No preprocessor_config.json required for run.

#1
by dorya - opened

code:
from transformers import AutoProcessor, AutoModelForSpeechSeq2Seq
processor = AutoProcessor.from_pretrained("ivrit-ai/whisper-v2-d4")
model = AutoModelForSpeechSeq2Seq.from_pretrained("ivrit-ai/whisper-v2-d4")

error:
OSError: ivrit-ai/whisper-v2-d4 does not appear to have a file named preprocessor_config.json. Checkout 'https://huggingface.co/ivrit-ai/whisper-v2-d4/tree/main' for available files.

ivrit.ai org

That's not the way to use the model.
You can look here: https://github.com/ivrit-ai/ivrit.ai/blob/master/run_whisper.py, or wait for the model instructions to be uploaded.

Generally speaking, best to use the faster-whisper-v2-d4 variant, which runs faster and works better on long segments (> 30 seconds).

benderrodriguez changed discussion status to closed

Sign up or log in to comment