# Colossal-AI
< div id = "top" align = "center" >
[](https://www.colossalai.org/)
一个整合高效并行技术的 AI 大模型训练系统。
< h3 > < a href = "https://arxiv.org/abs/2110.14883" > 论文 < / a > |
< a href = "https://www.colossalai.org/" > 文档 < / a > |
< a href = "https://github.com/hpcaitech/ColossalAI-Examples" > 例程 < / a > |
< a href = "https://github.com/hpcaitech/ColossalAI/discussions" > 论坛 < / a > |
< a href = "https://medium.com/@hpcaitech" > 博客 < / a > < / h3 >
[](https://github.com/hpcaitech/ColossalAI/actions/workflows/build.yml)
[](https://colossalai.readthedocs.io/en/latest/?badge=latest)
[](https://www.codefactor.io/repository/github/hpcaitech/colossalai)
[](https://huggingface.co/hpcai-tech)
[](https://join.slack.com/t/colossalaiworkspace/shared_invite/zt-z7b26eeb-CBp7jouvu~r0~lcFzX832w)
[](https://raw.githubusercontent.com/hpcaitech/public_assets/main/colossalai/img/WeChat.png)
| [English ](README.md ) | [中文 ](README-zh-Hans.md ) |
< / div >
## 目录
< ul >
< li > < a href = "#为何选择-Colossal-AI" > 为何选择 Colossal-AI< / a > < / li >
< li > < a href = "#特点" > 特点< / a > < / li >
< li >
< a href = "#并行样例展示" > 并行样例展示< / a >
< ul >
< li > < a href = "#ViT" > ViT< / a > < / li >
< li > < a href = "#GPT-3" > GPT-3< / a > < / li >
< li > < a href = "#GPT-2" > GPT-2< / a > < / li >
< li > < a href = "#BERT" > BERT< / a > < / li >
< li > < a href = "#PaLM" > PaLM< / a > < / li >
< / ul >
< / li >
< li >
< a href = "#单GPU样例展示" > 单GPU样例展示< / a >
< ul >
< li > < a href = "#GPT-2-Single" > GPT-2< / a > < / li >
< li > < a href = "#PaLM-Single" > PaLM< / a > < / li >
< / ul >
< / li >
< li >
< a href = "#安装" > 安装< / a >
< ul >
< li > < a href = "#PyPI" > PyPI< / a > < / li >
< li > < a href = "#从源代码安装" > 从源代码安装< / a > < / li >
< / ul >
< / li >
< li > < a href = "#使用-Docker" > 使用 Docker< / a > < / li >
< li > < a href = "#社区" > 社区< / a > < / li >
< li > < a href = "#做出贡献" > 做出贡献< / a > < / li >
< li > < a href = "#快速预览" > 快速预览< / a > < / li >
< ul >
< li > < a href = "#几行代码开启分布式训练" > 几行代码开启分布式训练< / a > < / li >
< li > < a href = "#构建一个简单的2维并行模型" > 构建一个简单的2维并行模型< / a > < / li >
< / ul >
< li > < a href = "#引用我们" > 引用我们< / a > < / li >
< / ul >
## 为何选择 Colossal-AI
< div align = "center" >
< a href = "https://youtu.be/KnXSfjqkKN0" >
< img src = "https://raw.githubusercontent.com/hpcaitech/public_assets/main/colossalai/img/JamesDemmel_Colossal-AI.png" width = "600" / >
< / a >
James Demmel 教授 (加州大学伯克利分校): Colossal-AI 让分布式训练高效、易用、可扩展。
< / div >
< p align = "right" > (< a href = "#top" > 返回顶端< / a > )< / p >
## 特点
Colossal-AI 为您提供了一系列并行训练组件。我们的目标是让您的分布式 AI 模型训练像普通的单 GPU 模型一样简单。我们提供的友好工具可以让您在几行代码内快速开始分布式训练。
- 并行化策略
- 数据并行
- 流水线并行
- 1维, [2维 ](https://arxiv.org/abs/2104.05343 ), [2.5维 ](https://arxiv.org/abs/2105.14500 ), [3维 ](https://arxiv.org/abs/2105.14450 ) 张量并行
- [序列并行 ](https://arxiv.org/abs/2105.13120 )
- [零冗余优化器 (ZeRO) ](https://arxiv.org/abs/2108.05818 )
- 异构内存管理
- [PatrickStar ](https://arxiv.org/abs/2108.05818 )
- 使用友好
- 基于参数文件的并行化
< p align = "right" > (< a href = "#top" > 返回顶端< / a > )< / p >
## 并行样例展示
### ViT
< p align = "center" >
< img src = "https://raw.githubusercontent.com/hpcaitech/public_assets/main/colossalai/img/ViT.png" width = "450" / >
< / p >
- 14倍批大小和5倍训练速度（ ）
### GPT-3
< p align = "center" >
< img src = "https://raw.githubusercontent.com/hpcaitech/public_assets/main/colossalai/img/GPT3.png" width = 700/ >
< / p >
- 释放 50% GPU 资源占用, 或 10.7% 加速
### GPT-2
< img src = "https://raw.githubusercontent.com/hpcaitech/public_assets/main/colossalai/img/GPT2.png" width = 800/ >
- 降低11倍 GPU 显存占用，或超线性扩展（张量并行）
< img src = "https://raw.githubusercontent.com/hpcaitech/public_assets/main/colossalai/img/(updated)GPT-2.png" width = 800 >
- 用相同的硬件条件训练24倍大的模型
- 超3倍的吞吐量
### BERT
< img src = "https://raw.githubusercontent.com/hpcaitech/public_assets/main/colossalai/img/BERT.png" width = 800/ >
- 2倍训练速度，
### PaLM
- [PaLM-colossalai ](https://github.com/hpcaitech/PaLM-colossalai ): 可扩展的谷歌 Pathways Language Model ([PaLM](https://ai.googleblog.com/2022/04/pathways-language-model-palm-scaling-to.html)) 实现。
请访问我们的[文档和教程](https://www.colossalai.org/)以了解详情。
< p align = "right" > (< a href = "#top" > 返回顶端< / a > )< / p >
## 单GPU样例展示
### GPT-2
< p id = "GPT-2-Single" align = "center" >
< img src = "https://raw.githubusercontent.com/hpcaitech/public_assets/main/colossalai/img/GPT2-GPU1.png" width = 450/ >
< / p >
- 用相同的硬件条件训练20倍大的模型
### PaLM
< p id = "PaLM-Single" align = "center" >
< img src = "https://raw.githubusercontent.com/hpcaitech/public_assets/main/colossalai/img/PaLM-GPU1.png" width = 450/ >
< / p >
- 用相同的硬件条件训练34倍大的模型
< p align = "right" > (< a href = "#top" > back to top< / a > )< / p >
## 安装
### 从官方安装
您可以访问我们[下载](/download)页面来安装Colossal-AI，
### 从源安装
> 此文档将与版本库的主分支保持一致。如果您遇到任何问题，欢迎给我们提 issue :)
```shell
git clone https://github.com/hpcaitech/ColossalAI.git
cd ColossalAI
# install dependency
pip install -r requirements/requirements.txt
# install colossalai
pip install .
```
如果您不想安装和启用 CUDA 内核融合（使用融合优化器时强制安装）：
```shell
NO_CUDA_EXT=1 pip install .
```
< p align = "right" > (< a href = "#top" > 返回顶端< / a > )< / p >
## 使用 Docker
运行以下命令从我们提供的 docker 文件中建立 docker 镜像。
```bash
cd ColossalAI
docker build -t colossalai ./docker
```
运行以下命令从以交互式启动 docker 镜像.
```bash
docker run -ti --gpus all --rm --ipc=host colossalai bash
```
< p align = "right" > (< a href = "#top" > 返回顶端< / a > )< / p >
## 社区
欢迎通过[论坛](https://github.com/hpcaitech/ColossalAI/discussions),
[Slack ](https://join.slack.com/t/colossalaiworkspace/shared_invite/zt-z7b26eeb-CBp7jouvu~r0~lcFzX832w ),
或[微信](https://raw.githubusercontent.com/hpcaitech/public_assets/main/colossalai/img/WeChat.png "qrcode")加入 Colossal-AI 社区，与我们分享你的建议和问题。
## 做出贡献
欢迎为该项目做出贡献，请参阅[贡献指南](./CONTRIBUTING.md)。
真诚感谢所有贡献者！
< a href = "https://github.com/hpcaitech/ColossalAI/graphs/contributors" > < img src = "https://raw.githubusercontent.com/hpcaitech/public_assets/main/colossalai/img/contributor_avatar.png" width = "800px" > < / a >
*贡献者头像的展示顺序是随机的。*
< p align = "right" > (< a href = "#top" > 返回顶端< / a > )< / p >
## 快速预览
### 几行代码开启分布式训练
```python
parallel = dict(
pipeline=2,
tensor=dict(mode='2.5d', depth = 1, size=4)
)
```
### 几行代码开启异构训练
```python
zero = dict(
model_config=dict(
tensor_placement_policy='auto',
shard_strategy=TensorShardStrategy(),
reuse_fp16_shard=True
),
optimizer_config=dict(initial_scale=2**5, gpu_margin_mem_ratio=0.2)
)
```
< p align = "right" > (< a href = "#top" > 返回顶端< / a > )< / p >
## 引用我们
```
@article {bian2021colossal,
title={Colossal-AI: A Unified Deep Learning System For Large-Scale Parallel Training},
author={Bian, Zhengda and Liu, Hongxin and Wang, Boxiang and Huang, Haichen and Li, Yongbin and Wang, Chuanrui and Cui, Fan and You, Yang},
journal={arXiv preprint arXiv:2110.14883},
year={2021}
}
```
< p align = "right" > (< a href = "#top" > 返回顶端< / a > )< / p >
