Repository navigation
fix: use model sampling defaults for compatible requests - #15
Merged
Merged
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Some compatible services reject the hard-coded
top_pparameter with HTTP 400. The shared generation/repair request path now omitstop_pandtemperature, using the selected model's default sampling settings. The single non-streaming answer and 10,000-token limit are retained.Validation:
cargo build -p utgen --lockedandcargo test -p utgen --lockedpassed (46 tests, one opt-in live test ignored). Local HTTP regressions verify that both sampling fields are omitted rather than null in generation and repair requests; proxy-isolation tests passed. Documentation describes the model-default sampling behavior.