There are tons of resources out there on simplified Training and optimized pre-trained Inferencing models. However, training something custom to optimize performance on readily available hardware with minimal effort still seemed far fetched!

In this article we will leverage the concept of transfer learning where a model trained to classify images is used to train our custom use-case (e.g. items in your pantry) in your device browser with Teachable Machine (GUI) and optimize CPU inferencing with Intel® OpenVINO™ Toolkit without any painful SW installation (in 10mins of-course!).

Is optimization necessary? Skip to the performance comparison at the end.

What setup do I need beforehand?

• 6th to 10th generation Intel Core or Intel Xeon processors
• (i.e. if you purchased an Intel device released 2016 onwards)
• Docker installed on a Linux System with internet access to hub.docker.com
• Access to a camera device (e.g. webcam)

#openvino #optimization-techniques #ai #image-classification #teachable-machine