huggingface / huggingface/accelerate

[FSDP] support activation offloading with FSDP

Open
#2,038 5 comments 0 reactions 0 assignees View on GitHub
enhancement feature request
Dominant language
Python
Stars
9.9k
Forks
1.5k
Avg merge
5d 2h
Merged PRs (30d)
27

Description

Support whole model activation offloading with FSDP - working in conjunction with activation checkpointing - via

https://github.com/pytorch/pytorch/blob/e9ebda29d87ce0916ab08c06ab26fd3766a870e5/torch/distributed/algorithms/_checkpoint/checkpoint_wrapper.py#L171-L191

As `apply_activation_checkpointing` does not wrap the overall root module, wrapping the overall root module with this could offload activation between layer, thus release more GPU memory. The diff should be small and i am happy to work on this.

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.