Tencent Releases WeVisOne-VL-2B for Unified Visual Perception
Tencent’s WeChat Vision released WeVisOne-VL-2B, a 2B-parameter image-text-to-text model built on Qwen3-VL and licensed under Apache 2.0. According to the model card, it combines object-level tasks such as open-vocabulary detection, referring localization, counting, pointing and OCR with pixel-level referring segmentation and promptable alpha matting. The card also warns that fine alpha boundaries can be sensitive to image resolution, occlusion, transparency and prompt phrasing.
The release brings object detection, grounding and pixel-level segmentation into one compact vision-language model, but its advertised capabilities are vendor-stated and have not been independently verified in the supplied reports.
tencent/WeVisOne-VL-2B Image-Text-to-Text • 2B • Updated 3 days ago • 8.82k • 64
license: apache-2.0 pipeline_tag: image-text-to-text library_name: transformers tags:
- vision-language
- object-detection
- referring-expression-comprehension
- object-counting
- point-grounding
- ocr
- referring-expression-segmentation
- image-matting
WeVisOne-VL: Unified Visual Perception from Objects to Pixels
Paper: WeVisOne-VL: Unified Visual Perception from Objects to Pixels | Project Page: WeVisOne-VL