Search for a command to run...
Co-training Transformer with Videos and Images Improves Action Recognition