Expose max_new_tokens run option#74
Open
vladkalinichencko wants to merge 1 commit into
Open
Conversation
Contributor
|
Thanks I will take a look tomorrow |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This exposes a per-run max_new_tokens option through the C API and the language bindings.
Some autoregressive ASR families can accept the audio input but still hit the fixed generation budget before emitting EOS. That currently returns OUTPUT_TRUNCATED with a partial transcript, which is correct, but callers have no way to request a larger output budget for longer transcripts.
The default behavior stays unchanged: -1/0 uses the existing family defaults. Qwen3-ASR and Canary just now have a positive max_new_tokens value for single and batch runs, bounded by the model decoder context.