-
### System information
- **Have I written custom code (as opposed to using a stock example script provided in TensorFlow)**: yes
- **OS Platform and Distribution (e.g., Linux Ubuntu 16.04)**: ArchLi…
-
The pixelCNN++ code [here](https://github.com/openai/pixel-cnn) has a "data" folder that this code also seems to rely on (I just copied it from that repo and it seemed to work). That code has imageio…
-
Have you tried a L1/L2 reconstruction loss on the raw data? Or whats the reason you chose a cross entropy loss with mu law? The paper doesn't really say what kind of reconstruction loss is used.
-
Hi, friends.
As I have not a good GPU for heping you directly in this, I have use the baseline of the work in this repository to develop a WaveNet text generator (self-generator): [https://github.com…
-
Hi,
I was trying to benchmark the results of PixelCNN on CIFAR10 and I when I ran the script, I landed into this [error](http://dpaste.com/33R4231). I am using Theano 0.9.0 and Lasagne 0.2.dev. I wo…
-
-
This issue is for what is reported in [[theano-users] Error while compiling two theano functions](https://groups.google.com/d/topic/theano-users/2D1zj_cs-kQ/discussion)
To reproduce:
- install htt…
nouiz updated
7 years ago
-
It seems that you only implemented the gated convolutional layer, but not the two convolutional network stack?
-
https://github.com/dritchie/pixelCNN/blob/master/util.lua#L15
If imgData is exactly 1, the pixel quantizes to `quantLevels+1`
-
Hi,
I have an issue with model performance at test time. When running the provided code I get the following log:
```
Iteration 0, time = 2875s, train bits_per_dim = 4.3207, test bits_per_dim = 9.…