading/processing images
+def preprocess_image(image_path):
+    img = load_img(image_path, target_size=(img_nrows, img_ncols))
+    img = img_to_array(img)
+    img = np.expand_dims(img, axis=0)
+    img = vgg19.preprocess_input(img)
+    return img
+
+
+def deprocess_image(x):
+    if K.image_dim_ordering() == 'th':
+        x = x.reshape((3, img_nrows, img_ncols))
+        x = x.transpose((1, 2, 0))
+    else:
+        x = x.reshape((img_nrows, img_ncols, 3))
+    x = x[:, :, ::-1]
+    x[:, :, 0] += 103.939
+    x[:, :, 1] += 116.779
+    x[:, :, 2] += 123.68
+    x = np.clip(x, 0, 255).astype('uint8')
+    return x
+
+
+def kmeans(xs, k):
+    assert xs.ndim == 2
+    try:
+        from sklearn.cluster import k_means
+        _, labels, _ = k_means(xs.astype("float64"), k)
+    except ImportError:
+        from scipy.cluster.vq import kmeans2
+        _, labels = kmeans2(xs, k, missing='raise')
+    return labels
+
+
+def load_mask_labels():
+    '''Load both target and style masks.
+    A mask image (nr x nc) with m labels/colors will be loaded
+    as a 4D boolean tensor: (1, m, nr, nc) for 'th' or (1, nr, nc, m) for 'tf'
+    '''
+    target_mask_img = load_img(target_mask_path,
+                               target_size=(img_nrows, img_ncols))
+    target_mask_img = img_to_array(target_mask_img)
+    style_mask_img = load_img(style_mask_path,
+                              target_size=(img_nrows, img_ncols))
+    style_mask_img = img_to_array(style_mask_img)
+    if K.image_dim_ordering() == 'th':
+        mask_vecs = np.vstack([style_mask_img.reshape((3, -1)).T,
+                               target_mask_img.reshape((3, -1)).T])
+    else:
+        mask_vecs = np.vstack([style_mask_img.reshape((-1, 3)),
+                               target_mask_img.reshape((-1, 3))])
+
+    labels = kmeans(mask_vecs, nb_labels)
+    style_mask_label = labels[:img_nrows *
+                              img_ncols].reshape((img_nrows, img_ncols))
+    target_mask_label = labels[img_nrows *
+                               img_ncols:].reshape((img_nrows, img_ncols))
+
+    stack_axis = 0 if K.image_dim_ordering() == 'th' else -1
+    style_mask = np.stack([style_mask_label == r for r in xrange(nb_labels)],
+                          axis=stack_axis)
+    target_mask = np.stack([target_mask_label == r for r in xrange(nb_labels)],
+                           axis=stack_axis)
+
+    return (np.expand_dims(style_mask, axis=0),
+            np.expand_dims(target_mask, axis=0))
+
+# Create tensor variables for images
+if K.image_dim_ordering() == 'th':
+    shape = (1, nb_colors, img_nrows, img_ncols)
+else:
+    shape = (1, img_nrows, img_ncols, nb_colors)
+
+style_image = K.variable(preprocess_image(style_img_path))
+target_image = K.placeholder(shape=shape)
+if use_content_img:
+    content_image = K.variable(preprocess_image(content_img_path))
+else:
+    content_image = K.zeros(shape=shape)
+
+images = K.concatenate([style_image, target_image, content_image], axis=0)
+
+# Create tensor variables for masks
+raw_style_mask, raw_target_mask = load_mask_labels()
+style_mask = K.variable(raw_style_mask.astype("float32"))
+target_mask = K.variable(raw_target_mask.astype("float32"))
+masks = K.concatenate([style_mask, target_mask], axis=0)
+
+# index constants for images and tasks variables
+STYLE, TARGET, CONTENT = 0, 1, 2
+
+# Build image model, mask model and use layer outputs as features
+# image model as VGG19
+image_model = vgg19.VGG19(include_top=False, input_tensor=images)
+
+# mask model as a series of pooling
+mask_input = Input(tensor=masks, shape=(None, None, None), name="mask_input")
+x = mask_input
+for layer in image_model.layers[1:]:
+    name = 'mask_%s' % layer.name
+    if 'conv' in layer.name:
+        x = AveragePooling2D((3, 3), strides=(
+            1, 1), name=name, border_mode="same")(x)
+    elif 'pool' in layer.name:
+        x = AveragePooling2D((2, 2), name=name)(x)
+mask_model = Model(mask_input, x)
+
+# Collect features from image_model and task_model
+image_features = {}
+mask_features = {}
+for img_layer, mask_layer in zip(image_model.layers, mask_model.layers):
+    if 'conv' in img_layer.name:
+        assert 'mask_' + img_layer.name == mask_layer.name
+        layer_name = img_layer.name
+        img_feat, mask_feat = img_layer.output, mask_layer.output
+        image_features[layer_name] = img_feat
+        mask_features[layer_name] = mask_feat
+
+
+# Define loss functions
+def gram_matrix(x):
+    assert K.ndim(x) == 3
+    features = K.batch_flatten(x)
+    gram = K.dot(features, K.transpose(features))
+    return gram
+
+
+def region_style_loss(style_image, target_image, style_mask, target_mask):
+    '''Calculate style loss between style_image and target_image,
+    for one common region specified by their (boolean) masks
+    '''
+    assert 3 == K.ndim(style_image) == K.ndim(target_image)
+    assert 2 == K.ndim(style_mask) == K.ndim(target_mask)
+    if K.image_dim_ordering() == 'th':
+        masked_style = style_image * style_mask
+        masked_target = target_image * target_mask
+        nb_channels = K.shape(style_image)[0]
+    else:
+        masked_style = K.permute_dimensions(
+            style_image, (2, 0, 1)) * style_mask
+        masked_target = K.permute_dimensions(
+            target_image, (2, 0, 1)) * target_mask
+        nb_channels = K.shape(style_image)[-1]
+    s = gram_matrix(masked_style) / K.mean(style_mask) / nb_channels
+    c = gram_matrix(masked_target) / K.mean(target_mask) / nb_channels
+    return K.mean(K.square(s - c))
+
+
+def style_loss(style_image, target_image, style_masks, target_masks):
+    '''Calculate style loss between style_image and target_image,
+    in all regions.
+    '''
+    assert 3 == K.ndim(style_image) == K.ndim(target_image)
+    assert 3 == K.ndim(style_masks) == K.ndim(target_masks)
+    loss = K.variable(0)
+    for i in xrange(nb_labels):
+        if K.image_dim_ordering() == 'th':
+            style_mask = style_masks[i, :, :]
+            target_mask = target_masks[i, :, :]
+        else:
+            style_mask = style_masks[:, :, i]
+            target_mask = target_masks[:, :, i]
+        loss += region_style_loss(style_image,
+                                  target_image, style_mask, target_mask)
+    return loss
+
+
+def content_loss(content_image, target_image):
+    return K.sum(K.square(target_image - content_image))
+
+
+def total_variation_loss(x):
+    assert 4 == K.ndim(x)
+    if K.image_dim_ordering() == 'th':
+        a = K.square(x[:, :, :img_nrows - 1, :img_ncols - 1] -
+                     x[:, :, 1:, :img_ncols - 1])
+        b = K.square(x[:, :, :img_nrows - 1, :img_ncols - 1] -
+                     x[:, :, :img_nrows - 1, 1:])
+    else:
+        a = K.square(x[:, :img_nrows - 1, :img_ncols - 1, :] -
+                     x[:, 1:, :img_ncols - 1, :])
+        b = K.square(x[:, :img_nrows - 1, :img_ncols - 1, :] -
+                     x[:, :img_nrows - 1, 1:, :])
+    return K.sum(K.pow(a + b, 1.25))
+
+# Overall loss is the weighted sum of content_loss, style_loss and tv_loss
+# Each individual loss uses features from image/mask models.
+loss = K.variable(0)
+for layer in content_feature_layers:
+    content_feat = image_features[layer][CONTENT, :, :, :]
+    target_feat = image_features[layer][TARGET, :, :, :]
+    loss += content_weight * content_loss(content_feat, target_feat)
+
+for layer in style_feature_layers:
+    style_feat = image_features[layer][STYLE, :, :, :]
+    target_feat = image_features[layer][TARGET, :, :, :]
+    style_masks = mask_features[layer][STYLE, :, :, :]
+    target_masks = mask_features[layer][TARGET, :, :, :]
+    sl = style_loss(style_feat, target_feat, style_masks, target_masks)
+    loss += (style_weight / len(style_feature_layers)) * sl
+
+loss += total_variation_weight * total_variation_loss(target_image)
+loss_grads = K.gradients(loss, target_image)
+
+# Evaluator class for computing efficiency
+outputs = [loss]
+if type(loss_grads) in {list, tuple}:
+    outputs += loss_grads
+else:
+    outputs.append(loss_grads)
+
+f_outputs = K.function([target_image], outputs)
+
+
+def eval_loss_and_grads(x):
+    if K.image_dim_ordering() == 'th':
+        x = x.reshape((1, 3, img_nrows, img_ncols))
+    else:
+        x = x.reshape((1, img_nrows, img_ncols, 3))
+    outs = f_outputs([x])
+    loss_value = outs[0]
+    if len(outs[1:]) == 1:
+        grad_values = outs[1].flatten().astype('float64')
+    else:
+        grad_values = np.array(outs[1:]).flatten().astype('float64')
+    return loss_value, grad_values
+
+
+class Evaluator(object):
+
+    def __init__(self):
+        self.loss_value = None
+        self.grads_values = None
+
+    def loss(self, x):
+        assert self.loss_value is None
+        loss_value, grad_values = eval_loss_and_grads(x)
+        self.loss_value = loss_value
+        self.grad_values = grad_values
+        return self.loss_value
+
+    def grads(self, x):
+        assert self.loss_value is not None
+        grad_values = np.copy(self.grad_values)
+        self.loss_value = None
+        self.grad_values = None
+        return grad_values
+
+evaluator = Evaluator()
+
+# Generate images by iterative optimization
+if K.image_dim_ordering() == 'th':
+    x = np.random.uniform(0, 255, (1, 3, img_nrows, img_ncols)) - 128.
+else:
+    x = np.random.uniform(0, 255, (1, img_nrows, img_ncols, 3)) - 128.
+
+for i in range(50):
+    print('Start of iteration', i)
+    start_time = time.time()
+    x, min_val, info = fmin_l_bfgs_b(evaluator.loss, x.flatten(),
+                                     fprime=evaluator.grads, maxfun=20)
+    print('Current loss value:', min_val)
+    # save current generated image
+    img = deprocess_image(x.copy())
+    fname = target_img_prefix + '_at_iteration_%d.png' % i
+    imsave(fname, img)
+    end_time = time.time()
+    print('Image saved as', fname)
+    print('Iteration %d completed in %ds' % (i, end_time - start_time))

examples/neural_style_transfer.py

@@ -0,0 +1,259 @@
+'''Neural style transfer with Keras.
+
+Run the script with:
+```
+python neural_style_transfer.py path_to_your_base_image.jpg path_to_your_reference.jpg prefix_for_results
+```
+e.g.:
+```
+python neural_style_transfer.py img/tuebingen.jpg img/starry_night.jpg results/my_result
+```
+
+It is preferable to run this script on GPU, for speed.
+
+Example result: https://twitter.com/fchollet/status/686631033085677568
+
+# Details
+
+Style transfer consists in generating an image
+with the same "content" as a base image, but with the
+"style" of a different picture (typically artistic).
+
+This is achieved through the optimization of a loss function
+that has 3 components: "style loss", "content loss",
+and "total variation loss":
+
+- The total variation loss imposes local spatial continuity between
+the pixels of the combination image, giving it visual coherence.
+
+- The style loss is where the deep learning keeps in --that one is defined
+using a deep convolutional neural network. Precisely, it consists in a sum of
+L2 distances between the Gram matrices of the representations of
+the base image and the style reference image, extracted from
+different layers of a convnet (trained on ImageNet). The general idea
+is to capture color/texture information at different spatial
+scales (fairly large scales --defined by the depth of the layer considered).
+
+ - The content loss is a L2 distance between the features of the base
+image (extracted from a deep layer) and the features of the combination image,
+keeping the generated image close enough to the original one.
+
+# References
+    - [A Neural Algorithm of Artistic Style](http://arxiv.org/abs/1508.06576)
+'''
+
+from __future__ import print_function
+from keras.preprocessing.image import load_img, img_to_array
+from scipy.misc import imsave
+import numpy as np
+from scipy.optimize import fmin_l_bfgs_b
+import time
+import argparse
+
+from keras.applications import vgg16
+from keras import backend as K
+
+parser = argparse.ArgumentParser(description='Neural style transfer with Keras.')
+parser.add_argument('base_image_path', metavar='base', type=str,
+                    help='Path to the image to transform.')
+parser.add_argument('style_reference_image_path', metavar='ref', type=str,
+                    help='Path to the style reference image.')
+parser.add_argument('result_prefix', metavar='res_prefix', type=str,
+                    help='Prefix for the saved results.')
+
+args = parser.parse_args()
+base_image_path = args.base_image_path
+style_reference_image_path = args.style_reference_image_path
+result_prefix = args.result_prefix
+
+# these are the weights of the different loss components
+total_variation_weight = 1.
+style_weight = 1.
+content_weight = 0.025
+
+# dimensions of the generated picture.
+img_nrows = 400
+img_ncols = 400
+assert img_ncols == img_nrows, 'Due to the use of the Gram matrix, width and height must match.'
+
+# util function to open, resize and format pictures into appropriate tensors
+def preprocess_image(image_path):
+    img = load_img(image_path, target_size=(img_nrows, img_ncols))
+    img = img_to_array(img)
+    img = np.expand_dims(img, axis=0)
+    img = vgg16.preprocess_input(img)
+    return img
+
+# util function to convert a tensor into a valid image
+def deprocess_image(x):
+    if K.image_dim_ordering() == 'th':
+        x = x.reshape((3, img_nrows, img_ncols))
+        x = x.transpose((1, 2, 0))
+    else:
+        x = x.reshape((img_nrows, img_ncols, 3))
+    x = x[:, :, ::-1]
+    x[:, :, 0] += 103.939
+    x[:, :, 1] += 116.779
+    x[:, :, 2] += 123.68
+    x = np.clip(x, 0, 255).astype('uint8')
+    return x
+
+# get tensor representations of our images
+base_image = K.variable(preprocess_image(base_image_path))
+style_reference_image = K.variable(preprocess_image(style_reference_image_path))
+
+# this will contain our generated image
+if K.image_dim_ordering() == 'th':
+    combination_image = K.placeholder((1, 3, img_nrows, img_ncols))
+else:
+    combination_image = K.placeholder((1, img_nrows, img_ncols, 3))
+
+# combine the 3 images into a single Keras tensor
+input_tensor = K.concatenate([base_image,
+                              style_reference_image,
+                              combination_image], axis=0)
+
+# build the VGG16 network with our 3 images as input
+# the model will be loaded with pre-trained ImageNet weights
+model = vgg16.VGG16(input_tensor=input_tensor,
+                    weights='imagenet', include_top=False)
+print('Model loaded.')
+
+# get the symbolic outputs of each "key" layer (we gave them unique names).
+outputs_dict = dict([(layer.name, layer.output) for layer in model.layers])
+
+# compute the neural style loss
+# first we need to define 4 util functions
+
+# the gram matrix of an image tensor (feature-wise outer product)
+def gram_matrix(x):
+    assert K.ndim(x) == 3
+    if K.image_dim_ordering() == 'th':
+        features = K.batch_flatten(x)
+    else:
+        features = K.batch_flatten(K.permute_dimensions(x, (2, 0, 1)))
+    gram = K.dot(features, K.transpose(features))
+    return gram
+
+# the "style loss" is designed to maintain
+# the style of the reference image in the generated image.
+# It is based on the gram matrices (which capture style) of
+# feature maps from the style reference image
+# and from the generated image
+def style_loss(style, combination):
+    assert K.ndim(style) == 3
+    assert K.ndim(combination) == 3
+    S = gram_matrix(style)
+    C = gram_matrix(combination)
+    channels = 3
+    size = img_nrows * img_ncols
+    return K.sum(K.square(S - C)) / (4. * (channels ** 2) * (size ** 2))
+
+# an auxiliary loss function
+# designed to maintain the "content" of the
+# base image in the generated image
+def content_loss(base, combination):
+    return K.sum(K.square(combination - base))
+
+# the 3rd loss function, total variation loss,
+# designed to keep the generated image locally coherent
+def total_variation_loss(x):
+    assert K.ndim(x) == 4
+    if K.image_dim_ordering() == 'th':
+        a = K.square(x[:, :, :img_nrows-1, :img_ncols-1] - x[:, :, 1:, :img_ncols-1])
+        b = K.square(x[:, :, :img_nrows-1, :img_ncols-1] - x[:, :, :img_nrows-1, 1:])
+    else:
+        a = K.square(x[:, :img_nrows-1, :img_ncols-1, :] - x[:, 1:, :img_ncols-1, :])
+        b = K.square(x[:, :img_nrows-1, :img_ncols-1, :] - x[:, :img_nrows-1, 1:, :])
+    return K.sum(K.pow(a + b, 1.25))
+
+# combine these loss functions into a single scalar
+loss = K.variable(0.)
+layer_features = outputs_dict['block4_conv2']
+base_image_features = layer_features[0, :, :, :]
+combination_features = layer_features[2, :, :, :]
+loss += content_weight * content_loss(base_image_features,
+                                      combination_features)
+
+feature_layers = ['block1_conv1', 'block2_conv1',
+                  'block3_conv1', 'block4_conv1',
+                  'block5_conv1']
+for layer_name in feature_layers:
+    layer_features = outputs_dict[layer_name]
+    style_reference_features = layer_features[1, :, :, :]
+    combination_features = layer_features[2, :, :, :]
+    sl = style_loss(style_reference_features, combination_features)
+    loss += (style_weight / len(feature_layers)) * sl
+loss += total_variation_weight * total_variation_loss(combination_image)
+
+# get the gradients of the generated image wrt the loss
+grads = K.gradients(loss, combination_image)
+
+outputs = [loss]
+if type(grads) in {list, tuple}:
+    outputs += grads
+else:
+    outputs.append(grads)
+
+f_outputs = K.function([combination_image], outputs)
+
+def eval_loss_and_grads(x):
+    if K.image_dim_ordering() == 'th':
+        x = x.reshape((1, 3, img_nrows, img_ncols))
+    else:
+        x = x.reshape((1, img_nrows, img_ncols, 3))
+    outs = f_outputs([x])
+    loss_value = outs[0]
+    if len(outs[1:]) == 1:
+        grad_values = outs[1].flatten().astype('float64')
+    else:
+        grad_values = np.array(outs[1:]).flatten().astype('float64')
+    return loss_value, grad_values
+
+# this Evaluator class makes it possible
+# to compute loss and gradients in one pass
+# while retrieving them via two separate functions,
+# "loss" and "grads". This is done because scipy.optimize
+# requires separate functions for loss and gradients,
+# but computing them separately would be inefficient.
+class Evaluator(object):
+    def __init__(self):
+        self.loss_value = None
+        self.grads_values = None
+
+    def loss(self, x):
+        assert self.loss_value is None
+        loss_value, grad_values = eval_loss_and_grads(x)
+        self.loss_value = loss_value
+        self.grad_values = grad_values
+        return self.loss_value
+
+    def grads(self, x):
+        assert self.loss_value is not None
+        grad_values = np.copy(self.grad_values)
+        self.loss_value = None
+        self.grad_values = None
+        return grad_values
+
+evaluator = Evaluator()
+
+# run scipy-based optimization (L-BFGS) over the pixels of the generated image
+# so as to minimize the neural style loss
+if K.image_dim_ordering() == 'th':
+    x = np.random.uniform(0, 255, (1, 3, img_nrows, img_ncols)) - 128.
+else:
+    x = np.random.uniform(0, 255, (1, img_nrows, img_ncols, 3)) - 128.
+
+for i in range(10):
+    print('Start of iteration', i)
+    start_time = time.time()
+    x, min_val, info = fmin_l_bfgs_b(evaluator.loss, x.flatten(),
+                                     fprime=evaluator.grads, maxfun=20)
+    print('Current loss value:', min_val)
+    # save current generated image
+    img = deprocess_image(x.copy())
+    fname = result_prefix + '_at_iteration_%d.png' % i
+    imsave(fname, img)
+    end_time = time.time()
+    print('Image saved as', fname)
+    print('Iteration %d completed in %ds' % (i, end_time - start_time))

examples/neural_turing_machine_copy.py

@@ -1,168 +0,0 @@
-from __future__ import absolute_import
-from __future__ import print_function
-import numpy as np
-np.random.seed(123)
-import matplotlib.pyplot as plt
-
-from theano import function
-
-from keras.models import Sequential
-from keras.layers.core import TimeDistributedDense, Activation
-from keras.layers.recurrent import LSTM
-from keras.optimizers import Adam
-from keras.utils import generic_utils
-
-from keras.layers.ntm import NeuralTuringMachine as NTM
-
-"""
-Copy Problem defined in Graves et. al [0]
-
-Training data is made of sequences with length 1 to 20. 
-Test data are sequences of length 100.
-The model is tested every 500 weight updates.
-After about 3500 updates, the accuracy jumps from around 50% to >90%.
-
-Estimated compile time: 12 min
-Estimated time to train Neural Turing Machine and 3 layer LSTM on an NVidia GTX 680: 2h
-
-[0]: http://arxiv.org/pdf/1410.5401v2.pdf
-"""
-
-batch_size = 100
-
-h_dim = 128
-n_slots = 128
-m_length = 20
-input_dim = 8
-lr = 1e-3
-clipvalue = 10
-
-##### Neural Turing Machine ######
-
-ntm = NTM(h_dim, n_slots=n_slots, m_length=m_length, shift_range=3,
-          inner_rnn='lstm', return_sequences=True, input_dim=input_dim)
-model = Sequential()
-model.add(ntm)
-model.add(TimeDistributedDense(input_dim))
-model.add(Activation('sigmoid'))
-
-sgd = Adam(lr=lr, clipvalue=clipvalue)
-model.compile(loss='binary_crossentropy', optimizer=sgd)
-
-# LSTM - Run this for comparison
-
-sgd2 = Adam(lr=lr, clipvalue=clipvalue)
-lstm = Sequential()
-lstm.add(LSTM(input_dim=input_dim, output_dim=h_dim*2, return_sequences=True))
-lstm.add(LSTM(output_dim=h_dim*2, return_sequences=True))
-lstm.add(LSTM(output_dim=h_dim*2, return_sequences=True))
-lstm.add(TimeDistributedDense(input_dim))
-lstm.add(Activation('sigmoid'))
-
-lstm.compile(loss='binary_crossentropy', optimizer=sgd)
-
-###### DATASET ########
-
-def get_sample(batch_size=128, n_bits=8, max_size=20, min_size=1):
-    # generate samples with random length
-    inp = np.zeros((batch_size, 2*max_size-1, n_bits))
-    out = np.zeros((batch_size, 2*max_size-1, n_bits))
-    sw = np.zeros((batch_size, 2*max_size-1, 1))
-    for i in range(batch_size):
-        t = np.random.randint(low=min_size, high=max_size)
-        x = np.random.uniform(size=(t, n_bits)) > .5
-        for j,f in enumerate(x.sum(axis=-1)): # remove fake flags
-            if f>=n_bits:
-                x[j, :] = 0.
-        del_flag = np.ones((1, n_bits))
-        inp[i, :t+1] = np.concatenate([x, del_flag], axis=0)
-        out[i, t+1:(2*t+1)] = x
-        sw[i, t+1:(2*t+1)] = 1
-    return inp, out, sw
-
-def show_pattern(inp, out, sw, file_name='ntm_output.png'):
-    ''' Helper function to visualize results '''
-    plt.figure(figsize=(10, 10))
-    plt.subplot(131)
-    plt.imshow(inp>.5)
-    plt.subplot(132)
-    plt.imshow(out>.5)
-    plt.subplot(133)
-    plt.imshow(sw>.5)
-    plt.savefig(file_name)
-    plt.close()
-
-# Show data example:
-inp, out, sw = get_sample(1, 8, 20)
-
-plt.subplot(131)
-plt.title('input')
-plt.imshow(inp[0], cmap='gray')
-plt.subplot(132)
-plt.title('desired')
-plt.imshow(out[0], cmap='gray')
-plt.subplot(133)
-plt.title('sample_weight')
-plt.imshow(sw[0], cmap='gray')
-
-# training uses sequences of length 1 to 20. Test uses series of length 100.
-def test_model(model, file_name, min_size=100):
-    I, V, sw = get_sample(batch_size=500, n_bits=input_dim, max_size=min_size+1, min_size=min_size)
-    Y = np.asarray(model.predict(I, batch_size=100) > .5).astype('float64')
-    acc = (V[:, -min_size:, :] == Y[:, -min_size:, :]).mean() * 100
-    show_pattern(Y[0], V[0], sw[0], file_name)
-    return acc
-
-##### TRAIN ######
-nb_epoch = 4000
-progbar = generic_utils.Progbar(nb_epoch)
-for e in range(nb_epoch):
-    I, V, sw = get_sample(n_bits=input_dim, max_size=20, min_size=1, batch_size=100)
-
-    loss1 = model.train_on_batch(I, V, sample_weight=sw)
-    loss2 = lstm.train_on_batch(I, V, sample_weight=sw)
-
-    progbar.add(1, values=[("NTM", loss1), ("LSTM", loss2)])
-
-    if e % 500 == 0:
-        print("")
-        acc1 = test_model(model, 'ntm.png')
-        acc2 = test_model(lstm, 'lstm.png')
-        print("NTM  test acc: {}".format(acc1))
-        print("LSTM test acc: {}".format(acc2))
-
-##### VISUALIZATION #####
-X = model.get_input()
-Y = ntm.get_full_output()[0:3]  # (memory over time, read_vectors, write_vectors)
-F = function([X], Y, allow_input_downcast=True)
-
-inp, out, sw = get_sample(1, 8, 21, 20)
-mem, read, write = F(inp.astype('float32'))
-Y = model.predict(inp)
-
-plt.figure(figsize=(15, 12))
-
-plt.subplot(221)
-plt.imshow(write[0])
-plt.xlabel('memory location')
-plt.ylabel('time')
-plt.title('write')
-
-plt.subplot(222)
-plt.imshow(read[0])
-plt.title('read')
-
-plt.subplot(223)
-plt.title('desired')
-plt.imshow(out[0])
-
-plt.subplot(224)
-plt.imshow(Y[0]>.5)
-plt.title('output')
-
-plt.figure(figsize=(15, 10))
-plt.subplot(325)
-plt.ylabel('time')
-plt.xlabel('location')
-plt.title('memory evolving in time (avg value per location)')
-plt.imshow(mem[0].mean(axis=-1))

examples/pretrained_word_embeddings.py

@@ -0,0 +1,144 @@
+'''This script loads pre-trained word embeddings (GloVe embeddings)
+into a frozen Keras Embedding layer, and uses it to
+train a text classification model on the 20 Newsgroup dataset
+(classication of newsgroup messages into 20 different categories).
+
+GloVe embedding data can be found at:
+http://nlp.stanford.edu/data/glove.6B.zip
+(source page: http://nlp.stanford.edu/projects/glove/)
+
+20 Newsgroup data can be found at:
+http://www.cs.cmu.edu/afs/cs.cmu.edu/project/theo-20/www/data/news20.html
+'''
+
+from __future__ import print_function
+import os
+import numpy as np
+np.random.seed(1337)
+
+from keras.preprocessing.text import Tokenizer
+from keras.preprocessing.sequence import pad_sequences
+from keras.utils.np_utils import to_categorical
+from keras.layers import Dense, Input, Flatten
+from keras.layers import Conv1D, MaxPooling1D, Embedding
+from keras.models import Model
+import sys
+
+BASE_DIR = ''
+GLOVE_DIR = BASE_DIR + '/glove.6B/'
+TEXT_DATA_DIR = BASE_DIR + '/20_newsgroup/'
+MAX_SEQUENCE_LENGTH = 1000
+MAX_NB_WORDS = 20000
+EMBEDDING_DIM = 100
+VALIDATION_SPLIT = 0.2
+
+# first, build index mapping words in the embeddings set
+# to their embedding vector
+
+print('Indexing word vectors.')
+
+embeddings_index = {}
+f = open(os.path.join(GLOVE_DIR, 'glove.6B.100d.txt'))
+for line in f:
+    values = line.split()
+    word = values[0]
+    coefs = np.asarray(values[1:], dtype='float32')
+    embeddings_index[word] = coefs
+f.close()
+
+print('Found %s word vectors.' % len(embeddings_index))
+
+# second, prepare text samples and their labels
+print('Processing text dataset')
+
+texts = []  # list of text samples
+labels_index = {}  # dictionary mapping label name to numeric id
+labels = []  # list of label ids
+for name in sorted(os.listdir(TEXT_DATA_DIR)):
+    path = os.path.join(TEXT_DATA_DIR, name)
+    if os.path.isdir(path):
+        label_id = len(labels_index)
+        labels_index[name] = label_id
+        for fname in sorted(os.listdir(path)):
+            if fname.isdigit():
+                fpath = os.path.join(path, fname)
+                if sys.version_info < (3,):
+                    f = open(fpath)
+                else:
+                    f = open(fpath, encoding='latin-1')
+                texts.append(f.read())
+                f.close()
+                labels.append(label_id)
+
+print('Found %s texts.' % len(texts))
+
+# finally, vectorize the text samples into a 2D integer tensor
+tokenizer = Tokenizer(nb_words=MAX_NB_WORDS)
+tokenizer.fit_on_texts(texts)
+sequences = tokenizer.texts_to_sequences(texts)
+
+word_index = tokenizer.word_index
+print('Found %s unique tokens.' % len(word_index))
+
+data = pad_sequences(sequences, maxlen=MAX_SEQUENCE_LENGTH)
+
+labels = to_categorical(np.asarray(labels))
+print('Shape of data tensor:', data.shape)
+print('Shape of label tensor:', labels.shape)
+
+# split the data into a training set and a validation set
+indices = np.arange(data.shape[0])
+np.random.shuffle(indices)
+data = data[indices]
+labels = labels[indices]
+nb_validation_samples = int(VALIDATION_SPLIT * data.shape[0])
+
+x_train = data[:-nb_validation_samples]
+y_train = labels[:-nb_validation_samples]
+x_val = data[-nb_validation_samples:]
+y_val = labels[-nb_validation_samples:]
+
+print('Preparing embedding matrix.')
+
+# prepare embedding matrix
+nb_words = min(MAX_NB_WORDS, len(word_index))
+embedding_matrix = np.zeros((nb_words + 1, EMBEDDING_DIM))
+for word, i in word_index.items():
+    if i > MAX_NB_WORDS:
+        continue
+    embedding_vector = embeddings_index.get(word)
+    if embedding_vector is not None:
+        # words not found in embedding index will be all-zeros.
+        embedding_matrix[i] = embedding_vector
+
+# load pre-trained word embeddings into an Embedding layer
+# note that we set trainable = False so as to keep the embeddings fixed
+embedding_layer = Embedding(nb_words + 1,
+                            EMBEDDING_DIM,
+                            weights=[embedding_matrix],
+                            input_length=MAX_SEQUENCE_LENGTH,
+                            trainable=False)
+
+print('Training model.')
+
+# train a 1D convnet with global maxpooling
+sequence_input = Input(shape=(MAX_SEQUENCE_LENGTH,), dtype='int32')
+embedded_sequences = embedding_layer(sequence_input)
+x = Conv1D(128, 5, activation='relu')(embedded_sequences)
+x = MaxPooling1D(5)(x)
+x = Conv1D(128, 5, activation='relu')(x)
+x = MaxPooling1D(5)(x)
+x = Conv1D(128, 5, activation='relu')(x)
+x = MaxPooling1D(35)(x)
+x = Flatten()(x)
+x = Dense(128, activation='relu')(x)
+preds = Dense(len(labels_index), activation='softmax')(x)
+
+model = Model(sequence_input, preds)
+model.compile(loss='categorical_crossentropy',
+              optimizer='rmsprop',
+              metrics=['acc'])
+
+# happy learning!
+model.fit(x_train, y_train, validation_data=(x_val, y_val),
+          nb_epoch=2, batch_size=128)

examples/reuters_mlp.py

@@ -1,28 +1,22 @@
-from __future__ import absolute_import
+'''Trains and evaluate a simple MLP
+on the Reuters newswire topic classification task.
+'''
+
 from __future__ import print_function
 import numpy as np
 np.random.seed(1337)  # for reproducibility
 
 from keras.datasets import reuters
 from keras.models import Sequential
-from keras.layers.core import Dense, Dropout, Activation
-from keras.layers.normalization import BatchNormalization
+from keras.layers import Dense, Dropout, Activation
 from keras.utils import np_utils
 from keras.preprocessing.text import Tokenizer
 
-'''
-    Train and evaluate a simple MLP on the Reuters newswire topic classification task.
-    GPU run command:
-        THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 python examples/reuters_mlp.py
-    CPU run command:
-        python examples/reuters_mlp.py
-'''
-
 max_words = 1000
 batch_size = 32
 nb_epoch = 5
 
-print("Loading data...")
+print('Loading data...')
 (X_train, y_train), (X_test, y_test) = reuters.load_data(nb_words=max_words, test_split=0.2)
 print(len(X_train), 'train sequences')
 print(len(X_test), 'test sequences')
@@ -30,20 +24,20 @@ print(len(X_test), 'test sequences')
 nb_classes = np.max(y_train)+1
 print(nb_classes, 'classes')
 
-print("Vectorizing sequence data...")
+print('Vectorizing sequence data...')
 tokenizer = Tokenizer(nb_words=max_words)
-X_train = tokenizer.sequences_to_matrix(X_train, mode="binary")
-X_test = tokenizer.sequences_to_matrix(X_test, mode="binary")
+X_train = tokenizer.sequences_to_matrix(X_train, mode='binary')
+X_test = tokenizer.sequences_to_matrix(X_test, mode='binary')
 print('X_train shape:', X_train.shape)
 print('X_test shape:', X_test.shape)
 
-print("Convert class vector to binary class matrix (for use with categorical_crossentropy)")
+print('Convert class vector to binary class matrix (for use with categorical_crossentropy)')
 Y_train = np_utils.to_categorical(y_train, nb_classes)
 Y_test = np_utils.to_categorical(y_test, nb_classes)
 print('Y_train shape:', Y_train.shape)
 print('Y_test shape:', Y_test.shape)
 
-print("Building model...")
+print('Building model...')
 model = Sequential()
 model.add(Dense(512, input_shape=(max_words,)))
 model.add(Activation('relu'))
@@ -51,9 +45,14 @@ model.add(Dropout(0.5))
 model.add(Dense(nb_classes))
 model.add(Activation('softmax'))
 
-model.compile(loss='categorical_crossentropy', optimizer='adam')
+model.compile(loss='categorical_crossentropy',
+              optimizer='adam',
+              metrics=['accuracy'])
 
-history = model.fit(X_train, Y_train, nb_epoch=nb_epoch, batch_size=batch_size, verbose=1, show_accuracy=True, validation_split=0.1)
-score = model.evaluate(X_test, Y_test, batch_size=batch_size, verbose=1, show_accuracy=True)
+history = model.fit(X_train, Y_train,
+                    nb_epoch=nb_epoch, batch_size=batch_size,
+                    verbose=1, validation_split=0.1)
+score = model.evaluate(X_test, Y_test,
+                       batch_size=batch_size, verbose=1)
 print('Test score:', score[0])
 print('Test accuracy:', score[1])

examples/stateful_lstm.py

@@ -0,0 +1,84 @@
+'''Example script showing how to use stateful RNNs
+to model long sequences efficiently.
+'''
+from __future__ import print_function
+import numpy as np
+import matplotlib.pyplot as plt
+from keras.models import Sequential
+from keras.layers import Dense, LSTM
+
+
+# since we are using stateful rnn tsteps can be set to 1
+tsteps = 1
+batch_size = 25
+epochs = 25
+# number of elements ahead that are used to make the prediction
+lahead = 1
+
+
+def gen_cosine_amp(amp=100, period=1000, x0=0, xn=50000, step=1, k=0.0001):
+    """Generates an absolute cosine time series with the amplitude
+    exponentially decreasing
+
+    Arguments:
+        amp: amplitude of the cosine function
+        period: period of the cosine function
+        x0: initial x of the time series
+        xn: final x of the time series
+        step: step of the time series discretization
+        k: exponential rate
+    """
+    cos = np.zeros(((xn - x0) * step, 1, 1))
+    for i in range(len(cos)):
+        idx = x0 + i * step
+        cos[i, 0, 0] = amp * np.cos(2 * np.pi * idx / period)
+        cos[i, 0, 0] = cos[i, 0, 0] * np.exp(-k * idx)
+    return cos
+
+
+print('Generating Data')
+cos = gen_cosine_amp()
+print('Input shape:', cos.shape)
+
+expected_output = np.zeros((len(cos), 1))
+for i in range(len(cos) - lahead):
+    expected_output[i, 0] = np.mean(cos[i + 1:i + lahead + 1])
+
+print('Output shape')
+print(expected_output.shape)
+
+print('Creating Model')
+model = Sequential()
+model.add(LSTM(50,
+               batch_input_shape=(batch_size, tsteps, 1),
+               return_sequences=True,
+               stateful=True))
+model.add(LSTM(50,
+               batch_input_shape=(batch_size, tsteps, 1),
+               return_sequences=False,
+               stateful=True))
+model.add(Dense(1))
+model.compile(loss='mse', optimizer='rmsprop')
+
+print('Training')
+for i in range(epochs):
+    print('Epoch', i, '/', epochs)
+    model.fit(cos,
+              expected_output,
+              batch_size=batch_size,
+              verbose=1,
+              nb_epoch=1,
+              shuffle=False)
+    model.reset_states()
+
+print('Predicting')
+predicted_output = model.predict(cos, batch_size=batch_size)
+
+print('Plotting Results')
+plt.subplot(2, 1, 1)
+plt.plot(expected_output)
+plt.title('Expected')
+plt.subplot(2, 1, 2)
+plt.plot(predicted_output)
+plt.title('Predicted')
+plt.show()

examples/variational_autoencoder.py

@@ -0,0 +1,97 @@
+'''This script demonstrates how to build a variational autoencoder with Keras.
+
+Reference: "Auto-Encoding Variational Bayes" https://arxiv.org/abs/1312.6114
+'''
+import numpy as np
+import matplotlib.pyplot as plt
+
+from keras.layers import Input, Dense, Lambda
+from keras.models import Model
+from keras import backend as K
+from keras import objectives
+from keras.datasets import mnist
+
+batch_size = 100
+original_dim = 784
+latent_dim = 2
+intermediate_dim = 256
+nb_epoch = 50
+
+x = Input(batch_shape=(batch_size, original_dim))
+h = Dense(intermediate_dim, activation='relu')(x)
+z_mean = Dense(latent_dim)(h)
+z_log_var = Dense(latent_dim)(h)
+
+
+def sampling(args):
+    z_mean, z_log_var = args
+    epsilon = K.random_normal(shape=(batch_size, latent_dim), mean=0.)
+    return z_mean + K.exp(z_log_var / 2) * epsilon
+
+# note that "output_shape" isn't necessary with the TensorFlow backend
+z = Lambda(sampling, output_shape=(latent_dim,))([z_mean, z_log_var])
+
+# we instantiate these layers separately so as to reuse them later
+decoder_h = Dense(intermediate_dim, activation='relu')
+decoder_mean = Dense(original_dim, activation='sigmoid')
+h_decoded = decoder_h(z)
+x_decoded_mean = decoder_mean(h_decoded)
+
+
+def vae_loss(x, x_decoded_mean):
+    xent_loss = original_dim * objectives.binary_crossentropy(x, x_decoded_mean)
+    kl_loss = - 0.5 * K.sum(1 + z_log_var - K.square(z_mean) - K.exp(z_log_var), axis=-1)
+    return xent_loss + kl_loss
+
+vae = Model(x, x_decoded_mean)
+vae.compile(optimizer='rmsprop', loss=vae_loss)
+
+# train the VAE on MNIST digits
+(x_train, y_train), (x_test, y_test) = mnist.load_data()
+
+x_train = x_train.astype('float32') / 255.
+x_test = x_test.astype('float32') / 255.
+x_train = x_train.reshape((len(x_train), np.prod(x_train.shape[1:])))
+x_test = x_test.reshape((len(x_test), np.prod(x_test.shape[1:])))
+
+vae.fit(x_train, x_train,
+        shuffle=True,
+        nb_epoch=nb_epoch,
+        batch_size=batch_size,
+        validation_data=(x_test, x_test))
+
+# build a model to project inputs on the latent space
+encoder = Model(x, z_mean)
+
+# display a 2D plot of the digit classes in the latent space
+x_test_encoded = encoder.predict(x_test, batch_size=batch_size)
+plt.figure(figsize=(6, 6))
+plt.scatter(x_test_encoded[:, 0], x_test_encoded[:, 1], c=y_test)
+plt.colorbar()
+plt.show()
+
+# build a digit generator that can sample from the learned distribution
+decoder_input = Input(shape=(latent_dim,))
+_h_decoded = decoder_h(decoder_input)
+_x_decoded_mean = decoder_mean(_h_decoded)
+generator = Model(decoder_input, _x_decoded_mean)
+
+# display a 2D manifold of the digits
+n = 15  # figure with 15x15 digits
+digit_size = 28
+figure = np.zeros((digit_size * n, digit_size * n))
+# we will sample n points within [-15, 15] standard deviations
+grid_x = np.linspace(-15, 15, n)
+grid_y = np.linspace(-15, 15, n)
+
+for i, yi in enumerate(grid_x):
+    for j, xi in enumerate(grid_y):
+        z_sample = np.array([[xi, yi]])
+        x_decoded = generator.predict(z_sample)
+        digit = x_decoded[0].reshape(digit_size, digit_size)
+        figure[i * digit_size: (i + 1) * digit_size,
+               j * digit_size: (j + 1) * digit_size] = digit
+
+plt.figure(figsize=(10, 10))
+plt.imshow(figure)
+plt.show()

examples/variational_autoencoder_deconv.py

@@ -0,0 +1,124 @@
+'''This script demonstrates how to build a variational autoencoder with Keras and deconvolution layers.
+
+Reference: "Auto-Encoding Variational Bayes" https://arxiv.org/abs/1312.6114
+'''
+import numpy as np
+import matplotlib.pyplot as plt
+
+from keras.layers import Input, Dense, Lambda, Flatten, Reshape
+from keras.layers import Convolution2D, Deconvolution2D, MaxPooling2D
+from keras.models import Model
+from keras import backend as K
+from keras import objectives
+from keras.datasets import mnist
+
+# input image dimensions
+img_rows, img_cols, img_chns = 28, 28, 1
+# number of convolutional filters to use
+nb_filters = 32
+# convolution kernel size
+nb_conv = 3
+
+batch_size = 16
+original_dim = (img_chns, img_rows, img_cols)
+latent_dim = 2
+intermediate_dim = 128
+epsilon_std = 0.01
+nb_epoch = 5
+
+
+x = Input(batch_shape=(batch_size,) + original_dim)
+c = Convolution2D(nb_filters, nb_conv, nb_conv, border_mode='same', activation='relu')(x)
+f = Flatten()(c)
+h = Dense(intermediate_dim, activation='relu')(f)
+
+z_mean = Dense(latent_dim)(h)
+z_log_var = Dense(latent_dim)(h)
+
+
+def sampling(args):
+    z_mean, z_log_var = args
+    epsilon = K.random_normal(shape=(batch_size, latent_dim),
+                              mean=0., std=epsilon_std)
+    return z_mean + K.exp(z_log_var) * epsilon
+
+# note that "output_shape" isn't necessary with the TensorFlow backend
+# so you could write `Lambda(sampling)([z_mean, z_log_var])`
+z = Lambda(sampling, output_shape=(latent_dim,))([z_mean, z_log_var])
+
+# we instantiate these layers separately so as to reuse them later
+decoder_h = Dense(intermediate_dim, activation='relu')
+decoder_f = Dense(nb_filters*img_rows*img_cols, activation='relu')
+decoder_c = Reshape((nb_filters, img_rows, img_cols))
+decoder_mean = Deconvolution2D(img_chns, nb_conv, nb_conv,
+                               (batch_size, img_chns, img_rows, img_cols),
+                               border_mode='same')
+
+h_decoded = decoder_h(z)
+f_decoded = decoder_f(h_decoded)
+c_decoded = decoder_c(f_decoded)
+x_decoded_mean = decoder_mean(c_decoded)
+
+
+def vae_loss(x, x_decoded_mean):
+    # NOTE: binary_crossentropy expects a batch_size by dim for x and x_decoded_mean, so we MUST flatten these!
+    x = K.flatten(x)
+    x_decoded_mean = K.flatten(x_decoded_mean)
+    xent_loss = objectives.binary_crossentropy(x, x_decoded_mean)
+    kl_loss = - 0.5 * K.mean(1 + z_log_var - K.square(z_mean) - K.exp(z_log_var), axis=-1)
+    return xent_loss + kl_loss
+
+vae = Model(x, x_decoded_mean)
+vae.compile(optimizer='rmsprop', loss=vae_loss)
+vae.summary()
+
+# train the VAE on MNIST digits
+(x_train, y_train), (x_test, y_test) = mnist.load_data()
+
+x_train = x_train.astype('float32')[:, None, :, :] / 255.
+x_test = x_test.astype('float32')[:, None, :, :] / 255.
+
+vae.fit(x_train, x_train,
+        shuffle=True,
+        nb_epoch=nb_epoch,
+        batch_size=batch_size,
+        validation_data=(x_test, x_test))
+
+
+# build a model to project inputs on the latent space
+encoder = Model(x, z_mean)
+
+# display a 2D plot of the digit classes in the latent space
+x_test_encoded = encoder.predict(x_test, batch_size=batch_size)
+plt.figure(figsize=(6, 6))
+plt.scatter(x_test_encoded[:, 0], x_test_encoded[:, 1], c=y_test)
+plt.colorbar()
+plt.show()
+
+# build a digit generator that can sample from the learned distribution
+decoder_input = Input(shape=(latent_dim,))
+_h_decoded = decoder_h(decoder_input)
+_f_decoded = decoder_f(_h_decoded)
+_c_decoded = decoder_c(_f_decoded)
+_x_decoded_mean = decoder_mean(_c_decoded)
+generator = Model(decoder_input, _x_decoded_mean)
+
+# display a 2D manifold of the digits
+n = 15  # figure with 15x15 digits
+digit_size = 28
+figure = np.zeros((digit_size * n, digit_size * n))
+# we will sample n points within [-15, 15] standard deviations
+grid_x = np.linspace(-15, 15, n)
+grid_y = np.linspace(-15, 15, n)
+
+for i, yi in enumerate(grid_x):
+    for j, xi in enumerate(grid_y):
+        z_sample = np.array([[xi, yi]])
+        x_decoded = generator.predict(z_sample)
+        digit = x_decoded[0].reshape(digit_size, digit_size)
+        figure[i * digit_size: (i + 1) * digit_size,
+               j * digit_size: (j + 1) * digit_size] = digit
+
+plt.figure(figsize=(10, 10))
+plt.imshow(figure)
+plt.show()

keras/__init__.py

@@ -1,15 +1,18 @@
+from __future__ import absolute_import
+from . import backend
+from . import datasets
+from . import engine
+from . import layers
+from . import preprocessing
+from . import utils
+from . import wrappers
+from . import callbacks
+from . import constraints
+from . import initializations
+from . import metrics
+from . import models
+from . import objectives
+from . import optimizers
+from . import regularizers
 
-"""
-Keras: Theano-based Deep Learning library
-==================================
-Keras is a minimalist, highly modular neural network library in 
-the spirit of Torch, written in Python / Theano so as not to have 
-to deal with the dearth of ecosystem in Lua. It was developed with 
-a focus on enabling fast experimentation. Being able to go from 
-idea to result with the least possible delay is key to doing 
-good research.
-
-See http://keras.io/
-"""
-
-__version__ = '0.2.0'
+__version__ = '1.1.0'

keras/activations.py

@@ -7,11 +7,9 @@ def softmax(x):
     if ndim == 2:
         return K.softmax(x)
     elif ndim == 3:
-        # apply softmax to each timestep
-        def step(x, states):
-            return K.softmax(x), []
-        last_output, outputs, states = K.rnn(step, x, [], masking=False)
-        return outputs
+        e = K.exp(x - K.max(x, axis=-1, keepdims=True))
+        s = K.sum(e, axis=-1, keepdims=True)
+        return e / s
     else:
         raise Exception('Cannot apply softmax to a tensor that is not 2D or 3D. ' +
                         'Here, ndim=' + str(ndim))
@@ -21,6 +19,10 @@ def softplus(x):
     return K.softplus(x)
 
 
+def softsign(x):
+    return K.softsign(x)
+
+
 def relu(x, alpha=0., max_value=None):
     return K.relu(x, alpha=alpha, max_value=max_value)
 
@@ -39,11 +41,13 @@ def hard_sigmoid(x):
 
 def linear(x):
     '''
-    The function returns the variable that is passed in, so all types work
+    The function returns the variable that is passed in, so all types work.
     '''
     return x
 
 
 from .utils.generic_utils import get_from_module
 def get(identifier):
+    if identifier is None:
+        return linear
     return get_from_module(identifier, globals(), 'activation function')

keras/applications/__init__.py

@@ -0,0 +1,4 @@
+from .vgg16 import VGG16
+from .vgg19 import VGG19
+from .resnet50 import ResNet50
+from .inception_v3 import InceptionV3

keras/applications/imagenet_utils.py

@@ -0,0 +1,43 @@
+import numpy as np
+import json
+
+from ..utils.data_utils import get_file
+from .. import backend as K
+
+CLASS_INDEX = None
+CLASS_INDEX_PATH = 'https://s3.amazonaws.com/deep-learning-models/image-models/imagenet_class_index.json'
+
+
+def preprocess_input(x, dim_ordering='default'):
+    if dim_ordering == 'default':
+        dim_ordering = K.image_dim_ordering()
+    assert dim_ordering in {'tf', 'th'}
+
+    if dim_ordering == 'th':
+        x[:, 0, :, :] -= 103.939
+        x[:, 1, :, :] -= 116.779
+        x[:, 2, :, :] -= 123.68
+        # 'RGB'->'BGR'
+        x = x[:, ::-1, :, :]
+    else:
+        x[:, :, :, 0] -= 103.939
+        x[:, :, :, 1] -= 116.779
+        x[:, :, :, 2] -= 123.68
+        # 'RGB'->'BGR'
+        x = x[:, :, :, ::-1]
+    return x
+
+
+def decode_predictions(preds):
+    global CLASS_INDEX
+    assert len(preds.shape) == 2 and preds.shape[1] == 1000
+    if CLASS_INDEX is None:
+        fpath = get_file('imagenet_class_index.json',
+                         CLASS_INDEX_PATH,
+                         cache_subdir='models')
+        CLASS_INDEX = json.load(open(fpath))
+    indices = np.argmax(preds, axis=-1)
+    results = []
+    for i in indices:
+        results.append(CLASS_INDEX[str(i)])
+    return results

keras/applications/inception_v3.py

@@ -0,0 +1,312 @@
+# -*- coding: utf-8 -*-
+'''Inception V3 model for Keras.
+
+Note that the ImageNet weights provided are from a model that had not fully converged.
+Inception v3 should be able to reach 6.9% top-5 error, but our model
+only gets to 7.8% (same as a fully-converged ResNet 50).
+For comparison, VGG16 only gets to 9.9%, quite a bit worse.
+
+Also, do note that the input image format for this model is different than for
+other models (299x299 instead of 224x224), and that the input preprocessing function
+is also different.
+
+# Reference:
+
+- [Rethinking the Inception Architecture for Computer Vision](http://arxiv.org/abs/1512.00567)
+
+'''
+from __future__ import print_function
+from __future__ import absolute_import
+
+import warnings
+
+from ..models import Model
+from ..layers import Flatten, Dense, Input, BatchNormalization, merge
+from ..layers import Convolution2D, MaxPooling2D, AveragePooling2D
+from ..utils.layer_utils import convert_all_kernels_in_model
+from ..utils.data_utils import get_file
+from .. import backend as K
+from .imagenet_utils import decode_predictions
+
+
+TH_WEIGHTS_PATH = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.2/inception_v3_weights_th_dim_ordering_th_kernels.h5'
+TF_WEIGHTS_PATH = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.2/inception_v3_weights_tf_dim_ordering_tf_kernels.h5'
+TH_WEIGHTS_PATH_NO_TOP = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.2/inception_v3_weights_th_dim_ordering_th_kernels_notop.h5'
+TF_WEIGHTS_PATH_NO_TOP = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.2/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5'
+
+
+def conv2d_bn(x, nb_filter, nb_row, nb_col,
+              border_mode='same', subsample=(1, 1),
+              name=None):
+    '''Utility function to apply conv + BN.
+    '''
+    if name is not None:
+        bn_name = name + '_bn'
+        conv_name = name + '_conv'
+    else:
+        bn_name = None
+        conv_name = None
+    if K.image_dim_ordering() == 'th':
+        bn_axis = 1
+    else:
+        bn_axis = 3
+    x = Convolution2D(nb_filter, nb_row, nb_col,
+                      subsample=subsample,
+                      activation='relu',
+                      border_mode=border_mode,
+                      name=conv_name)(x)
+    x = BatchNormalization(axis=bn_axis, name=bn_name)(x)
+    return x
+
+
+def InceptionV3(include_top=True, weights='imagenet',
+                input_tensor=None):
+    '''Instantiate the Inception v3 architecture,
+    optionally loading weights pre-trained
+    on ImageNet. Note that when using TensorFlow,
+    for best performance you should set
+    `image_dim_ordering="tf"` in your Keras config
+    at ~/.keras/keras.json.
+
+    The model and the weights are compatible with both
+    TensorFlow and Theano. The dimension ordering
+    convention used by the model is the one
+    specified in your Keras config file.
+
+    Note that the default input image size for this model is 299x299.
+
+    # Arguments
+        include_top: whether to include the 3 fully-connected
+            layers at the top of the network.
+        weights: one of `None` (random initialization)
+            or "imagenet" (pre-training on ImageNet).
+        input_tensor: optional Keras tensor (i.e. output of `layers.Input()`)
+            to use as image input for the model.
+
+    # Returns
+        A Keras model instance.
+    '''
+    if weights not in {'imagenet', None}:
+        raise ValueError('The `weights` argument should be either '
+                         '`None` (random initialization) or `imagenet` '
+                         '(pre-training on ImageNet).')
+    # Determine proper input shape
+    if K.image_dim_ordering() == 'th':
+        if include_top:
+            input_shape = (3, 299, 299)
+        else:
+            input_shape = (3, None, None)
+    else:
+        if include_top:
+            input_shape = (299, 299, 3)
+        else:
+            input_shape = (None, None, 3)
+
+    if input_tensor is None:
+        img_input = Input(shape=input_shape)
+    else:
+        if not K.is_keras_tensor(input_tensor):
+            img_input = Input(tensor=input_tensor, shape=input_shape)
+        else:
+            img_input = input_tensor
+
+    if K.image_dim_ordering() == 'th':
+        channel_axis = 1
+    else:
+        channel_axis = 3
+
+    x = conv2d_bn(img_input, 32, 3, 3, subsample=(2, 2), border_mode='valid')
+    x = conv2d_bn(x, 32, 3, 3, border_mode='valid')
+    x = conv2d_bn(x, 64, 3, 3)
+    x = MaxPooling2D((3, 3), strides=(2, 2))(x)
+
+    x = conv2d_bn(x, 80, 1, 1, border_mode='valid')
+    x = conv2d_bn(x, 192, 3, 3, border_mode='valid')
+    x = MaxPooling2D((3, 3), strides=(2, 2))(x)
+
+    # mixed 0, 1, 2: 35 x 35 x 256
+    for i in range(3):
+        branch1x1 = conv2d_bn(x, 64, 1, 1)
+
+        branch5x5 = conv2d_bn(x, 48, 1, 1)
+        branch5x5 = conv2d_bn(branch5x5, 64, 5, 5)
+
+        branch3x3dbl = conv2d_bn(x, 64, 1, 1)
+        branch3x3dbl = conv2d_bn(branch3x3dbl, 96, 3, 3)
+        branch3x3dbl = conv2d_bn(branch3x3dbl, 96, 3, 3)
+
+        branch_pool = AveragePooling2D(
+            (3, 3), strides=(1, 1), border_mode='same')(x)
+        branch_pool = conv2d_bn(branch_pool, 32, 1, 1)
+        x = merge([branch1x1, branch5x5, branch3x3dbl, branch_pool],
+                  mode='concat', concat_axis=channel_axis,
+                  name='mixed' + str(i))
+
+    # mixed 3: 17 x 17 x 768
+    branch3x3 = conv2d_bn(x, 384, 3, 3, subsample=(2, 2), border_mode='valid')
+
+    branch3x3dbl = conv2d_bn(x, 64, 1, 1)
+    branch3x3dbl = conv2d_bn(branch3x3dbl, 96, 3, 3)
+    branch3x3dbl = conv2d_bn(branch3x3dbl, 96, 3, 3,
+                             subsample=(2, 2), border_mode='valid')
+
+    branch_pool = MaxPooling2D((3, 3), strides=(2, 2))(x)
+    x = merge([branch3x3, branch3x3dbl, branch_pool],
+              mode='concat', concat_axis=channel_axis,
+              name='mixed3')
+
+    # mixed 4: 17 x 17 x 768
+    branch1x1 = conv2d_bn(x, 192, 1, 1)
+
+    branch7x7 = conv2d_bn(x, 128, 1, 1)
+    branch7x7 = conv2d_bn(branch7x7, 128, 1, 7)
+    branch7x7 = conv2d_bn(branch7x7, 192, 7, 1)
+
+    branch7x7dbl = conv2d_bn(x, 128, 1, 1)
+    branch7x7dbl = conv2d_bn(branch7x7dbl, 128, 7, 1)
+    branch7x7dbl = conv2d_bn(branch7x7dbl, 128, 1, 7)
+    branch7x7dbl = conv2d_bn(branch7x7dbl, 128, 7, 1)
+    branch7x7dbl = conv2d_bn(branch7x7dbl, 192, 1, 7)
+
+    branch_pool = AveragePooling2D((3, 3), strides=(1, 1), border_mode='same')(x)
+    branch_pool = conv2d_bn(branch_pool, 192, 1, 1)
+    x = merge([branch1x1, branch7x7, branch7x7dbl, branch_pool],
+              mode='concat', concat_axis=channel_axis,
+              name='mixed4')
+
+    # mixed 5, 6: 17 x 17 x 768
+    for i in range(2):
+        branch1x1 = conv2d_bn(x, 192, 1, 1)
+
+        branch7x7 = conv2d_bn(x, 160, 1, 1)
+        branch7x7 = conv2d_bn(branch7x7, 160, 1, 7)
+        branch7x7 = conv2d_bn(branch7x7, 192, 7, 1)
+
+        branch7x7dbl = conv2d_bn(x, 160, 1, 1)
+        branch7x7dbl = conv2d_bn(branch7x7dbl, 160, 7, 1)
+        branch7x7dbl = conv2d_bn(branch7x7dbl, 160, 1, 7)
+        branch7x7dbl = conv2d_bn(branch7x7dbl, 160, 7, 1)
+        branch7x7dbl = conv2d_bn(branch7x7dbl, 192, 1, 7)
+
+        branch_pool = AveragePooling2D(
+            (3, 3), strides=(1, 1), border_mode='same')(x)
+        branch_pool = conv2d_bn(branch_pool, 192, 1, 1)
+        x = merge([branch1x1, branch7x7, branch7x7dbl, branch_pool],
+                  mode='concat', concat_axis=channel_axis,
+                  name='mixed' + str(5 + i))
+
+    # mixed 7: 17 x 17 x 768
+    branch1x1 = conv2d_bn(x, 192, 1, 1)
+
+    branch7x7 = conv2d_bn(x, 192, 1, 1)
+    branch7x7 = conv2d_bn(branch7x7, 192, 1, 7)
+    branch7x7 = conv2d_bn(branch7x7, 192, 7, 1)
+
+    branch7x7dbl = conv2d_bn(x, 160, 1, 1)
+    branch7x7dbl = conv2d_bn(branch7x7dbl, 192, 7, 1)
+    branch7x7dbl = conv2d_bn(branch7x7dbl, 192, 1, 7)
+    branch7x7dbl = conv2d_bn(branch7x7dbl, 192, 7, 1)
+    branch7x7dbl = conv2d_bn(branch7x7dbl, 192, 1, 7)
+
+    branch_pool = AveragePooling2D((3, 3), strides=(1, 1), border_mode='same')(x)
+    branch_pool = conv2d_bn(branch_pool, 192, 1, 1)
+    x = merge([branch1x1, branch7x7, branch7x7dbl, branch_pool],
+              mode='concat', concat_axis=channel_axis,
+              name='mixed7')
+
+    # mixed 8: 8 x 8 x 1280
+    branch3x3 = conv2d_bn(x, 192, 1, 1)
+    branch3x3 = conv2d_bn(branch3x3, 320, 3, 3,
+                          subsample=(2, 2), border_mode='valid')
+
+    branch7x7x3 = conv2d_bn(x, 192, 1, 1)
+    branch7x7x3 = conv2d_bn(branch7x7x3, 192, 1, 7)
+    branch7x7x3 = conv2d_bn(branch7x7x3, 192, 7, 1)
+    branch7x7x3 = conv2d_bn(branch7x7x3, 192, 3, 3,
+                            subsample=(2, 2), border_mode='valid')
+
+    branch_pool = AveragePooling2D((3, 3), strides=(2, 2))(x)
+    x = merge([branch3x3, branch7x7x3, branch_pool],
+              mode='concat', concat_axis=channel_axis,
+              name='mixed8')
+
+    # mixed 9: 8 x 8 x 2048
+    for i in range(2):
+        branch1x1 = conv2d_bn(x, 320, 1, 1)
+
+        branch3x3 = conv2d_bn(x, 384, 1, 1)
+        branch3x3_1 = conv2d_bn(branch3x3, 384, 1, 3)
+        branch3x3_2 = conv2d_bn(branch3x3, 384, 3, 1)
+        branch3x3 = merge([branch3x3_1, branch3x3_2],
+                          mode='concat', concat_axis=channel_axis,
+                          name='mixed9_' + str(i))
+
+        branch3x3dbl = conv2d_bn(x, 448, 1, 1)
+        branch3x3dbl = conv2d_bn(branch3x3dbl, 384, 3, 3)
+        branch3x3dbl_1 = conv2d_bn(branch3x3dbl, 384, 1, 3)
+        branch3x3dbl_2 = conv2d_bn(branch3x3dbl, 384, 3, 1)
+        branch3x3dbl = merge([branch3x3dbl_1, branch3x3dbl_2],
+                             mode='concat', concat_axis=channel_axis)
+
+        branch_pool = AveragePooling2D(
+            (3, 3), strides=(1, 1), border_mode='same')(x)
+        branch_pool = conv2d_bn(branch_pool, 192, 1, 1)
+        x = merge([branch1x1, branch3x3, branch3x3dbl, branch_pool],
+                  mode='concat', concat_axis=channel_axis,
+                  name='mixed' + str(9 + i))
+
+    if include_top:
+        # Classification block
+        x = AveragePooling2D((8, 8), strides=(8, 8), name='avg_pool')(x)
+        x = Flatten(name='flatten')(x)
+        x = Dense(1000, activation='softmax', name='predictions')(x)
+
+    # Create model
+    model = Model(img_input, x)
+
+    # load weights
+    if weights == 'imagenet':
+        if K.image_dim_ordering() == 'th':
+            if include_top:
+                weights_path = get_file('inception_v3_weights_th_dim_ordering_th_kernels.h5',
+                                        TH_WEIGHTS_PATH,
+                                        cache_subdir='models',
+                                        md5_hash='b3baf3070cc4bf476d43a2ea61b0ca5f')
+            else:
+                weights_path = get_file('inception_v3_weights_th_dim_ordering_th_kernels_notop.h5',
+                                        TH_WEIGHTS_PATH_NO_TOP,
+                                        cache_subdir='models',
+                                        md5_hash='79aaa90ab4372b4593ba3df64e142f05')
+            model.load_weights(weights_path)
+            if K.backend() == 'tensorflow':
+                warnings.warn('You are using the TensorFlow backend, yet you '
+                              'are using the Theano '
+                              'image dimension ordering convention '
+                              '(`image_dim_ordering="th"`). '
+                              'For best performance, set '
+                              '`image_dim_ordering="tf"` in '
+                              'your Keras config '
+                              'at ~/.keras/keras.json.')
+                convert_all_kernels_in_model(model)
+        else:
+            if include_top:
+                weights_path = get_file('inception_v3_weights_tf_dim_ordering_tf_kernels.h5',
+                                        TF_WEIGHTS_PATH,
+                                        cache_subdir='models',
+                                        md5_hash='fe114b3ff2ea4bf891e9353d1bbfb32f')
+            else:
+                weights_path = get_file('inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5',
+                                        TF_WEIGHTS_PATH_NO_TOP,
+                                        cache_subdir='models',
+                                        md5_hash='2f3609166de1d967d1a481094754f691')
+            model.load_weights(weights_path)
+            if K.backend() == 'theano':
+                convert_all_kernels_in_model(model)
+    return model
+
+
+def preprocess_input(x):
+    x /= 255.
+    x -= 0.5
+    x *= 2.
+    return x

keras/applications/resnet50.py

@@ -0,0 +1,235 @@
+# -*- coding: utf-8 -*-
+'''ResNet50 model for Keras.
+
+# Reference:
+
+- [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385)
+
+Adapted from code contributed by BigMoyan.
+'''
+from __future__ import print_function
+from __future__ import absolute_import
+
+import warnings
+
+from ..layers import merge, Input
+from ..layers import Dense, Activation, Flatten
+from ..layers import Convolution2D, MaxPooling2D, ZeroPadding2D, AveragePooling2D
+from ..layers import BatchNormalization
+from ..models import Model
+from .. import backend as K
+from ..utils.layer_utils import convert_all_kernels_in_model
+from ..utils.data_utils import get_file
+from .imagenet_utils import decode_predictions, preprocess_input
+
+
+TH_WEIGHTS_PATH = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.2/resnet50_weights_th_dim_ordering_th_kernels.h5'
+TF_WEIGHTS_PATH = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.2/resnet50_weights_tf_dim_ordering_tf_kernels.h5'
+TH_WEIGHTS_PATH_NO_TOP = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.2/resnet50_weights_th_dim_ordering_th_kernels_notop.h5'
+TF_WEIGHTS_PATH_NO_TOP = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.2/resnet50_weights_tf_dim_ordering_tf_kernels_notop.h5'
+
+
+def identity_block(input_tensor, kernel_size, filters, stage, block):
+    '''The identity_block is the block that has no conv layer at shortcut
+
+    # Arguments
+        input_tensor: input tensor
+        kernel_size: defualt 3, the kernel size of middle conv layer at main path
+        filters: list of integers, the nb_filters of 3 conv layer at main path
+        stage: integer, current stage label, used for generating layer names
+        block: 'a','b'..., current block label, used for generating layer names
+    '''
+    nb_filter1, nb_filter2, nb_filter3 = filters
+    if K.image_dim_ordering() == 'tf':
+        bn_axis = 3
+    else:
+        bn_axis = 1
+    conv_name_base = 'res' + str(stage) + block + '_branch'
+    bn_name_base = 'bn' + str(stage) + block + '_branch'
+
+    x = Convolution2D(nb_filter1, 1, 1, name=conv_name_base + '2a')(input_tensor)
+    x = BatchNormalization(axis=bn_axis, name=bn_name_base + '2a')(x)
+    x = Activation('relu')(x)
+
+    x = Convolution2D(nb_filter2, kernel_size, kernel_size,
+                      border_mode='same', name=conv_name_base + '2b')(x)
+    x = BatchNormalization(axis=bn_axis, name=bn_name_base + '2b')(x)
+    x = Activation('relu')(x)
+
+    x = Convolution2D(nb_filter3, 1, 1, name=conv_name_base + '2c')(x)
+    x = BatchNormalization(axis=bn_axis, name=bn_name_base + '2c')(x)
+
+    x = merge([x, input_tensor], mode='sum')
+    x = Activation('relu')(x)
+    return x
+
+
+def conv_block(input_tensor, kernel_size, filters, stage, block, strides=(2, 2)):
+    '''conv_block is the block that has a conv layer at shortcut
+
+    # Arguments
+        input_tensor: input tensor
+        kernel_size: defualt 3, the kernel size of middle conv layer at main path
+        filters: list of integers, the nb_filters of 3 conv layer at main path
+        stage: integer, current stage label, used for generating layer names
+        block: 'a','b'..., current block label, used for generating layer names
+
+    Note that from stage 3, the first conv layer at main path is with subsample=(2,2)
+    And the shortcut should have subsample=(2,2) as well
+    '''
+    nb_filter1, nb_filter2, nb_filter3 = filters
+    if K.image_dim_ordering() == 'tf':
+        bn_axis = 3
+    else:
+        bn_axis = 1
+    conv_name_base = 'res' + str(stage) + block + '_branch'
+    bn_name_base = 'bn' + str(stage) + block + '_branch'
+
+    x = Convolution2D(nb_filter1, 1, 1, subsample=strides,
+                      name=conv_name_base + '2a')(input_tensor)
+    x = BatchNormalization(axis=bn_axis, name=bn_name_base + '2a')(x)
+    x = Activation('relu')(x)
+
+    x = Convolution2D(nb_filter2, kernel_size, kernel_size, border_mode='same',
+                      name=conv_name_base + '2b')(x)
+    x = BatchNormalization(axis=bn_axis, name=bn_name_base + '2b')(x)
+    x = Activation('relu')(x)
+
+    x = Convolution2D(nb_filter3, 1, 1, name=conv_name_base + '2c')(x)
+    x = BatchNormalization(axis=bn_axis, name=bn_name_base + '2c')(x)
+
+    shortcut = Convolution2D(nb_filter3, 1, 1, subsample=strides,
+                             name=conv_name_base + '1')(input_tensor)
+    shortcut = BatchNormalization(axis=bn_axis, name=bn_name_base + '1')(shortcut)
+
+    x = merge([x, shortcut], mode='sum')
+    x = Activation('relu')(x)
+    return x
+
+
+def ResNet50(include_top=True, weights='imagenet',
+             input_tensor=None):
+    '''Instantiate the ResNet50 architecture,
+    optionally loading weights pre-trained
+    on ImageNet. Note that when using TensorFlow,
+    for best performance you should set
+    `image_dim_ordering="tf"` in your Keras config
+    at ~/.keras/keras.json.
+
+    The model and the weights are compatible with both
+    TensorFlow and Theano. The dimension ordering
+    convention used by the model is the one
+    specified in your Keras config file.
+
+    # Arguments
+        include_top: whether to include the 3 fully-connected
+            layers at the top of the network.
+        weights: one of `None` (random initialization)
+            or "imagenet" (pre-training on ImageNet).
+        input_tensor: optional Keras tensor (i.e. xput of `layers.Input()`)
+            to use as image input for the model.
+
+    # Returns
+        A Keras model instance.
+    '''
+    if weights not in {'imagenet', None}:
+        raise ValueError('The `weights` argument should be either '
+                         '`None` (random initialization) or `imagenet` '
+                         '(pre-training on ImageNet).')
+    # Determine proper input shape
+    if K.image_dim_ordering() == 'th':
+        if include_top:
+            input_shape = (3, 224, 224)
+        else:
+            input_shape = (3, None, None)
+    else:
+        if include_top:
+            input_shape = (224, 224, 3)
+        else:
+            input_shape = (None, None, 3)
+
+    if input_tensor is None:
+        img_input = Input(shape=input_shape)
+    else:
+        if not K.is_keras_tensor(input_tensor):
+            img_input = Input(tensor=input_tensor, shape=input_shape)
+        else:
+            img_input = input_tensor
+    if K.image_dim_ordering() == 'tf':
+        bn_axis = 3
+    else:
+        bn_axis = 1
+
+    x = ZeroPadding2D((3, 3))(img_input)
+    x = Convolution2D(64, 7, 7, subsample=(2, 2), name='conv1')(x)
+    x = BatchNormalization(axis=bn_axis, name='bn_conv1')(x)
+    x = Activation('relu')(x)
+    x = MaxPooling2D((3, 3), strides=(2, 2))(x)
+
+    x = conv_block(x, 3, [64, 64, 256], stage=2, block='a', strides=(1, 1))
+    x = identity_block(x, 3, [64, 64, 256], stage=2, block='b')
+    x = identity_block(x, 3, [64, 64, 256], stage=2, block='c')
+
+    x = conv_block(x, 3, [128, 128, 512], stage=3, block='a')
+    x = identity_block(x, 3, [128, 128, 512], stage=3, block='b')
+    x = identity_block(x, 3, [128, 128, 512], stage=3, block='c')
+    x = identity_block(x, 3, [128, 128, 512], stage=3, block='d')
+
+    x = conv_block(x, 3, [256, 256, 1024], stage=4, block='a')
+    x = identity_block(x, 3, [256, 256, 1024], stage=4, block='b')
+    x = identity_block(x, 3, [256, 256, 1024], stage=4, block='c')
+    x = identity_block(x, 3, [256, 256, 1024], stage=4, block='d')
+    x = identity_block(x, 3, [256, 256, 1024], stage=4, block='e')
+    x = identity_block(x, 3, [256, 256, 1024], stage=4, block='f')
+
+    x = conv_block(x, 3, [512, 512, 2048], stage=5, block='a')
+    x = identity_block(x, 3, [512, 512, 2048], stage=5, block='b')
+    x = identity_block(x, 3, [512, 512, 2048], stage=5, block='c')
+
+    x = AveragePooling2D((7, 7), name='avg_pool')(x)
+
+    if include_top:
+        x = Flatten()(x)
+        x = Dense(1000, activation='softmax', name='fc1000')(x)
+
+    model = Model(img_input, x)
+
+    # load weights
+    if weights == 'imagenet':
+        if K.image_dim_ordering() == 'th':
+            if include_top:
+                weights_path = get_file('resnet50_weights_th_dim_ordering_th_kernels.h5',
+                                        TH_WEIGHTS_PATH,
+                                        cache_subdir='models',
+                                        md5_hash='1c1f8f5b0c8ee28fe9d950625a230e1c')
+            else:
+                weights_path = get_file('resnet50_weights_th_dim_ordering_th_kernels_notop.h5',
+                                        TH_WEIGHTS_PATH_NO_TOP,
+                                        cache_subdir='models',
+                                        md5_hash='f64f049c92468c9affcd44b0976cdafe')
+            model.load_weights(weights_path)
+            if K.backend() == 'tensorflow':
+                warnings.warn('You are using the TensorFlow backend, yet you '
+                              'are using the Theano '
+                              'image dimension ordering convention '
+                              '(`image_dim_ordering="th"`). '
+                              'For best performance, set '
+                              '`image_dim_ordering="tf"` in '
+                              'your Keras config '
+                              'at ~/.keras/keras.json.')
+                convert_all_kernels_in_model(model)
+        else:
+            if include_top:
+                weights_path = get_file('resnet50_weights_tf_dim_ordering_tf_kernels.h5',
+                                        TF_WEIGHTS_PATH,
+                                        cache_subdir='models',
+                                        md5_hash='a7b3fe01876f51b976af0dea6bc144eb')
+            else:
+                weights_path = get_file('resnet50_weights_tf_dim_ordering_tf_kernels_notop.h5',
+                                        TF_WEIGHTS_PATH_NO_TOP,
+                                        cache_subdir='models',
+                                        md5_hash='a268eb855778b3df3c7506639542a6af')
+            model.load_weights(weights_path)
+            if K.backend() == 'theano':
+                convert_all_kernels_in_model(model)
+    return model

keras/applications/vgg16.py

@@ -0,0 +1,149 @@
+# -*- coding: utf-8 -*-
+'''VGG16 model for Keras.
+
+# Reference:
+
+- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://arxiv.org/abs/1409.1556)
+
+'''
+from __future__ import print_function
+from __future__ import absolute_import
+
+import warnings
+
+from ..models import Model
+from ..layers import Flatten, Dense, Input
+from ..layers import Convolution2D, MaxPooling2D
+from ..utils.layer_utils import convert_all_kernels_in_model
+from ..utils.data_utils import get_file
+from .. import backend as K
+from .imagenet_utils import decode_predictions, preprocess_input
+
+
+TH_WEIGHTS_PATH = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.1/vgg16_weights_th_dim_ordering_th_kernels.h5'
+TF_WEIGHTS_PATH = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.1/vgg16_weights_tf_dim_ordering_tf_kernels.h5'
+TH_WEIGHTS_PATH_NO_TOP = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.1/vgg16_weights_th_dim_ordering_th_kernels_notop.h5'
+TF_WEIGHTS_PATH_NO_TOP = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.1/vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5'
+
+
+def VGG16(include_top=True, weights='imagenet',
+          input_tensor=None):
+    '''Instantiate the VGG16 architecture,
+    optionally loading weights pre-trained
+    on ImageNet. Note that when using TensorFlow,
+    for best performance you should set
+    `image_dim_ordering="tf"` in your Keras config
+    at ~/.keras/keras.json.
+
+    The model and the weights are compatible with both
+    TensorFlow and Theano. The dimension ordering
+    convention used by the model is the one
+    specified in your Keras config file.
+
+    # Arguments
+        include_top: whether to include the 3 fully-connected
+            layers at the top of the network.
+        weights: one of `None` (random initialization)
+            or "imagenet" (pre-training on ImageNet).
+        input_tensor: optional Keras tensor (i.e. output of `layers.Input()`)
+            to use as image input for the model.
+
+    # Returns
+        A Keras model instance.
+    '''
+    if weights not in {'imagenet', None}:
+        raise ValueError('The `weights` argument should be either '
+                         '`None` (random initialization) or `imagenet` '
+                         '(pre-training on ImageNet).')
+    # Determine proper input shape
+    if K.image_dim_ordering() == 'th':
+        if include_top:
+            input_shape = (3, 224, 224)
+        else:
+            input_shape = (3, None, None)
+    else:
+        if include_top:
+            input_shape = (224, 224, 3)
+        else:
+            input_shape = (None, None, 3)
+
+    if input_tensor is None:
+        img_input = Input(shape=input_shape)
+    else:
+        if not K.is_keras_tensor(input_tensor):
+            img_input = Input(tensor=input_tensor, shape=input_shape)
+        else:
+            img_input = input_tensor
+    # Block 1
+    x = Convolution2D(64, 3, 3, activation='relu', border_mode='same', name='block1_conv1')(img_input)
+    x = Convolution2D(64, 3, 3, activation='relu', border_mode='same', name='block1_conv2')(x)
+    x = MaxPooling2D((2, 2), strides=(2, 2), name='block1_pool')(x)
+
+    # Block 2
+    x = Convolution2D(128, 3, 3, activation='relu', border_mode='same', name='block2_conv1')(x)
+    x = Convolution2D(128, 3, 3, activation='relu', border_mode='same', name='block2_conv2')(x)
+    x = MaxPooling2D((2, 2), strides=(2, 2), name='block2_pool')(x)
+
+    # Block 3
+    x = Convolution2D(256, 3, 3, activation='relu', border_mode='same', name='block3_conv1')(x)
+    x = Convolution2D(256, 3, 3, activation='relu', border_mode='same', name='block3_conv2')(x)
+    x = Convolution2D(256, 3, 3, activation='relu', border_mode='same', name='block3_conv3')(x)
+    x = MaxPooling2D((2, 2), strides=(2, 2), name='block3_pool')(x)
+
+    # Block 4
+    x = Convolution2D(512, 3, 3, activation='relu', border_mode='same', name='block4_conv1')(x)
+    x = Convolution2D(512, 3, 3, activation='relu', border_mode='same', name='block4_conv2')(x)
+    x = Convolution2D(512, 3, 3, activation='relu', border_mode='same', name='block4_conv3')(x)
+    x = MaxPooling2D((2, 2), strides=(2, 2), name='block4_pool')(x)
+
+    # Block 5
+    x = Convolution2D(512, 3, 3, activation='relu', border_mode='same', name='block5_conv1')(x)
+    x = Convolution2D(512, 3, 3, activation='relu', border_mode='same', name='block5_conv2')(x)
+    x = Convolution2D(512, 3, 3, activation='relu', border_mode='same', name='block5_conv3')(x)
+    x = MaxPooling2D((2, 2), strides=(2, 2), name='block5_pool')(x)
+
+    if include_top:
+        # Classification block
+        x = Flatten(name='flatten')(x)
+        x = Dense(4096, activation='relu', name='fc1')(x)
+        x = Dense(4096, activation='relu', name='fc2')(x)
+        x = Dense(1000, activation='softmax', name='predictions')(x)
+
+    # Create model
+    model = Model(img_input, x)
+
+    # load weights
+    if weights == 'imagenet':
+        if K.image_dim_ordering() == 'th':
+            if include_top:
+                weights_path = get_file('vgg16_weights_th_dim_ordering_th_kernels.h5',
+                                        TH_WEIGHTS_PATH,
+                                        cache_subdir='models')
+            else:
+                weights_path = get_file('vgg16_weights_th_dim_ordering_th_kernels_notop.h5',
+                                        TH_WEIGHTS_PATH_NO_TOP,
+                                        cache_subdir='models')
+            model.load_weights(weights_path)
+            if K.backend() == 'tensorflow':
+                warnings.warn('You are using the TensorFlow backend, yet you '
+                              'are using the Theano '
+                              'image dimension ordering convention '
+                              '(`image_dim_ordering="th"`). '
+                              'For best performance, set '
+                              '`image_dim_ordering="tf"` in '
+                              'your Keras config '
+                              'at ~/.keras/keras.json.')
+                convert_all_kernels_in_model(model)
+        else:
+            if include_top:
+                weights_path = get_file('vgg16_weights_tf_dim_ordering_tf_kernels.h5',
+                                        TF_WEIGHTS_PATH,
+                                        cache_subdir='models')
+            else:
+                weights_path = get_file('vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5',
+                                        TF_WEIGHTS_PATH_NO_TOP,
+                                        cache_subdir='models')
+            model.load_weights(weights_path)
+            if K.backend() == 'theano':
+                convert_all_kernels_in_model(model)
+    return model

keras/applications/vgg19.py

@@ -0,0 +1,152 @@
+# -*- coding: utf-8 -*-
+'''VGG19 model for Keras.
+
+# Reference:
+
+- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://arxiv.org/abs/1409.1556)
+
+'''
+from __future__ import print_function
+from __future__ import absolute_import
+
+import warnings
+
+from ..models import Model
+from ..layers import Flatten, Dense, Input
+from ..layers import Convolution2D, MaxPooling2D
+from ..utils.layer_utils import convert_all_kernels_in_model
+from ..utils.data_utils import get_file
+from .. import backend as K
+from .imagenet_utils import decode_predictions, preprocess_input
+
+
+TH_WEIGHTS_PATH = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.1/vgg19_weights_th_dim_ordering_th_kernels.h5'
+TF_WEIGHTS_PATH = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.1/vgg19_weights_tf_dim_ordering_tf_kernels.h5'
+TH_WEIGHTS_PATH_NO_TOP = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.1/vgg19_weights_th_dim_ordering_th_kernels_notop.h5'
+TF_WEIGHTS_PATH_NO_TOP = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.1/vgg19_weights_tf_dim_ordering_tf_kernels_notop.h5'
+
+
+def VGG19(include_top=True, weights='imagenet',
+          input_tensor=None):
+    '''Instantiate the VGG19 architecture,
+    optionally loading weights pre-trained
+    on ImageNet. Note that when using TensorFlow,
+    for best performance you should set
+    `image_dim_ordering="tf"` in your Keras config
+    at ~/.keras/keras.json.
+
+    The model and the weights are compatible with both
+    TensorFlow and Theano. The dimension ordering
+    convention used by the model is the one
+    specified in your Keras config file.
+
+    # Arguments
+        include_top: whether to include the 3 fully-connected
+            layers at the top of the network.
+        weights: one of `None` (random initialization)
+            or "imagenet" (pre-training on ImageNet).
+        input_tensor: optional Keras tensor (i.e. output of `layers.Input()`)
+            to use as image input for the model.
+
+    # Returns
+        A Keras model instance.
+    '''
+    if weights not in {'imagenet', None}:
+        raise ValueError('The `weights` argument should be either '
+                         '`None` (random initialization) or `imagenet` '
+                         '(pre-training on ImageNet).')
+    # Determine proper input shape
+    if K.image_dim_ordering() == 'th':
+        if include_top:
+            input_shape = (3, 224, 224)
+        else:
+            input_shape = (3, None, None)
+    else:
+        if include_top:
+            input_shape = (224, 224, 3)
+        else:
+            input_shape = (None, None, 3)
+
+    if input_tensor is None:
+        img_input = Input(shape=input_shape)
+    else:
+        if not K.is_keras_tensor(input_tensor):
+            img_input = Input(tensor=input_tensor, shape=input_shape)
+        else:
+            img_input = input_tensor
+    # Block 1
+    x = Convolution2D(64, 3, 3, activation='relu', border_mode='same', name='block1_conv1')(img_input)
+    x = Convolution2D(64, 3, 3, activation='relu', border_mode='same', name='block1_conv2')(x)
+    x = MaxPooling2D((2, 2), strides=(2, 2), name='block1_pool')(x)
+
+    # Block 2
+    x = Convolution2D(128, 3, 3, activation='relu', border_mode='same', name='block2_conv1')(x)
+    x = Convolution2D(128, 3, 3, activation='relu', border_mode='same', name='block2_conv2')(x)
+    x = MaxPooling2D((2, 2), strides=(2, 2), name='block2_pool')(x)
+
+    # Block 3
+    x = Convolution2D(256, 3, 3, activation='relu', border_mode='same', name='block3_conv1')(x)
+    x = Convolution2D(256, 3, 3, activation='relu', border_mode='same', name='block3_conv2')(x)
+    x = Convolution2D(256, 3, 3, activation='relu', border_mode='same', name='block3_conv3')(x)
+    x = Convolution2D(256, 3, 3, activation='relu', border_mode='same', name='block3_conv4')(x)
+    x = MaxPooling2D((2, 2), strides=(2, 2), name='block3_pool')(x)
+
+    # Block 4
+    x = Convolution2D(512, 3, 3, activation='relu', border_mode='same', name='block4_conv1')(x)
+    x = Convolution2D(512, 3, 3, activation='relu', border_mode='same', name='block4_conv2')(x)
+    x = Convolution2D(512, 3, 3, activation='relu', border_mode='same', name='block4_conv3')(x)
+    x = Convolution2D(512, 3, 3, activation='relu', border_mode='same', name='block4_conv4')(x)
+    x = MaxPooling2D((2, 2), strides=(2, 2), name='block4_pool')(x)
+
+    # Block 5
+    x = Convolution2D(512, 3, 3, activation='relu', border_mode='same', name='block5_conv1')(x)
+    x = Convolution2D(512, 3, 3, activation='relu', border_mode='same', name='block5_conv2')(x)
+    x = Convolution2D(512, 3, 3, activation='relu', border_mode='same', name='block5_conv3')(x)
+    x = Convolution2D(512, 3, 3, activation='relu', border_mode='same', name='block5_conv4')(x)
+    x = MaxPooling2D((2, 2), strides=(2, 2), name='block5_pool')(x)
+
+    if include_top:
+        # Classification block
+        x = Flatten(name='flatten')(x)
+        x = Dense(4096, activation='relu', name='fc1')(x)
+        x = Dense(4096, activation='relu', name='fc2')(x)
+        x = Dense(1000, activation='softmax', name='predictions')(x)
+
+    # Create model
+    model = Model(img_input, x)
+
+    # load weights
+    if weights == 'imagenet':
+        if K.image_dim_ordering() == 'th':
+            if include_top:
+                weights_path = get_file('vgg19_weights_th_dim_ordering_th_kernels.h5',
+                                        TH_WEIGHTS_PATH,
+                                        cache_subdir='models')
+            else:
+                weights_path = get_file('vgg19_weights_th_dim_ordering_th_kernels_notop.h5',
+                                        TH_WEIGHTS_PATH_NO_TOP,
+                                        cache_subdir='models')
+            model.load_weights(weights_path)
+            if K.backend() == 'tensorflow':
+                warnings.warn('You are using the TensorFlow backend, yet you '
+                              'are using the Theano '
+                              'image dimension ordering convention '
+                              '(`image_dim_ordering="th"`). '
+                              'For best performance, set '
+                              '`image_dim_ordering="tf"` in '
+                              'your Keras config '
+                              'at ~/.keras/keras.json.')
+                convert_all_kernels_in_model(model)
+        else:
+            if include_top:
+                weights_path = get_file('vgg19_weights_tf_dim_ordering_tf_kernels.h5',
+                                        TF_WEIGHTS_PATH,
+                                        cache_subdir='models')
+            else:
+                weights_path = get_file('vgg19_weights_tf_dim_ordering_tf_kernels_notop.h5',
+                                        TF_WEIGHTS_PATH_NO_TOP,
+                                        cache_subdir='models')
+            model.load_weights(weights_path)
+            if K.backend() == 'theano':
+                convert_all_kernels_in_model(model)
+    return model

keras/backend/__init__.py

@@ -2,38 +2,72 @@ from __future__ import absolute_import
 from __future__ import print_function
 import os
 import json
-from .common import epsilon, floatx, set_epsilon, set_floatx
+import sys
+from .common import epsilon
+from .common import floatx
+from .common import set_epsilon
+from .common import set_floatx
+from .common import get_uid
+from .common import cast_to_floatx
+from .common import image_dim_ordering
+from .common import set_image_dim_ordering
+from .common import is_keras_tensor
+from .common import legacy_weight_ordering
+from .common import set_legacy_weight_ordering
 
-_keras_dir = os.path.expanduser(os.path.join('~', '.keras'))
+_keras_base_dir = os.path.expanduser('~')
+if not os.access(_keras_base_dir, os.W_OK):
+    _keras_base_dir = '/tmp'
+
+_keras_dir = os.path.join(_keras_base_dir, '.keras')
 if not os.path.exists(_keras_dir):
     os.makedirs(_keras_dir)
 
-_BACKEND = 'theano'
-_config_path = os.path.expanduser(os.path.join('~', '.keras', 'keras.json'))
+_BACKEND = 'tensorflow'
+_config_path = os.path.expanduser(os.path.join(_keras_dir, 'keras.json'))
 if os.path.exists(_config_path):
     _config = json.load(open(_config_path))
     _floatx = _config.get('floatx', floatx())
-    assert _floatx in {'float32', 'float64'}
+    assert _floatx in {'float16', 'float32', 'float64'}
     _epsilon = _config.get('epsilon', epsilon())
     assert type(_epsilon) == float
     _backend = _config.get('backend', _BACKEND)
     assert _backend in {'theano', 'tensorflow'}
+    _image_dim_ordering = _config.get('image_dim_ordering', image_dim_ordering())
+    assert _image_dim_ordering in {'tf', 'th'}
 
     set_floatx(_floatx)
     set_epsilon(_epsilon)
+    set_image_dim_ordering(_image_dim_ordering)
     _BACKEND = _backend
-else:
-    # save config file, for easy edition
+
+# save config file
+if not os.path.exists(_config_path):
     _config = {'floatx': floatx(),
                'epsilon': epsilon(),
-               'backend': _BACKEND}
-    json.dump(_config, open(_config_path, 'w'))
+               'backend': _BACKEND,
+               'image_dim_ordering': image_dim_ordering()}
+    with open(_config_path, 'w') as f:
+        f.write(json.dumps(_config, indent=4))
 
+if 'KERAS_BACKEND' in os.environ:
+    _backend = os.environ['KERAS_BACKEND']
+    assert _backend in {'theano', 'tensorflow'}
+    _BACKEND = _backend
+
+# import backend
 if _BACKEND == 'theano':
-    print('Using Theano backend.')
+    sys.stderr.write('Using Theano backend.\n')
     from .theano_backend import *
 elif _BACKEND == 'tensorflow':
-    print('Using TensorFlow backend.')
+    sys.stderr.write('Using TensorFlow backend.\n')
     from .tensorflow_backend import *
 else:
-    raise Exception('Unknown backend: ' + str(backend))
+    raise Exception('Unknown backend: ' + str(_BACKEND))
+
+
+def backend():
+    '''Publicly accessible method
+    for determining the current backend.
+    '''
+    return _BACKEND

keras/backend/common.py

@@ -1,31 +1,89 @@
 import numpy as np
 
+from collections import defaultdict
+
 # the type of float to use throughout the session.
 _FLOATX = 'float32'
 _EPSILON = 10e-8
+_UID_PREFIXES = defaultdict(int)
+_IMAGE_DIM_ORDERING = 'tf'
+_LEGACY_WEIGHT_ORDERING = False
 
 
 def epsilon():
+    '''Returns the value of the fuzz
+    factor used in numeric expressions.
+    '''
     return _EPSILON
 
 
 def set_epsilon(e):
+    '''Sets the value of the fuzz
+    factor used in numeric expressions.
+    '''
     global _EPSILON
     _EPSILON = e
 
 
 def floatx():
+    '''Returns the default float type, as a string
+    (e.g. 'float16', 'float32', 'float64').
+    '''
     return _FLOATX
 
 
 def set_floatx(floatx):
     global _FLOATX
-    if floatx not in {'float32', 'float64'}:
+    if floatx not in {'float16', 'float32', 'float64'}:
         raise Exception('Unknown floatx type: ' + str(floatx))
-    _FLOATX = floatx
+    _FLOATX = str(floatx)
 
 
 def cast_to_floatx(x):
     '''Cast a Numpy array to floatx.
     '''
     return np.asarray(x, dtype=_FLOATX)
+
+
+def image_dim_ordering():
+    '''Returns the image dimension ordering
+    convention ('th' or 'tf').
+    '''
+    return _IMAGE_DIM_ORDERING
+
+
+def set_image_dim_ordering(dim_ordering):
+    '''Sets the value of the image dimension
+    ordering convention ('th' or 'tf').
+    '''
+    global _IMAGE_DIM_ORDERING
+    if dim_ordering not in {'tf', 'th'}:
+        raise Exception('Unknown dim_ordering:', dim_ordering)
+    _IMAGE_DIM_ORDERING = str(dim_ordering)
+
+
+def get_uid(prefix=''):
+    _UID_PREFIXES[prefix] += 1
+    return _UID_PREFIXES[prefix]
+
+
+def reset_uids():
+    global _UID_PREFIXES
+    _UID_PREFIXES = defaultdict(int)
+
+
+def is_keras_tensor(x):
+    if hasattr(x, '_keras_shape'):
+        return True
+    else:
+        return False
+
+
+def set_legacy_weight_ordering(value):
+    global _LEGACY_WEIGHT_ORDERING
+    assert value in {True, False}
+    _LEGACY_WEIGHT_ORDERING = value
+
+
+def legacy_weight_ordering():
+    return _LEGACY_WEIGHT_ORDERING

keras/callbacks.py

@@ -8,6 +8,8 @@ import warnings
 
 from collections import deque
 from .utils.generic_utils import Progbar
+from keras import backend as K
+from pkg_resources import parse_version
 
 
 class CallbackList(object):
@@ -43,21 +45,26 @@ class CallbackList(object):
             callback.on_batch_begin(batch, logs)
         self._delta_ts_batch_begin.append(time.time() - t_before_callbacks)
         delta_t_median = np.median(self._delta_ts_batch_begin)
-        if self._delta_t_batch > 0. and delta_t_median > 0.95 * self._delta_t_batch and delta_t_median > 0.1:
+        if self._delta_t_batch > 0. and delta_t_median > 0.95 * \
+           self._delta_t_batch and delta_t_median > 0.1:
             warnings.warn('Method on_batch_begin() is slow compared '
-                          'to the batch update (%f). Check your callbacks.' % delta_t_median)
+                          'to the batch update (%f). Check your callbacks.'
+                          % delta_t_median)
         self._t_enter_batch = time.time()
 
     def on_batch_end(self, batch, logs={}):
+        if not hasattr(self, '_t_enter_batch'):
+            self._t_enter_batch = time.time()
         self._delta_t_batch = time.time() - self._t_enter_batch
         t_before_callbacks = time.time()
         for callback in self.callbacks:
             callback.on_batch_end(batch, logs)
         self._delta_ts_batch_end.append(time.time() - t_before_callbacks)
         delta_t_median = np.median(self._delta_ts_batch_end)
-        if self._delta_t_batch > 0. and delta_t_median > 0.95 * self._delta_t_batch and delta_t_median > 0.1:
+        if self._delta_t_batch > 0. and (delta_t_median > 0.95 * self._delta_t_batch and delta_t_median > 0.1):
             warnings.warn('Method on_batch_end() is slow compared '
-                          'to the batch update (%f). Check your callbacks.' % delta_t_median)
+                          'to the batch update (%f). Check your callbacks.'
+                          % delta_t_median)
 
     def on_train_begin(self, logs={}):
         for callback in self.callbacks:
@@ -69,7 +76,31 @@ class CallbackList(object):
 
 
 class Callback(object):
+    '''Abstract base class used to build new callbacks.
 
+    # Properties
+        params: dict. Training parameters
+            (eg. verbosity, batch size, number of epochs...).
+        model: instance of `keras.models.Model`.
+            Reference of the model being trained.
+
+    The `logs` dictionary that callback methods
+    take as argument will contain keys for quantities relevant to
+    the current batch or epoch.
+
+    Currently, the `.fit()` method of the `Sequential` model class
+    will include the following quantities in the `logs` that
+    it passes to its callbacks:
+
+        on_epoch_end: logs include `acc` and `loss`, and
+            optionally include `val_loss`
+            (if validation is enabled in `fit`), and `val_acc`
+            (if validation and accuracy monitoring are enabled).
+        on_batch_begin: logs include `size`,
+            the number of samples in the current batch.
+        on_batch_end: logs include `loss`, and optionally `acc`
+            (if accuracy monitoring is enabled).
+    '''
     def __init__(self):
         pass
 
@@ -99,6 +130,36 @@ class Callback(object):
 
 
 class BaseLogger(Callback):
+    '''Callback that accumulates epoch averages of
+    the metrics being monitored.
+
+    This callback is automatically applied to
+    every Keras model.
+    '''
+    def on_epoch_begin(self, epoch, logs={}):
+        self.seen = 0
+        self.totals = {}
+
+    def on_batch_end(self, batch, logs={}):
+        batch_size = logs.get('size', 0)
+        self.seen += batch_size
+
+        for k, v in logs.items():
+            if k in self.totals:
+                self.totals[k] += v * batch_size
+            else:
+                self.totals[k] = v * batch_size
+
+    def on_epoch_end(self, epoch, logs={}):
+        for k in self.params['metrics']:
+            if k in self.totals:
+                # make value available to next callbacks
+                logs[k] = self.totals[k] / self.seen
+
+
+class ProgbarLogger(Callback):
+    '''Callback that prints metrics to stdout.
+    '''
     def on_train_begin(self, logs={}):
         self.verbose = self.params['verbose']
         self.nb_epoch = self.params['nb_epoch']
@@ -109,7 +170,6 @@ class BaseLogger(Callback):
             self.progbar = Progbar(target=self.params['nb_sample'],
                                    verbose=self.verbose)
         self.seen = 0
-        self.totals = {}
 
     def on_batch_begin(self, batch, logs={}):
         if self.seen < self.params['nb_sample']:
@@ -119,82 +179,96 @@ class BaseLogger(Callback):
         batch_size = logs.get('size', 0)
         self.seen += batch_size
 
-        for k, v in logs.items():
-            if k in self.totals:
-                self.totals[k] += v * batch_size
-            else:
-                self.totals[k] = v * batch_size
         for k in self.params['metrics']:
             if k in logs:
                 self.log_values.append((k, logs[k]))
 
-        # skip progbar update for the last batch; will be handled by on_epoch_end
+        # skip progbar update for the last batch;
+        # will be handled by on_epoch_end
         if self.verbose and self.seen < self.params['nb_sample']:
             self.progbar.update(self.seen, self.log_values)
 
     def on_epoch_end(self, epoch, logs={}):
         for k in self.params['metrics']:
-            if k in self.totals:
-                self.log_values.append((k, self.totals[k] / self.seen))
             if k in logs:
                 self.log_values.append((k, logs[k]))
         if self.verbose:
-            self.progbar.update(self.seen, self.log_values)
+            self.progbar.update(self.seen, self.log_values, force=True)
 
 
 class History(Callback):
+    '''Callback that records events
+    into a `History` object.
 
+    This callback is automatically applied to
+    every Keras model. The `History` object
+    gets returned by the `fit` method of models.
+    '''
     def on_train_begin(self, logs={}):
         self.epoch = []
         self.history = {}
 
-    def on_epoch_begin(self, epoch, logs={}):
-        self.seen = 0
-        self.totals = {}
-
-    def on_batch_end(self, batch, logs={}):
-        batch_size = logs.get('size', 0)
-        self.seen += batch_size
-        for k, v in logs.items():
-            if k in self.totals:
-                self.totals[k] += v * batch_size
-            else:
-                self.totals[k] = v * batch_size
-
     def on_epoch_end(self, epoch, logs={}):
         self.epoch.append(epoch)
-        for k, v in self.totals.items():
-            if k not in self.history:
-                self.history[k] = []
-            self.history[k].append(v / self.seen)
-
         for k, v in logs.items():
-            if k not in self.history:
-                self.history[k] = []
-            self.history[k].append(v)
+            self.history.setdefault(k, []).append(v)
 
 
 class ModelCheckpoint(Callback):
-    def __init__(self, filepath, monitor='val_loss', verbose=0, save_best_only=False, mode='auto'):
+    '''Save the model after every epoch.
 
-        super(Callback, self).__init__()
+    `filepath` can contain named formatting options,
+    which will be filled the value of `epoch` and
+    keys in `logs` (passed in `on_epoch_end`).
+
+    For example: if `filepath` is `weights.{epoch:02d}-{val_loss:.2f}.hdf5`,
+    then multiple files will be save with the epoch number and
+    the validation loss.
+
+    # Arguments
+        filepath: string, path to save the model file.
+        monitor: quantity to monitor.
+        verbose: verbosity mode, 0 or 1.
+        save_best_only: if `save_best_only=True`,
+            the latest best model according to
+            the quantity monitored will not be overwritten.
+        mode: one of {auto, min, max}.
+            If `save_best_only=True`, the decision
+            to overwrite the current save file is made
+            based on either the maximization or the
+            minimization of the monitored quantity. For `val_acc`,
+            this should be `max`, for `val_loss` this should
+            be `min`, etc. In `auto` mode, the direction is
+            automatically inferred from the name of the monitored quantity.
+        save_weights_only: if True, then only the model's weights will be
+            saved (`model.save_weights(filepath)`), else the full model
+            is saved (`model.save(filepath)`).
+
+    '''
+    def __init__(self, filepath, monitor='val_loss', verbose=0,
+                 save_best_only=False, save_weights_only=False,
+                 mode='auto'):
+        super(ModelCheckpoint, self).__init__()
         self.monitor = monitor
         self.verbose = verbose
         self.filepath = filepath
         self.save_best_only = save_best_only
-        
+        self.save_weights_only = save_weights_only
+
         if mode not in ['auto', 'min', 'max']:
-            warnings.warn("ModelCheckpoint mode %s is unknown, fallback to auto mode" % (self.mode), RuntimeWarning)
+            warnings.warn('ModelCheckpoint mode %s is unknown, '
+                          'fallback to auto mode.' % (mode),
+                          RuntimeWarning)
             mode = 'auto'
-            
-        if mode == "min":
+
+        if mode == 'min':
             self.monitor_op = np.less
             self.best = np.Inf
-        elif mode == "max":
+        elif mode == 'max':
             self.monitor_op = np.greater
             self.best = -np.Inf
         else:
-            if "acc" in self.monitor:
+            if 'acc' in self.monitor:
                 self.monitor_op = np.greater
                 self.best = -np.Inf
             else:
@@ -206,90 +280,255 @@ class ModelCheckpoint(Callback):
         if self.save_best_only:
             current = logs.get(self.monitor)
             if current is None:
-                warnings.warn("Can save best model only with %s available, skipping." % (self.monitor), RuntimeWarning)
+                warnings.warn('Can save best model only with %s available, '
+                              'skipping.' % (self.monitor), RuntimeWarning)
             else:
                 if self.monitor_op(current, self.best):
                     if self.verbose > 0:
-                        print("Epoch %05d: %s improved from %0.5f to %0.5f, saving model to %s"
-                              % (epoch, self.monitor, self.best, current, filepath))
+                        print('Epoch %05d: %s improved from %0.5f to %0.5f,'
+                              ' saving model to %s'
+                              % (epoch, self.monitor, self.best,
+                                 current, filepath))
                     self.best = current
-                    self.model.save_weights(filepath, overwrite=True)
+                    if self.save_weights_only:
+                        self.model.save_weights(filepath, overwrite=True)
+                    else:
+                        self.model.save(filepath, overwrite=True)
                 else:
                     if self.verbose > 0:
-                        print("Epoch %05d: %s did not improve" % (epoch, self.monitor))
+                        print('Epoch %05d: %s did not improve' %
+                              (epoch, self.monitor))
         else:
             if self.verbose > 0:
-                print("Epoch %05d: saving model to %s" % (epoch, filepath))
-            self.model.save_weights(filepath, overwrite=True)
+                print('Epoch %05d: saving model to %s' % (epoch, filepath))
+            if self.save_weights_only:
+                self.model.save_weights(filepath, overwrite=True)
+            else:
+                self.model.save(filepath, overwrite=True)
 
 
 class EarlyStopping(Callback):
-    def __init__(self, monitor='val_loss', patience=0, verbose=0):
-        super(Callback, self).__init__()
+    '''Stop training when a monitored quantity has stopped improving.
+
+    # Arguments
+        monitor: quantity to be monitored.
+        patience: number of epochs with no improvement
+            after which training will be stopped.
+        verbose: verbosity mode.
+        mode: one of {auto, min, max}. In `min` mode,
+            training will stop when the quantity
+            monitored has stopped decreasing; in `max`
+            mode it will stop when the quantity
+            monitored has stopped increasing; in `auto`
+            mode, the direction is automatically inferred
+            from the name of the monitored quantity.
+    '''
+    def __init__(self, monitor='val_loss', patience=0, verbose=0, mode='auto'):
+        super(EarlyStopping, self).__init__()
 
         self.monitor = monitor
         self.patience = patience
         self.verbose = verbose
-        self.best = np.Inf
         self.wait = 0
 
+        if mode not in ['auto', 'min', 'max']:
+            warnings.warn('EarlyStopping mode %s is unknown, '
+                          'fallback to auto mode.' % (self.mode),
+                          RuntimeWarning)
+            mode = 'auto'
+
+        if mode == 'min':
+            self.monitor_op = np.less
+        elif mode == 'max':
+            self.monitor_op = np.greater
+        else:
+            if 'acc' in self.monitor:
+                self.monitor_op = np.greater
+            else:
+                self.monitor_op = np.less
+
+    def on_train_begin(self, logs={}):
+        self.wait = 0       # Allow instances to be re-used
+        self.best = np.Inf if self.monitor_op == np.less else -np.Inf
+
     def on_epoch_end(self, epoch, logs={}):
         current = logs.get(self.monitor)
         if current is None:
-            warnings.warn("Early stopping requires %s available!" % (self.monitor), RuntimeWarning)
+            warnings.warn('Early stopping requires %s available!' %
+                          (self.monitor), RuntimeWarning)
 
-        if current < self.best:
+        if self.monitor_op(current, self.best):
             self.best = current
             self.wait = 0
         else:
             if self.wait >= self.patience:
                 if self.verbose > 0:
-                    print("Epoch %05d: early stopping" % (epoch))
+                    print('Epoch %05d: early stopping' % (epoch))
                 self.model.stop_training = True
             self.wait += 1
 
 
 class RemoteMonitor(Callback):
-    def __init__(self, root='http://localhost:9000'):
+    '''Callback used to stream events to a server.
+
+    Requires the `requests` library.
+
+    # Arguments
+        root: root url to which the events will be sent (at the end
+            of every epoch). Events are sent to
+            `root + '/publish/epoch/end/'` by default. Calls are
+            HTTP POST, with a `data` argument which is a
+            JSON-encoded dictionary of event data.
+    '''
+
+    def __init__(self,
+                 root='http://localhost:9000',
+                 path='/publish/epoch/end/',
+                 field='data'):
+        super(RemoteMonitor, self).__init__()
         self.root = root
-
-    def on_epoch_begin(self, epoch, logs={}):
-        self.seen = 0
-        self.totals = {}
-
-    def on_batch_end(self, batch, logs={}):
-        batch_size = logs.get('size', 0)
-        self.seen += batch_size
-        for k, v in logs.items():
-            if k in self.totals:
-                self.totals[k] += v * batch_size
-            else:
-                self.totals[k] = v * batch_size
+        self.path = path
+        self.field = field
 
     def on_epoch_end(self, epoch, logs={}):
         import requests
         send = {}
         send['epoch'] = epoch
-
-        for k, v in self.totals.items():
-            send[k] = v / self.seen
         for k, v in logs.items():
             send[k] = v
-
         try:
-            r = requests.post(self.root + '/publish/epoch/end/', {'data': json.dumps(send)})
+            requests.post(self.root + self.path,
+                          {self.field: json.dumps(send)})
         except:
-            print('Warning: could not reach RemoteMonitor root server at ' + str(self.root))
+            print('Warning: could not reach RemoteMonitor '
+                  'root server at ' + str(self.root))
 
 
 class LearningRateScheduler(Callback):
-    '''LearningRateScheduler
-    schedule is a function that gets an epoch number as input and returns a new
-    learning rate as output.
+    '''Learning rate scheduler.
+
+    # Arguments
+        schedule: a function that takes an epoch index as input
+            (integer, indexed from 0) and returns a new
+            learning rate as output (float).
     '''
     def __init__(self, schedule):
         super(LearningRateScheduler, self).__init__()
         self.schedule = schedule
 
     def on_epoch_begin(self, epoch, logs={}):
-        self.model.optimizer.lr.set_value(self.schedule(epoch))
+        assert hasattr(self.model.optimizer, 'lr'), \
+            'Optimizer must have a "lr" attribute.'
+        lr = self.schedule(epoch)
+        assert type(lr) == float, 'The output of the "schedule" function should be float.'
+        K.set_value(self.model.optimizer.lr, lr)
+
+
+class TensorBoard(Callback):
+    ''' Tensorboard basic visualizations.
+
+    This callback writes a log for TensorBoard, which allows
+    you to visualize dynamic graphs of your training and test
+    metrics, as well as activation histograms for the different
+    layers in your model.
+
+    TensorBoard is a visualization tool provided with TensorFlow.
+
+    If you have installed TensorFlow with pip, you should be able
+    to launch TensorBoard from the command line:
+    ```
+    tensorboard --logdir=/full_path_to_your_logs
+    ```
+    You can find more information about TensorBoard
+    [here](https://www.tensorflow.org/versions/master/how_tos/summaries_and_tensorboard/index.html).
+
+    # Arguments
+        log_dir: the path of the directory where to save the log
+            files to be parsed by Tensorboard
+        histogram_freq: frequency (in epochs) at which to compute activation
+            histograms for the layers of the model. If set to 0,
+            histograms won't be computed.
+        write_graph: whether to visualize the graph in Tensorboard.
+            The log file can become quite large when
+            write_graph is set to True.
+    '''
+
+    def __init__(self, log_dir='./logs', histogram_freq=0, write_graph=True, write_images=False):
+        super(TensorBoard, self).__init__()
+        if K._BACKEND != 'tensorflow':
+            raise Exception('TensorBoard callback only works '
+                            'with the TensorFlow backend.')
+        self.log_dir = log_dir
+        self.histogram_freq = histogram_freq
+        self.merged = None
+        self.write_graph = write_graph
+        self.write_images = write_images
+
+    def _set_model(self, model):
+        import tensorflow as tf
+        import keras.backend.tensorflow_backend as KTF
+
+        self.model = model
+        self.sess = KTF.get_session()
+        if self.histogram_freq and self.merged is None:
+            for layer in self.model.layers:
+
+                for weight in layer.weights:
+                    tf.histogram_summary(weight.name, weight)
+
+                    if self.write_images:
+                        w_img = tf.squeeze(weight)
+
+                        shape = w_img.get_shape()
+                        if len(shape) > 1 and shape[0] > shape[1]:
+                            w_img = tf.transpose(w_img)
+
+                        if len(shape) == 1:
+                            w_img = tf.expand_dims(w_img, 0)
+
+                        w_img = tf.expand_dims(tf.expand_dims(w_img, 0), -1)
+
+                        tf.image_summary(weight.name, w_img)
+
+                if hasattr(layer, 'output'):
+                    tf.histogram_summary('{}_out'.format(layer.name),
+                                         layer.output)
+        self.merged = tf.merge_all_summaries()
+        if self.write_graph:
+            if parse_version(tf.__version__) >= parse_version('0.8.0'):
+                self.writer = tf.train.SummaryWriter(self.log_dir,
+                                                     self.sess.graph)
+            else:
+                self.writer = tf.train.SummaryWriter(self.log_dir,
+                                                     self.sess.graph_def)
+        else:
+            self.writer = tf.train.SummaryWriter(self.log_dir)
+
+    def on_epoch_end(self, epoch, logs={}):
+        import tensorflow as tf
+
+        if self.model.validation_data and self.histogram_freq:
+            if epoch % self.histogram_freq == 0:
+                # TODO: implement batched calls to sess.run
+                # (current call will likely go OOM on GPU)
+                if self.model.uses_learning_phase:
+                    cut_v_data = len(self.model.inputs)
+                    val_data = self.model.validation_data[:cut_v_data] + [0]
+                    tensors = self.model.inputs + [K.learning_phase()]
+                else:
+                    val_data = self.model.validation_data
+                    tensors = self.model.inputs
+                feed_dict = dict(zip(tensors, val_data))
+                result = self.sess.run([self.merged], feed_dict=feed_dict)
+                summary_str = result[0]
+                self.writer.add_summary(summary_str, epoch)
+
+        for name, value in logs.items():
+            if name in ['batch', 'size']:
+                continue
+            summary = tf.Summary()
+            summary_value = summary.value.add()
+            summary_value.simple_value = value
+            summary_value.tag = name
+            self.writer.add_summary(summary, epoch)
+        self.writer.flush()

keras/constraints.py

@@ -7,39 +7,87 @@ class Constraint(object):
         return p
 
     def get_config(self):
-        return {"name": self.__class__.__name__}
+        return {'name': self.__class__.__name__}
 
 
 class MaxNorm(Constraint):
-    def __init__(self, m=2):
+    '''Constrain the weights incident to each hidden unit to have a norm less than or equal to a desired value.
+
+    # Arguments
+        m: the maximum norm for the incoming weights.
+        axis: integer, axis along which to calculate weight norms. For instance,
+            in a `Dense` layer the weight matrix has shape (input_dim, output_dim),
+            set `axis` to `0` to constrain each weight vector of length (input_dim).
+            In a `MaxoutDense` layer the weight tensor has shape (nb_feature, input_dim, output_dim),
+            set `axis` to `1` to constrain each weight vector of length (input_dim),
+            i.e. constrain the filters incident to the `max` operation.
+            In a `Convolution2D` layer with the Theano backend, the weight tensor
+            has shape (nb_filter, stack_size, nb_row, nb_col), set `axis` to `[1,2,3]`
+            to constrain the weights of each filter tensor of size (stack_size, nb_row, nb_col).
+            In a `Convolution2D` layer with the TensorFlow backend, the weight tensor
+            has shape (nb_row, nb_col, stack_size, nb_filter), set `axis` to `[0,1,2]`
+            to constrain the weights of each filter tensor of size (nb_row, nb_col, stack_size).
+
+    # References
+        - [Dropout: A Simple Way to Prevent Neural Networks from Overfitting Srivastava, Hinton, et al. 2014](http://www.cs.toronto.edu/~rsalakhu/papers/srivastava14a.pdf)
+    '''
+    def __init__(self, m=2, axis=0):
         self.m = m
+        self.axis = axis
 
     def __call__(self, p):
-        norms = K.sqrt(K.sum(K.square(p), axis=0))
+        norms = K.sqrt(K.sum(K.square(p), axis=self.axis, keepdims=True))
         desired = K.clip(norms, 0, self.m)
-        p = p * (desired / (1e-7 + norms))
+        p = p * (desired / (K.epsilon() + norms))
         return p
 
     def get_config(self):
-        return {"name": self.__class__.__name__,
-                "m": self.m}
+        return {'name': self.__class__.__name__,
+                'm': self.m,
+                'axis': self.axis}
 
 
 class NonNeg(Constraint):
+    '''Constrain the weights to be non-negative.
+    '''
     def __call__(self, p):
         p *= K.cast(p >= 0., K.floatx())
         return p
 
 
 class UnitNorm(Constraint):
-    def __call__(self, p):
-        return p / K.sqrt(K.sum(K.square(p), axis=-1, keepdims=True))
+    '''Constrain the weights incident to each hidden unit to have unit norm.
+
+    # Arguments
+        axis: integer, axis along which to calculate weight norms. For instance,
+            in a `Dense` layer the weight matrix has shape (input_dim, output_dim),
+            set `axis` to `0` to constrain each weight vector of length (input_dim).
+            In a `MaxoutDense` layer the weight tensor has shape (nb_feature, input_dim, output_dim),
+            set `axis` to `1` to constrain each weight vector of length (input_dim),
+            i.e. constrain the filters incident to the `max` operation.
+            In a `Convolution2D` layer with the Theano backend, the weight tensor
+            has shape (nb_filter, stack_size, nb_row, nb_col), set `axis` to `[1,2,3]`
+            to constrain the weights of each filter tensor of size (stack_size, nb_row, nb_col).
+            In a `Convolution2D` layer with the TensorFlow backend, the weight tensor
+            has shape (nb_row, nb_col, stack_size, nb_filter), set `axis` to `[0,1,2]`
+            to constrain the weights of each filter tensor of size (nb_row, nb_col, stack_size).
+    '''
+    def __init__(self, axis=0):
+        self.axis = axis
+
+    def __call__(self, p):
+        return p / (K.epsilon() + K.sqrt(K.sum(K.square(p), axis=self.axis, keepdims=True)))
+
+    def get_config(self):
+        return {'name': self.__class__.__name__,
+                'axis': self.axis}
+
 
-identity = Constraint
 maxnorm = MaxNorm
 nonneg = NonNeg
 unitnorm = UnitNorm
 
 from .utils.generic_utils import get_from_module
 def get(identifier, kwargs=None):
-    return get_from_module(identifier, globals(), 'constraint', instantiate=True, kwargs=kwargs)
+    return get_from_module(identifier, globals(), 'constraint',
+                           instantiate=True, kwargs=kwargs)

keras/datasets/cifar.py

@@ -2,7 +2,7 @@
 from __future__ import absolute_import
 import sys
 from six.moves import cPickle
-from six.moves import range
+
 
 def load_batch(fpath, label_key='labels'):
     f = open(fpath, 'rb')

keras/datasets/cifar10.py

@@ -1,6 +1,7 @@
 from __future__ import absolute_import
 from .cifar import load_batch
-from .data_utils import get_file
+from ..utils.data_utils import get_file
+from .. import backend as K
 import numpy as np
 import os
 
@@ -10,7 +11,6 @@ def load_data():
     origin = "http://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz"
     path = get_file(dirname, origin=origin, untar=True)
 
-    nb_test_samples = 10000
     nb_train_samples = 50000
 
     X_train = np.zeros((nb_train_samples, 3, 32, 32), dtype="uint8")
@@ -19,8 +19,8 @@ def load_data():
     for i in range(1, 6):
         fpath = os.path.join(path, 'data_batch_' + str(i))
         data, labels = load_batch(fpath)
-        X_train[(i-1)*10000:i*10000, :, :, :] = data
-        y_train[(i-1)*10000:i*10000] = labels
+        X_train[(i - 1) * 10000: i * 10000, :, :, :] = data
+        y_train[(i - 1) * 10000: i * 10000] = labels
 
     fpath = os.path.join(path, 'test_batch')
     X_test, y_test = load_batch(fpath)
@@ -28,4 +28,8 @@ def load_data():
     y_train = np.reshape(y_train, (len(y_train), 1))
     y_test = np.reshape(y_test, (len(y_test), 1))
 
+    if K.image_dim_ordering() == 'tf':
+        X_train = X_train.transpose(0, 2, 3, 1)
+        X_test = X_test.transpose(0, 2, 3, 1)
+
     return (X_train, y_train), (X_test, y_test)