Spaces:
Running
Running
| title: SonicMaster | |
| emoji: 💎 | |
| colorFrom: gray | |
| colorTo: blue | |
| sdk: gradio | |
| sdk_version: 6.24.0 | |
| app_file: app.py | |
| pinned: false | |
| license: other | |
| license_name: stability-ai-community | |
| license_link: https://huggingface.co/stabilityai/stable-audio-open-1.0/blob/main/LICENSE.md | |
| short_description: Text-controlled music restoration and mastering | |
| tags: | |
| - category:production | |
| - subcategory:post-processing | |
| - domain:music | |
| - input:audio | |
| - input:text | |
| - output:audio | |
| - text-guided | |
| - zerogpu | |
| Text-controlled music restoration and mastering with | |
| [SonicMaster](https://github.com/AMAAI-Lab/SonicMaster), accessible in HARP. | |
| This Space is a frontend. The model runs on the authors' Space, | |
| [amaai-lab/SonicMaster](https://huggingface.co/spaces/amaai-lab/SonicMaster), on ZeroGPU, and each | |
| request is forwarded there with the caller's `x-ip-token`, so it spends the caller's own ZeroGPU | |
| allowance. SonicMaster's own code and weights are Apache 2.0, but it decodes through Stable Audio | |
| Open's VAE, whose weights are gated by the [Stability AI Community | |
| License](https://huggingface.co/stabilityai/stable-audio-open-1.0/blob/main/LICENSE.md), so the | |
| model is not served from here. See "Calling Another Space" in HARP's deployment guidelines. | |