Repository navigation
chat: accept null stream and use greedy decoding at zero temperature - #192
Merged
SearchSavior merged 1 commit intoOct 6, 2026
Merged
Conversation
Closes SearchSavior#185. temperature 0 maps to do_sample=False rather than rewriting the temperature, so the client's value is preserved on an OpenAI-compatible endpoint.
adas0x
marked this pull request as ready for review
October 6, 2026 13:27
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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 repro suggested clamping to temperature = 1.0 + do_sample = False. I kept the do_sample = False (that's the bit that actually stops OpenVINO's temperature > 0 assert) but left the temperature at 0, because on an OpenAI-compatible endpoint temperature: 0 is a perfectly valid request that just means "greedy" IIRC. and quietly rewriting the caller's value to 1 felt wrong, and it isn't needed once sampling is off.
Verified on an Arc A770 with a small model: omit stream, stream: null, and temperature: 0 all come back 200, and the model no longer gets unloaded on the temp-0 path. Pulled the temperature logic into a shared apply_temperature helper so llm and vlm don't drift.
Closes #185.