Accuracy for Office-Home dataset is quite low

[ltz0120]

Hi,

Thank you for releasing the code.

I run the code on the Office-Home dataset with the default parameters, but I find the accuracy is very low. I also try to tune the parameters, but it seems does not work. Can you provide the parameters that you use for the Office-Home dataset? Thank you.

[NaJaeMin92]

Hi @ltz0120

Here is my recipe for Office-Home dataset.

data_transforms = {
'train': transforms.Compose([
    transforms.Resize(256),
    transforms.RandomCrop(224),
    transforms.RandomHorizontalFlip(),
    transforms.ToTensor(),
    transforms.Normalize(
        mean=[0.485, 0.456, 0.406], 
        std=[0.229, 0.224, 0.225])
]),
'test': transforms.Compose([
    transforms.Resize(224),
    transforms.CenterCrop(224),
    transforms.ToTensor(),
    transforms.Normalize(
        mean=[0.485, 0.456, 0.406], 
        std=[0.229, 0.224, 0.225])
])}

I suggest you to tune the hyper-parameters such according to your baseline model.

[ltz0120]

The method works well in Office 31. For the Office-Home dataset, I tried to tune the parameters, but the results are still quite bad. Could you please share your pre-trained model for Office-Home dataset? Thank you.

[Arsiuuu]

@ltz0120
Hi, would you like to share the pretrained models or codes : )