YOLOv8-training-inference

MORE TO DO

How to Use

1. First git clone this file

git clone https://github.com/mangoggul/YOLOv8-training-inference.git

2. You need to download Training Dataset / Dataset Architecture needs to be

Dataset Tree

├── dataName
│   ├── train
│         ├── images
│         └── labels
│   ├── val
│         ├── images
│         └── labels

So For example

├── wheatCOCO
│   ├── train
│         ├── images
│         └── labels
│   ├── val
│         ├── images
│         └── labels

Important Thing : YOLO doesnt support json Annotation Files so you need to change json to txt

if you only have json file use json2txt.py to convert json file to txt

3. Start Training

first you need to make yaml file.

names:
  0: backGround
  1: wheat
nc : 2

path: C:/Users/user/Desktop/vscodeOD/YOLO_v8/wheatCOCO/  # Base dataset directory
train: train/images
val: val/images

this one is my yaml file.
names : class name

<YOLO detect background also So if you want to detect 1 object you need to write Object name and background>

nc : class number

path : dataset Path

train : dataset/train Path

val : dataset/val Path

---

use train.ipynb file to train your own custom data
you can choose epochs, workers, etc.
consider below documents to know more about parameters

https://docs.ultralytics.com/modes/train/#train-settings

4. Inference

After Training you can use inference.ipynb file. from ultralytics import YOLO

# Load a model
model = YOLO("runs/detect/train17/weights/best.pt")  # pretrained YOLOv8n model

In this part, put your pt file to run inference!

Enjoy Detection!