| 1 | # Third-party notices |
| 2 | |
| 3 | ## LTX-2 |
| 4 | |
| 5 | This project is based on LTX-2 by Lightricks Ltd. Original copyright, license, |
| 6 | patent, trademark, and attribution notices are retained. See `LICENSE` for the |
| 7 | terms applicable to this repository. |
| 8 | |
| 9 | ## MSST-WebUI |
| 10 | |
| 11 | Echo 1.5 voice-memory inference depends on |
| 12 | [MSST-WebUI](https://github.com/SUC-DriverOld/MSST-WebUI), pinned by the setup |
| 13 | script to commit `43e30b860c611b516ed9b67c75a56792a67ec902`. |
| 14 | |
| 15 | MSST-WebUI is a separate work licensed under the GNU Affero General Public |
| 16 | License v3.0. Its source checkout, license, and Git history live under |
| 17 | `third_party/MSST-WebUI` after setup and are not vendored into this repository. |
| 18 | |
| 19 | The Bandit checkpoint and configuration are distributed by the MSST projects |
| 20 | and remain subject to their upstream terms: |
| 21 | |
| 22 | - Model: `model_bandit_plus_dnr_sdr_11.47.chpt` |
| 23 | - Model source: <https://huggingface.co/Sucial/MSST-WebUI> |
| 24 | - Original model project: <https://github.com/ZFTurbo/Music-Source-Separation-Training> |
| 25 |