Public_prob_regression_with_huber_distributions
This repository contains the code used for the implementation of the paper "Probabilistic Regression with HuberDistributions"
Requirements:
-
conda (download from conda.io)
-
liblapack-dev (sudo apt-get install liblapack-dev)
-
blas
-
make
-
gcc
setup:
conda install environment.yaml
cd mathlib; make; cd ..
cd ImageTools; make; cd ..
How to use:
scripts folder contains all significant main files most important script is scripts/train.py run by $python -m scripts.train <CONFIG_FILE> --run_name example config files are in the folder configurations/ tensorflow logs and saved weights will be saved in the folder logs/