Why is saveconfig insufficient to reproduce predictions from an already trained XGBoost model?
Classic MLLevel: IntermediateXGBoost · Model IO
Why is save_config insufficient to reproduce predictions from an already trained XGBoost model?
Hints
- Hyperparameters do not determine fitted trees without data/training.
- Inspect which artifact contains learned structure.
Answer
The configuration records parameters, not the fitted tree ensemble itself. Preserve the model separately, and retain the configuration as experiment metadata rather than treating it as learned weights.
Source and licence: XGBoost · Model IO · Apache-2.0 · answer: Quantoros editorial