Generate somewhat realistic ML Flow data without actually training a model.
To generate mock ML Flow data, execute python mock-train.py with the following optional parameters:
experiment_name: The default is test. Name of experiment. If creating multiple experiments, unique names will be generated from string.num_experiments: The default is 1. Number of experiments to generate.num_runs: The default is 1. Number of runs to generate per experiment and/or parent".nested_runs: The default is False. Whether to generated nested runs or not.
To generate a single run:
python mock-train.py
To customize the experiment name:
python mock-train.py --experiment_name customExperiment
To generate multiple runs within an experiment:
python mock-train.py --num_runs 4
To generate nested runs within an experiment:
python mock-train.py --num_runs 4 --nested_runs True
To generate multiple experiments with multiple runs:
python mock-train.py --num_experiments 2 --num_runs 4
To generate multiple experiments with nested runs:
python mock-train.py --num_experiments 2 --num_runs 4 nested_runs True