model-tracking

model-tracking is a tool for tracking the results of machine learning services.

Running model-tracking

Prerequisites

model-tracking requires a PostgreSQL database in which to store results. During development, this database can be provided using a PostgreSQL Docker container:

docker run --rm -it -p 5432:5432 -e POSTGRES_USER=model-tracking -e POSTGRES_PASSWORD=model-tracking -e POSTGRES_DB=model-tracking postgres

To prepare the database, execute the following Makefile target:

# Run the migrations against the database.
# Note: the database parameters can be overwritten by specifying the DATABASE_URL environment variable.
make migrate