Thank you very much for the great paper, code and models. I have a basic question -
When I run the feature extraction script(https://github.com/facebook/C3D/blob/master/C3D-v1.0/examples/c3d_feature_extraction/extract_C3D_feature.py), I see that the feature_layer parameter is set to 'fc6-1'.
From the paper, I understand that the output dimensions of this layer are (1,4096). However, when I want to run read_binary_blob.m, it expects a 5-dimensional tensor to be in the extracted feature file.
Where can I get the 5 dimensions, for the output of fc6-1? Or should I take the conv-layer output, before fc6-1 and use it?
Thank you very much for the great paper, code and models. I have a basic question - When I run the feature extraction script(https://github.com/facebook/C3D/blob/master/C3D-v1.0/examples/c3d_feature_extraction/extract_C3D_feature.py), I see that the feature_layer parameter is set to 'fc6-1'. From the paper, I understand that the output dimensions of this layer are (1,4096). However, when I want to run read_binary_blob.m, it expects a 5-dimensional tensor to be in the extracted feature file.
Where can I get the 5 dimensions, for the output of fc6-1? Or should I take the conv-layer output, before fc6-1 and use it?
Cheers, Kumar