明略科技
www.mininglamp.com
2
Leaving SiteNav
External Link Disclaimer
You are about to visit www.mininglamp.com. This website is not operated by us. We are not responsible for its content or privacy practices.
About this website
Mininglamp Technology is a Chinese enterprise-grade data intelligence software provider with over 19 years of deep engagement in the enterprise service sector. The company focuses on helping organizations unlock business value from data, particularly in marketing and operational scenarios during digital transformation. Mininglamp specializes in multimodal data intelligence, enterprise knowledge graphs, and data privacy technologies, enabling businesses to process and analyze complex, heterogeneous data sources at scale. One of the company’s flagship offerings is Octo, an open-source, trusted agent collaboration network designed for both human and AI agent cooperation. Octo functions as a work platform where multiple AI agents can operate simultaneously in a shared chat group. Users pose a question, and the agents—each with distinct analytical perspectives—collaborate in real time, cross-checking each other’s reasoning, correcting errors, and synthesizing answers. The final output is presented for human evaluation, reducing manual oversight while improving response accuracy and reliability. This approach mimics a dynamic, multi-expert discussion rather than relying on a single monolithic AI model. In addition to Octo, Mininglamp has recently open-sourced two key technical components: Cider, an on-device inference acceleration framework, and Mano-P, an end-side GUI agent model. Cider is a lightweight SDK that optimizes computation operators and hardware-level mechanisms, allowing large language models to run efficiently on macOS local hardware with lower memory consumption and higher throughput. Mano-P is a compact model designed for graphical user interface (GUI) automation on end devices. Combined with the earlier open-sourced Mano-CUA skill set, Mano-P demonstrates pra
Tags & Categories
Categories
Tags
Statistics
2
Views
0
Clicks
0
Like
0
Dislike