We advance the development of AGI and foster open source collaboration towards a smarter future.
LOVE-R1: Advancing Long Video Understanding with an Adaptive Zoom-in Mechanism via Multi-Step Reasoning