Hello. I have some troubles when using the pretrained model from torchvision.model. I evaluate the ResNet18 and ResNet34 on ImageNet validation dataset, but I didn’t get the accuracy=0 after 1 iter batch.
i’m not sure i understand. Did you get accuracy=0, or did you not?
I get the accuracy =0. And I test these models using code in Pytorch examples.
please check your image preprocessing methods.