FastAPI ML Model Management Platform

This FastAPI application provides a comprehensive platform for managing and interacting with machine learning models, specifically YOLO models for computer vision tasks. The application serves as a complete model deployment solution, allowing users to upload, manage, and run predictions with various YOLO models through a clean RESTful API.

The platform features robust model management capabilities, allowing users to upload and organize multiple YOLO models with associated metadata including descriptions and preview images. Each model is stored in its own directory structure, making it easy to manage different versions and configurations. The session-based model selection system enables users to switch between models seamlessly for different prediction tasks.

One of the key technical achievements is the efficient video processing pipeline that handles predictions on video streams. The application processes videos at 2-second intervals, running YOLO inference on selected frames and returning comprehensive results including bounding boxes, confidence scores, and object classifications. The system also includes shared media management features, allowing users to store and retrieve processed images and videos with their prediction results.