Skip to main content

Tutorials

Getting Started with the HLO Feature Dataset​

  • Introduction: Learn how to use the dataset for predicting deep learning training times using metadata and graph features.
  • Prerequisites:
    • Python 3.x
    • pandas, numpy, xgboost
  • Steps:
    1. Download the dataset from Hugging Face: HLO Feature Dataset Repo
    2. Open the baseline notebook:
      Baseline_XGBoost_Resource_Estimation.ipynb
    3. Follow step-by-step instructions to train your first resource estimation model.
  • Outcome: A trained XGBoost model capable of predicting fit_time for DL workloads.