Only train layout raise error?

[trouble-maker007]

assert condition, message

AssertionError: Input height (512) doesn't match model (224).

the img_feature = raw_encoder.forward_features(img) img.size(): torch.Size([128, 3, 512, 512])

[whaohan]

Hi, please resize the background images to 224*224 for layout training.

[trouble-maker007]

@whaohan I have resize background images to 224*224, but the saliency detection images are 512 * 512