search

AliLotfi92 / SiGraphVAE_pytorch

Pytorch implementation of SiGraphVAE
MIT License
2 stars 0 forks source link

The results generated by this code are too far apart from the experimental results in the paper #1

Open liunianxuxie opened 3 years ago

liunianxuxie commented 3 years ago

Hi, thank you for your code. It is the pytorch implementation of the paper SIGVAE. However, the results I obtained by running the code you provided are too far from the experimental results in the paper. May I ask what part of the code is different from the tensorflow implementation or is it caused by other reasons?

AliLotfi92 commented 3 years ago

Hi, Sure thing! :)

Yes, you are right! I cannot say any significant difference in the implementations. So the only thing that I can tell you is that I saw that they put K (a hyperparameter) to be 2000, which is pretty far from today's CPUs or GPUs capabilities. Good Luck! Please let me know if you find another difference too. I just duplicate their codes.

liunianxuxie commented 3 years ago

Hi, Sure thing! :)

Yes, you are right! I cannot say any significant difference in the implementations. So the only thing that I can tell you is that I saw that they put K (a hyperparameter) to be 2000, which is pretty far from today's CPUs or GPUs capabilities. Good Luck! Please let me know if you find another difference too. I just duplicate their codes.

Thank you for your reply, I will compare the result by using different K.