Resume released trained model

GlassyWu / AECR-Net

Contrastive Learning for Compact Single Image Dehazing, CVPR2021

339 stars 48 forks source link

Resume released trained model #23

Open PaulTHong opened 2 years ago

PaulTHong commented 2 years ago

When I resume the released model, there is an error: "Unexpected key(s) in state_dict: "mix4.w", "mix5.w"". mix4 and mix5 don't appear in the network constructure. So the released model is not matched with the released code. Am I right? Thanks.

wangFanFanFan commented 2 years ago

have you figure out ? the net code has 'mix1' and 'mix2', and if i just rename the mix4 to mix1, mix5 to mix2, the psnr result seems well

PaulTHong commented 2 years ago

Oh, thanks! I will have a try. Have you tried to train from scratch? I can't reach the evaluation metrics as shown in the paper.

Bruce-WangGF commented 2 years ago

please， is the predict code named funtion test in train_aecrnet.py? How to revise it to apply a picture of JPG? Thanks!