How to deploy this on google collab?

#1
by awesome8 - opened

Hi,
I want to deploy your model on google colab and use it to transcribe Malayalam. I would be grateful if you could share how to do that. I'm a newbie, so if you can share a step by step tutorial that would be really helpful.

Thanks in advance.

Hey man the code he shared on the description should work

from transformers import pipeline, WhisperProcessor

processor = WhisperProcessor.from_pretrained("thennal/whisper-medium-ml")
forced_decoder_ids = processor.get_decoder_prompt_ids(language="ml", task="translate")
asr = pipeline(
        "automatic-speech-recognition", model="thennal/whisper-medium-ml", device=0,
    )
transcription = asr("audios/real_class.mp3", chunk_length_s=30, max_new_tokens=448, return_timestamps=False,  generate_kwargs={
        "forced_decoder_ids": forced_decoder_ids, 
        "do_sample": True,
    })

print(transcription)

with open("dump.txt", "w") as f:
     f.writelines(transcription["text"])

This is a script i used for my project.

Also make sure to install the correct dependencies as mentioned at the bottom of the description page ( the newer versions break this one )

Sign up or log in to comment