Alibaba’s Qwen Team has released a new open-source AI model named Qwen-Image Edit that allows users to modify images using simple text instructions. The tool is capable of performing a wide range of complex editing tasks that challenge established software like Adobe Photoshop.
According to an article by Carl Franzen for VentureBeat, users can upload an image and type a prompt, such as “make the man wear a tuxedo,” to generate an edited version in seconds. The model is built on the Qwen-Image foundation model and is designed to preserve the style and content of the original picture.
Qwen-Image Edit can handle both significant semantic changes, like transforming a city photo into a Lego scene, and delicate appearance edits, such as removing a single strand of hair. A key feature is its ability to add, remove, or alter text within images in both English and Chinese while maintaining the original font and style.
The tool is available on platforms like Qwen Chat and through an API for developers. As an open-source model under the Apache 2.0 license, companies can also deploy it on their own hardware. This could offer a significant cost saving compared to proprietary software.