Possible bug in the code

[abdulfatir]

UnsupervisedDeepLearning-Pytorch/udlp/clustering/vade.py

Line 119 in 6ea7b71

p_c_z = torch.exp(torch.log(theta_tensor2) - torch.sum(0.5*torch.log(2*math.pi*lambda_tensor3)+\

Hi!
In this line, why did you replace - (minus) with + at the end?

[eelxpeng]

Note that it is -torch.sum(first_term + second_term). I enclose the two terms inside the sum with a negative sign on the front. In math, it should be -log(1/2*pi*sigma^2) - (x-mu)^2/sigma^2.

[abdulfatir]

Thanks for the clarification. Didn't notice the parentheses.

[OmranKaddah]

Hi,
How about this? why are you adding pi?
qentropy = -0.5*torch.sum(1+z_log_var+math.log(2*math.pi), 1)

I think it is a wrong naming for logpzc variable. Better be KL divergence p_z_c and p_z_x
,and qentropy is included.
Also in the loss function, you are only considering binary cross-entropy case in the loss function.