amir0907
/

student-v5

speech-emotion-recognition

speech-to-speech-translation

Model card Files Files and versions

student-v5 / code /precompute_v4.py

Commit History

Oct 28 2 days ago

18:47
Surgical emotion module v5: 6.25 Hz + weak supervision from RAVDESS/CREMA-D intensity labels (pairwise ranking) - breaks the label-free proxy ceiling

amir0907

Verified signature
d29b895