256 pixel output, not 512, despite training on --load_size 512

[turian]

I've read the FAQ and training tips carefully. Nonetheless, my 512 height images are being output as 256 at test time. Why?

Train:

!python train.py --continue_train --gpu_ids 0 --dataroot ./datasets/joseph--name joseph --model pix2pix --preprocess crop --no_flip  --load_size 512 --crop_size 256 --display_winsize 512 --batch_size 8 #--direction BtoA

Test:

!python test.py --dataroot ./datasets/joseph --model pix2pix --name joseph--display_winsize 512 --load_size 512

[synthetica3d]

How about that --crop size 256? Doesn't sound like what you want.

[improgress]

Modify the parameters in the test.py， --load_size 512，and then you will get 512pix image.

[junyanz]

See training & test tips for more details.