best_hyperparameters.yaml

# 30/30 iterations complete ✅ (6414.87s)
# Results saved to blood_vessel_segmentation/tune
# Best fitness=0.34726 observed at iteration 1
# Best fitness metrics are {'metrics/precision(B)': 0.57028, 'metrics/recall(B)': 0.24676, 'metrics/mAP50(B)': 0.28238, 'metrics/mAP50-95(B)': 0.15513, 'metrics/precision(M)': 0.47646, 'metrics/recall(M)': 0.26639, 'metrics/mAP50(M)': 0.31251, 'metrics/mAP50-95(M)': 0.16461, 'val/box_loss': 1.35056, 'val/seg_loss': 1.14584, 'val/cls_loss': 1.13323, 'val/dfl_loss': 0.86719, 'fitness': 0.34726}
# Best fitness model is blood_vessel_segmentation/hyperparameter_tuning
# Best fitness hyperparameters are printed below.

lr0: 0.01
lrf: 0.01
momentum: 0.937
weight_decay: 0.0005
warmup_epochs: 3.0
warmup_momentum: 0.8
box: 7.5
cls: 0.5
dfl: 1.5
hsv_h: 0.015
hsv_s: 0.7
hsv_v: 0.4
degrees: 0.0
translate: 0.1
scale: 0.5
shear: 0.0
perspective: 0.0
flipud: 0.0
fliplr: 0.5
mosaic: 1.0
mixup: 0.0
copy_paste: 0.0