I am building a Bayesian Neural Network for multi-class classification by using PyMC3 and Theano. In the PyMC3 docs there’s a Variational Inference example by building a 2-class classifier neural network . In the docs example, the final output is modeled via a Bernoulli sample. How should I go about…

Multi-Class Classifier in PyMC3

vb690 January 16, 2022, 8:46pm 9

I don’t know if this applies to you case, but pm.Categorical as a likelyhood requires integer labels for the target rather than one-hot encoded.

This has caused issues to other users (see this), including myself.

1 Like

Topic		Replies	Views
Getting multinomial class probabilities during posterior prediction on test Data Questions	4	1450	January 31, 2018
Bayesian Classifier Combination Model (need help) Questions	3	2020	February 22, 2018
Probability estimation in pymc3 Questions	5	686	April 22, 2018
Posterior to Prior in Categorical distribution (or encapsulating multiple data sources to categorical analysis) Questions modeling	4	1565	February 28, 2022
How to implement the Bayesian inference correctly? Questions	2	809	December 21, 2021