Skip to content
Bayesian and ML Implementation of the Normalizing Flow Network (NFN)
Python Jupyter Notebook Makefile
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
data/local
estimators
evaluation
simulation
tests
.gitignore
.pre-commit-config.yaml
Makefile
README.md
config.py
demo.py
evaluate.py
nfn_outline.png
plot.py
pyproject.toml
pytest.ini
requirements.txt

README.md

(Bayesian) Normalizing Flow Network

A Normalizing Flow based model for Conditional Density Estimation. Outline of the Normalizing Flow Network More info about the NFN: paper, blog post

This repo implements:

Conditional Density Estimators:

  • Normalizing Flow Network (NFN)
  • Mixture Density Network (MDN)
  • Kernel Mixture Network (KMN)

Normalizing Flows

  • Radial Flow
  • Planar Flow
  • Affine Flow

Installing dependencies

We use the nightly releases of TensorFlow Probability to get the newest layers for Variational Inference. Python version is 3.6.

pip install -r requirements.txt
# Only necessary for running the evaluation scripts, not necessary for development
pip install --no-dependencies cde

Running tests

Tests are implemented using pytest

# run full test suit
make tests
# run tests without the slow integration tests
make quicktest
You can’t perform that action at this time.