Hello, I noticed in your paper that there is a comparative experiment on whether to add FSNet to the FAST network. My approach was to directly replace the original NAS architecture of FAST with FSNet as the backbone. However, the results of training in this manner were not as good as those of FAST itself. Could you please explain how you specifically implemented this? Also, would it be possible for you to share the code for this part?
Hello, I noticed in your paper that there is a comparative experiment on whether to add FSNet to the FAST network. My approach was to directly replace the original NAS architecture of FAST with FSNet as the backbone. However, the results of training in this manner were not as good as those of FAST itself. Could you please explain how you specifically implemented this? Also, would it be possible for you to share the code for this part?