Welcome to PyTabKit’s documentation!

Tabular ML models in pytabkit.models

• Overview of the models part
  • Scikit-learn interfaces
  • AlgInterface: more fine-grained control
  • Hyperparameter handling
  • Internal data representation
  • Data preprocessing (also available for other models)
  • NN implementation
  • Vectorization
• Scikit-learn interfaces
  • fit()
  • RealMLP
    • __init__()
    • __init__()
  • Boosted Trees
    • __init__()
    • __init__()
    • __init__()
  • Other NN baselines
    • __init__()
    • __init__()
    • __init__()
    • __init__()
    • __init__()
    • __init__()
  • Other methods
  • Saving and loading
• Hyperparameter optimization
  • Option 1: Using the HPO interface
  • Option 2: Performing your own HPO
• NN implementation
• Training directly with PyTorch Lightning
  • Using PyTorch Lightning
• (Multi)quantile regression with RealMLP

Tabular benchmarking using pytabkit.bench

• Overview and Installation of the Benchmarking code
  • Installation
  • Using Sphinx Documentation
• Running the benchmark
  • Configuration of data paths
  • Download datasets
  • Run experiments with slurm
  • Run experiments without slurm
  • Time measurements
  • Evaluating the benchmark results
    • show_eval()
  • Creating plots and tables
  • Single-task experiments
  • Other utilities
• Adding your own models to the benchmark
• Data format
  • Algs folder
  • Tasks folder
  • Task collections folder
  • Results folder
  • Result summaries folder
  • Other folders
• Code structure
  • Algorithm wrappers
  • Datasets
  • Scheduling code
  • Resource estimation
  • Evaluation and plotting
• Downloading the benchmark results
• Reproducing results of “Rethinking Early Stopping: Refine, Then Calibrate”
  • Installation
    • Original environment
  • Downloading datasets
  • Running experiments

Indices and tables

• Index

• Module Index

• Search Page