DPO #30
Open
DPO #30
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
Collaborator
shahbuland
commented
Jan 26, 2024
shahbuland
commented
- Add DPO LoRA support
- Should add non-lora support as well but this is WIP for now
tmabraham
reviewed
Jan 28, 2024
Comment on lines
+116
to
+118
| def time_per_1k(n_samples : int): | ||
| total_time = timer.hit() | ||
| return total_time * 1000 / n_samples |
There was a problem hiding this comment.
this func should probably be in utils
tmabraham
reviewed
Jan 28, 2024
| if not do_lora: | ||
| ref_model = deepcopy(self.accelerator.unwrap_model(self.model).unet) | ||
| ref_model.requires_grad = False | ||
| if self.config.method.ref_mem_strategy == "half": ref_model = ref_model.half() |
There was a problem hiding this comment.
hmm is it a problem that the ref model is in half? does the original DPO do this?
There was a problem hiding this comment.
I'm not aware of how it would cause issues but I figured it was just a nice addition for memory saving purposes. It's not necessary if you match the original papers hyperparameters though.
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Outdated suggestions cannot be applied.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.