Extract c3d features, given video


How to extract c3d features in Pytorch?

Could you explain, what c3d features are?

its a 3d convolution model for videos https://github.com/facebookarchive/C3D

Would you like to extract the output activations from the 3d conv layers?
If so, you could use forward_hooks as described here

1 Like

Hi Rafael: could you please give me a hand to extract c3d features, given video?