LSPatch-T
Tutorial: LSPatch-T
This project, LSPatch-T, provides a framework for time series forecasting using various neural network models, including specialized ones like LSPatch-T and PatchTST. It handles the entire experiment lifecycle, from loading and preparing data to training models, evaluating performance, and tracking results using MLflow for reproducibility and analysis. It leverages modular model architectures built from fundamental core layers and applies specific loss functions and helpful utility functions.