Pipeline Parallelism Python Packages

deepspeed

DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.

colossalai

Making large AI models cheaper, faster and more accessible

omniback

Serving Inside Pytorch

colossalai-nightly

Making large AI models cheaper, faster and more accessible

petals

Easy way to efficiently run 100B+ language models without high-end GPUs

torchgpipe

A GPipe implementation in PyTorch

hami-core

Serving Inside Pytorch

fleet-x

No description available

custom-colossalai

Making large AI models cheaper, faster and more accessible

torchslicer

Split learning framework for PyTorch — partition a model across devices and train end-to-end without any single node holding the full model.

elf-pipeline

Automated Memory-Aware Pipeline Parallel Training of Neural Networks

fleet-lightning

飞桨大模型开发套件，提供大语言模型、跨模态大模型、生物计算大模型等领域的全流程开发工具链。

adeepspeed

DeepSpeed library

paddle-fleet

飞桨大模型开发套件，提供大语言模型、跨模态大模型、生物计算大模型等领域的全流程开发工具链。

test-petals

🌸 Run LLMs at home, BitTorrent-style. Fine-tuning and inference up to 10x faster than offloading