Welcome to TensorLayer

Good News: We won the Best Open Source Software Award @ACM Multimedia (MM) 2017.

TensorLayer is a Deep Learning (DL) and Reinforcement Learning (RL) library extended from Google TensorFlow. It provides popular DL and RL modules that can be easily customized and assembled for tackling real-world machine learning problems. More details can be found here.

Note

If you got problem to read the docs online, you could download the repository on GitHub, then go to /docs/_build/html/index.html to read the docs offline. The _build folder can be generated in docs using make html.

User Guide

The TensorLayer user guide explains how to install TensorFlow, CUDA and cuDNN, how to build and train neural networks using TensorLayer, and how to contribute to the library as a developer.

• Installation
  • Step 1 : Install dependencies
  • Step 2 : TensorFlow
  • Step 3 : TensorLayer
  • Step 4 : GPU support
  • Windows User
  • Issue
• Tutorials
  • Before we start
  • TensorLayer is simple
  • Run the MNIST example
  • Understand the MNIST example
  • Run the Pong Game example
  • Understand Reinforcement learning
  • Run the Word2Vec example
  • Understand Word Embedding
  • Run the PTB example
  • Understand LSTM
  • More Tutorials
  • More info
• Examples
  • Basics
  • Computer Vision
  • Natural Language Processing
  • Adversarial Learning
  • Reinforcement Learning
  • Miscellaneous
• Development
  • What to contribute
  • How to contribute
• More
  • FQA
  • Recruitment

API Reference

If you are looking for information on a specific function, class or method, this part of the documentation is for you.

• API - Layers
  • Understanding the Basic Layer
  • Customizing Layers
  • Layer list
  • Name Scope and Sharing Parameters
  • Basic layer
  • Input layer
  • One-hot layer
  • Word Embedding Input layer
  • Dense layer
  • Noise layer
  • Convolutional layer (Pro)
  • Convolutional layer (Simplified)
  • Super-Resolution layer
  • Spatial Transformer
  • Pooling and Padding layers
  • Normalization layer
  • Object Detection
  • Time distributed layer
  • Fixed Length Recurrent layer
  • Recurrent Convolutional layer
  • Advanced Ops for Dynamic RNN
  • Dynamic RNN layer
  • Sequence to Sequence
  • Shape layer
  • Lambda layer
  • Merge layer
  • Extend layer
  • Stack layer
  • Connect TF-Slim
  • Parametric activation layer
  • Flow control layer
  • Helper functions
• API - Cost
  • Your cost function
  • Softmax cross entropy
  • Sigmoid cross entropy
  • Binary cross entropy
  • Mean squared error (L2)
  • Normalized mean square error
  • Absolute difference error (L1)
  • Dice coefficient
  • Hard Dice coefficient
  • IOU coefficient
  • Cross entropy for sequence
  • Cross entropy with mask for sequence
  • Cosine similarity
  • Regularization functions
• API - Preprocessing
  • Threading
  • Images
  • Object detection
  • Sequence
• API - Iteration
  • Non-time series
  • Time series
• API - Utility
  • Training, testing and predicting
  • Evaluation functions
  • Class balancing functions
  • Random functions
  • Dictionary and list
  • Close TF session and associated processes
  • Open TensorBoard
  • Clear TensorFlow placeholder
  • Set GPU functions
• API - Natural Language Processing
  • Iteration function for training embedding matrix
  • Sampling functions
  • Vector representations of words
  • Read words from file
  • Read analogy question file
  • Build vocabulary, word dictionary and word tokenization
  • Convert words to IDs and IDs to words
  • Functions for translation
  • Metrics
• API - Reinforcement Learning
  • Reward functions
  • Cost functions
  • Sampling functions
• API - Files
  • Load dataset functions
  • Load and save network
  • Load and save variables
  • Folder/File functions
  • Sort
  • Visualizing npz file
• API - Visualization
  • Save and read images
  • Visualize model parameters
  • Visualize images
  • Visualize embeddings
• API - Activations
  • Your activation
  • Identity
  • Ramp
  • Leaky Relu
  • Swish
  • Pixel-wise softmax
  • Parametric activation
• API - Distributed Training
  • Distributed training
  • Session hooks

Command-line Reference

TensorLayer provides a handy command-line tool tl to perform some common tasks.

• CLI - Command Line Interface
  • tl train

Indices and tables

• Index
• Module Index
• Search Page