drop_in_model_artifacts

Drop-In Environment

These model artifacts are meant to be used with the drop-in template environments in this repo

Artifact Details

For python, you will need to install the libraries from here

For R you will need an environment with the libraries installed in this Dockerfile

The Java model artifacts were made using codegen.

Datasets used

• Binary classification iris_binary_training.csv
  • This dataset does not have enough rows for DR. To get the Java Scoring Code models, I just copied the dataset.
• Regression juniors_3_year_stats_regression.csv

Generating new artifacts

• XGBoost, SKLearn, Keras: run the associated notebooks in this directory
• PyTorch: run PyTorch.py
  • This is a separate script because the custom model needs to include the tourch.nn class for the model as well as the model artifact
• R: run Rmodel.R
• Java:
  • create projects with the desired dataset (double the iris dataset)
  • train a model that generates scoring code
  • download scoring code
  • rename jar to java_bin.jar or java_reg.jar