Alibaba Cloud has introduced its latest AI image generation model, Tongyi Wanxiang (‘Wanxiang’ means ‘tens of thousands of images’). This model is designed to help businesses improve their creativity by generating high-quality images across different art styles.

Text-to-image generation

The company states that Tongyi Wanxiang is able to generate detailed images in response to text prompts in both Chinese and English. This includes watercolours, oil paintings, sketches, animations, and more. Additionally, It can also transform existing images into new ones with similar styles and apply style transfers to create visually appealing compositions..

Tongyi Wanxiang is powered by Alibaba Cloud’s advanced technologies and leverages multilingual materials for enhanced training. This AI image generation model optimizes the high-resolution diffusion process to strike a balance between composition accuracy, detail sharpness, and clean backgrounds.

Alibaba Cloud developed it using Composer, the company’s proprietary large model that offers greater control over the final image output, including spatial layout and colour palette.

ModelScopeGPT Launched for Sophisticated AI Tasks

The company has also launched ModelScopeGPT, a versatile framework that utilizes Large Language Models (LLMs) to accomplish complex AI tasks across language, vision, and speech domains. ModelScopeGPT connects with an extensive array of domain-specific expert models, allowing businesses and developers to access and execute the best-suited models for sophisticated AI tasks. The platform is available for free and offers a rich Model-as-a-Service (MaaS) ecosystem.