Why is the PSNR value of the training log different from the PSNR value tested with the final saved model with test.py ?

ykdai / Flare7K

Official Implementation of "Flare7K: A Phenomenological Nighttime Flare Removal Dataset"

https://ykdai.github.io/projects/Flare7K

Other

119 stars 16 forks source link

Closed qulishen closed 4 months ago

qulishen commented 5 months ago

The values in the log are always lower.

ykdai commented 4 months ago

I am not sure if you mean the net_last.pth. If so, please use the net_g_600000.pth or other checkpoints which have indexes.