rorydm/mito_gbdts repository overview

⁠Mitotic classification using gradient boosted decision trees

This repository contains a notebook lightgbm_mito.ipynb and some scaffolding to facilitate reproducibility.

Most importantly, the Dockerfile creates an environment in which the notebook can be run. You can either build the docker image yourself with

bash build.sh

or pull it from docker hub with

docker pull rorydm/mito_gbdts

To run the notebook (you might want to do this in a screen session):

bash run.sh <port number>

then copy the link jupyter gives you to your browser, replacing the default port 8888 with <port number>. If you're working remotely, you''l have to forward the remote <port number> to your local machine.

Once the docker process is running, open the notebook, and Kernel > Restart & Run All should work.