Face Wall Intelligence announced that its "Little Steel Cannon" MiniCPM series models are free for commercial use. This move will have a profound impact on academic research and commercial applications. The MiniCPM and MiniCPM-V models are now open for free commercial use to academic researchers and corporate individuals. This marks a major contribution from the open source community and reflects the determination of Wallface Intelligence, OpenBMB and Tsinghua NLP Laboratory to promote knowledge sharing and technological progress. This initiative aims to promote innovative development in the field of artificial intelligence and provide more users with powerful AI tools.
Wall-facing Intelligence announced that as a contributor and beneficiary of the open source community, Wall-facing Intelligence, OpenBMB & Tsinghua NLP Laboratory have carefully discussed and decided to commercialize MiniCPM for free. Including MiniCPM and MiniCPM-V models, they are open to academic researchers and corporate individuals for free commercial use from now on. This initiative will undoubtedly promote innovation and development in academia and business.
Both academic and commercial openings promote knowledge sharing and technological progress.
The weights of the MiniCPM series models will be completely open to academic researchers to promote the sharing of knowledge and the advancement of academic research. At the same time, companies and individuals can also start using these models commercially after completing a simple questionnaire registration. The open strategy not only reflects support for academic research, but also encourages business innovation.
Comply with the license agreement and ensure compliant use
It is worth noting that when community members use the MiniCPM series models, they need to comply with the Apache2.0 License Agreement and the "MiniCPM Model Community License Agreement". This ensures the compliance of model use and protects the rights of developers and users.
MiniCPM-Llama3-V2.5: Excellent performance, end-side multi-modal acceleration
MiniCPM-Llama3-V2.5 takes center stage with its excellent performance. This model surpasses Gemini Pro and GPT-4V in terms of end-side multi-modal comprehensive performance. In particular, it has reached the SOTA (State of the Art) level in terms of OCR capabilities and can accurately identify long texts in complex images.
High energy efficiency and multi-language support
Another highlight of MiniCPM-Llama3-V2.5 is its high energy efficiency. When running on the mobile phone, only 8GB of video memory is needed, and with the NVIDIA GeForce RTX4070 graphics card, rapid inference can be achieved and the running speed is extremely fast. In addition, the image encoding speed is increased by 150 times, achieving end-side system-level multi-modal acceleration for the first time. What’s even more impressive is that the model supports more than 30 languages, providing strong support for multi-language environments.
MiniCPM Series Models: Drivers of Innovation
With the free commercial use of the MiniCPM series models, we look forward to seeing more innovative applications being born. Whether it is academic research or commercial application, the MiniCPM series models will serve as a promoter of innovation and have a profound impact on all walks of life.
The free commercial use of MiniCPM series models will undoubtedly accelerate the application of artificial intelligence technology in various fields, promote technological progress and industrial upgrading, and it is worth looking forward to its future development.