What's the loss used in training VAE?

axium / Blind-Image-Deconvolution-using-Deep-Generative-Priors

Paper implemention of paper "Blind Image Deconvolution using Deep Generative Priors"

MIT License

30 stars 12 forks source link

What's the loss used in training VAE? #2

Closed JingyunLiang closed 3 years ago

JingyunLiang commented 4 years ago

Hi, I didn't find the training code of VAEs. Although there is a testing code, I don't know which loss is used in training, nor do I find the loss in your paper. What's the loss used in training a VAE for kernels and SVHN?

If possible, could you please provide the training code of VAE?

fahadshamshad commented 3 years ago

Hi, we have used standard VAE training loss and use the trained VAE in our deblurring algorithm. Feel free to open the issue again if you have any doubt.