Learn how to make your ML model available to end-users and optimize the inference process
What's included
7 videos6 readings3 quizzes3 app items
7 videos•Total 34 minutes
- Course Overview•4 minutes
- Introduction to Model Serving•6 minutes
- Introduction to Model Serving Infrastructure•5 minutes
- Deployment Options•3 minutes
- Improving Prediction Latency and Reducing Resource Costs•5 minutes
- Creating and deploying models to AI Prediction Platform•2 minutes
- Installing TensorFlow Serving•6 minutes
6 readings•Total 125 minutes
- Ungraded Labs - Best Practices•5 minutes
- Ungraded Lab - Introduction to Docker•20 minutes
- Optional: Build, train, and deploy an XGBoost model on Cloud AI Platform•45 minutes
- Ungraded Lab - Tensorflow Serving with Docker•20 minutes
- Ungraded Lab - Serve a model with TensorFlow Serving•30 minutes
- Lecture Notes Week 1•5 minutes
3 quizzes•Total 90 minutes
- Introduction to Model Serving•30 minutes
- Introduction to Model Serving Infrastructure•30 minutes
- TensorFlow Serving•30 minutes
3 app items•Total 130 minutes
- Ungraded Lab - Introduction to Docker (Google Cloud)•60 minutes
- [IMPORTANT] Have questions, issues or ideas? Join our Community!•10 minutes
- Ungraded Lab - Vertex AI Workbench Notebook: Qwikstart•60 minutes
Learn how to serve models and deliver batch and real-time inference results by building scalable and reliable infrastructure
What's included
8 videos8 readings6 quizzes3 app items
8 videos•Total 44 minutes
- Model Serving Architecture•4 minutes
- Model Servers: TensorFlow Serving•3 minutes
- Model Servers: Other Providers•5 minutes
- Scaling Infrastructure•10 minutes
- Online Inference•6 minutes
- Data Preprocessing•4 minutes
- Batch Inference Scenarios•5 minutes
- Batch Processing with ETL•3 minutes
8 readings•Total 195 minutes
- Documentation on model servers•10 minutes
- Ungraded Lab - Deploy a ML model with FastAPI and Docker•60 minutes
- Learn about scaling with boy bands•10 minutes
- Model Management and Deployment Infrastructure•30 minutes
- 2 app items•Total 180 minutes
- TFX on Google Cloud Vertex Pipelines•90 minutes
- Implementing Canary Releases of TensorFlow Model Deployments with Kubernetes and Anthos Service Mesh•90 minutes
1 ungraded lab•Total 45 minutes
- Ungraded Lab: Developing TFX Custom Components•45 minutes
- Establish procedures to detect model decay and prevent reduced accuracy in a continuously operating production system
- What's included
- 13 videos9 readings3 quizzes1 app item
- 13 videos•Total 66 minutes
- Why Monitoring Matters•6 minutes
Observability in ML•4 minutes
- Monitoring Targets in ML•4 minutes
- Logging for ML Monitoring•7 minutes
- Tracing for ML Systems•3 minutes
What is Model Decay?•3 minutes
Model Decay Detection•2 minutes
Ways to Mitigate Model Decay•5 minutes
Responsible AI•5 minutes
- Legal Requirements for Secure and Private AI•9 minutes
- Anonymization and Pseudonymisation •4 minutes
- Right to be Forgotten•8 minutes
- Specialization recap and farewell•1 minute
- 9 readings•Total 77 minutes
- Monitoring Machine Learning Models in Production•10 minutes
- [IMPORTANT] Reminder about end of access to Lab Notebooks•2 minutes
- Addressing Model Decay•10 minutes
- Responsible AI•10 minutes
GDPR and CCPA•10 minutes
- Lecture Notes Week 4•5 minutes
- Course 4 Optional References•10 minutes
- Acknowledgements•10 minutes
- (Optional) Opportunity to Mentor Other Learners•10 minutes
- 3 quizzes•Total 90 minutes
- Model Monitoring and Logging•30 minutes
- Model Decay•30 minutes
- GDPR and Privacy•30 minutes
- 1 app item•Total 120 minutes
- Data Loss Prevention: Qwik Start - JSON•120 minutes
3 quizzes•Total 90 minutes
- ML Experiments Management and Workflow Automation•30 minutes
- MLOps Methodology•30 minutes
- Model Management and Deployment Infrastructure•30 minutes
2 app items•Total 180 minutes
- TFX on Google Cloud Vertex Pipelines•90 minutes
- Implementing Canary Releases of TensorFlow Model Deployments with Kubernetes and Anthos Service Mesh•90 minutes
1 ungraded lab•Total 45 minutes
- Ungraded Lab: Developing TFX Custom Components•45 minutes
Establish procedures to detect model decay and prevent reduced accuracy in a continuously operating production system
What's included
13 videos9 readings3 quizzes1 app item
13 videos•Total 66 minutes
- Why Monitoring Matters•6 minutes
- Observability in ML•4 minutes
- Monitoring Targets in ML•4 minutes
- Logging for ML Monitoring•7 minutes
- Tracing for ML Systems•3 minutes
- What is Model Decay?•3 minutes
- Model Decay Detection•2 minutes
- Ways to Mitigate Model Decay•5 minutes
- Responsible AI•5 minutes
- Legal Requirements for Secure and Private AI•9 minutes
- Anonymization and Pseudonymisation •4 minutes
- Right to be Forgotten•8 minutes
- Specialization recap and farewell•1 minute
9 readings•Total 77 minutes
- Monitoring Machine Learning Models in Production•10 minutes
- [IMPORTANT] Reminder about end of access to Lab Notebooks•2 minutes
- Addressing Model Decay•10 minutes
- Responsible AI•10 minutes
- GDPR and CCPA•10 minutes
- Lecture Notes Week 4•5 minutes
- Course 4 Optional References•10 minutes
- Acknowledgements•10 minutes
- (Optional) Opportunity to Mentor Other Learners•10 minutes
3 quizzes•Total 90 minutes
- Model Monitoring and Logging•30 minutes
- Model Decay•30 minutes
- GDPR and Privacy•30 minutes
1 app item•Total 120 minutes
- Data Loss Prevention: Qwik Start - JSON•120 minutes
