This topic describes the use scenarios, customer requirements, architecture, and references for implementing elastic computing-based AI inference.
Use scenarios
You can use elastic computing-based AI inference in Container Service for Kubernetes (ACK) clusters that contain GPU-accelerated nodes. You must use trained models for AI inference. Facial recognition payment is an application of AI inference. Other applications of AI inference include image classification, object detection, speech recognition, and semantic analysis.
Customer requirements
Build environments for AI inference on GPU-accelerated Elastic Compute Service (ECS) instances
Build environments for AI inference in ACK clusters
Store model data in File Storage NAS (NAS) file systems
Use Apsara AI acceleration tools to accelerate AI inference
Architecture
