Algorithm info: Trained with https://github.com/ZFTurbo/Music-Source-Separation-Training
Datasets: MUSDB18HQ and custom dataset (419 tracks from https://multitracksearch.cambridge-mt.com/ms-mtk-search-ads.htm)
Model name: model_mdx23c_ep_2_sdr_13.4715.ckpt (2 epoch because it hasn't retained a better model than this one in all its training.)
Date added: 2023-12-31 |
