run whisper speech to text in the colab

whisper_runner.py

!ffmpeg -i file.m4a file.wav

import os
os.environ["PYTORCH_CUDA_ALLOC_CONF"] = "expandable_segments:True"
# torch must be loaded after the env setup
import torch
torch.cuda.empty_cache()
from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
from datasets import load_dataset


device = "cuda:0" if torch.cuda.is_available() else "cpu"
torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32

model_id = "openai/whisper-large-v3-turbo"

model = AutoModelForSpeechSeq2Seq.from_pretrained(
    model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, use_safetensors=True
)
model.to(device)

processor = AutoProcessor.from_pretrained(model_id)

pipe = pipeline(
    "automatic-speech-recognition",
    model=model,
    tokenizer=processor.tokenizer,
    feature_extractor=processor.feature_extractor,
    torch_dtype=torch_dtype,
    device=device,
)


result = pipe('file.wav', return_timestamps=True, generate_kwargs={"language": "persian"})
print(result["text"])