take or find vehicle images for create a special dataset for fine-tuning.
Train : 70%
Validition : 20%
Test : 10%
config dataset.yaml for the address and information of your dataset.
path: Dataset/dataset-vehicles # dataset root dir
train: images/train # train images (relative to 'path')
val: images/val # val images (relative to 'path')
test: # test images (optional)
# Classes
nc: 5 # number of classes
names: [ 'Car', 'Motorcycle', 'Truck', 'Bus', 'Bicycle'] # class names
git clone https://github.com/MaryamBoneh/Vehicle-Detection
cd Vehicle-Detection
pip install -r requirements.txt
to have mAP, loss, confusion matrix, and other metrics, sign in www.wandb.ai.
pip install wandb
fine-tuning on a pre-trained model of yolov5.
python train.py --img 640 --batch 16 --epochs 50 --data dataset.yaml --weights yolov5m.pt
after train, gives you weights of train and you should use them for test.
python detect.py --weights runs/train/exp12/weights/best.pt --source test_images/imtest13.JPG
you can also use the weight file in path 'runs/train/exp12/weights/best.pt' without the train. this weight is the result of 128 epoch train on the following dataset.
https://b2n.ir/vehicleDataset