app.py

import torch
from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
import gradio as gr
import torchaudio

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

model_id = "distil-whisper/distil-large-v3"

model = AutoModelForSpeechSeq2Seq.from_pretrained(
    model_id, torch_dtype=torch_dtype, 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,
)

def speech_to_text(audio_file):
    try:
        waveform, sample_rate = torchaudio.load(audio_file)
        if waveform.size(0) > 1:
            resample = torchaudio.transforms.Resample(sample_rate, sample_rate)
            waveform = resample(waveform)
        waveform_np = waveform.numpy()
        print("pass to pipe")
        result = pipe(waveform_np[0])  
        print("result",result)
        return result["text"]
    except Exception as e:
        print(f"Error: {str(e)}")


iface = gr.Interface(fn=speech_to_text, inputs="file", outputs="text", title="Speech-to-Text")

if __name__ == "__main__":
    iface.launch()