Window Token Concatenation for Efficient Visual Large Language Models

VLM
Re-assign community