# tensorpack/tensorpack

**Attribution required: if you use, quote, or summarise this content, you must credit and link back to [awesome-repositories.com](https://awesome-repositories.com/repository/tensorpack-tensorpack).**

6,287 stars · 1,779 forks · Python · Apache-2.0

## Links

- GitHub: https://github.com/tensorpack/tensorpack
- awesome-repositories: https://awesome-repositories.com/repository/tensorpack-tensorpack.md

## Description

Tensorpack is a high-level TensorFlow neural network framework and research library designed for building and training deep learning models. It provides a collection of reproducible neural network architectures for computer vision, generative tasks, reinforcement learning, and natural language processing.

The project distinguishes itself through a specialized deep learning data pipeline that uses pure Python for parallel data loading and streaming. It includes a multi-GPU training orchestrator for distributing workloads via data-parallel strategies and a dedicated interpretability toolkit for visualizing model saliency and activation maps.

The framework covers a broad range of capabilities, including computer vision pipelines for object detection and semantic segmentation, sequence modeling for speech and text, and reinforcement learning agent development. It also provides model optimization tools for weight quantization and low bitwidth training, alongside utilities for reproducing academic research papers and converting legacy Caffe model weights.

## Tags

### Artificial Intelligence & ML

- [Neural Network Training Frameworks](https://awesome-repositories.com/f/artificial-intelligence-ml/neural-network-training-frameworks.md) — Provides a high-level framework for building and training diverse deep learning architectures across vision, speech, and NLP. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples))
- [TensorFlow Model Development](https://awesome-repositories.com/f/artificial-intelligence-ml/tensorflow-model-development.md) — Provides a high-level framework for designing, building, and training deep learning models within the TensorFlow ecosystem.
- [Actor-Critic Architectures](https://awesome-repositories.com/f/artificial-intelligence-ml/actor-critic-architectures.md) — Implements Actor-Critic and Q-Learning architectures for decision-making and gaming environments. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples))
- [Research Architecture Libraries](https://awesome-repositories.com/f/artificial-intelligence-ml/deep-learning-libraries/research-architecture-libraries.md) — Offers a curated collection of reproducible neural network architectures for vision, NLP, and reinforcement learning research.
- [Deep Learning Research](https://awesome-repositories.com/f/artificial-intelligence-ml/deep-learning-research.md) — Provides a library for implementing and reproducing complex neural network architectures from academic research papers.
- [Deep Learning Training Pipelines](https://awesome-repositories.com/f/artificial-intelligence-ml/deep-learning-training-pipelines.md) — Implements a specialized data pipeline using pure Python for parallel loading and streaming of large datasets.
- [Distributed GPU Training](https://awesome-repositories.com/f/artificial-intelligence-ml/distributed-gpu-training.md) — Distributes the computational load of neural network training across multiple graphics cards to reduce training time.
- [Multi-GPU Parallelism Strategies](https://awesome-repositories.com/f/artificial-intelligence-ml/distributed-inference-services/multi-gpu-parallelism-strategies.md) — Distributes training batches across multiple GPUs using data-parallel strategies to increase throughput.
- [Computer Vision Pipelines](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/frameworks/computer-vision/computer-vision-pipelines.md) — Ships automated workflows for processing and manipulating visual data for classification, detection, and generative tasks.
- [Parallel Data Loading](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/model-training-and-tuning/data-and-checkpointing/model-loading/parallel-data-loading.md) — Maximizes throughput by processing large datasets with automated parallelization strategies in pure Python. ([source](https://github.com/tensorpack/tensorpack/blob/master/README.md))
- [Multi-GPU Agent Training](https://awesome-repositories.com/f/artificial-intelligence-ml/model-optimization/inference-deployment/model-deployment-toolkits/distributed-deployment-utilities/multi-gpu-execution-scaling/multi-gpu-agent-training.md) — Trains agents on complex environments using the Asynchronous Advantage Actor-Critic algorithm with multi-GPU support. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples/A3C-Gym))
- [Data Pipeline Multi-GPU Distribution](https://awesome-repositories.com/f/artificial-intelligence-ml/model-optimization/inference-deployment/model-deployment-toolkits/distributed-deployment-utilities/multi-gpu-workload-distribution/data-pipeline-multi-gpu-distribution.md) — Implements a specialized system for distributing data pipeline execution across multiple GPUs to maximize training throughput. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples/ResNet))
- [Vision Model Training](https://awesome-repositories.com/f/artificial-intelligence-ml/model-training-frameworks/vision-model-training.md) — Supports training architectures for image classification, object detection, and super-resolution using generative adversarial networks. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples))
- [Modular Architectures](https://awesome-repositories.com/f/artificial-intelligence-ml/neural-network-architectures/modular-architectures.md) — Implements a library of reproducible academic architectures using interchangeable blocks for flexible model construction.
- [Neural Network Frameworks](https://awesome-repositories.com/f/artificial-intelligence-ml/neural-network-frameworks.md) — Offers a modular framework for defining and scaling complex neural network architectures with a high-level interface.
- [Neural Network Model Implementations](https://awesome-repositories.com/f/artificial-intelligence-ml/neural-network-model-implementations.md) — Provides practical and reproducible implementations of diverse neural network architectures across multiple data domains. ([source](https://github.com/tensorpack/tensorpack#readme))
- [Agent Architectures](https://awesome-repositories.com/f/artificial-intelligence-ml/reinforcement-learning-algorithms/agent-architectures.md) — Implements structural agent architectures for reinforcement learning, including Deep Q-Networks and Actor-Critic models. ([source](https://github.com/tensorpack/tensorpack/tree/master/examples))
- [Reinforcement Learning Training](https://awesome-repositories.com/f/artificial-intelligence-ml/reinforcement-learning-training.md) — Provides frameworks for training RL agents to solve complex games and simulation environments. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples/DeepQNetwork))
- [Research Reproductions](https://awesome-repositories.com/f/artificial-intelligence-ml/research-papers/research-reproductions.md) — Enables the creation of executable implementations designed to reproduce the results and logic of academic research papers. ([source](https://github.com/tensorpack/tensorpack/blob/master/README.md))
- [Object Detection](https://awesome-repositories.com/f/artificial-intelligence-ml/computer-vision-systems/computer-vision/object-detection-tracking/object-detection.md) — Performs object detection and instance segmentation using architectures like Faster R-CNN and Mask R-CNN. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples/FasterRCNN))
- [ResNet Variants](https://awesome-repositories.com/f/artificial-intelligence-ml/convolutional-neural-network-architectures/resnet-variants.md) — Applies ResNet architectures and training recipes for ImageNet and CIFAR10, including Squeeze-and-Excitation variants. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples/ResNet))
- [Deep Reinforcement Learning Implementations](https://awesome-repositories.com/f/artificial-intelligence-ml/deep-q-learning-implementations/deep-reinforcement-learning-implementations.md) — Provides functional implementations of reinforcement learning agents using Deep Q-Networks and actor-critic architectures.
- [Domain-to-Domain Translation](https://awesome-repositories.com/f/artificial-intelligence-ml/domain-to-domain-translation.md) — Converts images from one domain to another, such as transforming urban scenes into semantic maps. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples/GAN))
- [Fully Convolutional Architectures](https://awesome-repositories.com/f/artificial-intelligence-ml/fully-convolutional-architectures.md) — Produces dense pixel-level predictions for edge detection and semantic segmentation using fully-convolutional architectures. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples/HED))
- [Generative Adversarial Networks](https://awesome-repositories.com/f/artificial-intelligence-ml/generative-adversarial-networks.md) — Supports training Generative Adversarial Networks with conditional labels and mutual information maximization for synthetic data generation. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples/GAN))
- [High-Level Model APIs](https://awesome-repositories.com/f/artificial-intelligence-ml/high-level-model-apis.md) — Provides declarative interfaces and high-level abstractions for constructing neural networks with reduced complexity. ([source](https://github.com/tensorpack/tensorpack/blob/master/setup.py))
- [Pre-trained Weight Loading](https://awesome-repositories.com/f/artificial-intelligence-ml/large-scale-model-training/vision-transformer-pre-training/pre-trained-model-checkpoints/pre-trained-weight-loading.md) — Allows importing weights from existing architectures to jumpstart training or perform immediate inference. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples))
- [Recurrent State Managers](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/frameworks/model-construction/sequential-containers/recurrent-state-managers.md) — Manages recurrent neural network hidden states directly within the computation graph to preserve temporal context. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples/PennTreebank))
- [GPU Training Accelerators](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/machine-learning-training/distributed-and-accelerated-compute/training-acceleration-tools/gpu-training-accelerators.md) — Provides tools that use parallelization strategies to increase the overall speed of deep learning model training. ([source](https://github.com/tensorpack/tensorpack#readme))
- [Sequence-to-Sequence Tasks](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/speech-processing/sequence-to-sequence-tasks.md) — Develops Long Short-Term Memory networks for sequence-to-sequence tasks in speech recognition and natural language processing. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples))
- [Neural Network Interpretability](https://awesome-repositories.com/f/artificial-intelligence-ml/neural-network-interpretability.md) — Includes a toolkit for analyzing convolutional networks via saliency maps and activation visualizations to understand model decisions.
- [Class Activation Map Generation](https://awesome-repositories.com/f/artificial-intelligence-ml/prediction-visualization/predictive-density-heatmaps/class-activation-map-generation.md) — Provides spatial heatmaps to localize image regions that drive specific neural network class predictions. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples/Saliency))
- [Weight Quantization](https://awesome-repositories.com/f/artificial-intelligence-ml/quantized-inference-runtimes/weight-quantization.md) — Reduces model memory footprint by implementing binary and ternary weight representations during training.
- [Character-Level Language Models](https://awesome-repositories.com/f/artificial-intelligence-ml/recurrent-neural-networks/character-level-language-models.md) — Trains recurrent neural networks on text corpora to learn and predict character-level linguistic patterns. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples/Char-RNN))
- [RNN State Management](https://awesome-repositories.com/f/artificial-intelligence-ml/rnn-state-management.md) — Manages recurrent neural network hidden states directly within the TensorFlow computation graph to preserve temporal context.
- [Pixel Saliency Maps](https://awesome-repositories.com/f/artificial-intelligence-ml/saliency-mapping/pixel-saliency-maps.md) — Visualizes model decision making by backpropagating activations from the output layer to input image pixels.
- [Training Dataset Management](https://awesome-repositories.com/f/artificial-intelligence-ml/training-dataset-management.md) — Automates the downloading and augmentation of image datasets to improve model training quality. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples/HED))
- [Unpaired Image Translation](https://awesome-repositories.com/f/artificial-intelligence-ml/unpaired-image-translation.md) — Transforms images between unpaired domains using adversarial loss to map style without direct examples.

### Data & Databases

- [Parallel Data Loading](https://awesome-repositories.com/f/data-databases/python-data-pipeline-frameworks/parallel-data-loading.md) — Uses pure Python multiprocessing to stream datasets into the computation graph, bypassing framework-specific pipeline constraints.
- [Training Data Pipelines](https://awesome-repositories.com/f/data-databases/data-processing-pipelines/data-processing/ml-data-pipelines/training-data-pipelines.md) — Streams datasets into models using specialized reader pipelines to improve training and inference efficiency. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples/PennTreebank))
- [Low Bitwidth Training](https://awesome-repositories.com/f/data-databases/memory-optimization-strategies/training-memory-optimizers/low-bitwidth-training.md) — Reduces memory usage by training convolutional neural networks with limited bitwidth for weights and gradients. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples/DoReFa-Net))
- [Training State Restoration](https://awesome-repositories.com/f/data-databases/model-state-restoration/checkpoint-based-state-restoration/training-state-restoration.md) — Continues training from a previous stop by restoring model weights and epoch counters from checkpoints. ([source](https://github.com/tensorpack/tensorpack/blob/master/CHANGES.md))

### Part of an Awesome List

- [Model Weight Checkpoints](https://awesome-repositories.com/f/awesome-lists/ai/model-training-and-fine-tuning/checkpoint-saving-and-restoration/model-weight-checkpoints.md) — Provides mechanisms to save and restore neural network weights and optimizer states to ensure training continuity.
- [ImageNet Classifier Training](https://awesome-repositories.com/f/awesome-lists/ai/model-training-and-fine-tuning/imagenet-classifier-training.md) — Builds standard architectures like ResNet and VGG using customizable scripts for the ImageNet dataset. ([source](https://github.com/tensorpack/tensorpack/blob/master/examples/ImageNetModels))
- [Deep Learning Ecosystems](https://awesome-repositories.com/f/awesome-lists/ai/deep-learning-ecosystems.md) — Neural network training interface for TensorFlow.
- [Deep Learning Frameworks](https://awesome-repositories.com/f/awesome-lists/ai/deep-learning-frameworks.md) — Neural network training interface for TensorFlow.
- [Machine Learning Frameworks](https://awesome-repositories.com/f/awesome-lists/ai/machine-learning-frameworks.md) — Training interface for TensorFlow focused on efficiency and speed.
- [Quantization Frameworks](https://awesome-repositories.com/f/awesome-lists/ai/quantization-frameworks.md) — Training low bitwidth networks with low bitwidth gradients.