Notebooks¶
The notebooks are Colab examples grouped by topic. Each entry links to the related documentation page.
Getting Started¶
01 Optimization Model: build an
optModelfrom a GurobiPy / COPT / Pyomo / OR-Tools / MPAX backend. Pairs with Optimization Models.02 Optimization Dataset: generate synthetic data and wrap it in
optDataset. Pairs with Data and Datasets.03 Training and Testing: train method families on a shortest-path dataset. Pairs with Training Methods and Evaluation.
Method Deep Dives¶
04 CaVE for Binary Linear Programs: train with the cone-aligned CaVE loss on TSP. Pairs with the Cone-Aligned Estimation section of Training Methods.
08 kNN Robust Losses: train with the kNN robust loss via
optDatasetKNN. Pairs with the optDatasetKNN section of Data and Datasets.
GPU Acceleration¶
09 Solving on MPAX with PDHG: batch-solve LPs on GPU via MPAX, end-to-end without CPU round-trips. See the MPAX backend (
optMpaxModel) in the Solver Backend Subclass section of Optimization Models.10 JAX Frontend: train PyEPO losses in JAX/Flax with
jax.grad. MPAX is GPU-native and jittable; non-JAX backends run throughjax.pure_callback. Pairs with JAX Frontend.
Applied Examples¶
05 2D Knapsack Solution Visualization: visualize selected items for a 2D knapsack instance to inspect what the trained predictor is doing.
06 Warcraft Shortest Path: train an image-based shortest-path predictor on the Warcraft terrain dataset, using a CNN encoder feeding into PyEPO.
07 Real-World Energy Scheduling: apply PyEPO to an energy scheduling dataset with measured demand data.