search

GuHongyang / VaDE-pytorch

the reproduce of Variational Deep Embedding : A Generative Approach to Clustering Requirements by pytorch
123 stars 25 forks source link

multivariate Gaussian distribution #2

Open ThibaultGROUEIX opened 4 years ago

ThibaultGROUEIX commented 4 years ago

Hi @GuHongyang , Thanks for the reimplementation! they mention in the paper that As for the generative process in Section 3.1, the multivariate Bernoulli distribution is used for MNIST dataset, and the multivariate Gaussian distribution is used for the others . In your code, you've implemented the multivariate Gaussian distribution right? Sincerely, Thibault

GuHongyang commented 4 years ago

My setting is the same as the original paper. As you can see in the Decoder of model.py, the output is the parameters of the bernoulli distribution.

ThibaultGROUEIX commented 4 years ago

Thanks ! Sorry i didn't dare dive in the details of the ELBO loss, i am not familiar with it and I would need comments to follow through. I just saw a lot of gaussians in there so I assumed you were using a multivariate Gaussian distribution. Thanks again

ThibaultGROUEIX commented 4 years ago

Related question: I am trying the code on AFFNIST and it systematically crashes. It seems that z_sigma_2log becomes too large here which gives NaN after the exponential. Do you have an idea why that might happen? Thanks again.

huyong1369 commented 4 years ago

hi, I run your code in my computer, but I didn't get the good results as your figure (acc about 81%), I'm not sure whether I need to reset the parameters?

goodman1204 commented 4 years ago

hi, I run your code in my computer, but I didn't get the good results as your figure (acc about 81%), I'm not sure whether I need to reset the parameters?

Hi, when I set the hid_dim as 50, the accuracy I got was is worse. I don't know what happened here? Any ideas?

wzm2256 commented 4 years ago

@huyong1369 I'm running related experiments using this codes. The average accuracy is about 81% for mnist. Try to repeat the experiment to achieve better results.

ferb2015 commented 3 years ago

Related question: I am trying the code on AFFNIST and it systematically crashes. It seems that z_sigma_2log becomes too large here which gives NaN after the exponential. Do you have an idea why that might happen? Thanks again.

Hello, this happened to me today. Were you able to solve this issue? Thank you.

ThibaultGROUEIX commented 3 years ago

No I did not solve it. As you can see from table 2 here, we reported the divergence on problem on Affnist, but the code ran on our other toy datasets.