app.py

import gradio as gr
import torchaudio
from audiocraft.models import MusicGen
from audiocraft.data.audio import audio_write
import spaces

@spaces.GPU(duration=120)
@spaces.GPU(duration=120) 
def generate_music(description, melody_audio):

    model = MusicGen.get_pretrained('nateraw/musicgen-songstarter-v0.2')
    model.set_generation_params(duration=8)  

    if description:
        description = [description]
        if melody_audio:
            melody, sr = torchaudio.load(melody_audio)
            wav = model.generate_with_chroma(description, melody[None], sr) 
        else:
            wav = model.generate(description)  
    else:
        wav = model.generate_unconditional(1)  

    output_path = 'output.wav'
    audio_write(output_path, wav[0].cpu(), model.sample_rate, strategy="loudness", loudness_compressor=True)
    
    return output_path


description = gr.Textbox(label="Description", placeholder="acoustic, guitar, melody, trap, d minor, 90 bpm")
melody_audio = gr.Audio(label="Melody Audio (optional)", type="filepath")
output_path = gr.Audio(label="Generated Music", type="filepath")

gr.Interface(
    fn=generate_music,
    inputs=[description, melody_audio],
    outputs=output_audio,
    title="MusicGen Demo",
    description="Generate music using the MusicGen model.",
    examples=[
        ["trap, synthesizer, songstarters, dark, G# minor, 140 bpm", "./assets/kalhonaho.mp3"],
        ["upbeat, electronic, synth, dance, 120 bpm", None]
    ]
).launch()