Whisper parameters?

#52
by Megatron17 - opened

Where can i add the whisper parameters like length_penalty, no_speech_threshold, etc

from /usr/local/lib/python3.9/dist-packages/transformers/pipelines/automatic_speech_recognition.py, the only parameters i have found are:

  chunk_length_s=None,
    stride_length_s=None,
    ignore_warning=None,
    decoder_kwargs=None,
    return_timestamps=None,
    return_language=None,
    generate_kwargs=None,
    max_new_tokens=None,

Sign up or log in to comment