Project Architecture and Status Overview

Directory Structure and Core Components

• Model Architectures (codebase/): This directory houses the model pipeline classes, which are designed to support modular model implementations, including architectures like iTransformer. Each pipeline class standardizes workflows by exposing four primary methods:
  • preprocess_splits (for training)
  • preprocess_inference
  • train
  • inference
• Data Management (data/): Contains the underlying datasets used for training and evaluation.
• Preprocessing Utilities (preprocess/): Contains utility and helper functions utilized across the pipeline classes to ensure consistent data transformation.

---

Development Environment

• Runtime: Python 3.12.3
• Environment: Virtual environment (venv) deployed on a Linux infrastructure.

---

Execution and Verification Status

• Testing Workflow: Functional verification of the pipeline classes is currently conducted via execution scripts in main.py.
• Testing Coverage: While the training and preprocessing pipelines are functional, the inference capabilities have not yet undergone rigorous testing or validation.