self-distillation really works?

[LASTLINEK]

Hi, I recently downloaded your code and did some experiments. I found that the performance of the self-distillation was about the same as just adding label loss at each stage. Have you ever encountered such a situation?

[SmoothJing]

Hi, I recently downloaded your code and did some experiments. I found that the performance of the self-distillation was about the same as just adding label loss at each stage. Have you ever encountered such a situation?嗨，我最近下载了您的代码并做了一些实验。我发现自蒸馏的性能与在每个阶段添加标签损失大致相同。你有没有遇到过这样的情况？

I am also looking at this code, easy to add friends exchange