-
1. Is there any reason (either empirical or theoretical) to choose Instance normalization instead of Batch normalization? The paper refer 'resnet' paper for explaining RESBLK but I think 'resnet' does…
-
Why not add Convolutional Autoencoders?
-
Throughout our initial review of papers, our guiding question has been: _What would need to be true for deep learning to transform how we categorize, study, and treat individuals to maintain or restor…
-
I can't build a convolutinal Autoencder that has been introduced [here](https://blog.keras.io/building-autoencoders-in-keras.html)
this is part of the error:
[/home/hoda/anaconda2/lib/python2.7/si…
ehfo0 updated
7 years ago
-
Hi,
I have tweaked around with Lasagne and I found it an excellent library. My experiments revolve around building an autoencoder for the MNIST dataset. Unfortunately I didn't manage to train it prop…
-
Hi Kipf,
Thanks a lot for your excellent paper. I have two questions on GCNModelAE model.
1. InnerProductDecoder can be seen as an inverse operation of the encoder operation (GraphConvolutionSpa…
-
I have a very large number of dimensions in my training data (5.4e06), which is very sparse. I wanted to try using the data raw as I could not find any embedding examples for continuous non-NLP tasks.…
-
Is there a way to reshape a TF tensor inside of a custom Keras loss function? I'm defining this custom loss function for a convolutional neural network?
def custom_loss(x, x_hat):
"""
…
-
Hello, I am currently trying to build a convolutional autoencoder with Keras for 3d volumetric images with one channel only. My images are of size 10x10x10 and stored in a numpy array of shape (n_samp…
-
I am trying to create convolutional autoencoder for temporal sequences using Keras.
Here's the code
``` python
ae = Sequential()
encoder = containers.Sequential([Convolution1D(5, 5, border_mode='vali…