This walkthrough includes the following steps:

1. Collect and prepare your own dataset to feed into an ML algorithm
2. Train a model with Amazon SageMaker, a fully managed service that provides the ability to build, train, and deploy machine learning (ML) models quickly
3. Running the model locally on AWS DeepLens to predict types of trash without sending any data to the cloud
4. Optionally, after AWS DeepLens makes its prediction, you can set up AWS DeepLens to send a message to a Raspberry Pi via AWS IoT Greengrass to show you which bin to throw the item in. The following diagram illustrates the solution architecture.