app.py

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

# Setup device
DEVICE = "cuda:0" if torch.cuda.is_available() else "cpu"
torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
WHISPER_SAMPLE_RATE = 16000

# Load model and processor
model_id = "ixxan/whisper-small-thugy20"
processor = AutoProcessor.from_pretrained("ixxan/whisper-small-thugy20")
model = AutoModelForSpeechSeq2Seq.from_pretrained(
    model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, use_safetensors=True
)
model.to(DEVICE)

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

def preprocess_audio(audio_path: str) -> torch.Tensor:
    try:
        audio, sample_rate = torchaudio.load(audio_path)
        # Resample if necessary
        if sample_rate != WHISPER_SAMPLE_RATE:
            resampler = torchaudio.transforms.Resample(
                orig_freq=sample_rate, new_freq=WHISPER_SAMPLE_RATE
            )
            audio = resampler(audio)
        # Convert to mono
        if audio.shape[0] > 1:
            audio = torch.mean(audio, dim=0)
        return audio.squeeze()
    except Exception as e:
        raise RuntimeError(f"Error processing audio file: {str(e)}")

def transcribe(audio_path):
    try:
        if audio_path is None:
            return "No audio provided. Please record or upload an audio file."
        
        audio_input = preprocess_audio(audio_path)
        input_features = processor(
            audio_input,
            sampling_rate=WHISPER_SAMPLE_RATE,
            return_tensors="pt",
        ).input_features.to(DEVICE)
        predicted_ids = model.generate(input_features)
        transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)[0]
        return transcription
    except Exception as e:
        return f"Error transcribing audio: {str(e)}"

# Create Gradio interface
iface = gr.Interface(
    fn=transcribe,
    inputs=gr.Audio(type="filepath"),
    outputs="text",
    title="Uyghur Speech Recognition",
    description="Upload or record audio in Uyghur to get its transcription.",
    examples=[],  # You can add example audio files here if you have them
)

# Launch the app
if __name__ == "__main__":
    iface.launch()