search

mjkwon2021 / CAT-Net

Official code for CAT-Net: Compression Artifact Tracing Network. Image manipulation detection and localization.
210 stars 25 forks source link

About validate performance #48

Open Mike-Xiao opened 6 months ago

Mike-Xiao commented 6 months ago

Hi, I'm kind of confused about the validation performance after each training epoch, I am now traing on my own dataset , it seems that the training loss is continuing dropping after each epoch , but the numbers in validation keeps exactly the same after each epoch, but i think it should at least change a bit. may i know what cause that?Thank you! image image image image

Mike-Xiao commented 6 months ago

I just directly used CAT_full.yaml for training settings as suggested in train.py, and at first I thought it was becasues loading the two pretrained models maks the output of validation stable, however after commenting out loading them I still get exactly the same validation numbers after each epoch, the numbers associated with iou are all close to 50 even at the first epoch, it bothers me a lot, could you please give me any suggestion about that? thanks alot!