This repository was archived by the owner on Dec 16, 2022. It is now read-only.
add gradient checkpointing for transformer token embedders #4544
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Awesome. I just noticed that the support for checkpointing in Huggingface came originally from AI2's own @ibeltagy. Thanks for driving that! |
dirkgr
approved these changes
Aug 10, 2020
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Gradient checkpointing has been released with huggingface/transformers v3.0.0. Gradient checkpointing reduces memory by 5x which makes it possible to process longer sequences on smaller GPUs. This PR allows to fine-tune transformer token embedders on a more regular GPU with high(er) max_length values.
Sorry for not providing any tests, I have no clue how to create a simple test case and no resources at the moment to figure that out. But I tried it succesfully with Longformer (max_length: 4096) and SciBert (max_length: 512) on a single RTX 2080 Ti.