01_model_training.md

Lecture 1

Model training and experimentation with Comet ML

Steps

1. Create the virtual environment with Poetry
2. Generate training data
3. Build a baseline model
4. Build Machine Learning models

1. Create the virtual environment with Poetry

Watch the video 🎬

1. Create a Python virtual environment with all project dependencies with

   $ curl -sSL https://install.python-poetry.org | python3 -
   $ poetry install
   

   or simply use the Makefile

   $ make init
   

2. Activate the virtual environment you just created

   $ poetry shell
   

3. Sign up for Comet ML for FREE, create a workspace and copy your API key from the dashboard.

4. Set your API key and workspace name variables in set_environment_variables_template.sh, rename the file and run it

   $ . ./set_environment_variables.sh
   

Your local development environment is ready. Let's now generate some training data.

2. Generate training data

Watch the video 🎬

Download historical data from Coinbase and save it locally to disk

• Run either $ python src/data.py, or
• Simply $ make data

3. Build a baseline model

Watch the video 🎬

• Establish a baseline performance using a very dummy (yet powerful) baseline model

  $ python src/baseline_model.py
  

4. Build Machine Learning models

Watch the video 🎬

• Here is the list of experiments I ran

  $ python src/baseline_model.py
  $ python src/train.py --model lasso
  $ python src/train.py --model lasso --tune-hyperparams --hyperparam-trials 3
  $ python src/train.py --model lightgbm
  $ python src/train.py --model lightgbm --tune-hyperparams --hyperparam-trials 3
  

• Feel free to try adding more features, using other technical indicators, or experiment with other ML models.

➡️ Go to the next lecture