Fetching the paper…

MA-LMM: Memory-Augmented Large Multimodal Model for Long-Term Video Understanding · Around