Update README.md

README.md

@@ -1,3 +1,63 @@
 ---
+library_name: transformers.js
 license: other
----
+tags:
+- mobileclip
+---
+
+https://github.com/apple/ml-mobileclip with ONNX weights to be compatible with Transformers.js.
+
+## Usage (Transformers.js)
+
+If you haven't already, you can install the [Transformers.js](https://huggingface.co/docs/transformers.js) JavaScript library from [NPM](https://www.npmjs.com/package/@xenova/transformers) using:
+```bash
+npm i @xenova/transformers
+```
+
+**Example:** Perform zero-shot image classification.
+```js
+import {
+  AutoTokenizer,
+  CLIPTextModelWithProjection,
+  AutoProcessor,
+  CLIPVisionModelWithProjection,
+  RawImage,
+  dot,
+  softmax,
+} from '@xenova/transformers';
+
+const model_id = 'Xenova/mobileclip_s0';
+
+// Load tokenizer and text model
+const tokenizer = await AutoTokenizer.from_pretrained(model_id);
+const text_model = await CLIPTextModelWithProjection.from_pretrained(model_id);
+
+// Load processor and vision model
+const processor = await AutoProcessor.from_pretrained(model_id);
+const vision_model = await CLIPVisionModelWithProjection.from_pretrained(model_id, {
+  quantized: false, // NOTE: vision model is sensitive to quantization.
+});
+
+// Run tokenization
+const texts = ['cats', 'dogs', 'birds'];
+const text_inputs = tokenizer(texts, { padding: 'max_length', truncation: true });
+
+// Compute text embeddings
+const { text_embeds } = await text_model(text_inputs);
+const normalized_text_embeds = text_embeds.normalize().tolist();
+
+// Read image and run processor
+const url = 'https://huggingface.co/datasets/Xenova/transformers.js-docs/resolve/main/cats.jpg';
+const image = await RawImage.read(url);
+const image_inputs = await processor(image);
+
+// Compute vision embeddings
+const { image_embeds } = await vision_model(image_inputs);
+const normalized_image_embeds = image_embeds.normalize().tolist();
+
+// Compute probabilities
+const probabilities = normalized_image_embeds.map(
+  x => softmax(normalized_text_embeds.map(y => 100 * dot(x, y)))
+);
+console.log(probabilities); // [[ 0.999993040175817, 0.000006828091823929405, 1.3173235896278122e-7 ]]
+```