Home > Technology peripherals > AI > body text

The Quark large model application is the first to support the Quark network disk and dig into the usage scenarios of the photo album.

WBOY
Release: 2024-02-07 10:57:36
forward
1302 people have browsed it

2024 is widely considered to be the breakthrough year for the implementation of large model applications. Since the Quark large model was released in November last year, it has been successfully used in multiple scenarios on the Quark App through small steps and fast running. Recently, Quark Network Disk has updated several intelligent image processing tools based on Spring Festival scenes and large model technology.

Quark Netdisk’s upcoming “Spring Festival Photo Story” function aims to provide intelligent filtering and compilation services for photos taken and uploaded by users during the Spring Festival. In addition to filtering based on time and location, this function will also use people to intelligently filter and delete photos that do not meet the requirements such as being too bright or too dark. In addition, Quark Network Disk will also use AI algorithms to intelligently generate copywriting for picture collections. For example, the copywriting for a flower album can be "Flowers and the Wonderful World."

AI technology has previously been applied to intelligent search and classification in Quark Network Disk photo albums. Quark Network Disk uses AI face understanding technology to classify and store different portrait pictures, and you can find pictures of corresponding people through their avatars. In addition, Quark Network Disk can also classify pictures according to location, and further classify them according to shooting location, people, time and picture characteristics.

Quark Network Disk supports AI natural language search function. Users can quickly find pictures by entering key information such as fuzzy words and adjectives. For example, when a user searches for "sleeping cat", Quark Network Disk will list pictures that meet the conditions. In contrast, searching for the same keyword using the phone's local photo album will only show "no results."

The Quark large model application is the first to support the Quark network disk and dig into the usage scenarios of the photo album.

Search for "Sleeping Cat" in Quark Network Disk (left) and iPhone Photos App (right)

According to reports, AI natural language search The function is implemented using the technology of Quark visual self-supervised large model and graphic multi-modal large model. The core principle is to convert search terms into feature vectors, then compare them with images, and recall images that meet the requirements by calculating spatial distance. This vector encoding is based on the analysis of massive image information data across the entire network for AI training and learning. Therefore, the user's privacy data will not be infringed, and the function of intelligent image search is realized.

Quark Big Model continues to improve the experience of using the photo album function through continuous upgrading of technology. Users can intelligently search and find pictures on the Quark network disk, and can quickly edit and share pictures, including image editing and Quark quick transfer functions.

The Quark model’s positioning in the vertical field is consistent with this description. The technical director of Quark Network Disk explained that the Quark large model is an application-oriented large model specially designed for search, productivity tools and asset management assistants. The smart tools of Quark Network Disk, especially the photo album function, demonstrate the capabilities of Quark Large Model as an asset management assistant.

The above is the detailed content of The Quark large model application is the first to support the Quark network disk and dig into the usage scenarios of the photo album.. For more information, please follow other related articles on the PHP Chinese website!

Related labels:
source:jiqizhixin.com
Statement of this Website
The content of this article is voluntarily contributed by netizens, and the copyright belongs to the original author. This site does not assume corresponding legal responsibility. If you find any content suspected of plagiarism or infringement, please contact admin@php.cn
Popular Tutorials
More>
Latest Downloads
More>
Web Effects
Website Source Code
Website Materials
Front End Template