| license: agpl-3.0 | |
| datasets: | |
| - Yusen/Sovits_ATRI | |
| ### abstruct | |
| "speech_encoder": "vec768l12". | |
| more trainning paramaters please find in ATRI_config.json | |
| sovits,diffusion,kmeans moddels included, take it as you need. | |
| ### performance | |
| a vocal only demo is in the folder. |