Convert YOLOv9 Model to TensorFlow Lite

This repository offers scripts and instructions for converting a YOLOv9 model to TensorFlow Lite format. TensorFlow Lite is a lightweight solution for deploying machine learning models on mobile and edge devices, making it ideal for applications that require real-time object detection, such as mobile apps or embedded systems.

New

We now provide the model weights of TFLite (quantized INT8)
Link: YOLOv9-e
Link: YOLOv9-e-int8

Requirements

• Python 3.8.10
• TensorFlow 2.13.1
• Other dependencies (refer to requirements.txt)

Installation

1. Create a Conda environment:

   conda create --name yolo9-tflite python=3.8.10

2. Activate the environment:

   conda activate yolo9-tflite

3. Install required packages:

   pip install -r requirements.txt

Convert

1. To convert to TFLite, run the provided script:

   convert_tflite.sh

Inference

1. I have provided the config to run yolov9 (config/yolov9.yaml)

2. You run to test the model
   

python inference.py

Output

Contact

Email : [email protected]