How to train on large images?

[universewill]

I change the load_size and crop_size to 512 in train_facades.sh, and get error.
I want to train on large image size, how to do that?

Traceback (most recent call last):
  File "./train.py", line 48, in <module>
    model.optimize_parameters()   # calculate loss functions, get gradients, update network weights
  File "gxl/BicycleGAN/models/bicycle_gan_model.py", line 209, in optimize_parameters
    self.forward()
  File "gxl/BicycleGAN/models/bicycle_gan_model.py", line 106, in forward
    self.z_encoded, self.mu, self.logvar = self.encode(self.real_B_encoded)
  File "gxl/BicycleGAN/models/bicycle_gan_model.py", line 82, in encode
    mu, logvar = self.netE.forward(input_image)
  File "miniconda3/lib/python3.7/site-packages/torch/nn/parallel/data_parallel.py", line 159, in forward
    return self.module(*inputs[0], **kwargs[0])
  File "miniconda3/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
    result = self.forward(*input, **kwargs)
  File "gxl/BicycleGAN/models/networks.py", line 647, in forward
    output = self.fc(conv_flat)
  File "miniconda3/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
    result = self.forward(*input, **kwargs)
  File "miniconda3/lib/python3.7/site-packages/torch/nn/modules/container.py", line 117, in forward
    input = module(input)
  File "miniconda3/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
    result = self.forward(*input, **kwargs)
  File "miniconda3/lib/python3.7/site-packages/torch/nn/modules/linear.py", line 93, in forward
    return F.linear(input, self.weight, self.bias)
  File "miniconda3/lib/python3.7/site-packages/torch/nn/functional.py", line 1690, in linear
    ret = torch.addmm(bias, input, weight.t())
RuntimeError: mat1 dim 1 must match mat2 dim 0