Discussion

[Deepa985]

Hi, thank you for uploading the codes and datasets. Now I'm doing this project, but I'm facing some errors. Could you please help me in resolving the error?

[liuyiqing2018]

Thank you for your interest. It seems that you have used your own model, of which the structure may differ from the given one. So try to use the given pre-trained model first and see whether it works:
trained_model = torch.load('resnet18_0.9326.model')
and then you can try your own model. Note that your model should be the same structure as 'class Net', which has a few differences from the original ResNet18:

1. Our model has no average pooling layer;
2. The length of the input feature in the FC layer is 51222 to keep consistent with the input size (N, 3, 64, 64);
3. The length of the output feature in the FC layer is 3 since the task is a three-class classification problem.