-
Notifications
You must be signed in to change notification settings - Fork 1.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
BCOTrainer
conversational dataset support
#2107
Conversation
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. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for the nice refactor! I left a comment about using this as an opportunity to switch the model/dataset to something more standard, but happy to leave that for a follow-up PR if you wish
@@ -18,6 +18,8 @@ | |||
# Full training: | |||
python examples/scripts/bco.py \ | |||
--model_name_or_path=nnheui/stablelm-2-1_6b-sft-full \ | |||
--trust_remote_code \ | |||
--dataset_name trl-lib/ultrafeedback-gpt-3.5-turbo-helpfulness \ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I wonder if we can switch this to an unpaired version of trl-lib/ultrafeedback_binarized
? That way we have two simple datasets that should "just" work
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Good idea. How trl-lib/ultrafeedback_binarized
is generated? Which "aspect" (helpfulness, ...) is used?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I merge the PR, but we can work on it in a following PR
What does this PR do?
Part of #2071
Before submitting
Pull Request section?
to it if that's the case.
documentation guidelines.
Who can review?
Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.