Loss divergence problem in train_PBAFN_stage1 using 1024 x 768 dataset.

geyuying / PF-AFN

Official code for "Parser-Free Virtual Try-on via Distilling Appearance Flows", CVPR 2021.

545 stars 138 forks source link

Open jjxxmiin opened 2 years ago

jjxxmiin commented 2 years ago

Thank you for your good work.

Currently, I do training 1024x768 images.

But I got a loss divergence problem.

How I can solve this problem?

If there are any ideas please tell me.

Thank you.

KevinLiu0716 commented 6 months ago

@jjxxmiin Hi, Can you share how you implement training for 1024x768 images? Thank you