mistralai/megablocks-public

mistralai/megablocks-public

License: Apache-2.0

Stars: 868

Forks: 62

Open issues: 0

Created: 2023-12-08T08:35:30Z

Pushed: 2023-12-08T09:56:39Z

Default branch: main

Fork: yes

Parent repository: databricks/megablocks

Archived: yes

README:

:robot: MegaBlocks

MegaBlocks is a light-weight library for mixture-of-experts (MoE) training. The core of the system is efficient "dropless-MoE" ([dMoE](megablocks/layers/dmoe.py), paper) and standard [MoE](megablocks/layers/moe.py) layers.

MegaBlocks is built on top of Megatron-LM, where we support data, expert and pipeline parallel training of MoEs. We're working on extending more frameworks to support MegaBlocks.

:rocket: Performance


MegaBlocks dMoEs outperform MoEs trained with Tutel by up to 40% compared to Tutel's best performing capacity_factor configuration. MegaBlocks dMoEs use a reformulation of MoEs in terms of block-sparse operations, which allows us to avoid token dropping without sacrificing hardware efficiency. In addition to being faster, MegaBlocks simplifies MoE training by removing the capacity_factor hyperparameter alltogether. Compared to dense Transformers trained with Megatron-LM, MegaBlocks dMoEs can accelerate training by as much as 2.4x. Check out our paper for more details!

:building_construction: Installation

Note: this assumes you have numpy and torch installed

Training models with Megatron-LM: We recommend using NGC's `nvcr.io/nvidia/pytorch:23.01-py3` PyTorch container. The [Dockerfile](Dockerfile) builds on this image with additional dependencies. To build the image, run docker build . -t megablocks-dev and then bash docker.sh to launch the container. Once inside the container, install MegaBlocks with pip install .. See [Usage](#steam_locomotive-usage) for instructions on training MoEs with MegaBlocks + Megatron-LM.

Using MegaBlocks in other packages: To install the MegaBlocks package for use in other frameworks, run pip install megablocks.

:steam_locomotive: Usage

We provide scripts for pre-training Transformer MoE and dMoE language models under the [top-level directory](megablocks/). The quickest way to get started is to use one of the [experiment launch scripts](exp/). These scripts require a dataset in Megatron-LM's format, which can be created by following their instructions.

:writing_hand: Citation

@article{megablocks,
title={{MegaBlocks: Efficient Sparse Training with Mixture-of-Experts}},
author={Trevor Gale and Deepak Narayanan and Cliff Young and Matei Zaharia},
journal={Proceedings of Machine Learning and Systems},
volume={5},
year={2023}
}