Using RandStainNA with Resnet18

[SaraAlthubaiti]

Hello all,

Thanks for sharing this nice work. I tried to use RandStainNA on whole slide images and as a model I used resnet18 (pre-trained) but it gave:

RuntimeError: Given groups=1, weight of size [64, 3, 7, 7], expected input[16, 224, 224, 3] to have 3 channels, but got 224 channels instead

And this is what I used:

transforms_list = [
        transforms.Resize((224, 224)),
        RandStainNA(yaml_file='/kaggle/input/kaggleinputdatasetmine/CRC_LAB_randomTrue_n0.yaml', std_hyper=-0.3, probability=1.0,distribution='normal', is_train=True)
    ]
    
transform = transforms.Compose(transforms_list)

# I used 16 as a batch size

And for the model, I changed last layer based on the number of classes I have (3 classes):

# Load pre-trained ResNet-18 model
model = models.resnet18(pretrained=True)

# Replace the last fully connected layer to match the number of classes
num_ftrs = model.fc.in_features
model.fc = nn.Linear(num_ftrs, 3)

model = model.to(device)

And the error happen when I call the model:

---> 73 outputs = model(image)

Can you help me to solve that, please?