README.md 516 Bytes
Newer Older
Jiarui Fang's avatar
Jiarui Fang committed
1
## Overview
2
This example shows how to use Colossal-AI to run huggingface GPT training in distributed manners.
Jiarui Fang's avatar
Jiarui Fang committed
3

4
## GPT
5
We use the GPT2 model from huggingface transformers. The input data is randonly generated.
Jiarui Fang's avatar
Jiarui Fang committed
6

7
## Our Modifications
8
9
The `train_gpt_demo.py` provides three distributed plans, i.e. Colossal-AI, PyTorch DDP and ZeRO.
The Colossal-AI leverages Tensor Parallel and Gemini.
Jiarui Fang's avatar
Jiarui Fang committed
10

11
## Quick Start
12
You can launch training by using the following bash script.
Jiarui Fang's avatar
Jiarui Fang committed
13
14

```bash
15
16
pip install -r requirements.txt
bash run.sh
Jiarui Fang's avatar
Jiarui Fang committed
17
```