Skip to main content
Version: v0.68.0

Understand experiments in MLOps

In H2O MLOps, an experiment is defined as the output of a training job. Many different experiments can be rapidly created by modifying specific parameters and hyperparameters. Experiments can be imported in the following formats:

  • Driverless AI MOJO zip file. Import directly through the DAI interface or by dragging and dropping the zip file.

  • H2O-3 open-source MOJO zip file. Import by dragging and dropping the zip file.

  • Large language model (LLM) experiments. A zip file with the following content:

    • Filename: artifacts/vllm.json (Note that this is a file within the artifacts directory.)
    • Sample file content:
      {
          "model": "mistralai/Mistral-7B-Instruct-v0.2"
      }
  • Third-party model frameworks. This includes scikit-learn, PyTorch, XGBoost, LightGBM, and TensorFlow. Import by dragging and dropping an MLflow packaged file.
note

Before an experiment can be deployed, it must first be registered in the H2O MLOps Model Registry. For more information, see Understand model registration and versioning and Register an experiment as a model.

To learn more about adding experiments to H2O MLOps, refer to the following pages:

Feedback