app.py

import torch
import gradio as gr
from sgmse.model import ScoreModel

# Load your model
model_path = "https://huggingface.co/sp-uhh/speech-enhancement-sgmse/resolve/main/pretrained_checkpoints/speech_enhancement/train_vb_29nqe0uh_epoch%3D115.ckpt"
model = SGMSE()  # Initialize your model class
model.load_state_dict(torch.load(model_path, map_location=torch.device('cpu')))
model.eval()  # Set the model to evaluation mode

def enhance_audio(input_audio):
    import torchaudio

    # Load the input audio file
    waveform, sample_rate = torchaudio.load(input_audio)

    with torch.no_grad():
        enhanced_waveform = model(waveform)

    output_path = "enhanced_audio.wav"
    torchaudio.save(output_path, enhanced_waveform.cpu(), sample_rate)
    return output_path

# Create the Gradio interface
iface = gr.Interface(
    fn=enhance_audio,
    inputs=gr.Audio(source="upload", type="filepath"),
    outputs=gr.Audio(type="file"),
    title="Speech Enhancement Model",
    description="Upload a noisy audio file to enhance it using the SGMSE model."
)

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