Sam Vit Base

🚀 Segment Anything Model (SAM) with Transformers.js

This project adapts the Segment Anything Model (SAM) with ONNX weights to be compatible with Transformers.js, enabling seamless image segmentation in the browser.

🚀 Quick Start

This repository contains the ONNX weights of the Segment Anything Model (SAM) to make it compatible with Transformers.js.

📦 Installation

If you haven't already, you can install the Transformers.js JavaScript library from NPM using:

npm i @huggingface/transformers

💻 Usage Examples

Basic Usage

Perform mask generation with Xenova/sam-vit-base.

import { SamModel, AutoProcessor, RawImage } from "@huggingface/transformers";

// Load model and processor
const model = await SamModel.from_pretrained("Xenova/sam-vit-base");
const processor = await AutoProcessor.from_pretrained("Xenova/sam-vit-base");

// Prepare image and input points
const img_url = "https://huggingface.co/datasets/Xenova/transformers.js-docs/resolve/main/corgi.jpg";
const raw_image = await RawImage.read(img_url);
const input_points = [[[340, 250]]];

// Process inputs and perform mask generation
const inputs = await processor(raw_image, { input_points });
const outputs = await model(inputs);

// Post-process masks
const masks = await processor.post_process_masks(outputs.pred_masks, inputs.original_sizes, inputs.reshaped_input_sizes);
console.log(masks);
// [
//   Tensor {
//     dims: [ 1, 3, 410, 614 ],
//     type: 'bool',
//     data: Uint8Array(755220) [ ... ],
//     size: 755220
//   }
// ]
const scores = outputs.iou_scores;
console.log(scores);
// Tensor {
//   dims: [ 1, 1, 3 ],
//   type: 'float32',
//   data: Float32Array(3) [
//     0.9466127157211304,
//     0.9890615344047546,
//     0.8316894769668579
//   ],
//   size: 3
// }

Advanced Usage

You can then visualize the generated mask with:

const image = RawImage.fromTensor(masks[0][0].mul(255));
image.save('mask.png');

Next, select the channel with the highest IoU score, which in this case is the second (green) channel. Intersecting this with the original image gives us an isolated version of the subject:

📚 Documentation

Demo

We've also got an online demo, which you can try out here.

⚠️ Important Note

Having a separate repo for ONNX weights is intended to be a temporary solution until WebML gains more traction. If you would like to make your models web-ready, we recommend converting to ONNX using 🤗 Optimum and structuring your repo like this one (with ONNX weights located in a subfolder named onnx).