This week, the Alibaba Cloud Tongyi Wansiang team launched a brand new image editing model, ACE, aimed at providing users with a more convenient and intelligent image generation and editing service. Users can generate or modify images simply by using straightforward spoken commands, significantly simplifying the complexity of image editing. This tool supports a wide range of application scenarios, including stylized portraits, storyboard creation, and interior design.
The ACE model not only supports text-to-image generation but also features powerful image editing capabilities. Users can perform controllable visual editing, element modification, area repainting, and layered editing tasks through conversational input. For instance, users can easily achieve Photoshop-like functionalities by entering commands such as "change the background of the ID photo" or "remove watermark with one click." Notably, the local stylization feature of ACE has already been launched in the Tongyi App, further enhancing user experience.
According to official sources, the core innovation of the ACE model lies in its unique Long-context Condition Unit (LCU) module. The LCU supports multimodal conditional input to meet the needs of various general editing tasks and establishes a complete editing data construction link and instruction set generation link, ensuring the precision and high quality of image editing results.
The launch of ACE not only broadens the application of AI in the field of image editing but also provides open-source support for developers worldwide, marking a further breakthrough for Alibaba Cloud in the realm of intelligent image processing technology. This new tool will greatly enhance the efficiency and convenience of image processing, offering users a more creative and personalized editing experience.
Open-source address:
https://github.com/ali-vilab/ACE
Experience address:
https://huggingface.co/spaces/scepter-studio/ACE-Chat
Magic Dock Community:
https://www.modelscope.cn/models/iic/ACE-0.6B-1024px
