tool-call: fix non-tool-calling grammar crashes w/ Qwen / Hermes 2 templates
#12900
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
…re aren't tools (or when there's a schema)
ochafik
changed the title
tool-call: don't call common_chat_params_init_hermes_2_pro when there's no toolstool-call: fix non-tool-calling grammar crashes w/ Qwen / Hermes 2 templates
ngxson
approved these changes
Apr 11, 2025
|
I encountered this problem while evaluating Gemma 3 it 12b over the MMLU test set, this PR seems to have fixed the issue. Thank you! |
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Outdated suggestions cannot be applied.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.
Adds missing fence of tool call support logic for hermes 2 / qwen format (+ missing tests that ensure the tool call logic is only triggered when expected; for a few models it's always, e.g. R1, CR7B, Functionary v3.2)
This fixes the empty grammar stack crash reported in #12379
cc/ @edmcman