Kensho Hara

I removed nn.DataParallel from the latest file, and if no_cuda == True, nn.DataParallel does not be used. So I think there are other problems. But I currently confirmed the problem...

I'm sorry for the long delay. I executed the code using the same options as yours. But I can execute correctly in my environment. I'm not sure what is the...

I calculated the mean and std using a randomly selected small subset of Kinetics. In addition, I calculated the means for each frame and averaged them for the over all...

Yes, it's OK. In addition, you have to change sample_duration in main.py to use 64 frame inputs. https://github.com/kenshohara/video-classification-3d-cnn-pytorch/blob/master/main.py#L19

See classify.py. This code calculates scores of each segment. You can use the score and theshold it. https://github.com/kenshohara/video-classification-3d-cnn-pytorch/blob/master/classify.py