Alibaba Cloud Launches Tongyi Wanxiang, AI-Based Image Producer

Image Of Artificial Results Tongyi Wanxiang (ANTARA)

JAKARTA - Alibaba Cloud, the backbone of Alibaba Group's digital technology and intelligence, launched the latest artificial intelligence (AI)-based image-producing model Tongyi Wanxiang (Wanxiang means tens of thousands of images) at the World Artificial Intelligence Conference 2023.

This cutting-edge version of AI's generative model is now available to companies in China for limited beta testing.

In addition, the cloud pioneer announced the launch of the ModelScopeGPT, a versatile framework designed to assist users in completing complex and specific AI functions in various fields ranging from language, vision and conversation by utilizing various AI models in the ModelScope.

ModelScope is an open source (open-source) Model-as-a-Service (MaaS) platform introduced by Alibaba Cloud last year, which is equipped with more than 900 AI models.

"With the release of Tongyi Wanxiang, high-quality image illustrations from the generative AI will be more accessible, contributing to the development of innovative AI art and creative expression for businesses in various sectors, including e-commerce, gaming, design, and advertising," said CTO Alibaba Cloud Intelligence Jingren Zhou quoted from ANTARA, Saturday, July 8.

Tongyi Wanxiang was introduced as Image Generation. This AI-generative model is claimed to be good at carrying out various functions, responding to text orders in Mandarin and English to produce detailed images in various styles, including water paint, oil paint, and paintings to animation, sketches, flat illustrations, and 3D cartoons.

In addition, this model can convert any image into a new image with a similar style and organize the image through a force transfer process by maintaining original image material while applying another visual image style.

Supported by Alibaba Cloud's innovative technology in insight management, AI visuals, and natural language processing (NLP), this model utilizes various language materials to increase trials.

This model has qualified semantics understanding capabilities, thus producing more accurate and contextually relevant images.

In addition, by optimizing the high-resolution combinatorial process based on the signal-to-noise ratio, this model can balance between composition accuracy and detail sharpness, while still improving its performance in order to produce contrasting and visually stunning images with a clean background.

Tongyi Wanxiang was developed using Composer, Alibaba Cloud's large-scale model by providing flexibility in controlling the final results of the image, such as spatial layouts and pallets, while maintaining the synthesis quality of images and creativity.

In addition, Alibaba Cloud also launched the ModelScopeGPT, a framework designed to harness the power of the Large Language Model (Large Language Models) (LLM) available on this platform.

The ScopeGPT model will use LLMs as controllers to connect models to certain field expertise in the open-source ModelScope community.

Developed in a more complete Model-as-a-Service ecosystem, the ModelScopeGPT maximizes the various AI capabilities possessed by Alibaba Cloud.

Companies and developers can take advantage of the ModelScopeGPT for free to access and run the most appropriate models to perform complex AI tasks according to user needs, such as developing multilanguage videos.

Alibaba Cloud launched an NGO called Tongyi Qianwen in April, and plans to integrate LLM with Alibaba's various businesses to improve the quality of user experience in the future.

Subscribers and company developers will also have access to this model to create a modified AI feature with cost-effective methods.

Since the launch of this model, more than 300,000 limited beta testing requests have been received from various companies from various sectors, including fintech, electronics, transportation, fashion, and dairy products.

Tongyi Qianwen is also said to have integrated into Alibaba Cloud's intelligent assistant, Tingwu, which provides the ability for assistants to understand and analyze multimedia content with a high level of accuracy and efficiency. More than 360,000 users have been accessing this AI-powered assistant since its launch.