Beautiful picturesThe company disclosed its 2023 performance report yesterday, achieving total revenue of 2.7 billion yuan for the whole year, a year-on-year increase of 29.3%. The adjusted net profit attributable to the equity holders of the parent company was 370 million yuan, a year-on-year increase of 233.2%.
According to Sina Technology, Wu Xinhong, founder, chairman and CEO of Meitu, said at the performance meeting that AI native workflow will be a huge opportunity in the future, and Meitu will focus on AI native applicationsAt the same time, he also emphasized thatMeitu will not make universal large modelsHowever, Meitu will train and fine-tune the model according to the needs of different vertical scenarios.
Wu Xinhong revealed,Sora This has given the industry a great psychological shock. Originally, it was thought that such capabilities would only be realized in 2-3 years.Sora also made Meitu reflect on the underlying architecture of the large visual model, urgently upgraded the underlying framework of the model, and reinvested in the training of image and video models. He believes that there is still a window of about 2 years for vertical scenarios to support innovative AI native applications.
OpenAI Sora The model can produce a one-minute video strictly based on the prompt words entered by the user, maintaining high visual quality. The model can deeply simulate the real physical world, marking a major leap in the ability of artificial intelligence to understand and interact with real-world scenes.
Mira Murati, OpenAI’s chief technology officer, told the Wall Street Journal:Vincent video tool Sora will be released to the public in a few months, intends to support audio generation.
In recent years, Meitu has also been seeking transformation and launched the 4.0 version of the MiracleVision AI model. The model is currently being gradually applied to Meitu products such as Meitu XiuXiu, Beauty Camera, Wink, Meitu Design Studio, and WHEE. The model has various functions such as "vector graphics", "text effects", "intelligent layering", "intelligent typesetting", and using prompt words and pictures to generate videos.