Metadata-Version: 2.1
Name: yolov8-onnx
Version: 1.0.4
Summary: YoloV8 Package Tools
Author: LauNT
Author-email: ttruongllau@gmail.com
License: Apache License 2.0
Requires-Python: >=3.7
Description-Content-Type: text/markdown
License-File: LICENSE

# YOLOV8 ONNX

[![License](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](LICENSE)
[![Python 3.7](https://img.shields.io/badge/python-3.7+-aff.svg)](https://www.python.org/downloads/release/python-370/)
[![ONNX Compatible](https://img.shields.io/badge/ONNX-Compatible-brightgreen)](https://onnx.ai/)
[![Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.google/)

## Description

This package is compatible with [YoloV8](https://github.com/ultralytics/ultralytics) for object detection program, using [ONNX](https://onnx.ai/) format model (CPU & GPU speed can be **x2 Times Faster**). This code is referenced from [this awesome repo](https://github.com/ibaiGorordo/ONNX-YOLOv8-Object-Detection).


## Usage

Step 1: Convert your pre-trained model to ONNX format.

```python
from ultralytics import YOLO

# Load your pre-trained model
model = YOLO('your-trained-model.pt')

# Export the model
model.export(format='onnx', 
            batch=1, 
            device='cpu', 
            simplify=True, 
            imgsz=640, 
            dynamic=True)
```

Step 2: Use in your code.

```python
from yolov8_onnx import DetectEngine

engine = DetectEngine(model_path= str | 'your-model.onnx',
                        image_size = int | 640,
                        conf_thres= float | 0.5, 
                        iou_thres= float | 0.1)

output = engine(image) # cv2 image
```

**Note**: akaOCR (Transform documents into useful data with AI-based IDP - Intelligent Document Processing) - helps make inefficient manual entry a thing of the past—and reliable data insights a thing of the present. Details at: https://app.akaocr.io
