You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
ColossalAI/colossalai/nn/optimizer
Frank Lee 40d376c566
[setup] support pre-build and jit-build of cuda kernels (#2374)
2 years ago
..
README.md add optimizer README for tutorials (#1707) 2 years ago
__init__.py [hotfix] fix CPUAdam kernel nullptr (#1410) 2 years ago
colossalai_optimizer.py [checkpoint] add ColoOptimizer checkpointing (#1316) 2 years ago
cpu_adam.py [setup] support pre-build and jit-build of cuda kernels (#2374) 2 years ago
fused_adam.py [setup] support pre-build and jit-build of cuda kernels (#2374) 2 years ago
fused_lamb.py [setup] support pre-build and jit-build of cuda kernels (#2374) 2 years ago
fused_sgd.py [setup] support pre-build and jit-build of cuda kernels (#2374) 2 years ago
gemini_optimizer.py [Gemini] add GeminiAdamOptimizer (#1960) 2 years ago
hybrid_adam.py [setup] support pre-build and jit-build of cuda kernels (#2374) 2 years ago
lamb.py [hotfix] adapt ProcessGroup and Optimizer to ColoTensor (#1388) 2 years ago
lars.py Fixed docstring in colossalai (#171) 3 years ago
nvme_optimizer.py [nvme] CPUAdam and HybridAdam support NVMe offload (#1360) 2 years ago
zero_optimizer.py [optimizer] add div_scale for optimizers (#2117) 2 years ago

README.md

Colossal-AI Optimization Techniques

Introduction

Welcome to the large-scale deep learning optimization techniques of Colossal-AI, which has been accepted as official tutorials by top conference AAAI, PPoPP, etc.

Colossal-AI, a unified deep learning system for the big model era, integrates many advanced technologies such as multi-dimensional tensor parallelism, sequence parallelism, heterogeneous memory management, large-scale optimization, adaptive task scheduling, etc. By using Colossal-AI, we could help users to efficiently and quickly deploy large AI model training and inference, reducing large AI model training budgets and scaling down the labor cost of learning and deployment.

Colossal-AI | Paper | Documentation | Forum | Slack

Table of Content

Large transformer models display promising performance on a wide spectrum of AI applications. Both academia and industry are scaling DL training on larger clusters. However, degrading generalization performance, non-negligible communication overhead, and increasing model size prevent DL researchers and engineers from exploring large-scale AI models.

We aim to provide a clear sketch of the optimizations for large-scale deep learning with regard to model accuracy and model efficiency. One way to achieve the goal of maintaining or improving the model accuracy in the large-scale setting while maintaining compute efficiency is to design algorithms that are less communication and memory hungry. Notably, they are not mutually exclusive but can be optimized jointly to further speed up training.

  1. Model Accuracy

    • Gradient Descent Optimization
      • Gradient Descent Variants
      • Momentum
      • Adaptive Gradient
    • Large Batch Training Optimization
      • LARS
      • LAMB
      • Generalization Gap
    • Second-Order Optimization
      • Hessian-Free
      • K-FAC
      • Shampoo
  2. Model Accuracy

    • Communication Efficiency
      • Reduce Volumn of Comm.
      • Reduce Frequency of Comm.
    • Memory Efficiency
      • Mix-Precision Training
      • Memory-Efficient Methods, e.g. ZeRO, Gemini, etc.

Some of the above are still under development. If you wish to make a contribution to this repository, please read the Contributing section below.

Discussion

Discussion about the Colossal-AI project is always welcomed! We would love to exchange ideas with the community to better help this project grow. If you think there is a need to discuss anything, you may jump to our Slack.

If you encounter any problem while running these optimizers, you may want to raise an issue in this repository.

Contributing

This project welcomes constructive ideas and implementations from the community.

Update an Optimizer

If you find that an optimizer is broken (not working) or not user-friendly, you may put up a pull request to this repository and update this optimizer.

Add a New Optimizer

If you wish to add an optimizer for a specific application, please follow the steps below.

  1. create the new optimizer file in the current folder
  2. Prepare the corresponding example files in the Examples repository to prove effectiveness of the new optimizer
  3. Prepare a detailed readme on environment setup, dataset preparation, code execution, etc. in your example folder
  4. Update the table of content (last section above) in this readme file

If your PR is accepted, we may invite you to put up a tutorial or blog in ColossalAI Documentation.