Closed DRJ2016 closed 3 years ago
Sorry to disturb you. Really excellent work! May I ask what Conv models to use for extracting frame features from raw videos?
For Charades, we use VGG features obtained from here. For ActivityNet, we use C3D features obtained from here. For TACoS, we use use C3D features obtained from here. You can refer to their websites for more details about extraction.
Really thank you for your reply.
Sorry to disturb you. Really excellent work! May I ask what Conv models to use for extracting frame features from raw videos?