Emmanuel Benazera

 This is to illustrate my argument: synth data has no trees/natural life basically, so the discriminator is pushing for them on fake images. Which is not what the goal...

Hi, my understanding from Pytorch/Google TPU doc is that it requires importing XLA and creating a device. So I believe the devic ``` # imports the torch_xla package import torch_xla...

` NameError: name 'y_t' is not defined`

Breaks with `--G_netG mobile_resnet_attn` ``` File "/data1/beniz/code/joliGAN/train.py", line 87, in train_gpu model = create_model(opt, rank) # create a model given opt.model and other options File "/data1/beniz/code/joliGAN/models/__init__.py", line 84, in create_model...

Hi, you can easily implement it here: https://github.com/jolibrain/joliGEN/blob/master/data/unaligned_dataset.py from looking at https://github.com/jolibrain/joliGEN/blob/master/data/unaligned_labeled_mask_dataset.py If it works you can PR it :)

Thanks. this might be due to using a new protocol buffer version. What version are you using ?

Thanks, I'll take time to review this.

My apologies I was taken elsewhere and missed this PR. Is this still relevant ? Are you using Seeks ?

/seeks already exists, right ? Now, this project is to complicated, old and ill-structured for a reboot or even maintenance. At least this is my point of view... What would...

> Maybe so, maybe not; either way, I think we should provide an official domain (a GitHub organization) to archive this repository and have an ability to join others to...