Huawei Cloud Launches Industry's First Large Model Hybrid Cloud, Providing Full Scenario Cross-Cloud Deployment Capabilities

todayHuawei CloudIndustry Summit 2023 on.Huawei (brand)Cloud relaunched the industry's first big model capability for hybrid cloud-based deployments, releasing the "Deep Cloud Vision 2025" whitepaper and Deep Cloud Action Plan. The release of the Big Model Hybrid Cloud is designed to provide the industry's most completeAIThe production chain helps government and enterprise customers to establish exclusive big model capability in one-stop.

Huawei Cloud Launches Industry's First Large Model Hybrid Cloud, Providing Full Scenario Cross-Cloud Deployment Capabilities

Huawei (brand)CloudStack 8.3 is the first in the industry to realize the big model capability based on hybrid cloud deployment, extending the big model from local to edge and public cloud through multi-cloud collaboration architecture. This cross-cloud deployment solution can give full play to the advantages of each cloud platform and realize intelligent applications in all scenarios.

For large model scenarios, Huawei Cloud Stack performs four major optimizations:

  • With native hybrid cloud-based capabilities, users can extend big models from local to edge and public clouds, enabling full-scenario cross-cloud deployment. This multi-cloud collaborative architecture allows industry big models to be trained on the public cloud, fine-tuned with enterprise local data on the hybrid cloud, and then pushed to the edge cloud for inference, thus better meeting the needs of different scenarios.
  • The optimization of arithmetic fusion and mixing accuracy is achieved through software-hardware collaboration, and the model training performance is improved by 451 TP3 T. This optimization enables large model training tasks to be completed faster, saving time and resources for the enterprise.
  • Huawei Cloud Stack provides 3 types of tool suites, including tools for data cleansing, model development, and application development, to help organizations efficiently complete the full process of large model training and application development. These tool suites lower the development threshold, enabling more enterprises to complete the development and application of large models on their own.
  • While large model training tasks usually need to be executed continuously for several days, Huawei Cloud Stack enables AI training tasks to run long and steadily for more than 30 days through the ability of sensorless breakpoint renewal. This capability ensures the stability and continuity of large model training tasks, providing enterprises with a more reliable large model training solution.

Currently, Huawei Cloud Stack has joined forces with more than 10 government and enterprise customers to explore industry big models based on hybrid cloud. These customers have realized more efficient data processing, model training, and application development through Huawei Cloud Stack's big model hybrid cloud platform. Meanwhile, Huawei Cloud Stack also provides professional technical support and services to help customers better apply big model technology to solve practical problems.

Huawei Cloud Stack's Big Model Hybrid Cloud Platform has the industry's most complete AI production chain, helping government and enterprise customers establish exclusive Big Model capabilities in a one-stop shop. This hybrid cloud deployment approach makes big model technology closer to the actual needs of enterprises, providing smarter and more efficient solutions for various industries. In the future, withartificial intelligence (AI)Technology continues to evolve, and large model hybrid cloud platforms will become the preferred solution for more and more organizations.

This article comes from users or anonymous contributions, does not represent the position of Mass Intelligence; all content (including images, videos, etc.) in this article are copyrighted by the original author. Please refer to this site for the relevant issues involvedstatement denying or limiting responsibilityPlease contact the operator of this website for any infringement of rights (Contact Us) We will handle this as stated. Link to this article: https://dzzn.com/en/2023/1850.html

Like (1)
Previous December 1, 2023 at 12:38 pm
Next December 1, 2023 at 12:42 pm

Recommended

Leave a Reply

Please Login to Comment