Fix custom token in train.py #246
Open
Conversation
This file contains hidden or 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
After the LLaMA model finetuning using the existing training code, I realized that the model never outputs the EOS token, which causes the generation never stop until max_new_token is reached. I tried to debug the code and found that `tokenizer.eos_token`, `tokenizer.bos_token`, and `tokenizer.unk_token` are all `'' (empty string).` Since `'' (empty string)` is not equal to `None`, the custom tokens in the training code will not be added. So I would suggest fixing using the current code changes. I have tested that after the training using the modified code, the model can output EOS token correctly.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.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.
After the LLaMA model finetuning using the existing training code, I realized that the model never outputs the EOS token, which causes the generation never stop until max_new_token is reached.
I tried to debug the code and found that
tokenizer.eos_token,tokenizer.bos_token, andtokenizer.unk_tokenare all'' (empty string).Since
'' (empty string)is not equal toNone, the custom tokens in the training code will not be added. So I would suggest fixing using the current code changes.I have tested that after the training using the modified code, the model can output EOS token correctly.