Completely refactor and test YaRN finetuning #78
Merged
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
…can go a little further)
loubbrad
reviewed
Dec 13, 2023
| else: | ||
| self.rotary_emb = RotaryEmbedding(self.d_head) | ||
| cfg = model_config.yarn_config or YaRNConfig() | ||
| self.rotary_emb = YaRNScaledRotaryEmbedding( |
There was a problem hiding this comment.
Are YaRN rotary embeddings always being used? What about when we are pre-training? Does the default YaRNConfig() encapsulate the situation where YaRNScaledRotaryEmbedding is the same as normal rotary embeddings?
There was a problem hiding this comment.
YaRN embedding is always used. The default is equivalent to the usual RoPE (scale=1.0)
loubbrad
approved these changes
Dec 13, 2023
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Outdated suggestions cannot be applied.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.
The code and the config can still be improved. Still WIPRefactored the codes.use_cacheparameter.model.max_seq_lenbecause dynamic yarn can go a little longer in principle.