Show HN: I made a computer vision tool for evaluating deadlift form (github.com)

A new computer vision tool has been launched to evaluate deadlift form using AI, which counts repetitions from side-on video clips, records pull timing from the barbell, and assesses back posture as either straight or rounded. The tool utilizes three distinct models: ViTPose for measuring hip angles, SAM 3.1 for tracking bar height and counts, and another model to analyze back position in every frame. This tool provides live probability metrics alongside the video, enhancing the feedback loop for athletes and trainers. This innovation is significant for the AI/ML community as it represents a practical application of machine learning in sports science, promoting better training practices and injury prevention among weightlifters. The software's design includes features like adjustable settings for precise rep counting and posture evaluation, leveraging the capabilities of modern computer vision techniques and user-friendly deployment with a simple API key setup. By allowing users to run analysis directly from their own videos, it encourages wider accessibility of advanced performance analysis tools.