TensorLayer

TensorLayer is an open-source deep learning and reinforcement learning library for scientists and engineers, built to make TensorFlow development faster and more modular. It has around 7,400 stars on GitHub and has been used in research and teaching since the early days of the TensorFlow ecosystem.

Overview

TensorLayer wraps low-level TensorFlow code in a clean, layer-centric API: networks are assembled from reusable layers with sensible defaults, while still allowing full access to the underlying framework when you need it. Alongside the core deep learning functionality — covering convolutional networks, recurrent models, GANs, object detection and ImageNet-scale training — the library ships a substantial reinforcement learning component with implementations of algorithms such as A3C and DQN and environments to train them in. Extensive tutorials and examples accompany the code, which made TensorLayer a popular choice in university courses and research labs. The project was last updated in 2023; its successor effort continues under the TensorLayerX name. The library itself is free and open source.

Key Features

• High-level, modular layer API that simplifies TensorFlow model building
• Covers CNNs, RNNs, GANs, object detection and large-scale image training
• Built-in reinforcement learning algorithms including A3C and DQN
• Rich set of tutorials, examples and pretrained-model recipes
• Designed for both rapid prototyping and research-grade flexibility
• Long-running community project from the Imperial College research ecosystem

💳 Pricing

TensorLayer is completely free and open source. You can download it from GitHub, use it in research or commercial projects, and pay nothing beyond your own compute costs.

Who Is It For

TensorLayer is for researchers, students and engineers working with TensorFlow who want a cleaner abstraction for building and training neural networks, and for anyone learning deep learning or reinforcement learning through hands-on examples.