AudioCraft

AudioCraft is Meta's open-source deep learning library for audio generation, featuring three state-of-the-art AI models: MusicGen for text-to-music generation, AudioGen for text-to-sound effects, and EnCodec for high-fidelity neural audio compression. With over 23,000 GitHub stars, AudioCraft provides a unified framework for training and deploying generative audio models. MusicGen can generate high-quality music from text descriptions in various genres and styles, supporting both mono and stereo output. AudioGen creates environmental sounds and effects from text prompts. The library supports custom model training, fine-tuning on domain-specific data, and runs on consumer GPUs with 8GB+ VRAM. AudioCraft is the foundation behind many AI music generation services and tools.