Hugging Face's logo

openai
/
whisper-large-v2

Automatic Speech Recognition
Transformers
PyTorch
TensorFlow
JAX
Safetensors
whisper
audio
hf-asr-leaderboard
Eval Results
Model card Files
xet
 Community
122

Using pipeline with language and task set

#54
by jonfv - opened
Discussion
jonfv

Hello!

How to use
pipe = pipeline("automatic-speech-recognition", model="openai/whisper-large-v2")
and set
model.config.forced_decoder_ids = WhisperProcessor.get_decoder_prompt_ids(language="portuguese", task="transcribe")

when I try to do this, return this error:
TypeError: WhisperProcessor.get_decoder_prompt_ids() missing 1 required positional argument: 'self'

Thx a lot!!!
jonfv

Solution:

pipe = pipeline(
    "automatic-speech-recognition",
    model="openai/whisper-large-v2",
    generate_kwargs={"language": "br", "task": "transcribe"},
    device="cpu",
    chunk_length_s=60
)
jonfv changed discussion status to closed

Sign up or log in to comment