Determining hyperparameters for data augmentation

[klock18]

❔Question

Hi!

Just wondering if someone could please point me towards code/resources to find out where/how YOLOv5 determines the following parameters for data augmentation:

Additional context

[glenn-jocher]

yolov5/data/hyp.finetune.yaml

Lines 1 to 27 in 0c01afc

	# Hyperparameters for VOC fine-tuning
	# python train.py --batch 64 --cfg '' --weights yolov5m.pt --data voc.yaml --img 512 --epochs 50
	# See tutorials for hyperparameter evolution https://github.com/ultralytics/yolov5#tutorials
	lr0: 0.01 # initial learning rate (SGD=1E-2, Adam=1E-3)
	momentum: 0.94 # SGD momentum/Adam beta1
	weight_decay: 0.0005 # optimizer weight decay 5e-4
	giou: 0.05 # GIoU loss gain
	cls: 0.4 # cls loss gain
	cls_pw: 1.0 # cls BCELoss positive_weight
	obj: 0.5 # obj loss gain (scale with pixels)
	obj_pw: 1.0 # obj BCELoss positive_weight
	iou_t: 0.20 # IoU training threshold
	anchor_t: 4.0 # anchor-multiple threshold
	fl_gamma: 0.0 # focal loss gamma (efficientDet default gamma=1.5)
	hsv_h: 0.015 # image HSV-Hue augmentation (fraction)
	hsv_s: 0.7 # image HSV-Saturation augmentation (fraction)
	hsv_v: 0.4 # image HSV-Value augmentation (fraction)
	degrees: 1.0 # image rotation (+/- deg)
	translate: 0.1 # image translation (+/- fraction)
	scale: 0.6 # image scale (+/- gain)
	shear: 1.0 # image shear (+/- deg)
	perspective: 0.0 # image perspective (+/- fraction), range 0-0.001
	flipud: 0.01 # image flip up-down (probability)
	fliplr: 0.5 # image flip left-right (probability)
	mixup: 0.2 # image mixup (probability)

[klock18]

Thanks for your quick reply, found this which is a great read for anyone going down a similar line of inquiry https://docs.ultralytics.com/yolov5/tutorials/hyperparameter_evolution