README.md
| 1 | --- |
| 2 | base_model: nvidia/segformer-b0-finetuned-ade-512-512 |
| 3 | library_name: transformers.js |
| 4 | pipeline_tag: image-segmentation |
| 5 | --- |
| 6 | |
| 7 | https://huggingface.co/nvidia/segformer-b0-finetuned-ade-512-512 with ONNX weights to be compatible with Transformers.js. |
| 8 | |
| 9 | ## Usage (Transformers.js) |
| 10 | |
| 11 | 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/@huggingface/transformers) using: |
| 12 | ```bash |
| 13 | npm i @huggingface/transformers |
| 14 | ``` |
| 15 | |
| 16 | **Example:** Image segmentation with `Xenova/segformer-b0-finetuned-ade-512-512`. |
| 17 | |
| 18 | ```js |
| 19 | import { pipeline } from '@huggingface/transformers'; |
| 20 | |
| 21 | // Create an image segmentation pipeline |
| 22 | const segmenter = await pipeline('image-segmentation', 'Xenova/segformer-b0-finetuned-ade-512-512'); |
| 23 | |
| 24 | // Segment an image |
| 25 | const url = 'https://huggingface.co/datasets/Xenova/transformers.js-docs/resolve/main/house.jpg'; |
| 26 | const output = await segmenter(url); |
| 27 | console.log(output) |
| 28 | // [ |
| 29 | // { |
| 30 | // score: null, |
| 31 | // label: 'wall', |
| 32 | // mask: RawImage { ... } |
| 33 | // }, |
| 34 | // { |
| 35 | // score: null, |
| 36 | // label: 'building', |
| 37 | // mask: RawImage { ... } |
| 38 | // }, |
| 39 | // ... |
| 40 | // ] |
| 41 | ``` |
| 42 | |
| 43 | You can visualize the outputs with: |
| 44 | ```js |
| 45 | for (const l of output) { |
| 46 | l.mask.save(`${l.label}.png`); |
| 47 | } |
| 48 | ``` |
| 49 | |
| 50 | --- |
| 51 | |
| 52 | 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](https://huggingface.co/docs/optimum/index) and structuring your repo like this one (with ONNX weights located in a subfolder named `onnx`). |