Diverse Video Captioning by Adaptive Spatio-temporal Attention | Read Paper on Bytez