VISion On Request: Enhanced VLLM efficiency with sparse, dynamically selected, vision-language interactions

Add code
Mar 24, 2026

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: