Instructions to use jignesh-work/vitaltools-birefnet-lite with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers.js
How to use jignesh-work/vitaltools-birefnet-lite with Transformers.js:
// npm i @huggingface/transformers import { pipeline } from '@huggingface/transformers'; // Allocate pipeline const pipe = await pipeline('image-segmentation', 'jignesh-work/vitaltools-birefnet-lite');
Pinned mirror of onnx-community/BiRefNet_lite-ONNX (fp16 ONNX only) used by vitaltools.top's Background Remover.
BiRefNet by ZhengPeng7, MIT licence. Only change: Concat nodes with more than 6 inputs (up to 1024) are split into a tree of Concats, and Split nodes with more than 6 outputs are replaced by Slice nodes, so every op fits WebGPU's limit of 10 storage buffers per shader. Output is bit-identical (checked with onnxruntime CPU).
- Downloads last month
- 30
Model tree for jignesh-work/vitaltools-birefnet-lite
Base model
ZhengPeng7/BiRefNet_lite