Tencent Releases WeVisOne-VL, One Model for Many Image Tasks
Tencent's WeChat Vision team has released WeVisOne-VL, an open-licence vision model in two sizes, 2B and 4B. Instead of needing separate tools, it can spot objects, read text in pictures, count items, cut out subjects from backgrounds, and handle fine edge separation in one system.
Quick glossaryTencent, vision language model, alpha matting, OCR
- Tencent
- A large Chinese technology company best known for the WeChat messaging app and popular online games.
- vision language model
- An AI model that understands both pictures and written words, and can answer questions or follow instructions about images.
- alpha matting
- A technique that cuts a subject out of its background while keeping soft, fine edges such as hair intact.
- OCR
- Software that reads printed or handwritten words from a photo or scanned page and turns them into editable text.
Burning Potato score: 63%