Backup of music source separation models
Preservation copy of the models used by the Colab Inference notebook that today only survive as third-party re-uploads (their original accounts are gone). Files are byte-for-byte copies; nothing was retrained or modified.
None of these models are ours. They were trained and released by the people credited below; all credit and all rights are theirs. The goal of this project is preservation and keeping a free tool working, not taking anyone's work. Models are always downloaded from their author's own link first; the backup is used only when that link is gone. If you are the author of a model and want it removed, renamed or credited differently, open an issue or a discussion and it will be done.
Ninguno de estos modelos es nuestro. Los entrenaron y publicaron las personas acreditadas abajo; todo el crédito y todos los derechos son suyos. El objetivo de este proyecto es la conservación y mantener funcionando una herramienta gratuita, no apropiarse del trabajo de nadie. Los modelos se descargan siempre primero del enlace de su autor; el respaldo solo se usa cuando ese enlace ya no existe. Si eres autor de un modelo y quieres que se retire, se renombre o se acredite de otra forma, abre un issue o una discusión y se hará.
| Model | Author | File | Size | sha256 | Copied from |
|---|---|---|---|---|---|
| VOCALS-BS-RoformerLargev1 (by unwa) | unwa | BS-Roformer_LargeV1.ckpt |
706 MiB | 3d1334565bb9e4c0… |
https://huggingface.co/Sucial/MSST-WebUI |
| VOCALS-BS-RoformerLargev1 (by unwa) | unwa | BS-Roformer_LargeV1.yaml |
0 MiB | a88e3df7d62cd352… |
https://huggingface.co/baicai1145/pymss |
| VOCALS-BS-Roformer_1297 (by viperx) | viperx | model_bs_roformer_ep_317_sdr_12.9755.yaml |
0 MiB | 42e5635ceab7287b… |
https://raw.githubusercontent.com/ZFTurbo/Music-Source-Separation-Training |
| VOCALS-BS-Roformer_1297 (by viperx) | viperx | model_bs_roformer_ep_317_sdr_12.9755.ckpt |
610 MiB | 5b84f37e8d444c8c… |
https://github.com/TRvlvr/model_repo |
| VOCALS-BS-Roformer_1296 (by viperx) | viperx | model_bs_roformer_ep_368_sdr_12.9628.ckpt |
610 MiB | f6c94864adfb73bb… |
https://github.com/TRvlvr/model_repo |
| VOCALS-BS-Roformer_1296 (by viperx) | viperx | model_bs_roformer_ep_368_sdr_12.9628.yaml |
0 MiB | aea599b3f9bd4892… |
https://raw.githubusercontent.com/TRvlvr/application_data |
| KARAOKE-MelBand-Roformer (by aufr33 & viperx) | aufr33 & viperx | mel_band_roformer_karaoke_aufr33_viperx_sdr_10.1956.ckpt |
871 MiB | 1de20d459332fe88… |
https://huggingface.co/Eddycrack864/audio-separator-models |
| KARAOKE-MelBand-Roformer (by aufr33 & viperx) | aufr33 & viperx | config_mel_band_roformer_karaoke.yaml |
0 MiB | 1ad4ebd15653108e… |
https://github.com/deton24/Colab-for-new-MDX_UVR_models |
| OTHER-BS-Roformer_1053 (by viperx) | viperx | model_bs_roformer_ep_937_sdr_10.5309.ckpt |
375 MiB | a2e825a03bc908cb… |
https://github.com/TRvlvr/model_repo |
| OTHER-BS-Roformer_1053 (by viperx) | viperx | model_bs_roformer_ep_937_sdr_10.5309.yaml |
0 MiB | 302b6cee54adf397… |
https://raw.githubusercontent.com/TRvlvr/application_data |
| 6STEMS-BS-Roformer-SW | unknown ("shared weights") | BS-Rofo-SW-Fixed.yaml |
0 MiB | f9fada9f94e5ba2d… |
https://huggingface.co/enerjazzer/BS-ROFO-SW-Fixed |
| 6STEMS-BS-Roformer-SW | unknown ("shared weights") | BS-Rofo-SW-Fixed.ckpt |
667 MiB | 24e7d35ee9c64415… |
https://huggingface.co/enerjazzer/BS-ROFO-SW-Fixed |
| CINEMATIC-BandIt_v2 multi (by kwatcharasupat) | kwatcharasupat | config_dnr_bandit_v2_mus64.yaml |
0 MiB | b7062645d3fcd996… |
https://raw.githubusercontent.com/ZFTurbo/Music-Source-Separation-Training |
| CINEMATIC-BandIt_v2 multi (by kwatcharasupat) | kwatcharasupat | checkpoint-multi_fixed.ckpt |
142 MiB | 20bcd513dc7eb054… |
https://huggingface.co/Politrees/UVR_resources |
| DRUMSEP-MDX23C_DrumSep_5stem_new (by jarredou) | jarredou | drumsep_5stems_mdx23c_jarredou.ckpt |
417 MiB | 1f8e636fb674b88a… |
https://huggingface.co/xavriley/source_separation_mirror |
| DRUMSEP-MDX23C_DrumSep_5stem_new (by jarredou) | jarredou | config_mdx23c_drumsep2025.yaml |
0 MiB | d2962f43d6682e8e… |
https://huggingface.co/xavriley/source_separation_mirror |
| DRUMSEP-MDX23C_DrumSep_6stem (by aufr33 & jarredou) | aufr33 & jarredou | aufr33-jarredou_DrumSep_model_mdx23c_ep_141_sdr_10.8059.ckpt |
417 MiB | d2a4aa53eb584d21… |
https://huggingface.co/Sucial/MSST-WebUI |
| DRUMSEP-MDX23C_DrumSep_6stem (by aufr33 & jarredou) | aufr33 & jarredou | config_drumsep_mdx23c.yaml |
0 MiB | 17d1649a227f8411… |
https://huggingface.co/Politrees/UVR_resources |
| DE-REVERB-MDX23C (by aufr33 & jarredou) | aufr33 & jarredou | dereverb_mdx23c_sdr_6.9096.ckpt |
427 MiB | eae2471b707758d7… |
https://huggingface.co/Sucial/MSST-WebUI |
| DE-REVERB-MDX23C (by aufr33 & jarredou) | aufr33 & jarredou | config_dereverb_mdx23c.yaml |
0 MiB | a0cf11216913ab89… |
https://huggingface.co/Eddycrack864/audio-separator-models |
| DEBLEED-MelBand-Roformer (by unwa/97chris) | unwa/97chris | mel_band_roformer_bleed_suppressor_v1.ckpt |
871 MiB | a9a9d10faa7f8997… |
https://huggingface.co/Eddycrack864/audio-separator-models |
| DEBLEED-MelBand-Roformer (by unwa/97chris) | unwa/97chris | config_mel_band_roformer_bleed_suppressor_v1.yaml |
0 MiB | bca5755de9946ded… |
https://huggingface.co/Eddycrack864/audio-separator-models |
| VOCALS-Male Female-BS-RoFormer Male Female Beta 7_2889 (by aufr33) | aufr33 | bs_roformer_male_female_by_aufr33_sdr_7.2889.ckpt |
503 MiB | 3cf11736d1b42a11… |
https://huggingface.co/RareSirMix/AIModelRehosting |
| VOCALS-Male Female-BS-RoFormer Male Female Beta 7_2889 (by aufr33) | aufr33 | config_chorus_male_female_bs_roformer.yaml |
0 MiB | e952a90412ed3e9d… |
https://huggingface.co/Sucial/Chorus_Male_Female_BS_Roformer |
| PHANTOM-CENTER-HTDemucs (by wesleyr36) | wesleyr36 | model_htdemucs_ep_21_sdr_13.6970.ckpt |
160 MiB | b9970dca36a15c0d… |
https://huggingface.co/baicai1145/pymss |
| PHANTOM-CENTER-HTDemucs (by wesleyr36) | wesleyr36 | model_htdemucs_ep_21_sdr_13.6970.yaml |
0 MiB | cdc9f166e42ef308… |
https://huggingface.co/baicai1145/pymss |
Licenses: each model keeps whatever terms its author set. Where the author stated none, none is claimed here.