transcribe.py

import torch
from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
from datasets import load_dataset


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

model_id = "openai/whisper-large-v3"

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,
    max_new_tokens=128,
    chunk_length_s=30,
    batch_size=16,
    return_timestamps=True,
    torch_dtype=torch_dtype,
    device=device,
)


if __name__ == "__main__":
    from tkinter import filedialog, simpledialog

    language = simpledialog.askstring("Language?", "Which language (e.g. german, english, ...)")
    if language:
        filename = filedialog.askopenfilename(title="Choose file to transcribe")
        if filename:
            result = pipe(filename, generate_kwargs={"language": language})
            with filedialog.asksaveasfile(title="Choose file to save transcript to") as f:
                f.write(result["text"])