MoTE: Mixture of Ternary Experts for Memory-efficient Large Multimodal Models

Add code
Jun 17, 2025

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: