Fetching the paper…

B-VLLM: A Vision Large Language Model with Balanced Spatio-Temporal Tokens · Around