This document outlines the core technologies, infrastructure, and machine learning components utilized in this project.
- Vehicle Classification Data: Infinite Dataset Hub - Vehicle Classifier — Utilized for categorizing visual automotive inputs and model validation.
- Traffic Speed & Time-Series Data: witgaw/METR-LA — Sourced for spatial-temporal graph modeling and highway sensor performance tracking.
- NVIDIA Kaolin 3D Library Wheel Index:
https://nvidia-kaolin.s3.us-east-2.amazonaws.com/torch-2.0.0_cu118.html- Target Build: PyTorch 2.0.0 with CUDA 11.8 support for accelerated 3D deep learning and differentiable rendering pipelines.
- Servers: WSGI, Gunicorn process manager, Render Paas
- Infrastructure: Multiple server processes
- Data Strategy: Local MongoDB layer implementation ensuring zero data export policy.
- MongoDB: Atlas Version 8.0.29,
myAtlasClusterEDU - PostgreSQL: Aiven PaaS Version 18.4
- ML Frameworks:
- PyTorch: LSTM models utilized for time-series horizon forecasting.
- scikit-learn: Implementation of Logistic Regression, SVMs (Support Vector Machines), and Isolation Forests for anomaly detection.