tensorflow / tensorflow/probability

Bayesian Neural Networks Tensorflow 2.4.1

Open
#1,305 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Jupyter Notebook
Stars
4.4k
Forks
1.1k
PR merge metrics
No merged PRs in 30d

Description

Hello and thank you for your work!
I was trying to implement a Bayesian Neural Network for classifying images with color depth. I am using Tensorflow 2.4.1 and Tensorflow Probability 0.12.1
No matter my architecture, learning rate, and loss function, my accuracy does not improve, is goes between 0.1 and 0.18, up and down.
I think it is a problem with the gradients, that are not being calculated well for the probabilistic layers, like DenseFlipout(). Would you help me please?

Thank you, Lucia

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

The issue provides no files, tests, entry points, or reproducible example to inspect. Start by obtaining the model, data, loss, and training code, then isolate whether DenseFlipout gradients fail in TensorFlow 2.4.1 with TensorFlow Probability 0.12.1. Done means identifying a reproducible cause and documenting or fixing it.

Written by the indexing model from the issue text.

Assessment

Tech stack
tensorflow
Domain
machine-learning
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.