Application Scenarios
The Atlas data center training solution is mainly used in enterprise data centers or supercomputing center equipment rooms. The solution consists of software and hardware and provides end-to-end AI training solutions for customers in different application scenarios. Table 3-1 describes the application scenarios.
Industry |
Scenario |
Description |
Form Factor |
|---|---|---|---|
Traditional industry |
Quality detection, OCR, and deep learning system (DLS) |
A customer does not have an in-house DLS. The customer needs an easy-to-use DLS that integrates software and hardware. |
Atlas 800–based software-hardware integrated solution |
Internet/Security protection industry |
Content moderation, video and image analysis, and DLS |
A customer has an in-house DLS. The customer needs open-source plugins that can seamlessly adapt to the customer's DLS to quickly roll out training services of Ascend AI Processors. |
|
Supercomputing center and public cloud industry |
Cloud services for scientific research and AI training |
A customer does not have an AI training cluster. The customer needs large-scale AI training clusters supporting ultra-high-density deployment, and full-cabinet delivery to shorten the project delivery period, accelerate service rollout, and reduce installation, deployment, and commissioning costs. |
Atlas 900–based cluster delivery solution |
Other industries |
Algorithm research |
CLI-based AI training and algorithm research The customer needs a distributed training CLI tool to simplify the distributed cluster training mode and improve usability. |
Open hardware ecosystem based on the Atlas 300T 9000 training card |
