Frequently Asked Questions
Learn more about the Intel Geti platform through some of the most frequently asked questions.
How can the Intel® Geti™ platform be deployed? Local on-prem machine? Or cloud instances?
It can be deployed on a local on-premises machine or an AWS or Azure cloud virtual machine instance.
How can a customer deploy models developed using the Intel Geti platform? What is the role of OpenVINO™ toolkit?
The Intel Geti platform enables you to produce your trained model either in native framework, or as an optimized model for OpenVINO toolkit. OpenVINO toolkit is tightly integrated into the platform to easily quantize trained models for scaled deployment on Intel hardware. You can use the models built using the Intel Geti platform and deploy them into your deployment pipeline.
What computer vision tasks does the Intel Geti platform support? What about deep learning frameworks?
The Intel Geti platform supports a variety of commonly employed computer vision tasks as well as numerous deep learning frameworks. Please refer to Intel Geti AI Platform Overview: Learn What Is Under the Hood for information on what is currently supported.
What type of data formats are supported?
The Intel Geti platform supports 2D visual data in either image or video formats.
Are you leveraging pre-trained models in the platform?
Model architectures in the Intel Geti platform are pre-trained to a certain extent in that they understand how to extract features from images, and then learn the key features from the data provided for specific use cases.
What model architectures are supported?
The Intel Geti platform supports a range of model architectures today, and support for additional architectures will be coming in future releases. Please refer to Intel® Geti™ AI Platform Overview: Learn What Is Under the Hood for the current model architectures supported.
What is active learning? How does it enable the Intel Geti platform to train on less data?
Active learning is an intelligent algorithm that selects the best sample inputs from the data to help a model learn quickly from a small amount of data. Please refer to this blog post on active learning for a deeper dive into active learning within the Intel Geti platform.
