use VGGM model extract features

[wentaoxandry]

Hallo everyone,

I'm trying use VGGM.py extract video features. I have done with only one frame (1,3,224,224:1 frame, 3 channels, each frames 224224 ) feature extraction, now I want try with multi frames (5,3,224,224: 5 frames, 3 channels, each frames 224224). I'm trying put those data into cnn, but I got this error: TypeError: conv2d(): argument 'input' (position 1) must be Tensor, not Tensor. I want to ask if there is anyway use VGGM model extract multi frames features?

Thank you for your help. I greatly appreciate it.