Code accompanying the paper Least Squares Auto-Tuning.
We use conda for python package management.
To install our lsat environment, run the following commands while in the repository:
$ conda env create -f environment.yml
$ conda activate lsat
To run tests, run (with the lsat environment activated):
$ python lstsq.py
You should see "All tests passed!".
To run the MNIST example described in the paper,
type into a terminal (with the lsat environment activated):
$ python mnist.py # to download the data
$ python mnist_example.py # to run the script