tensorpack/tensorpack

Health v2 · maintenance only

Adoption not part of the score

6286 stars · 1778 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Tensorpack is a high-level neural network training interface built on graph-mode TensorFlow, focused on training speed and flexibility for research. It provides efficient data loading via its dataflow library and off-the-shelf multi-GPU/distributed training, with reproducible implementations of well-known papers.

Use cases

• train resnet on imagenet fast
• train faster r-cnn on coco
• speed up tensorflow training over keras
• multi-gpu distributed training in tensorflow
• reproduce research papers in computer vision
• efficient python data loading pipeline for deep learning
• reinforcement learning training framework

When to choose

• you need maximum training throughput with graph-mode TensorFlow
• you want reproducible, research-quality implementations of vision papers
• you need scalable multi-GPU or distributed training out of the box
• your data pipeline is the bottleneck and you need fast Python data loading

When to avoid

• you are starting a new project with PyTorch or modern TF2/Keras eager execution
• you need a simple high-level model wrapper for quick prototyping
• you require active community development and frequent updates

Facets

library · maturity maintenance

machine-learning deep-learning llm-training data-science deep-learning machine-learning computer-vision reinforcement-learning python cross-platform tensorflow training-interface dataflow multi-gpu model-zoo research gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="tensorpack/tensorpack")