news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

According to Fast Technology on August 16, the Global Open Source Technology Summit (GOTC2024), jointly organized by Open Source China and Shanghai Pudong Software Park, opened in Shanghai yesterday.

Zhang Qunhui, chief expert of Huawei's data storage software, was invited to this summit to share the practical application of ModelEngine in the field of vertical large models.

According to reports, ModelEngine is the full-process tool chain for AI training and promotion in Huawei's DCS (Data Center Stack), covering everything from data processing and knowledge generation to model fine-tuning and deployment, as well as RAG application development and application deployment. It aims to provide data processing engineers, model engineers, and application developers with an efficient, easy-to-use, open, flexible, and out-of-the-box full-process AI development experience.

Zhang Qunhui further pointed out that ModelEngine is based on open data-enabling, model-enabling and application-enabling tool chains, and supports open data processing operators, mainstream models and open application operators.This openness reduces the time required to generate corpus from several months to just a few days.

For developers, ModelEngine provides one-stop AI application development, evaluation, optimization, and deployment services, enabling rapid construction of high-accuracy AI applications.In addition, ModelEngine also supports plug-ins such as LangChain and LlamaIndex to protect existing AI assets.

Looking ahead, ModelEngine will continue to enhance its scalability and customization capabilities, optimize performance and efficiency, and provide more application examples to meet the diverse needs of developers in different scenarios.

Zhang Qunhui, Chief Expert of Huawei Data Storage Software