Support Knowledge Base

Why Is the Frame per Second (FPS) in Deep Learning (DL) Workbench Higher than OpenVINO™ Demo Inference Script?

Content Type: Maintenance & Performance | Article ID: 000059581 | Last Reviewed: 05/20/2022

Description Resolution

Description

Ran YOLOv4 Object Detection model with Microsoft Common Object in Context (MS COCO) dataset on DL Workbench and get 50 - 60 FPS.
Ran OpenVINO™ Demo Inference Script and got higher FPS.

Resolution

Inference Results may vary depending on many factors:

Load of local machine
Usage of OpenVINO™ Toolkit API
Tools used

Related Products

This article applies to 2 products.

OpenVINO™ toolkit

Discontinued Products

Intel® Developer Cloud for the Edge

Need more help?

Contact support

Contact support