AIM 2020 Efficient Super-Resolution Challenge Forum

Go back to competition Back to thread list Post in this thread

> should we consider the improvement of PSNR in the final testing phase?

Is the improvement of PSNR compared to the base model matters in the final ranking?
For example, I got a modelA and modelB, where modelA is lighter than B in terms of parameter size, running time, FLOPS, etc. But for higher validation PSNR: valB > valA > 29.00dB. Which model we should submit as the final entry?

Posted by: hcwang @ July 7, 2020, 2:44 p.m.

This is not a challenge for PSNR improvement. You should submit modelA.

Posted by: cszn @ July 7, 2020, 3:16 p.m.
Post in this thread