app.py

import streamlit as st
import torch
import torchaudio
import os
import numpy as np
import base64
from audiocraft.models import MusicGen

# Before
batch_size = 64

# After
batch_size = 32 
torch.cuda.empty_cache()

genres = ["Pop", "Rock", "Jazz", "Electronic", "Hip-Hop", "Classical", "Lofi", "Chillpop"]

@st.cache_resource()
def load_model():
    model = MusicGen.get_pretrained('facebook/musicgen-small')
    return model

def generate_music_tensors(description, duration: int):
    model = load_model()

    model.set_generation_params(
        use_sampling=True,
        top_k=250,
        duration=duration
    )

    with st.spinner("Generating Music..."):
        output = model.generate(
            descriptions=description,
            progress=True,
            return_tokens=True
        )

    st.success("Music Generation Complete!")
    return output


def save_audio(samples: torch.Tensor):
    sample_rate = 30000
    save_path = "audio_output" 
    assert samples.dim() == 2 or samples.dim() == 3

    samples = samples.detach().cpu()
    if samples.dim() == 2:
        samples = samples[None, ...]

    for idx, audio in enumerate(samples):
        audio_path = os.path.join(save_path, f"audio_{idx}.wav")
        torchaudio.save(audio_path, audio, sample_rate)
        return audio_path

def get_binary_file_downloader_html(bin_file, file_label='File'):
    with open(bin_file, 'rb') as f:
        data = f.read()
    bin_str = base64.b64encode(data).decode()
    href = f'<a href="data:application/octet-stream;base64,{bin_str}" download="{os.path.basename(bin_file)}">Download {file_label}</a>'
    return href

st.set_page_config(
    page_icon= "musical_note",
    page_title= "Music Gen"
)

def main():
    st.title("🎧 AI Composer Small-Model 🎧")

    st.subheader("Craft your perfect melody!")
    bpm = st.number_input("Enter Speed in BPM", min_value=60)

    text_area = st.text_area('Ex : 80s rock song with guitar and drums')
    st.text('')
    # Dropdown for genres
    selected_genre = st.selectbox("Select Genre", genres)

    st.subheader("2. Select time duration (In Seconds)")
    time_slider = st.slider("Select time duration (In Seconds)", 0, 30, 10)

    if st.button('Let\'s Generate 🎶'):
        st.text('\n\n')
        st.subheader("Generated Music")

        description = f"{text_area} {selected_genre} {bpm} BPM"
        
        # Clear CUDA memory cache before generating music
        torch.cuda.empty_cache()

        music_tensors = generate_music_tensors(description, time_slider)

        # Only play the full audio for index 0
        idx = 0
        music_tensor = music_tensors[idx]
        audio_filepath = save_audio(music_tensor)
        audio_file = open(audio_filepath, 'rb')
        audio_bytes = audio_file.read()

        # Play the full audio
        st.audio(audio_bytes, format='audio/wav')
        st.markdown(get_binary_file_downloader_html(audio_filepath, f'Audio_{idx}'), unsafe_allow_html=True)


if __name__ == "__main__":
    main()