YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/model-cards#model-card-metadata)

v2_inst--------------------------------------------------------------------
https://mvsep.com/quality_checker/entry/9475
Metric sdr for instrum: 17.4007 Metric si_sdr for instrum: 17.3154 Metric l1_freq for instrum: 39.9287 Metric log_wmse for instrum: 14.1031 Metric aura_stft for instrum: 16.5670 Metric aura_mrstft for instrum: 22.8199 Metric bleedless for instrum: 37.8745 Metric fullness for instrum: 38.0255 Currently, this model holds the highest aura_mrstft score on the instrumental side of the Multisong dataset.

v2_voc--------------------------------------------------------------------
https://mvsep.com/quality_checker/entry/9470
Metric sdr for vocals: 11.3957 Metric si_sdr for vocals: 10.9629 Metric l1_freq for vocals: 39.5419 Metric log_wmse for vocals: 14.3935 Metric aura_stft for vocals: 10.2444 Metric aura_mrstft for vocals: 12.8111 Metric bleedless for vocals: 34.0758 Metric fullness for vocals: 19.0952

v1--------------------------------------------------------------------
https://mvsep.com/quality_checker/entry/9303
Metric sdr for instrum: 17.2702 Metric si_sdr for instrum: 17.1860 Metric l1_freq for instrum: 39.6100 Metric log_wmse for instrum: 13.9687 Metric aura_stft for instrum: 16.7484 Metric aura_mrstft for instrum: 22.7086 Metric bleedless for instrum: 38.7761 Metric fullness for instrum: 36.9074

To use this model, replace the model file with one from this repository.

Some components in the SegmModel module were implemented based on this paper: https://arxiv.org/abs/2506.17733.

This weight is based on the following weights. Thank you, anvuew!
https://huggingface.co/anvuew/BS-RoFormer

Downloads last month: 486

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support