Pin pyannoteAI pipelines to precision-3 - #104
Merged
Merged
Conversation
Add an explicit `model` field to PyannoteAIApi and the pyannote-api, pyannote-orchestration, and pyannote-transcription pipelines, defaulting to precision-3. Previously no model was sent, so runs tracked whatever pyannoteAI's API default was at the time.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What does this PR do?
pyannoteAI released Precision-3 on 2026-09-21. Our
PyannoteAIApiengine neversent a
modelfield, so every job ran whatever the API default happened to be.That default is Precision-2 today, flips to Precision-3 on October 1, and
Precision-2 is deprecated on October 15. Without a pin, benchmark runs a week
apart would silently use different models.
This PR makes the model an explicit config field and pins all pyannoteAI
pipelines to
precision-3.Changes
Engine (
src/openbench/engine/pyannote_engine.py)PyannoteAIApitakes amodelargument (defaultprecision-3) and sends iton every
/v1/diarizerequest.PyannoteAIModelliteral type restricting values toprecision-3,precision-2, andcommunity-1, matching the current diarize API reference.Pipelines
PyannoteApiConfig,PyannoteOrchestrationPipelineConfig, andPyannoteTranscriptionPipelineConfiggain amodelfield defaulting toprecision-3and pass it through to the engine.Aliases and configs
pyannote-api,pyannote-orchestration, andpyannote-transcriptionpinmodel: precision-3in bothpipeline_aliases.pyand their YAML configs.pyannote-apidescription, which told users to setPYANNOTE_API_KEY. The engine readsPYANNOTE_TOKEN.Usage
Default (Precision-3):
openbench-cli evaluate -p pyannote-api -d <dataset-name> -m derPrecision-2 baseline, available until 2026-10-15:
openbench-cli inference -p pyannote-api --audio-path <file> --pipeline-config '{"model": "precision-2"}'