lw0517

Thank you for your question. We have revised run.sh. Please refer to the latest version. Please utilize the same splits of data and the same environment for re-implementations. This open-source...

In fact, it is a special case of MetaFed. In some situations, threshold>1 can bring better performance and there still exist some differences between MetaFed and FedAvg. There also exist...

You can try some other datasets, e.g. PAMAP2. For this situation, the best results are fixed in the fixed environment. MetaFed is a framework. The threshold is only a hyperparameter...

We utilized 3090 to train FedCLIP on PACS and it cost 323s for 10 rounds.  

> when threshold=1.1 maybe it likes an iid suituation, not a non-iid problem? In fact, it is still for a non-iid problem. Besides the common knowledge accumulation stage, there also...

> 在用自己的数据集运行DG库中的MMD算法时，出现了矩阵形状不一样导致无法相加的问题。下图是github中MMD计算的代码：  我的数据，x1和x2的形状均为(32, 600, 64）。考虑到batch问题，我在计算时将`addmm`换为了`addbmm`，但是我发现，无论是用二维数据使用`addmm`还是三维数据使用`addbmm`，在计算时存在矩阵形状不一样导致无法相加的问题，各个张量的形状如下图所示：  可以看到，`x2_norm.transpose(-2, -1)`的形状与`matmul(x1, x2.transpose(-2, -1))`的形状不一致，二者是没办法做和相加的。麻烦大佬们看一下，是哪里有了问题？感谢 建议的做法是将三维的转化成二维的进行MMD距离计算。如果用torch.addbmm，最好参考一下https://pytorch.org/docs/1.10/generated/torch.addmm.html?highlight=addmm#torch.addmm， 可以看出来维度是不一致的。