Hello, Dr chen, thank you very much for being able to open the source code, it is a very rewarding work.
Do you facilitate the provision of l2 parameters in CMSDNet? Because I found that the default parameter in the code (1e-8) caused some overfitting when I try to reproduce it.
Thank you very much!
Hello, Dr chen, thank you very much for being able to open the source code, it is a very rewarding work. Do you facilitate the provision of l2 parameters in CMSDNet? Because I found that the default parameter in the code (1e-8) caused some overfitting when I try to reproduce it. Thank you very much!