Metrics calculating
fzuzyb opened this issue · 1 comments
fzuzyb commented
Excellent work!!, I follow the README to infer results, however, I found the MANIQA is lower than the original paper (0.5050 vs 0.6198). I would greatly appreciate it if you could provide the calculation code.
cswry commented
Hello, this is likely a difference in the model version being tested.
We calculate the MANIQA metric with MANIQA_PIPAL-ae6d356b.pth.
You can refer the pyiqa for the calculate code.
The crop_boarder is set to 4 following previous work at X4SR task.
Please feel free to contact if you have any other questions.