OPPO’s first huge AI model
On December 27, Kuai Technology said that OPPO held a technical communication conference for the Find X7 series today, citing OPPO official news. The Andes big model AndesGPT, OPPO’s first AI large model, was formally unveiled on this occasion.
According to the official introduction, AndesGPT has a comprehensive and powerful generative visual model that can accurately understand the semantics of images and, as a result, can support the recognition and segmentation of more than 120 different types of subjects.
Not only does the segmentation accuracy reach the hairline level, but it can also support the segmentation and extraction of up to 6 subjects, fill and generate images of very large areas, and perfectly generate natural light and shadow in pictures.
The local image generation time is as low as 6 seconds, which is 60% faster than other models on the same platform, allowing the client to achieve generation accuracy and quality comparable to those of the cloud.
Furthermore, the AndesGPT 7 billion parameter language model has more “intelligent” comprehension capabilities than the 1 billion parameter huge model.
Using the call summary function as an example, AndesGPT can assist users in sorting out the essential points of the call content and properly generate a call summary containing subjects, key points, and to-do items. The precision is far superior to the 1 billion-level device-side model.
source: Kuai Technology
Pingback: OPPO's satellite communications are supported by the Find X7 series -