3D-ResNets-PyTorch
3D-ResNets-PyTorch copied to clipboard
train error--num_samples=0
"value, but got num_samples={}".format(self.num_samples)) ValueError: num_samples should be a positive integer value, but got num_samples=0
same problem! did anyone solved this one?
same problem!
it happens cause dataset is not arranged as per imagenet/activitynet/ucf101 that is data/classes/vid_name/images doing this solved this problem for me. cheers
I removed the "/" at the end of the paths that i gave to the function and it worked for me because during the code it also add "/" to the path
same problem and I am using my own datasets