Fixed a bug in test.py that caused float 0 regarded as 1.
The float 0 can be a very small positive number, so using >0 to distinguish it from 1 can cause problems.
This bug probably caused incorrect PSNR and SSIM values in testing and validation. Currently not sure if training is affected.
Fixed a bug in test.py that caused float 0 regarded as 1. The float 0 can be a very small positive number, so using >0 to distinguish it from 1 can cause problems. This bug probably caused incorrect PSNR and SSIM values in testing and validation. Currently not sure if training is affected.