Donut or Not! - Deploying Deep Learning Inference as Serverless Functions
There is a common myth that to perform deep learning, one needs high compute, GPU enabled devices. While this is true to a degree, when training deep learning models, it is often just as possible to perform inference using simple CPU based architectures in compute and memory constrained environments - such as serverless functions. This blog takes you through my journey of deploying a simple donut vs bagel vs vada classifier as a AWS Lambda function!
