fix polish model removed by groq, add empty guard - #735
Open
Omarinho12345 wants to merge 1 commit into
Open
Omarinho12345 wants to merge 1 commit into
Omarinho12345 wants to merge 1 commit into
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Hi, thanks for the project.
I saw issue #720. The polish step uses llama-3.3-70b-versatile but Groq removed it, so polish always fails with 404 now.
This change sets the default to qwen/qwen3-32b with POLISH_MODEL env to override. It also stops the recursion when the model returns empty content, so text is not eaten.
This covers P0-2 and part of P1-1 from the issue. The Windows GBK and TPM parts are left for later.
Please let me know if you prefer another default model. Thanks.