Train the model using CelebAMask-HQ dataset: Just run the train script: $ CUDA_VISIBLE_DEVICES=0,1 python -m torch.distributed.launch --nproc_per_node=2 train.py If you do not wish to train the model, ...