Support Rank Stabilized LoRA in the ModelConfig/LoraConfig #1877
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
|
@JohnGiorgi do we need to pin peft to some version for this? |
kashif
approved these changes
Jul 28, 2024
|
yes i believe we will need to pin peft to that for this PR, also could you kindly also add a test? |
|
@kashif Okay I bumped the peft dependency to >=0.8.0. I also added a test for |
JohnGiorgi
changed the title
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
|
@JohnGiorgi i believe its failing for one of the older python with: AttributeError: 'str' object has no attribute 'removeprefix' |
|
Fixed! Let me know if anything else is blocking this |
qgallouedec
reviewed
Aug 6, 2024
qgallouedec
reviewed
Aug 6, 2024
| "pytest", | ||
| "pytest-xdist", | ||
| "accelerate", |
|
Merged, thank you @JohnGiorgi |
Thank you! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Tiny PR to support Rank Stabilized LoRA. To do this we:
ModelConfig. It is copied straight from https://huggingface.co/docs/peft/en/package_reference/lora#peft.LoraConfig.use_rsloraLoraConfiginutils.pyOf course, a user could always do this outside TRL, but my motivation for enabling it this way was that I am working with a slightly modified version of the
sft.pyscript, so it would be nice if I could enable Rank Stabilized LoRA with a single flag.