Breed Identifier

Vision Transformer (ViT)

This tool uses a Vision Transformer (ViT) image classifier loaded via Transformers.js, which brings HuggingFace models directly to the browser using WebAssembly. The model was trained on ImageNet, which includes 120+ dog breeds and several dozen cat breeds.

Image Classification Pipeline

When you click "Identify Breed", the tool:

1. Loads the ViT model weights from HuggingFace CDN (first run only; then cached in your browser).
2. Resizes and normalises your image to 224x224 pixels.
3. Runs a forward pass through the neural network to produce a probability distribution over 1,000 ImageNet classes.
4. Filters results to dog and cat breed classes and shows the top 5 by confidence.

Confidence Scores

Each score is the softmax probability the model assigned to that breed class. A score near 100% means the model is very certain; a spread across several breeds suggests a mixed heritage or ambiguous photo.

Privacy

Your photo is processed entirely in your browser using WebAssembly. It is never uploaded to any server.

You can embed this util on your own site as a widget. Adding ?embed=1 to the URL loads a compact version with just the tool itself; no header, menu, or documentation. Paste this snippet into your HTML:

Copy snippet Adjust the height to taste.