Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
46 changes: 34 additions & 12 deletions README-zh-Hans.md
Original file line number Diff line number Diff line change
Expand Up @@ -5,10 +5,10 @@

Colossal-AI: 一个面向大模型时代的通用深度学习系统

<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> |
<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>

[![Build](https://github.com/hpcaitech/ColossalAI/actions/workflows/build.yml/badge.svg)](https://github.com/hpcaitech/ColossalAI/actions/workflows/build.yml)
Expand All @@ -35,7 +35,7 @@
<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="#并行训练样例展示">并行训练样例展示</a>
<a href="#并行训练样例展示">并行训练样例展示</a>
<ul>
<li><a href="#GPT-3">GPT-3</a></li>
<li><a href="#GPT-2">GPT-2</a></li>
Expand All @@ -47,22 +47,22 @@
</ul>
</li>
<li>
<a href="#单GPU训练样例展示">单GPU训练样例展示</a>
<a href="#单GPU训练样例展示">单GPU训练样例展示</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="#推理-Energon-AI-样例展示">推理 (Energon-AI) 样例展示</a>
<a href="#推理-Energon-AI-样例展示">推理 (Energon-AI) 样例展示</a>
<ul>
<li><a href="#GPT-3-Inference">GPT-3</a></li>
<li><a href="#OPT-Serving">1750亿参数OPT在线推理服务</a></li>
<li><a href="#BLOOM-Inference">1750亿参数 BLOOM</a></li>
</ul>
</li>
<li>
<a href="#Colossal-AI-in-the-Real-World">Colossal-AI 成功案例</a>
<a href="#Colossal-AI-in-the-Real-World">Colossal-AI 成功案例</a>
<ul>
<li><a href="#AIGC">AIGC: 加速 Stable Diffusion</a></li>
<li><a href="#生物医药">生物医药: 加速AlphaFold蛋白质结构预测</a></li>
Expand Down Expand Up @@ -131,7 +131,7 @@ Colossal-AI 为您提供了一系列并行组件。我们的目标是让您的
<img src="https://raw.githubusercontent.com/hpcaitech/public_assets/main/colossalai/img/(updated)GPT-2.png" width=800>

- 用相同的硬件训练24倍大的模型
- 超3倍的吞吐量
- 超3倍的吞吐量

### BERT
<img src="https://raw.githubusercontent.com/hpcaitech/public_assets/main/colossalai/img/BERT.png" width=800/>
Expand All @@ -145,7 +145,7 @@ Colossal-AI 为您提供了一系列并行组件。我们的目标是让您的
<img src="https://raw.githubusercontent.com/hpcaitech/public_assets/main/colossalai/img/OPT_update.png" width=800/>

- [Open Pretrained Transformer (OPT)](https://github.com/facebookresearch/metaseq), 由Meta发布的1750亿语言模型,由于完全公开了预训练参数权重,因此促进了下游任务和应用部署的发展。
- 加速45%,仅用几行代码以低成本微调OPT。[[样例]](https://github.com/hpcaitech/ColossalAI-Examples/tree/main/language/opt) [[在线推理]](https://service.colossalai.org/opt)
- 加速45%,仅用几行代码以低成本微调OPT。[[样例]](https://github.com/hpcaitech/ColossalAI-Examples/tree/main/language/opt) [[在线推理]](https://service.colossalai.org/opt)

请访问我们的 [文档](https://www.colossalai.org/) 和 [例程](https://github.com/hpcaitech/ColossalAI-Examples) 以了解详情。

Expand Down Expand Up @@ -255,6 +255,28 @@ Colossal-AI 为您提供了一系列并行组件。我们的目标是让您的

## 安装

### 从PyPI安装

您可以用下面的命令直接从PyPI上下载并安装Colossal-AI。我们默认不会安装PyTorch扩展包

```bash
pip install colossalai
```

但是,如果你想在安装时就直接构建PyTorch扩展,您可以设置环境变量`CUDA_EXT=1`.

```bash
CUDA_EXT=1 pip install colossalai
```

**否则,PyTorch扩展只会在你实际需要使用他们时在运行时里被构建。**

与此同时,我们也每周定时发布Nightly版本,这能让你提前体验到新的feature和bug fix。你可以通过以下命令安装Nightly版本。

```bash
pip install colossalai-nightly
```

### 从官方安装

您可以访问我们[下载](https://www.colossalai.org/download)页面来安装Colossal-AI,在这个页面上发布的版本都预编译了CUDA扩展。
Expand All @@ -274,10 +296,10 @@ pip install -r requirements/requirements.txt
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):
我们默认在`pip install`时不安装PyTorch扩展,而是在运行时临时编译,如果你想要提前安装这些扩展的话(在使用融合优化器时会用到),可以使用一下命令。

```shell
NO_CUDA_EXT=1 pip install .
CUDA_EXT=1 pip install .
```

<p align="right">(<a href="#top">返回顶端</a>)</p>
Expand Down
28 changes: 24 additions & 4 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -257,9 +257,32 @@ Acceleration of [AlphaFold Protein Structure](https://alphafold.ebi.ac.uk/)

## Installation

### Install from PyPI

You can easily install Colossal-AI with the following command. **By defualt, we do not build PyTorch extensions during installation.**

```bash
pip install colossalai
```

However, if you want to build the PyTorch extensions during installation, you can set `CUDA_EXT=1`.

```bash
CUDA_EXT=1 pip install colossalai
```

**Otherwise, CUDA kernels will be built during runtime when you actually need it.**

We also keep release the nightly version to PyPI on a weekly basis. This allows you to access the unreleased features and bug fixes in the main branch.
Installation can be made via

```bash
pip install colossalai-nightly
```

### Download From Official Releases

You can visit the [Download](https://www.colossalai.org/download) page to download Colossal-AI with pre-built CUDA extensions.
You can visit the [Download](https://www.colossalai.org/download) page to download Colossal-AI with pre-built PyTorch extensions.


### Download From Source
Expand All @@ -270,9 +293,6 @@ You can visit the [Download](https://www.colossalai.org/download) page to downlo
git clone https://github.com/hpcaitech/ColossalAI.git
cd ColossalAI

# install dependency
pip install -r requirements/requirements.txt

# install colossalai
pip install .
```
Expand Down