YOLOX OpenVINO Inference

This repository provides a step-by-step guide to run inference with the YOLOX model using Intel's OpenVINO toolkit.

Prerequisites

To install the required Python dependencies:

pip install -r requirements.txt

Download and Convert YOLOX-Tiny Model

1. Download the YOLOX-Tiny Model

Use the Open Model Zoo Downloader to fetch the YOLOX-Tiny model:

omz_downloader --name yolox-tiny

This command will download the YOLOX-Tiny model to the default Open Model Zoo directory.

2. Convert the Model to OpenVINO IR Format

To convert the YOLOX-Tiny model into OpenVINO's Intermediate Representation (IR) format:

omz_converter --name yolox-tiny --precisions FP16

The conversion generates .xml and .bin files that are optimized for OpenVINO inference.

Run Inference

1. Locate the Converted Model Files

The converted model files (yolox-tiny.xml and yolox-tiny.bin) will be located in the public/yolox-tiny/FP16 directory.

2. Inference Script

Refer to the infer_yolox_openvino.ipynb Jupyter notebook for detailed steps on running image and video inference.

This notebook provides a comprehensive guide to running inference and visualizing the results using OpenVINO.

References

• YOLOX-Tiny Model Documentation
• OpenVINO Toolkit Documentation
• Open Model Zoo Repository