Recently, the 2023 OSCAR Open-Source Industry Conference successfully concluded in Beijing. At the conference, the Trusted Open-Source Large Model Industry Promotion Matrix (TWOS-LM) initiated by the China Academy of Information and Communications Technology (hereinafter referred to as the "CAICT ") was officially established. DataCanvas became one of the first member units with the powerful open-source support capabilities and flexible AI application capabilities of the DataCanvas Alaya large model, empowering the open-source large model industry to develop upwards.

• DataCanvas has been selected as one of the first batch of members of the China Academy of Information and Communications Technology's Trusted Open-Source Large Model Industry Promotion Array

• The 2023 OSCAR Open Source Industry Conference is jointly hosted by the China Academy of Information and Communications Technology and the China Communications Standardization Association, aiming to promote the exchange of excellent open-source experiences in the industry, promote the prosperity of China's open-source ecosystem construction, and promote the high-quality development of the open-source industry. At the conference, CAICT officially launched a trusted open-source big model industry promotion campaign, gathering enterprises and talents related to open-source large models, exploring hot issues such as governance and commercialization models of open-source large models, and providing empowerment of the large model ecosystem. CAICT, Alibaba Cloud, Institute of Automation, Chinese Academy of Sciences, DataCanvas and other cutting-edge AI enterprises, universities, and research institutions have become the first founding members, working together to promote the ecological construction and industrial application of open-source large models.

• As a leading domestic AI foundation software supplier, DataCanvas has always adhered to the friendly and open-source concept, empowering enterprises with a rich series of AI open-source products to build independent AI capabilities and leading the wave of digital upgrading in the AI industry. The DataCanvas Alaya Large Model Matrix, independently developed by the company, follows the Apache 2.0 protocol for global open-source. It has core advantages such as multimodal data processing and upgraded model training mechanisms. By providing enterprises with an open and flexible "general +industry" white box large model matrix, it accelerates the derivation of a series of sub models from foundational large models to vertical scenes, with low threshold efficiently assist in the widespread implementation of industry models in diversified business scenarios.

• At the same time, DataCanvas provides enterprises with a series of engineering and automated large model development toolchains, including open-source tools such as DAT automatic machine learning toolkit and DingoDB multimodal vector database, further assisting in the development governance and large-scale application of open-source large models, empowering the construction and improvement of the upstream and downstream ecological chain of open-source large models.

• With the excellent creative model of equality, openness, collaboration, and sharing, open-source technology is becoming an important driving force for empowering innovation in the application of large model industries. In the future, DataCanvas will continue to adhere to the original concept of open-source and openness, empower the standard system construction and large-scale implementation of open-source large models with independent innovative AI technology and cutting-edge experience in implementing large models, and add momentum to the prosperous development of the AI large model open source ecosystem.

•