|
Download README.md from khaos-dev/README: direct link, hf CLI and curl.
- Browser
- Download file 1.26 kB
-
https://huggingface.co/spaces/khaos-dev/README/resolve/main/README.md
- Command line
-
hf download hf://spaces/khaos-dev/README/README.md
-
curl -L -o README.md https://huggingface.co/spaces/khaos-dev/README/resolve/main/README.md
1.26 kB
| title: README | |
| emoji: 🏢 | |
| colorFrom: indigo | |
| colorTo: blue | |
| sdk: static | |
| pinned: false | |
| # Khaos | |
| Artist-owned AI models for musicians. | |
| Khaos is a free-to-use toolkit that lets a musician fine-tune an audio generation model on their own catalogue, on their own computer, and own the resulting model outright. Training data is consented because it is the artist's own work, with writers, publishers and identifiers attached to every track. Nothing leaves the machine unless the artist chooses. | |
| Every output carries a provenance record and a per-output attribution estimate: how much each training track shaped it, and who holds the rights in that track. | |
| **Research** | |
| - Measured Influence: pre-registered validation and calibration of per-output training-data attribution for consented music adapters. Working paper, v3. https://doi.org/10.5281/zenodo.23618202 | |
| **Code and standards** | |
| - Attribution schemas and verification: https://github.com/khaos-foundation/Khaos_Attribution | |
| **Models** | |
| Khaos does not publish artists' models here. Each model belongs to the artist who trained it, and is shared only if they choose to. | |
| Founded by composer Chris Green. On Digital Catapult's Applied AI Growth Sprints, funded by Innovate UK. | |
| https://www.khaos.dev |