Tencent releases WeVisOne-VL: one VLM for detection, segmentation, OCR and matting
_akhaliq · x · 2026-10-23
Tencent released WeVisOne-VL on Hugging Face, a single VLM for unified visual perception spanning points, boxes, masks and alpha mattes.
It handles detection, dense detection, counting, OCR, referring/reasoning/multi-target segmentation and matting, and reliably rejects absent targets (e.g. answering "none" for a blue giraffe). Pixel-level perception (segmentation & matting) and object-level perception are unified as next-token prediction from a single <OBJ> token.
Related event: Tencent Open-Sources WeVisOne-VL for Unified Visual Perception(4 posts)→
More from Multimodal
- Dev builds Figma-vs-website visual QA tool on Glimpse API, catches all 5 planted bugs — AndrewDai · 2026-10-24
- Hi3D v1.0 refines rough 3D meshes into million-face high-res geometry from an image — linoy_tsaban · 2026-10-23
- New Royal Society study maps how facial movements change how speech is interpreted — Limor_Raviv · 2026-10-23
- Tencent releases Hy-Nomas-Bench: 280 cases testing video world model controllability — _akhaliq · 2026-10-23
- Kling 4.0 hits Runware API day one: 30s 1080p video with native audio from $0.133/s — aziz4ai · 2026-10-23
- Free realistic portrait model Jib Mix Moss Realistic ships with 8-step merged LoRA — jib_reddit · 2026-10-23