app.py

import spaces
import gradio as gr
# Use a pipeline as a high-level helper
import torch
from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
from datasets import load_dataset

@spaces.GPU(duration=120)
def transcribe_audio(audio):
    if audio is None:
        return "Please upload an audio file."

    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", "alvanlii/whisper-small-cantonese"]

    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=25,
        batch_size=16,
        torch_dtype=torch_dtype,
        device=device,
    )

    result = pipe(audio)
    return result["text"]


demo = gr.Interface(fn=transcribe_audio,
                    inputs=[gr.Audio(sources="upload", type="filepath"), gr.Dropdown(choices=["openai/whisper-large-v3", "alvanlii/whisper-small-cantonese"])],
                    outputs="text")
demo.launch()