TensorLayer is a deep learning and reinforcement learning library based on Google TensorFlow. It provides rich data processing, model training and serving modules to help researchers and engineers build practical machine learning workflows.

• Release ROI layer for Object Detection.
• Release Sub-pixel Convolution 1D for Audio Super-resolution.
• Release Flickr dataset loader, see load_flickr25k and load_flickr1M.
• Release SpatialTransformer2dAffineLayer for Spatial Transformer Networks see example code.
• Release Sub-pixel Convolution 2D for Super-resolution see SRGAN code.
• Join Slack Now.
• You can now use TensorLayer with TF-Slim and Keras together!

As deep learning practitioners, we have been looking for a library that can serve for various development phases. This library shall be easy for beginners by providing rich neural network reference implementations. Later, it can be extended to address real-world problems by controlling training backends to exhibit low-level cognitive behaviours. In the end, it shall be able to serve in challenging production environments.

TensorLayer is designed for beginning, intermediate and professional deep learning users with following goals:

• Simplicity : TensorLayer lifts the low-level dataflow abstraction of TensorFlow to high-level deep learning modules. A user often find it easy to bootstrap with TensorLayer, and then dive into low-level implementation only if need.
• Transparency : TensorLayer provides access to the native APIs of TensorFlow. This helps users achieve flexible controls within the training engine.
• Composability : If possible, deep learning modules are composed, not built. TensorLayer can glue existing pieces together (e.g., connected with TF-Slim and Keras).
• Performance : TensorLayer provides zero-cost compatibility for TensorFlow. It can run on distributed yet heterogeneous platforms.

A frequent question regarding TensorLayer is that why don't we use libraries like Keras and Tflearn. These libraries are comfortable to start with. They provide imperative abstractions to lower adoption barrier; but in turn mask the underlying engine from users. Though good for bootstrap, it becomes hard to tune and modify from the bottom, which is quite necessary in tackling many real-world problems.

Without compromise in simplicity, TensorLayer advocates a more flexible and composable paradigm: neural network libraries shall be used interchangeably with the native engine. This allows users to enjoy the ease of pre-built modules without losing visibility to the deep. This non-intrusive nature also makes it viable to consolidate with other TF's wrappers such as TF-Slim and Keras. However, flexibility does not sacrifice performance. TensorLayer allows seamless distributed and heterogeneous deployment.

TensorLayer is in an active development stage and has received numerous contributions from an open community. It has been widely used by researchers from Imperial College London, Carnegie Mellon University, Stanford University, Tsinghua University, UCLA, Linköping University and etc., as well as engineers from Google, Microsoft, Alibaba, Tencent, ReFULE4, Bloomberg and many others.

TensorLayer has install prerequisites including TensorFlow, numpy and matplotlib. For GPU support, CUDA and cuDNN are required. Please check here for detailed instructions.

If you already had the pre-requisites ready (numpy, scipy, scikit-image, matplotlib and nltk(optional)), the simplest way to install TensorLayer in your python program is:

[for master version] pip install git+https://github.com/zsdonghao/tensorlayer.git (Highly Recommended)
[for stable version] pip install tensorlayer

The documentation [Online] [PDF] [Epub] [HTML] describes the usages of TensorLayer APIs. It is also a self-contained document that walks through different types of deep neural networks, reinforcement learning and their applications in Natural Language Processing (NLP) problems.

We have included the corresponding modularized implementations of Google TensorFlow Deep Learning tutorial, so you can read the TensorFlow tutorial [en] [cn] along with our document.

Chinese documentation is also available.

The first program trains a multi-layer perception network to solve the MNIST problem. We use the well-known scikit-style functions such as fit() and test(). The program is self-explained.

import tensorflow as tf
import tensorlayer as tl

sess = tf.InteractiveSession()

# Prepare data
X_train, y_train, X_val, y_val, X_test, y_test = tl.files.load_mnist_dataset(shape=(-1,784))

# Define placeholder
x = tf.placeholder(tf.float32, shape=[None, 784], name='x')
y_ = tf.placeholder(tf.int64, shape=[None, ], name='y_')

# Define the neural network structure
network = tl.layers.InputLayer(x, name='input')
network = tl.layers.DropoutLayer(network, keep=0.8, name='drop1')
network = tl.layers.DenseLayer(network, 800, tf.nn.relu, name='relu1')
network = tl.layers.DropoutLayer(network, keep=0.5, name='drop2')
network = tl.layers.DenseLayer(network, 800, tf.nn.relu, name='relu2')
network = tl.layers.DropoutLayer(network, keep=0.5, name='drop3')

# The softmax is implemented internally in tl.cost.cross_entropy(y, y_) to
# speed up computation, so we use identity here.
# see tf.nn.sparse_softmax_cross_entropy_with_logits()
network = tl.layers.DenseLayer(network, n_units=10, act=tf.identity, name='output')
                                
# Define cost function and metric.
y = network.outputs
cost = tl.cost.cross_entropy(y, y_, 'cost')
correct_prediction = tf.equal(tf.argmax(y, 1), y_)
acc = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
y_op = tf.argmax(tf.nn.softmax(y), 1)

# Define the optimizer
train_params = network.all_params
train_op = tf.train.AdamOptimizer(learning_rate=0.0001).minimize(cost, var_list=train_params)

# Initialize all variables in the session
tl.layers.initialize_global_variables(sess)

# Print network information
network.print_params()
network.print_layers()

# Train the network, we recommend to use tl.iterate.minibatches()
tl.utils.fit(sess, network, train_op, cost, X_train, y_train, x, y_,
            acc=acc, batch_size=500, n_epoch=500, print_freq=5,
            X_val=X_val, y_val=y_val, eval_train=False)

# Evaluation
tl.utils.test(sess, network, acc, X_test, y_test, x, y_, batch_size=None, cost=cost)

# Save the network to .npz file
tl.files.save_npz(network.all_params , name='model.npz')

sess.close()

We provide many helper functions (like fit() , test()) that is similar to Keras to facilitate your development; however, if you want to obtain a fine-grain control over the model or its training process, you can use TensorFlow’s methods like sess.run() in your program directly (tutorial_mnist.py provides more details about this). Many more DL and RL examples can be found here.

Tricks to use TL is also a good introduction to use TensorLayer.

Examples can be found in this repository and TensorLayer Topic.

• Multi-layer perceptron (MNIST). A multi-layer perceptron implementation for MNIST classification task, see tutorial_mnist_simple.py.
• Multi-layer perceptron (MNIST) classification using Iterator, see method1 and method2.

• Denoising Autoencoder (MNIST). A multi-layer perceptron implementation for MNIST classification task, see tutorial_mnist.py.
• Stacked Denoising Autoencoder and Fine-Tuning (MNIST). A multi-layer perceptron implementation for MNIST classification task, see tutorial_mnist.py.
• Convolutional Network (MNIST). A Convolutional neural network implementation for classifying MNIST dataset, see tutorial_mnist.py.
• Convolutional Network (CIFAR-10). A Convolutional neural network implementation for classifying CIFAR-10 dataset, see tutorial_cifar10.py and tutorial_cifar10_tfrecord.py.
• VGG 16 (ImageNet). A Convolutional neural network implementation for classifying ImageNet dataset, see tutorial_vgg16.py.
• VGG 19 (ImageNet). A Convolutional neural network implementation for classifying ImageNet dataset, see tutorial_vgg19.py.
• InceptionV3 (ImageNet). A Convolutional neural network implementation for classifying ImageNet dataset, see tutorial_inceptionV3_tfslim.py.
• Wide ResNet (CIFAR) by ritchieng.
• More CNN implementations of TF-Slim can be connected to TensorLayer via SlimNetsLayer.
• Spatial Transformer Networks by zsdonghao.
• U-Net for brain tumor segmentation by zsdonghao.
• Variational Autoencoder (VAE) for CelebA by yzwxx.

• Recurrent Neural Network (LSTM). Apply multiple LSTM to PTB dataset for language modeling, see tutorial_ptb_lstm.py and tutorial_ptb_lstm_state_is_tuple.py.
• Word Embedding - Word2vec. Train a word embedding matrix, see tutorial_word2vec_basic.py.
• Restore Embedding matrix. Restore a pre-train embedding matrix, see tutorial_generate_text.py.
• Text Generation. Generates new text scripts, using LSTM network, see tutorial_generate_text.py.
• Chinese Text Anti-Spam by pakrchen.

• DCGAN - Generating images by Deep Convolutional Generative Adversarial Networks by zsdonghao.
• Generative Adversarial Text to Image Synthesis by zsdonghao.
• Unsupervised Image to Image Translation with Generative Adversarial Networks by zsdonghao.
• Super Resolution GAN by zsdonghao.

• Policy Gradient / Network - Atari Ping Pong, see tutorial_atari_pong.py.
• Deep Q-Network - Frozen lake, see tutorial_frozenlake_dqn.py.
• Q-Table learning algorithm - Frozen lake, see tutorial_frozenlake_q_table.py.
• Asynchronous Policy Gradient using TensorDB - Atari Ping Pong by nebulaV.
• AC for discrete action space - Cartpole, see tutorial_cartpole_ac.py.
• A3C for continuous action space - Bipedal Walker, see tutorial_bipedalwalker_a3c*.py.
• DAGGER - Gym Torcs by zsdonghao.
• TRPO for continuous and discrete action space by jjkke88.

• Image Captioning - Reimplementation of Google's im2txt by zsdonghao.
• A simple web service - TensorFlask by JoelKronander.

• Merge TF-Slim into TensorLayer. tutorial_inceptionV3_tfslim.py.
• Merge Keras into TensorLayer. tutorial_keras.py.
• Data augmentation with TFRecord. Effective way to load and pre-process data, see tutorial_tfrecord*.py and tutorial_cifar10_tfrecord.py.
• Data augmentation with TensorLayer, see tutorial_image_preprocess.py.
• TensorDB by fangde see here.

• TensorLayer provides two set of Convolutional layer APIs, see (Professional) and (Simplified) on readthedocs website.
• If you get into trouble, you can start a discussion on Slack, Gitter, Help Wanted Issues, QQ group and Wechat group.

TensorLayer is released under the Apache 2.0 license.

TensorLayer is maintained by numerous Github contributors here.

• 🇬🇧 If you are in London, we can discuss in person. Drop us an email to organize a meetup: [email protected].
• 🇨🇳 我们有官方的 中文文档。另外, 我们建立了多种交流渠道，如QQ群和微信群.

If you find it is useful, please cite our paper in your project and paper.

@article{TensorLayer2017,
author = {Dong, Hao and Supratak, Akara and Mai, Luo and Liu, Fangde and Oehmichen, Axel and Yu, Simiao and Guo, Yike},
journal = {ACM Multimedia},
title = {{TensorLayer: A Versatile Library for Efficient Deep Learning Development}},
url = {http://tensorlayer.org},
year = {2017}
}