Urgent.News

What's breaking now, across thousands of outlets.

AI

Grass Bingo: a daily nature card your phone checks offline, with an open-weight model

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass What I Built Grass Bingo is a daily 4x4 bingo card of small things to find outside: a pinecone, a red leaf, moss, a snail, a cloud shaped like something. You tap a square, take a photo, and an open-weight CLIP model running in your browser checks whether the photo matches. Get four in a row and you have a…

Grass Bingo is a daily 4x4 nature bingo card app that encourages users to explore the outdoors. The app displays a bingo card with 16 small items such as pinecones, red leaves, moss, snails, and clouds. Users tap a square, take a photo, and an open-weight CLIP model runs in their browser to check if the photo matches the item. Four in a row results in a bingo.

The app is designed for anyone who wants a simple reason to go for a walk, including kids, parents, and busy professionals. The same card is provided to all users on the same day, allowing them to compare with friends without needing an account.

The app is built using open-weight CLIP model (Xenova/clip-vit-base-patch32) and Transformers.js, which allows the model to run entirely on the user's device without requiring any backend or server. This ensures that users' photos never leave their phone and are not uploaded or stored anywhere. The app is optimized for offline use, as it caches the model and app shell, enabling it to function without an internet connection.

Users can track their progress, share a text-only summary, and even use the demo mode by accessing the app with the '?demo=1' parameter.

The project was developed as part of the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass. The code is open-source and available on GitHub, allowing anyone to modify the labels, threshold, or model to create their own custom bingo card. The app has undergone extensive testing and is ready for use.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in AI

Train an 8B LLM on a 4GB Laptop GPU: Hands-On with Soup

Train an 8B LLM on a 4GB Laptop GPU: Hands-On with Soup TL;DR Soup eliminates the high VRAM barrier of LLM fine-tuning by introducing layer streaming, enabling you to fine-tune an 8B parameter model…

  • Soup model trains 8B LLM on 4GB laptop GPU using layer streaming
  • Declarative YAML config specifies training parameters like model, batch size, learning rate
  • Supports 4-bit and 8-bit quantization for model compression and speed

More from Saturday 10 October →