Questions tagged [densevariational]
8 questions
4
votes
3 answers
How to save a model with DenseVariational layer?
I'm trying to build a model with DenseVariational layer so that it can report epistemic uncertainties. Something like https://www.tensorflow.org/probability/examples/Probabilistic_Layers_Regression#figure_3_epistemic_uncertainty
The model training…
Edmond
- 614
- 2
- 11
- 26
1
vote
0 answers
Variational Inference with Normalizing Flows in Tensorflow Probability
In the last time I've read a little bit about using normalizing flows to improve variational inference f.e. Link1 Link2.
Tensorflow probability already offers RealNVP and MaskedAutoregressiveFlow in the bijector submodule as well as an…
Fermat
- 11
- 2
1
vote
2 answers
Regression Model with 3 Hidden DenseVariational Layers in Tensorflow-Probability returns nan as loss during training
I am getting acquainted with Tensorflow-Probability and here I am running into a problem. During training, the model returns nan as the loss (possibly meaning a huge loss that causes overflowing). Since the functional form of the synthetic data is…
user8270077
- 4,621
- 17
- 75
- 140
0
votes
0 answers
Problem regarding predicting uncertainties using Dense Variational
My model contains a DenseVariational layer that I want to use to predict the mean and standard deviation for each data point prediction. I am quite new to this and I am not able to understand whether my model is able to do that. Everytime I try to…
Rushat
- 1
0
votes
1 answer
validation loss goes up and down [variational inference]
i was training a mlp through variational inference for a regression task on a small dataset with 1 feature. The nn works and the training loss goes down but the validation loss has random spikes and i do not understand how to avoid them
import…
Alucard
- 187
- 7
0
votes
0 answers
How can I specify the prior in Bayesian Neural Networks?
I try to compare different priors, e.g., N(0,1), N(0,10),..., using a Bayesian neural network. Unfortunately, I cannot find a solution for specifying the MVN.
Do you have any clues?
N = len(y_train)*0.8 # number of samples
m = 2 # number of…
feviro
- 9
- 1
0
votes
1 answer
The parameter kl_use_exact in DenseVariational layer of TF
Trying to create a Bayesian neural network using DenseVariational layer in Tensorflow probability. My question is when we set the parameter kl_use_exact to False does by that we are not taking the prior function into consideration. I tried to look…
Lah Cen
- 27
- 1
- 8
0
votes
1 answer
Number of parameters in Tensorflow-Probability network using DenseVariational layers
I cannot figure out why the model returned has in the second layer 189 parameters. The way I calculate them they should be more. Why is this happening?
The code is the following:
# Define the prior weight distribution -- all N(0, 1) -- and not…
user8270077
- 4,621
- 17
- 75
- 140