File size: 1,818 Bytes
1d94dcd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
# YOLOv9 Configuration File

# Model architecture details
model:
  type: YOLOv9  # Model type
  backbone:
    type: CSPDarknet  # Backbone architecture type
    depth_multiple: 1.0  # Depth multiplier
    width_multiple: 1.0  # Width multiplier

# Input image settings
input_size: 640  # Input image size (square)

# Anchor parameters
anchors:
  - [10, 13, 16, 30, 33, 23]  # Anchor box 1
  - [30, 61, 62, 45, 59, 119]  # Anchor box 2
  - [116, 90, 156, 198, 373, 326]  # Anchor box 3

# Training hyperparameters
hyp:
  lr0: 0.01  # Initial learning rate
  lrf: 0.01  # Learning rate reduction factor
  momentum: 0.937  # Momentum
  weight_decay: 0.0005  # Weight decay
  warmup_epochs: 3.0  # Warmup epochs
  warmup_momentum: 0.8  # Warmup momentum
  warmup_bias_lr: 0.1  # Warmup bias learning rate
  box: 7.5  # Box loss gain
  cls: 0.5  # Class loss gain
  cls_pw: 1.0  # Class label smoothing
  dfl: 1.5  # DFL loss gain
  obj_pw: 1.0  # Objectness loss gain
  iou_t: 0.2  # IoU training threshold
  anchor_t: 5.0  # Anchor matching threshold
  fl_gamma: 0.0  # Focal loss gamma
  hsv_h: 0.015  # HSV hue gain
  hsv_s: 0.7  # HSV saturation gain
  hsv_v: 0.4  # HSV value gain
  degrees: 0.0  # Image rotation (degrees)
  translate: 0.1  # Image translation
  scale: 0.9  # Image scale
  shear: 0.0  # Image shear
  perspective: 0.0  # Image perspective transform
  flipud: 0.0  # Flip image vertically
  fliplr: 0.5  # Flip image horizontally
  mosaic: 1.0  # Mosaic augmentation
  mixup: 0.15  # Mixup augmentation
  copy_paste: 0.3  # Copy-Paste augmentation

# Other settings
other:
  multi_scale: false  # Use multi-scale training
  flip: true  # Use random horizontal flipping
  blur: false  # Use random image blurring
  letterbox: false  # Use letterbox resizing
  rect: false  # Use rectangular resizing