-
I'm playing around with Variational Autoencoders and have noticed that when running it on GPUs with PlaidML it doesn't learn as well as when I run it on CPU.
Here are outputs from my work computer:…
-
Hi,
I'm evaluating this framework for machine learning. So far, I've made several experiments with this framework very successfully, but I'm facing a strong wall when it comes tor RBM.
I'm try…
-
I am training a deep autoencoder 978->500->250->100->30 but it seems quite weird that during the pretraining phase the score (both for RMSE_XENT and MSE )is increasing. Changing random seeds does not…
-
I have set up `NeuralNetConfiguration` with a constant seed to be able to reproduce the results, and then I've tried modifying "learning rate" setting on both the RBM layer configuration builder and `…
-
https://github.com/deeplearning4j/deeplearning4j/blob/ed6547956b08df9a9477936d65db7dc80bc75874/deeplearning4j-nn/src/main/java/org/deeplearning4j/nn/layers/feedforward/rbm/RBM.java#L166-L182
Dependin…
-
RBM calculates scores by calling this https://github.com/deeplearning4j/deeplearning4j/blob/master/deeplearning4j-nn/src/main/java/org/deeplearning4j/nn/layers/feedforward/rbm/RBM.java#L192 (note that…
-
I am trying to modify the loss function in the VAE example available here: https://github.com/rstudio/keras/blob/master/vignettes/examples/variational_autoencoder.R
In the example, the KL divergenc…
-
디즈니와 칼텍에서 한 영화 관객들의 표정 데이터를 가지고 autoencoder와 matrix factorization으로 모델링 하는 논문인데 흥미로운 것 같습니다.
CVPR 2017 논문인데요,
기존 non-linear matrix factorization에 variational autoencoder을 적용하고 있습니다. 이를 통해 기존 baselin…
-
Hi all,
Foremost, my theano and keras are up-to-date.
I am trying to adapt the keras VAE template `variational_autoencoder_deconv.py` for a non-MNIST unlabeled dataset. I am using 38,585 256x256…
-
Refer to Learning Discourse-level Diversity for Neural Dialog Models using Conditional Variational Autoencoders