README.md

Asset Prediction Sagemaker Pipeline Example

• For quick start guide, see quickstart guide.

Services, features of the project

• Asset data
• Asset importer
• Model Training Concepts
• Model Training Pipeline
• System events
• User interface
• Security
• Cost estimation
• Uninstall

---

End to end flow

1. Create asset entries and upload CSVs through the UI
2. Create a model training template and set all the parameters
3. Create a model training execution and set the template to use
4. Trigger model training with the Start model training button
5. Lambda function will call startPipelineExecution with the right parameters
6. Processing step performs the feature engineering step, stores features/test/training data in S3
7. Training step trains the model
8. Model gets created and registered with Sagemaker
9. Create model endpoint (and config) from the UI using the Create endpoint button
10. Run inference against model with parameters set on the UI
11. Analyze output from inference on a chart in the UI
12. Delete model endpoint manually or leave it and it will be automatically cleaned up after 60 minutes

---