Share your thoughts, 1 month free Claude Pro on usSee more

Home/Benchmarks

Speech Separation on Libri3Mix simulated (test)

15SI-SNRi

USE-B

Updated 1mo ago

Evaluation Results

Method	Links
USE-B 2025.12		15
USE-B 2025.12		15

SOTA Paper

USE-B

USE: A Unified Model for Universal Sound Separation and Extraction

Dataset

LibriMix

Follow for update

@wizwand_team Discord

Related Benchmarks

Target Sound Extraction on AudioSet 2Mix Target Sound Extraction on AudioSet 3Mix Universal Sound Separation on FUSS (Dry) 2-mix (test)Universal Sound Separation on FUSS Dry 3-mix (test)Universal Sound Separation on FUSS Dry 4-mix (test)

© 2026 wizwand

Blog Contact Changelog Swarm

Privacy Policy Terms of Service FAQs Swarm Docs