Customization

SDK Guide

The microservice extensively uses the features provided by the DeepStream SDK.

You can build a new pipeline/app using DeepStream SDK (by modifying ours or creating one from scratch). For reference, refer to the README.md in the metropolis-apps-standalone-deployment/modules/detect-track-embedding/deepstream-fewshot-learning-app directory.

Pipeline Customization

Here are few ways to customize the provided DeepStream pipeline/application inside the microservice:

• Customize on messages that are sent to Metropolis Microservices.

• Modify source code.

• Using your custom AI model.

Custom DeepStream Pipeline

If you have your own custom DeepStream applications, you can add the NvMsgConv and NvMsgBroker plugin to your application to send metadata to Metropolis Microservices.

NvMsgBroker packages the payload into a format that can be read by Metropolis Microservices.