In this session, you will learn how to build a serverless image classification with deep learning from scratch and how to deploy it on AWS with the serverless framework. In the first part, we will create the lambda functions that store, resize and classify the uploaded image. We will persist the result of the classification as metadata in DynamoDB. For the orchestration, we will rely on AWS step functions to connect the single lambda functions. In the second part, we build a custom image classifier with PyTorch using transfer learning and run the model with the ONNX runtime in lambda function. At the end of the workshop participants will be able to write their own AWS lambda function with the serverless framework and python and know the basics of image classification with transfer learning.
You will learn:
Nico is working as a Machine Learning Engineer for codecentric and develops data-driven products and solutions. At the moment he is focusing on the deployment and scaling of machine learning models in production environments.