Environment Setup

This code is tested on python 3.9

Video tutorial: https://youtu.be/Dmv4EVBuCTQ

Step -1 : git clone https://github.com/THU-MIG/yolov10.git

Step -2 : cd yolov10

Step -3 : pip install .

Step -4 : Download yolov10 pretrained weights from official repo (https://github.com/THU-MIG/yolov10)

Step -5 : Test Pretrained YOLOv10 model using this command:

!yolo task=detect mode=predict conf=0.25 save=True model=../weights/yolov10n.pt source=test_images/1.jpg

Train Custom model:

1- Custom dataset is required for training the model. Sample dataset is in "custom_dataset" folder, Your dataset should have the same format.

2- Make changes in in custom_data.yaml file as per your dataset

3. Train model using this command:

!yolo task=detect mode=train epochs=100 batch=16 plots=True model=weights/yolov10n.pt data=custom_data.yaml

4. Inference using custom trained model:

!yolo task=detect mode=predict conf=0.25 save=True model=runs/detect/train/weights/best.pt source=test_images_1/veh2.jpg