Facebook AI Introduces ‘Anticipative Video Transformer’ (AVT): An End-To-End Attention-Based Model To Anticipate Future Actions In Videos
This model is able to predict the next line of action in a video sequence base on previously learnt video.
Paper: https://arxiv.org/abs/2106.02036
Code: https://github.com/facebookresearch/AVT
Project: https://facebookresearch.github.io/AVT/
