MNIST-SageMaker

Build, Train, Deploy, and Query an MNIST Image Classifier on AWS

Steps

1. Hyperparameter Optimization.

• Objective: Minimize Test Loss.
• Perfom hyperparameter search that satisfies this objective.

2. Training.

• Train a custom CNN based model using the best hyperparameters from HPO step above.

3. Deployment.

• Method 1 (Basic Model Loader then Query in Notebook)

• Method 2 (Extensive Process then Query with a Lambda Function). Tagged as _v2 in the code.

Built with ❤️ by martinoywa