TSP

Travelling salesman problem — edge classification (in/not in optimal tour). The primary catalog metric is F1.

Travelling salesman problem — edge classification (in/not in optimal tour).

The benchmark release supplies fixed graph tensors and task-specific node or edge inputs. Image graphs use superpixel appearance and position, synthetic node tasks use generated signals, TSP uses coordinates, and CSL is intentionally featureless.

The catalog records these protocols or variants: Standard split. Evaluation uses F1 (higher is better). Exact masks or folds must come from the cited release.

Use the released train/validation/test files rather than resplitting. These tasks were designed to expose specific architectural limitations, and preprocessing such as changing superpixels, adding positional encodings, or balancing edge labels changes the benchmark.