support R-DPO #25
Open
support R-DPO #25
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
Comment on lines
+100
to
+101
| chosen_lens: int | None, | ||
| rejected_lens: int | None, |
There was a problem hiding this comment.
не очень нравится идея добавлять штуку в общий абстрактный класс которая только в одном лоссе используется
хочется как-то внутри более изолировано и чисто внутри лосса это считать мб
There was a problem hiding this comment.
Согласен, что это не оч красиво. Делал по примеру precomputed_margins, ща мб попробую переписать без этого
|
|
||
| unscaled = self.beta * logits | ||
|
|
||
| if chosen_lens is not None and rejected_lens is not None: |
There was a problem hiding this comment.
обязательно же не None должно быть
| unscaled = self.beta * logits | ||
|
|
||
| if chosen_lens is not None and rejected_lens is not None: | ||
| unscaled -= self.alpha * rejected_lens - self.alpha * chosen_lens |
There was a problem hiding this comment.
мне визуально так приятней и проще, но и так как у тебя ок
Suggested change
| unscaled -= self.alpha * rejected_lens - self.alpha * chosen_lens | |
| unscaled += self.alpha * (chosen_lens - rejected_lens) |
| @@ -85,6 +86,12 @@ class APODownLossSettings(DPOLossSettings): | |||
| loss_type: Literal[DPOLossesType.APO_DOWN] | |||
|
|
|||
|
|
|||
| class RDPOLossSettings(DPOLossSettings): | |||
| loss_type: Literal[DPOLossesType.RDPO] | |||
| beta: float = 0.1 | |||
There was a problem hiding this comment.
Suggested change
| beta: float = 0.1 |
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Outdated suggestions cannot be applied.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.
No description provided.