KeyError : 'D_ema' : error when training with --resume options

[LeoAlethio]

Hi, thanks for your wonderful work.

I have some issues when I try to use my custom dataset.

First. the command that you mentioned on README works for my custom dataset. python train.py --gpus=8 \ --data=${DATA_PATH} \ --cfg=paper256 \ --outdir=training_example

However, when I try to use pretrained model(ffhq256) and retrain with my custom dataset it gave me some error message. The command that I tired is CUDA_VISIBLE_DEVICES=0,1 python3 train.py --gpus=2 --data=${DATA_PATH} --cfg=paper256 --resume=ffhq256 --outdir=training_example

And the error message that I got is this

I hope that I can get any solution!

Thanks!

[49xxy]

Have you solved this problem？

[49xxy]

@limbo0000 Hello, I encountered this problem when resume training, and I hope to get your help!