Octopus: History-Free Gradient Orthogonalization for Continual Learning in Multimodal Large Language Models

Add code
May 14, 2026

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: