Smooth knowledge distillation between different kinds of VLMs is now possible👇 @NVIDIA and @kaist_ai (it's always so interesting to read their papers) have created a universal method called GenRecal. It uses Recalibrator acting like a translator that adjusts how one model https://t.co/TvQG0ktE8e
— Ksenia Se (@Kseniase_) Jun 22, 2025
from Twitter https://twitter.com/Kseniase_
June 22, 2025 at 10:04PM
via IFTTT
No comments:
Post a Comment