Skip to content
nonparametric instrumental variables, like the DeepIV paper, but without any deep learning
Jupyter Notebook Python Dockerfile
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.
example
npiv
tests
.gitignore
Dockerfile
README.md
docker-compose.yml
setup.py

README.md

Implementation of the core idea behind the DeepIV paper, but with tree boosting and quantile regression instead of deep learning and mixtures of gaussians.

Just browsing:

Run via docker: docker-compose up

How to build: docker-compose build

Development:

  • dockerfile just installs the package via pip with the -e option, and docker-compose then mounts this directory to that install location, so that any changes made are directly reflected. Thus, just run the container, and make whatever changes as you like.
  • to run tests as you develop, do docker exec 270033469fd0 pytest -s, excep with whatever the hash of the container is (do docker container ls to see running containers)
  • or even better, just use bash in the container: docker exec 270033469fd0 bash
You can’t perform that action at this time.