Musique IA
torch-log-wmse
An audio quality metric and loss for training cleaner models.
★ 50 étoilesApache-2.0
Voir le guide →Stable Audio Tools est un projet de musique IA open source, créé par Stability-AI (Stability-AI/stable-audio-tools sur GitHub). Il compte 3 892 étoiles sur GitHub et 488 forks, il est publié sous la licence MIT et a été mis à jour pour la dernière fois sur GitHub le 24 octobre 2026.
Stable Audio Tools — training and inference code for Stability's audio generation models, including Stable Audio Open, MIT
Clone, then sync for inference only — or add the train and UI extras for the full toolkit.
For Stable Audio Open 1.0, accept the model terms on its Hugging Face page before downloading.
Launch the interface pointed at the pretrained model and generate audio from prompts.
Dataset training, inference and a Gradio UI live together, instead of three separate projects.
Generate music and sound with the open model, on your own hardware.
PyTorch 2.5+ with Flash Attention and Flex Attention for faster generation.
Prefer pip over uv? The package installs directly with its train extra.
The README install route with uv.
git clone https://github.com/Stability-AI/stable-audio-tools.git
cd stable-audio-tools
uv syncTraining plus the Gradio UI in one sync.
uv sync --extra train --extra uiPointed at Stable Audio Open 1.0 after accepting its Hugging Face terms.
uv run python run_gradio.py --pretrained-name stabilityai/stable-audio-open-1.0Stability AI, as the training and inference code for its audio generation models.
The code is MIT licensed; the pretrained Stable Audio Open model has its own Hugging Face terms to accept before download.
Python 3.10, PyTorch 2.5+ (for Flash/Flex Attention), and uv for the recommended install.
Yes — training is a first-class part of the repo, with dataset configs and a train extra in the install.
Clone, uv sync, accept the Stable Audio Open terms on Hugging Face, and run the Gradio interface.
An audio quality metric and loss for training cleaner models.
The hub for audio AI research: papers, open models, benchmarks & datasets across audio LLMs, speech recognition, TTS, music & audio generation.
Loss functions for training audio source separation models.
A Python inference package for music source separation models.
A curated MIDI dataset built for training precise music AI models.
Research code for instruction-based text-to-music editing of music models.
Les autres projets de la catégorie Musique IA avec le plus d'étoiles du catalogue LibreAI — la sélection réaliste d'alternatives si Stable Audio Tools ne vous convient pas.
YuE2: frontier music generation with symbolic planning, zero-shot covers, and agentic music editing.
🎵 The Ultimate Open Source Suno Alternative - Professional UI for ACE-Step 1.5 AI Music Generation. Free, local, unlimited. Stop paying for Suno!
ACE-Step — the open foundation model for music generation: full songs with vocals and lyrics in 19 languages, up to 4 minutes in about 20 seconds, Apache-2.0
Official DiffSinger code: diffusion-based singing voice synthesis and text-to-speech (AAAI 2022)
Use API to call the music generation AI of suno.ai, and easily integrate it into agents like GPTs.
Projets de la catégorie Musique IA proches dans le catalogue, les plus proches en étoiles de Stable Audio Tools — continuez à explorer de part et d'autre.
Official DiffSinger code: diffusion-based singing voice synthesis and text-to-speech (AAAI 2022)
Use API to call the music generation AI of suno.ai, and easily integrate it into agents like GPTs.
ACE-Step — the open foundation model for music generation: full songs with vocals and lyrics in 19 languages, up to 4 minutes in about 20 seconds, Apache-2.0
DiffRhythm — blazingly fast, embarrassingly simple full-length song generation with latent diffusion: complete songs with vocals from lyrics and a style prompt, Apache-2.0
🎵 The Ultimate Open Source Suno Alternative - Professional UI for ACE-Step 1.5 AI Music Generation. Free, local, unlimited. Stop paying for Suno!
A WebUI app for Music-Source-Separation-Training and we bundled UVR together!
Gratuit et ouvert : récupérez le code source, mettez une étoile au dépôt et commencez à créer.