v1.0.1 – 2025-07-29

Features

• ACGAN model implementation for fast time series generation
• Diffusion-TS model for high-fidelity time series generation
• Context-aware Normalizer for data preprocessing
• Pre-trained Watts model series available on Hugging Face Hub
• PyTorch Lightning training pipeline with Hydra configuration
• Comprehensive evaluation framework with discriminative and predictive scores
• PecanStreet dataset integration with preprocessing pipeline
• Model registry system with decorator-based registration
• DataGenerator API for easy synthetic data generation
• Improved documentation

Models Available on HF Hub

• Watts_1_1D: ACGAN-based 1D electricity consumption
• Watts_1_2D: ACGAN-based 2D consumption + generation
• Watts_2_1D: Diffusion-TS 1D consumption
• Watts_2_2D: Diffusion-TS 2D consumption + generation