This release includes the following changes and improvements:
Improved Training Speed:
The model training routines have been optimized to reduce training time while maintaining accuracy.
Cross‑Validation Bug Fix:
Fixed an issue in the cross‑validation splitting logic so that the splits now correctly cover the dataset without overlap errors.
Enhanced Scaling Methods:
New scaling options have been added to the preprocessing module to better support various data distributions (e.g., additional robust scaling techniques).
Evaluation and Metric Factory Updates:
The evaluation framework now offers a more flexible Metric Factory that allows users to easily add or modify evaluation metrics.
Logging Improvements:
The logging utility now ensures that log files are correctly created (with directory creation if necessary) and provides more informative output during both training and inference.
General Code Clean‑up:
Minor bug fixes and code refactoring have been performed to improve overall code readability and maintainability.
add Embedding layer custom for Time series
Thank you for using APDTFlow!