Sanath Narayan
Sanath Narayan
We haven't come across such an issue. The code has been tested and verified on Pytorch 0.3.1.
Unfortunately, this issue isn't reproducible in our systems. It runs fine for the default repo on both HMDB51 and UCF101.
Kindly check your Pytorch version and run the code with the default settings. We obtained the reported results with version 0.4.1.
Not sure what the problem is, since a few others could obtain the detection mAP around the reported numbers in the paper. We have not used any decay. It is...
Unfortunately, I don't have access to that particular code anymore.
We did the same steps as in paper (and detecionMAP code) to get the detections. Network output of tcam for a category -> find threshold from min and max values...
There is no threshold. Once the final T-CAM (t x num_class) is computed by the net, we do the top-k pooling over time and get a k x num_class vector,...
The softmax was for the mAP computation. For finding the classes present, we don't perform the softmax above. Instead take all the labels whose top-k mean is greater than 0...
There is no demo code as such. The OD setting can be run using shell script that is already shared in the repo.
There is no special preprocessing performed on the videos. The steps to extract are detailed in the experimental setup section of the paper. You can extract your own features using...