As a Solutions Architect on the team that leverages the platform to bring apps and services to life for customers every day, I can truly say that this isn’t an overstatement. To demonstrate this first-hand and train our own internal teams, several months ago, our team decided to create a sample solution to showcase just how easy application deployment on the platform really is. We figured a simple “Hello World” example was just too easy, so we “jumped right into the deep end,” with a machine learning app. The solution, we internally named “Facefeed,” builds upon the popular open source Facenet
project that utilizes the Inception ResNet v1 model and VGGFace2 dataset to identify faces. As with any solution on the platform, the application and its data pipelines are deployed directly to a physical Xi Edge device or VM running at the edge.
To showcase the solution and enable our teams, we created (with the help of our fantastic enablement team) a deployment guide that details, step-by-step, the simple process to deploy the solution. The end-to-end solution deployed in the tutorial ingests a sample video stream using the real time streaming protocol (RTSP), and uses machine learning to detect known faces.
After great feedback internally, we also decided to showcase the application as part of the Hands On Labs during our .Next London conference in November ’18. Now, we’ve taken the guide public so that anyone can start their journey to become a machine learning expert. As you explore the guide
, you’ll notice that all of the machine learning code (based in python) and application containers (used to serve a sample RTSP stream and interactive UI) are also available for public consumption. So what are you waiting for? Deploy your first machine learning solution, or discover how much easier it can be, today!