RuntimeError: size mismatch , when using my own 512x512 dataset about bicyclegan HOT 2 CLOSED

The current Encoder network (E_ResNet) only supports 128x128 (--which_model_netE resnet_128) and 256x256 (--which_model_netE resnet_256). But you can simiply add one line after this line.

 elif which_model_netE == 'resnet_512':
        netE = E_ResNet(input_nc, output_nc, ndf, n_blocks=6, norm_layer=norm_layer,
                        nl_layer=nl_layer, gpu_ids=gpu_ids, vaeLike=vaeLike)

from bicyclegan.

The current Encoder network (E_ResNet) only supports 128x128 (--which_model_netE resnet_128) and 256x256 (--which_model_netE resnet_256). But you can simiply add one line after this line.

 elif which_model_netE == 'resnet_512':
        netE = E_ResNet(input_nc, output_nc, ndf, n_blocks=6, norm_layer=norm_layer,
                        nl_layer=nl_layer, gpu_ids=gpu_ids, vaeLike=vaeLike)

Is this line up to date?
comment was on mar 2018

and last update was on 6 Jan 2019.
Its not working for me. I have the same issue 512x512 pixels - 800images dataset preprocess none

from bicyclegan.