rps_tfjs_demo and rock-paper-scissors
These are competitors—both implement gesture-recognized Rock Paper Scissors games using TensorFlow.js, with the main difference being that one focuses on training the model interactively in-browser while the other uses a pre-built hand pose detection library (FingerPose) for inference.
About rps_tfjs_demo
GantMan/rps_tfjs_demo
Training a Rock Paper Scissors model in the browser via TFJS - Learn along style
Built with React and TensorFlow.js, this project captures webcam input for real-time image classification to train and play Rock Paper Scissors against a browser-based neural network. The model uses transfer learning from a pre-trained base, enabling fast training directly in the client without server-side computation or data uploads.
About rock-paper-scissors
andypotato/rock-paper-scissors
Rock, Paper, Scissors game implemented with TensorFlow.js and FingerPose
Leverages MediaPipe Hands for real-time hand tracking combined with the FingerPose library for custom gesture recognition, enabling the game to classify rock, paper, and scissors poses from webcam input. The implementation uses TensorFlow.js for in-browser inference, ensuring no server-side processing is required. Includes a webpack-based development setup with hot-reload capabilities and builds to a standalone single-page application.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work