AI Music Remixer

AI-powered intelligent music remixer with optional neural transitions.

Features

• Intelligent audio remixing using Spleeter for stem separation
• Optional PyTorch-based neural transition model for smoother transitions
• Beat-synchronized remixing with automatic tempo detection
• Command-line interface for easy usage
• Safe fallback to deterministic smoothing if neural model is unavailable

Installation

pip install ai_music_remixer

For GPU support (optional):

pip install ai_music_remixer[gpu]

Usage

Command Line

Basic remix:

ai-remix input.mp3 -o output.wav

Beat-synchronized remix:

ai-remix input.mp3 --beats -o output.wav

With custom BPM:

ai-remix input.mp3 --beats --bpm 120 -o output.wav

Python API

from ai_music_remixer import ai_remix, ai_remix_beatsync

# Basic remix
ai_remix("input.mp3", "output.wav")

# Beat-synchronized remix
ai_remix_beatsync("input.mp3", "output.wav", beats_per_minute=120)

Neural Transition Model

The package includes an optional PyTorch-based neural transition model. If you place a trained checkpoint file at models/transition_ckpt.pt, the remixer will automatically use it for smoother transitions. Without a checkpoint, the package falls back to deterministic Gaussian smoothing.

To train a model (requires PyTorch):

python -m ai_music_remixer.neural.train_transition

Requirements

• Python >= 3.8
• pydub
• librosa
• soundfile
• spleeter
• numpy
• scipy

Optional:

• torch (for neural transitions)

License

MIT