tests/pipelines/test_pipelines_automatic_speech_recognition.py · 9c8979e35fc4d0f991214368b58054573b8747ce · chenpangpang / transformers

Word-level timestamps broken for short-form audio (#30325) · 9c8979e3

Kamil Akesbi authored May 07, 2024



* force chunk_length_s in AutomaticSpeechRecognitionPipeline

* compute num_frames even when stride is None

* add slow tests

* fix test

* Update src/transformers/pipelines/automatic_speech_recognition.py
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* Update tests/pipelines/test_pipelines_automatic_speech_recognition.py
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* add input validation

* fixup

* small fix

---------
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

9c8979e3

test_pipelines_automatic_speech_recognition.py 76.7 KB

Replace test_pipelines_automatic_speech_recognition.py