beatlesctr/libai

LiBai(李白): A Toolbox for Large-Scale Distributed Parallel Training

LiBai

Introduction

English | 简体中文

LiBai is a large-scale open-source model training toolbox based on OneFlow. The main branch works with OneFlow 0.7.0.

Highlights

• Support a collection of parallel training components

  LiBai provides multiple parallelisms such as Data Parallelism, Tensor Parallelism, and Pipeline Parallelism. It's also extensible for other new parallelisms.

• Varied training techniques

  LiBai provides many out-of-the-box training techniques such as Distributed Training, Mixed Precision Training, Activation Checkpointing, Recomputation, Gradient Accumulation, and Zero Redundancy Optimizer(ZeRO).

• Support for both CV and NLP tasks

  LiBai has predifined data process for both CV and NLP datasets such as CIFAR, ImageNet, and BERT Dataset.

• Easy to use

  LiBai's components are designed to be modular for easier usage as follows:

  • LazyConfig system for more flexible syntax and no predefined structures
  • Friendly trainer and engine
  • Used as a library to support building research projects on it. See projects/ for some projects that are built based on LiBai

• High Efficiency

Installation

See Installation instructions.

Getting Started

See Quick Run for the basic usage of LiBai.

Documentation

See LiBai's documentation for full API documentation and tutorials.

ChangeLog

Beta 0.1.0 was released in 22/03/2022, the main features and supported models in 0.1.0 version are as follows:

Features:

• Support Data Parallelism
• Support 1D Tensor Parallelism
• Support Pipeline Parallelism
• Unified distributed Layers for both single-GPU and multi-GPU training
• LazyConfig system for more flexible syntax and no predefined structures
• Easy-to-use trainer and engine
• Support both CV and NLP data processing
• Mixed Precision Training
• Activation Checkpointing
• Gradient Accumulation
• Gradient Clipping
• Zero Redundancy Optimizer (ZeRO)

Supported Models:

• Support 3D parallel BERT model
• Support 3D parallel GPT-2 model
• Support 3D parallel T5 model
• Support 3D parallel Vision Transformer model
• Support Data parallel Swin Transformer model
• Support finetune task in QQP project
• Support text classification task in text classification project
• Support Pathways Language Model (PaLM) in PaLM project
• Support MoCo_v3 in MOCOV3 project
• (experimental) Support MAE in MAE project

See changelog for details and release history.

Contributing

We appreciate all contributions to improve LiBai. See CONTRIBUTING for the contributing guideline.

License

This project is released under the Apache 2.0 license.

Citation

If you find this project useful for your research, consider cite:

@misc{of2021libai,
  author =       {Xingyu Liao and Peng Cheng and Tianhe Ren and Depeng Liang and
                  Kai Dang and Yi Wang and Xiaoyu Xu},
  title =        {LiBai},
  howpublished = {\url{https://github.com/Oneflow-Inc/libai}},
  year =         {2021}
}