codeslake/RefVSR

frame_num in other benchmark models

YoungRaeKimm opened this issue · 1 comments

YoungRaeKimm commented

Thank you for your great work.
I'm going to reproduce results in Table 2, but I'm confused in some configurations in other models.

  1. How did you configure the frame_num in other benchmark models? I guess this hyper-parameter would be important, but there is no information about this in the paper.
  2. Why the frame_num of each model is different? It varies from 7 to 13.

Thank you

codeslake commented

Thank you for your great work. I'm going to reproduce results in Table 2, but I'm confused in some configurations in other models.

  1. How did you configure the frame_num in other benchmark models? I guess this hyper-parameter would be important, but there is no information about this in the paper.
  2. Why the frame_num of each model is different? It varies from 7 to 13.

Thank you

Hi, @YoungRaeKimm.
Thanks for your interest in our work!

  1. frame_num actually did not affect the final performance that much, as long as it was greater than 5. I guess the model learns to properly propagate temporal features when frame_num is greater than 5.
  2. The numbers are chosen to max out the GPU memory, which differs by which model you train.
  • 👍1