/megablocks

Primary LanguagePythonApache License 2.0Apache-2.0

🤖 MegaBlocks

MegaBlocks is a light-weight library for mixture-of-experts (MoE) training. The core of the system is efficient "dropless-MoE" (dMoE, paper) and standard MoE 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.

🚀 Performance

MegaBlocks 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 altogether. 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!

🏗️ Installation

NOTE: This assumes you have numpy and torch installed.

Training models with Megatron-LM: We recommend using NGC's nvcr.io/nvidia/pytorch:23.09-py3 PyTorch container. The 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 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. For example, Mixtral-8x7B can be run with vLLM + MegaBlocks with this installation method.

Extras: MegaBlocks has optional dependencies that enable additional features.

Installing megablocks[quant] enables configurable quantization of saved activations in the dMoE layer to save memory during training. The degree of quantization is controlled via the quantize_inputs_num_bits, quantize_rematerialize_num_bits and quantize_scatter_num_bits arguments.

Installing megablocks[gg] enables dMoE computation with grouped GEMM. This feature is enabled by setting the mlp_impl argument to grouped. This is currently our recommended path for Hopper-generation GPUs.

MegaBlocks can be installed with all dependencies via the megablocks[all] package.

🚂 Usage

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

✍️ 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}
}