/ml4t-docker

A docker image to run coursework for ML4T without overhead of full ubuntu VM

Primary LanguageShell

ml4t-docker

This repository defines a docker image to run projects for CS 7646 (machine learning for trading).

basic usage

prerequisites

  1. install docker
  2. download the utility/grading modules (ML4T_2020Spring.zip) and project zips for your semester.
  3. unzip the items from 2.
    • first unzip the utility archive
    • next, unzip projects so they are in the ML4T_2020Spring directory
  4. modify run.sh to have the proper local path to your ML4T_2020Spring from (3)
    • replace /path/to/ML4T_2020Spring with your local path

getting started

  1. at your terminal, run sh build.sh
  2. at your terminal, run sh run.sh
  3. you will now be in the container in the /ml4t directory.
  4. test your image.
    • run cd optimize_something then python grade_optimization.py

Sample Setup

mkdir tmp
cd tmp
git clone https://github.com/ankona/ml4t-docker.git
cd ml4t-docker

curl http://quantsoftware.gatech.edu/images/c/ce/ML4T_2020Spring.zip --output ML4T_2020Spring.zip

curl http://quantsoftware.gatech.edu/images/9/92/20Spring_martingale.zip --output 20Spring_martingale.zip

unzip -q ML4T_2020Spring.zip -d ML4T_2020Spring
unzip -q 20Spring_martingale.zip -d ML4T_2020Spring

sh build.sh

// edit run.sh now to have your path.
sh run.sh

WARNING

The actual projects may not be shared and any images may not be published to dockerhub or any public repository without risking an honor code violation.

todo:

  1. document how to connect using visual studio code and debug 'remotely.'