AI Asset Manager¶
Do you see AI as a potential solution for your current challenges?¶
Are you concerned about how to manage a large number of AI models on the shop-floor, managing when and how they are deployed and updated in your production environments? Have you ever felt you need more data about your AI model execution?
AI Asset Manager is the Edge Application to support AI model management on Industrial Edge, easing model download, tracking and deployment over a large number of devices. It enables customers to upload models via the UI, via API or download them automatically from the cloud (AWS Sagemaker, AzureML, others). Moreover, AI Asset Manager provides comprehensive supervision of your AI executions by collecting, storing and aggregating relevant metrics from multiple sources.
Key features and benefits¶
Key features:
- Automated deployment from the cloud to the Edge
- Deployment of trained models to multiple Industrial Edge devices
- Orchestration of deployed models
- Tracking of AI Inference Server pipeline status
- Edge Infrastructure monitoring to understand resource constraints
- AI Model monitoring to validate model performance via custom metrics
- Control actions on running deployments on AI Inference Server (e.g. start, stop)
Benefits
- Reduced time and effort to deploy models to Industrial Edge using standard CI/CD tools and pipelines
- Manual import of AI models from user's computer or cloud file storage possible
- Maintain an overview of available devices and model deployments
- Visualize all deployments from a single management dashboard
- Keep track of inference performance, detecting HW and SW issues
- Receive automated alerts whenever your model or data are not behaving as expected
Available APIs¶
| Async APIs |
|---|
| AWS Cloud Package Delivery API |
| Azure Cloud Package Delivery API |
| Cloud API |
| OpenAPI versions |
|---|
| V2 |
| V3 |
