Yolo Onnx

Overview

This is project about detection using Yolov8.onnx

Table of Contents

• Yolo Onnx
  • Overview
  • Table of Contents
  • Demo
  • Features
  • Requirements
  • Installation
  • Train on custom data

Demo

It can detect object from image or link of image With 5 classes: {0: "bicycle", 1: "bus", 2: "car", 3: "motorbike", 4: "person"}

Datasets

Features

• Detects and identifies objects in images or links.
• Supports real-time object detection.
• Easily configurable and extensible for various use cases.

Requirements

Specify the software and hardware requirements needed to run the project. For example:

• Pytorch
• Python 3.x
• GPU with CUDA support

Installation

Provide detailed installation instructions. If possible, create a virtual environment for your project. Here's a general installation process:

1. Clone the repository:

   git clone https://github.com/dinhanit/YoloOnnx.git
   

2. Set up

   pip install -r requirements.txt

3. Demo on fastapi

   • Setup docker

     chmod +x install.sh
     ./install.sh

   • Open browser

     http://localhost:8501/docs
     

4. Demo with streamlit

   cd YoloOnnx/Web/
   streamlit run st.pt

Train on custom data

1. Model properties

   

2. Set up

   ├───Web
   │   └───weight
   └───Yolo
       └───datasets
           ├───test
           │   ├───images
           │   └───labels
           ├───train
           │   ├───images
           │   └───labels
           └───valid
               ├───images
               └───labels
   

3. Train

   cd Yolo
   python train.py
   

4. Convert to onnx model

   python convert_onnx.py