BEANS-Next asks bioacoustic models to do more than name the species
-
BEANS-Next asks bioacoustic models to do more than name the species sonicfield.org
BEANS-Next tests audio-language models on acoustic description, biological categories, scene structure and learning from reference recordings, while ROOTS supplies matching training material drawn from animal sound archives.
Sign in with your torc.art membership to vote and to talk. Membership starts free.