Nico Zhou
Nico Zhou
it is normal as the facades datasets is transfering B to A. Or You can interpret imgB = input , and imgA = output
i think you can change the if statement to enable loading from epoch 0.
Well, I think you can use BCELoss() too, but you need to add sigmoid layer in the discriminator to let the output in range 0 to 1
Oh i see, in my case i happened to have more positive samples, so i have an error when training with it. So, do you think i must change the...
will there be any step by step method to do this? Or do I just have to provide the human dataset in the coco format?
hi regarding this, how to do augmentation in patchcore. Seems like patchcore need heavy augmentation right?
I think you need to write it by yourself, here is the steps: 1. Prepare your testset_videos and run the two-GAN inference. 2. Obtain testset_features from two-GAN (e.g. you can...