Multiple mismatch errors

[rickkk856]

I'm trying to run your code with a 512x512 paired dataset (like in pix2pix models) so the train images are 1024x512, but every time I try

python train.py --dataroot ./datasets/08K_Rplan_Black --name 08K_RPlan_Black --model bicycle_gan --use_dropout --loadSize 512 --fineSize 512 --display_winsize 512

I GOT:

  File "C:\Users\Pichau\anaconda3\envs\py37\lib\site-packages\torch\nn\functional.py", line 1370, in linear
    ret = torch.addmm(bias, input, weight.t())
RuntimeError: size mismatch, m1: [1 x 1024], m2: [256 x 8] at C:/w/1/s/windows/pytorch/aten/src\THC/generic/THCTensorMathBlas.cu:290

Even when I specify other --which_model_netE resnet512 like junyanz sugested on the original source code, but then the code freezes and doesn't process any image.

Could you tell what networks were used to process results with the dataset 512x512 you used in the article... it seems this repo is the only one BicycleGAN repository that implemented networks that support bigger images in comparison to the 256x256...

NOTE: i've used the sabe dataset to train a model using the original source code, but it requires imgs to be 256* and the results are not satisfatory.

HELP PLZ <3