wxtu / dfcn Goto Github PK

您好，我想请问一下，跑10次取平均，这10次的参数设置有什么不一样吗，是随机种子的设置不一样吗，如果是的话，那具体取什么值

可以提供一下预训练的代码吗？

Hi, thanks a lot for the code!

As mentioned in the paper, the training procedure includes 3 steps: "First, we pre-train the AE and IGAE independently for 30 iterations by minimizing the reconstruction loss functions. Then, both subnetworks are integrated into a united framework for another 100 iterations. Finally, with the learned centers of different clusters and under the guidance of the triplet self-supervised strategy, we train the whole network for at least 200 itera- tions until convergence."

I have trained AE and IGAE independently, but how to integrate them into a united framework for training without using the triplet strategy, i.e. the second step ?

Thanks for your advice in advance!

When reproducing the SDCN model experiment, do you use the pre trained code provided by the author to obtain the pre trained parameter file on your own computer, and then use the file in the official code to obtain the experimental data, or do you directly use the optimal. pkl file provided by the SDCN author to obtain the experimental data in the official code? For advice, thank you

怎么得到预训练模型，几个py文件里面没有看到关于预训练的文件

hello，Which method is used to get the graph in the code？