All Questions
112 questions
0
votes
0
answers
12
views
custom callback "BetaAnnealing" isn't updating the beta value correctly for my VAE kl_loss
I have a custom layer in my VAE which handles the loss calculation. I have a custom callback which is meant to update the beta value at the beginning of every epoch. The calculation is correct because ...
- 1
3
votes
0
answers
371
views
Workaround for removal of add_loss()
I'm working through a Keras/Tensorflow course that uses Keras 2 to build a variational autoencoder and I'm trying to get it working in Keras 3. I've managed to overcome a lot of issues but I'm stuck ...
- 174
0
votes
1
answer
56
views
Autoencoder shaping issue
I am having an issue with my autoencoder as I am shaping the ouput incorrectly. Currently the autoencoder is coded lke this.
I Got This Error :
ValueError: Dimensions must be equal, but are 2000 and ...
2
votes
1
answer
386
views
Generating new data using VAE in keras
I have built the following function which takes as input some data and runs a VAE on them:
def VAE(data, original_dim, latent_dim, test_size, epochs):
x_train, x_test = train_test_split(data, ...
- 4,492
1
vote
1
answer
120
views
ML model fit with training data outperforms model fit with generator
My ultimate goal is to fit a ML autoencoder by inputting a data generator into the fit method within the keras API. However, I am finding that models fit with a generator are inferior to models fit ...
- 862
0
votes
1
answer
356
views
RGB Autoencoder outputs blank images
I am working on a college assignment, the assignment is to make a denoise autoencoder to encode and decode images with noise.
I've implemented everything right but when trying to predict it outputs ...
2
votes
1
answer
307
views
Autoencoder to convert 2d to 3d layers ambiguity
i got around some references and research papers and taking idea from one of them i thought to go ahead and implement the same the image reference-
So, here we are inputing a 2d input and the model ...
- 71
0
votes
0
answers
768
views
Why does my autoencoder not reconstruct RGB images as it should and in BGR form?
An autoencoder is a combination of two CNN's and can take an image as an input. One CNN reduces the image into a very small representation and the second CNN takes this small representation and tries ...
- 1
0
votes
0
answers
73
views
Autoencoder Custom Error Metric not working as intended
I have created a custom error metric to measure the success of an autoencoder implemented in Keras and sklearn. At a high level, it computes the loss for each of the output neurons separately (...
- 65
2
votes
2
answers
979
views
How can I connect the outputs of 2 keras models into one seperate layer?
I'm trying to solve a problem where I initially need to get 2 inputs, pass each of them through layers of autoencoders and then connect everything to a layer that will connect to both of the ...
- 21
1
vote
0
answers
117
views
How to use sparse_categorical_cross_entropy in an autoencoder?
I am trying to train an autoencoder and due to memory issues I decided that I would stop using one-hot-encoding, instead tryout the SparseCategoricalCrossentropy(). Unfortunetely after trying to train ...
- 326
1
vote
1
answer
750
views
How to match input and output shapes of Conv2D AutoEncoder
Having a set of black and white images with the following shape (1000, 11, 1). I'm trying to modify the keras mnist example to work with my data, so I've written the following code:
input_img = layers....
- 8,913
1
vote
1
answer
349
views
Reducing dimension of a DataSet with Autoencoders in Keras
I'm trying to reduce both instances and variables of a dataset. The shape of my dataset is, say, (x , y), with y being columns and x being rows. I want to reduce it to (k, m).
However, in Keras, I do ...
- 55
2
votes
0
answers
294
views
Variational autoencoder (VAE) - decoder weights in Keras: access and interpretation
I have 2 questions about the following VAE :
inputs = Input(shape=(input_size,), name='concat_input')
x = Dense(denselayer_size, activation=activ_fct, name="encoding")(inputs)
x = BN()(x) ...
- 363
2
votes
1
answer
2k
views
NotImplementedError: Layer ModuleWrapper has arguments in `__init__` and therefore must override `get_config`
I'm trying to save my autoencoder model (for classification) to disk but the following error appears when doing: model.save(model_name)
NotImplementedError: Layer ModuleWrapper has arguments in ...
- 383