You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm trying to replicate the results in your article, but I'm having trouble getting similar performance with my own R2U-Net trained from scratch. The default parameters in the code do not lead to the desired performance. What optimization process and hyper-parameters did you use to get the pre-trained weights in the Google Drive ?
Also, why is the interpolation mode in the alignment loss set to "nearest" since it is unlikely to lead to a differentiable loss?
Thank you very much.
The text was updated successfully, but these errors were encountered:
tcourat
changed the title
Optimization process
Optimization process to replicate the results & strange interpolation mode
Sep 8, 2022
Hello,
I'm trying to replicate the results in your article, but I'm having trouble getting similar performance with my own R2U-Net trained from scratch. The default parameters in the code do not lead to the desired performance. What optimization process and hyper-parameters did you use to get the pre-trained weights in the Google Drive ?
Also, why is the interpolation mode in the alignment loss set to "nearest" since it is unlikely to lead to a differentiable loss?
Thank you very much.
The text was updated successfully, but these errors were encountered: