2 ms·Co-Training Transformer with Videos and Images Improves Action Recognition2 points by RafelMri 5y ago