Algorithm info: Algorithm: https://github.com/Audio-AGI/AudioSep
Text Prompt: vocals
Use chunks: True
Metrics:
Metric sdr for vocals: 0.3292
Metric si_sdr for vocals: -2.3949
Metric l1_freq for vocals: 17.6153
Metric log_wmse for vocals: 6.7169
Metric aura_stft for vocals: 3.6598
Metric aura_mrstft for vocals: 5.5466
Metric bleedless for vocals: 9.2866
Metric fullness for vocals: 16.5756
Date added: 2025-06-11 |
