LocalMode
Transformers

Image Classification

Classify images into categories using Vision Transformer models.

Classify images into predefined categories using Vision Transformer (ViT) models. The model returns the top predicted labels with confidence scores.

For full API reference (classifyImage(), options, result types, and custom providers), see the Core Vision guide.

See it in action

Try the Photo Search block — it auto-classifies and organizes a photo library on-device.

ModelSizeCategoriesUse Case
Xenova/vit-base-patch16-224~86MB1000 ImageNet classesGeneral image classification

ImageNet Classes

ViT models trained on ImageNet classify into 1000 categories including animals, vehicles, food, and everyday objects. For classifying into custom categories, use Zero-Shot Image Classification with CLIP.

Composed Block

BlockDescriptionLinks
Smart GalleryAdaptive CLIP photo ingest with per-batch zero-shot categorization and a grid/list galleryLive · Install: npx shadcn add @localmode/ui/blocks/photo/smart-gallery

Next Steps

Core Vision API

Full API reference including classifyImage().

Zero-Shot Image

Classify images into custom categories without training.

Object Detection

Detect and locate objects in images.

Text-to-Speech

Generate speech audio from text with Kokoro TTS — 29 English voices, phonemizer-backed.

Zero-Shot Image Classification

Classify images into custom categories using CLIP — no fine-tuning required.

On this page

Recommended ModelsComposed BlockNext Steps