[Global Network Technology Report] On July 7, at the 2023 World Artificial Intelligence Conference, Alibaba Cloud announced that Tongyi Wanxiang, a large AI painting creation model, has opened targeted invitation testing. This model can assist humans in picture creation, and can be used in the future. It is used in application scenarios such as art design, e-commerce, games, and cultural creation. Zhou Jingren, CTO of Alibaba Cloud Intelligence Group, said that this is a key step for Alibaba Cloud's large model to fully master multi-modal capabilities, which will gradually be opened to industry customers.
Alibaba Cloud announced that the large-scale AI painting creation model Tongyi Wanxiang has opened a targeted invitation test
According to reports, Tongyi Wanxiang has excellent performance in both semantic understanding and image generation. The first batch of three major capabilities are online: in the basic Wensheng diagram function, watercolors, flat illustrations, two-dimensional illustrations, oil paintings, etc. can be generated based on text content. Images in styles such as Chinese paintings, 3D cartoons and sketches; in the similar image generation function, after users upload any image, they can diversify their creativity and generate AI paintings with similar content and style; this model is also the first in the industry to support image style migration, allowing users Upload the original image and style image, and the original image can be automatically processed into the specified style image.
Tongyi Wanxiang is based on Composer, a combined generation model developed by Alibaba. It can disassemble and combine image design elements such as color matching, layout, style, etc. to provide highly controllable and extremely free image generation effects. , only one model can support multi-type image generation task capabilities. The research results have been published at ICML 2023, the top international AI conference.
Alibaba Cloud demonstrated the effect of the model on the generated picture. When inputting "Immerse yourself in the charming world of full-page grayscale coloring, there is a tiger in the tranquil forest of mandala, and the picture uses lines and strokes." Afterwards, the system can quickly generate tiger pictures with rich line details. If you continue to use the style migration function, you can also switch picture styles at will.
At last year’s World Artificial Intelligence Conference, Alibaba Cloud’s “Tongyi” large model series made its debut. After that, models and products such as Tongyi Qianwen and Tongyi Tingwu were launched one after another, and Tongyi Wanxiang was launched. This means that Alibaba Cloud has the ability to process or generate modalities such as text, voice, and pictures in the field of large models. At present, Alibaba Cloud's large model is gradually being used in various industries. More than 300,000 companies have applied to test Tongyi Qianwen, and the cumulative number of Tongyi Tingwu users has reached 360,000.
We believe that multi-modality is an inevitable trend in the development of large-scale models. We hope to provide services to all walks of life by leveraging the capabilities of different modalities and help enterprises fully embrace the intelligent era. " Zhou Jingren said.
The above is the detailed content of Alibaba Cloud AI Painting Creation Large Model Tongyi Wanxiang Released Directed Invitation Testing Opened. For more information, please follow other related articles on the PHP Chinese website!