Quick Start Guide

This guide will help you train and test your first model in under 5 minutes.

Step 1: Prepare Your Data

Ensure you have:

• Problem instances in ./data/<your_folder>/
• (Optional) Gold solutions as json as ./golds/<your_folder>/gold.json

Step 2: Train a Model

We suggest to visit the other pages to have more informations about training routines:

Further references

• Bundle Network: A fully ML-based Bundle Method
• Hyper-Parameter Learning: Learning an Hyper-parameter in the (aggregated) Bundle method
• Baselines: Grid-Search based baselines including: classic (aggregated) Bundle method with different t-strategies (tuning strategies for the regularization parameter), Adam and Descent.