Model weights used by BG Studio — on-device background remover (https://bg-studio.skillsafe.ai/)

This app runs 2 models on your device. About 16.7 MB of weights are downloaded from https://models.skillsafe.ai/ on first use and cached by your browser; inference happens locally and nothing you enter is uploaded to load them.

- U2-Net-p background removal (rembg) (u2netp@cd3a3d67, fp32) — 4.4 MB — Apache-2.0
  U2-Net: Going Deeper with Nested U-Structure for Salient Object Detection (Qin et al.), lightweight u2netp weights as distributed by rembg; mirrored on Hugging Face by tomjackson2023. Licensed under the Apache License 2.0.
  Notice: https://huggingface.co/tomjackson2023/rembg
  Source: https://huggingface.co/tomjackson2023/rembg/tree/cd3a3d6767a7859efea31ef0f2f373582cf06d82
- MODNet portrait matting (modnet@fa2fa546, fp16) — 12.4 MB — Apache-2.0
  MODNet: Trimap-Free Portrait Matting in Real Time (Ke et al.). ONNX fp16 export published by Xenova on Hugging Face. Licensed under the Apache License 2.0.
  Notice: https://huggingface.co/Xenova/modnet
  Source: https://huggingface.co/Xenova/modnet/tree/fa2fa546052fba4c08921230a26cc69a333fca12

Catalogue: https://api.skillsafe.ai/v1/browser-models