evaluation-ml

What problem does it solve?

This Skill streamlines the process of training, calibrating, and evaluating machine learning models for highlight detection, ensuring accurate and reliable scoring.

Core Features & Use Cases

• Automated Training: Learns optimal weights and synergy boosts from labeled clip data.
• Score Calibration: Applies Platt scaling to ensure model predictions are well-calibrated.
• Performance Evaluation: Calculates key metrics like Precision@K and NDCG to assess model quality.
• Use Case: After gathering user feedback on generated clips, use this Skill to retrain the highlight detection model, improving its accuracy and relevance for future content.

Quick Start

Use the evaluation-ml skill to retrain the model with at least 50 samples for the gaming content type.