Hey, Pytorch’s (torchvision) Resnet 50 accuracy is declared to be 76.15.
But when I’m using the training script from PyTorch’s repo, which is mentioned in the official torchvision website(https://pytorch.org/docs/stable/torchvision/models.html#classification):
https://github.com/pytorch/examples/blob/master/imagenet/main.py
and the Resnet50 from torchvision:
[https://github.com/pytorch/vision/blob/master/torchvision/models/resnet.py]
When training it, after one epoch I’m getting an accuracy of 76.6, how can it be? isn’t the models fully trained?
Thanks!