Fast technology news on December 27, according to OPPO official news, OPPO held a find X7 series technical communication meeting today, at this meeting, OPPO's first AI large model - Andean large model AndesGPT was officially released.
According to the official introduction, AndesGPT has a comprehensive and powerful generative vision model, which can accurately understand the semantics of images, and on this basis, it realizes the recognition and segmentation support of more than 120 types of subjects.
Not only is the segmentation accuracy as high as the hairline, but it can also support the segmentation and extraction of up to 6 subjects, and can fill and generate images for large areas, and can even perfectly generate natural light and shadow in **.
The local generation time of images is as low as 6s, which is 60% faster than other models on the same platform, so that the generation accuracy and quality comparable to the cloud can also be achieved on the device side.
In addition, compared with the 1 billion parameter large model, the AndesGPT 7 billion parameter language model brings more "IQ" comprehension ability.
Taking the call summary function as an example, AndesGPT can help users sort out the key points of the call content and accurately output the call content summary with topics, key points, and to-do items, which is significantly more accurate than the 1 billion-level device-side model.