Training Dataset for SEST 2026 Submission: Influence of Training Data Selection on Data-Driven Power Quality Disturbance Classification Models

Minimal, self-contained bundle to reproduce the test metrics reported in Influence of Training Data Selection on Data-Driven Power Quality Disturbance Classification Models from the trained models, without the full training framework. It contains the datasets' loading + preprocessing + model code needed for inference only, the trained checkpoints, the training configs, and a single evaluate.py example. Training/benchmarking/DDP/hyperopt code is intentionally not included. For more info, please refer to the README.md