ModelVS1-VL
VS1 VL
A multi-task vision model open-sourced by the WeChat Vision team that integrates object detection, OCR, segmentation, and matting into a single model, with built-in high-accuracy rejection capability.
Timeline (last 90 days)
Nov 4
The VS1 VL model open-sourced by WeChat's vision team integrates object detection, OCR, segmentation, and matting, with a rejection accuracy of 96.4%.
Unverified50%