AWS sagemaker¶
About¶
End-to-end machine learning service that enables data scientists, developers, and machine learning experts to build, train, and deploy machine learning models quickly.
- Collect and prepare data
- Build and train machine learning models
- Deploy and monitor the performance of the predictions
Built-in algorithms (extract)¶
- Supervised algorithms
- Linear regression
- Classification
- KNN Algorithm
- Unsupervised algorithms
- Principal component analysis
- K-means
- Anomaly detection
- textual algorithms
- Natural language processing
- Summarization
- image processing
- Classification
- Detection
Model deployment e inference¶
- Deploy models in real-time or batch mode
- Monitor the performance of the deployed models
- Scale the deployed models
- Managed solution: reduce the operational overhead
- Real-time
- One prediction at a time
- Serverless
- No need to manage the infrastructure
Deployment types¶
- Real-time
- Fast
- Latency: low
- near instant predictions
- Batch
- Large datasets
- Latency: high
- High throughput
- Asynchronous
- Serverless
- No need to manage the infrastructure
- Latency: low
- Pay for what you use
- Auto-scaling
- Asynchronous
- Large datasets
- Latency: medium/high
- High throughput
- Asynchronous
SageMaker Studio¶
- Deploy ML models
- Team collaboration
- Tune and debug ML models
- Automate ML workflows
- End-to-end ML development