Colossal-AI

Colossal-AI: A Unified Deep Learning System for Big Model Era

Paper | Documentation | Examples | Forum | Blog

| English | 中文 |

Table of Contents

• Why Colossal-AI
• Features
• Parallel Training Demo
  • ViT
  • GPT-3
  • GPT-2
  • BERT
  • PaLM
  • OPT
• Single GPU Training Demo
  • GPT-2
  • PaLM
• Inference (Energon-AI) Demo
  • GPT-3
• Installation
  • PyPI
  • Install From Source
• Use Docker
• Community
• Contributing
• Quick View
• Start Distributed Training in Lines
• Write a Simple 2D Parallel Model
• Cite Us

Why Colossal-AI

Prof. James Demmel (UC Berkeley): Colossal-AI makes distributed training efficient, easy and scalable.

(back to top)

Features

Colossal-AI provides a collection of parallel components for you. We aim to support you to write your distributed deep learning models just like how you write your model on your laptop. We provide user-friendly tools to kickstart distributed training and inference in a few lines.

• Parallelism strategies

  • Data Parallelism
  • Pipeline Parallelism
  • 1D, 2D, 2.5D, 3D Tensor Parallelism
  • Sequence Parallelism
  • Zero Redundancy Optimizer (ZeRO)

• Heterogeneous Memory Management

  • PatrickStar

• Friendly Usage

  • Parallelism based on configuration file

• Inference

  • Energon-AI

(back to top)

Parallel Training Demo

ViT

• 14x larger batch size, and 5x faster training for Tensor Parallelism = 64

GPT-3

• Save 50% GPU resources, and 10.7% acceleration

GPT-2

• 11x lower GPU memory consumption, and superlinear scaling efficiency with Tensor Parallelism

• 24x larger model size on the same hardware
• over 3x acceleration

BERT

• 2x faster training, or 50% longer sequence length

PaLM

• PaLM-colossalai: Scalable implementation of Google's Pathways Language Model (PaLM).

OPT

• Open Pretrained Transformer (OPT), a 175-Billion parameter AI language model released by Meta, which stimulates AI programmers to perform various downstream tasks and application deployments because public pretrained model weights.
• 40% speedup fine-tuning OPT at low cost in lines. [Example]

Please visit our documentation and examples for more details.

(back to top)

Single GPU Training Demo

GPT-2

• 20x larger model size on the same hardware

PaLM

• 34x larger model size on the same hardware

(back to top)

Inference (Energon-AI) Demo

GPT-3

• Energon-AI: 50% inference acceleration on the same hardware

(back to top)

Installation

Download From Official Releases

You can visit the Download page to download Colossal-AI with pre-built CUDA extensions.

Download From Source

The version of Colossal-AI will be in line with the main branch of the repository. Feel free to raise an issue if you encounter any problem. :)

git clone https://github.com/hpcaitech/ColossalAI.git
cd ColossalAI

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

# install colossalai
pip install .

If you don't want to install and enable CUDA kernel fusion (compulsory installation when using fused optimizer):

NO_CUDA_EXT=1 pip install .

(back to top)

Use Docker

Pull from DockerHub

You can directly pull the docker image from our DockerHub page. The image is automatically uploaded upon release.

Build On Your Own

Run the following command to build a docker image from Dockerfile provided.

Building Colossal-AI from scratch requires GPU support, you need to use Nvidia Docker Runtime as the default when doing docker build. More details can be found here. We recommend you install Colossal-AI from our project page directly.

cd ColossalAI
docker build -t colossalai ./docker

Run the following command to start the docker container in interactive mode.

docker run -ti --gpus all --rm --ipc=host colossalai bash

(back to top)

Community

Join the Colossal-AI community on Forum, Slack, and WeChat to share your suggestions, feedback, and questions with our engineering team.

Contributing

If you wish to contribute to this project, please follow the guideline in Contributing.

Thanks so much to all of our amazing contributors!

The order of contributor avatars is randomly shuffled.

(back to top)

Quick View

Start Distributed Training in Lines

parallel = dict(
    pipeline=2,
    tensor=dict(mode='2.5d', depth = 1, size=4)
)

Start Heterogeneous Training in Lines

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)
)

(back to top)

Cite Us

@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}
}

(back to top)