https://github.com/zabir-nabil/vertexai-notebook-demos

Notebook demos + tutorials for Vertex AI (training, deployment, explainability, monitoring, etc.)
https://github.com/zabir-nabil/vertexai-notebook-demos

Last synced: 3 months ago
JSON representation

Notebook demos + tutorials for Vertex AI (training, deployment, explainability, monitoring, etc.)

• Host: GitHub
• URL: https://github.com/zabir-nabil/vertexai-notebook-demos
• Owner: zabir-nabil
• License: mit
• Created: 2022-12-27T06:38:32.000Z (almost 3 years ago)
• Default Branch: main
• Last Pushed: 2023-01-03T07:32:57.000Z (almost 3 years ago)
• Last Synced: 2025-03-24T15:22:12.785Z (7 months ago)
• Size: 2.93 KB
• Stars: 5
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# VertexAI-notebook-demos

Notebook demos + tutorials for Vertex AI (training, deployment, explainability, monitoring, etc.)

| **Feature / Component** | **Vertex AI**                                                                                                                                                                  |

|-------------------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|

| Data ingestion          | 👌 Monitoring jobs or SDK                                                                                                                                                      |

| Performance monitoring  | For online model performance, Vertex AI is pretty limited \(as per docs\)\. They offer only a ‘Prediction error percentage’ which is not well documented as of now\.           |

| Feature drift           | 👌 Vertex AI has reasonably good support for feature drift\. Alerts can be configured and sent periodically\.                                                                  |

| Training\-serving skew  | 🏆 Vertex AI becomes really useful when we have a continuous process from model training to deployment\. We can detect skewness in training and production data distribution\. |

| Monitors & Alerts       | 👌 Vertex AI provides basic monitoring and alerting for drift, and skew\. It’s available for the feature store too\.