darknet/cfg/yolov7.cfg

1025 lines
10 KiB
INI

[net]
# Testing
#batch=1
#subdivisions=1
# Training
batch=8
subdivisions=1
width=640
height=640
channels=3
momentum=0.9
decay=0.0005
angle=0
saturation = 1.5
exposure = 1.5
hue=.1
learning_rate=0.00261
burn_in=1000
max_batches = 2000200
policy=steps
steps=1600000,1800000
scales=.1,.1
# 0
[convolutional]
batch_normalize=1
filters=32
size=3
stride=1
pad=1
activation=swish
# 1
[convolutional]
batch_normalize=1
filters=64
size=3
stride=2
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=64
size=3
stride=1
pad=1
activation=swish
# 3
[convolutional]
batch_normalize=1
filters=128
size=3
stride=2
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=64
size=1
stride=1
pad=1
activation=swish
[route]
layers=-2
[convolutional]
batch_normalize=1
filters=64
size=1
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=64
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=64
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=64
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=64
size=3
stride=1
pad=1
activation=swish
[route]
layers = -1,-3,-5,-7
# 12
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=swish
[maxpool]
size=2
stride=2
[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=swish
[route]
layers=-3
[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=128
size=3
stride=2
pad=1
activation=swish
# 18
[route]
layers = -1,-4
[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=swish
[route]
layers=-2
[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=swish
[route]
layers = -1,-3,-5,-7
# 27
[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=swish
[maxpool]
size=2
stride=2
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=swish
[route]
layers=-3
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=256
size=3
stride=2
pad=1
activation=swish
# 33
[route]
layers = -1,-4
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=swish
[route]
layers=-2
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=swish
[route]
layers = -1,-3,-5,-7
# 42
[convolutional]
batch_normalize=1
filters=1024
size=1
stride=1
pad=1
activation=swish
[maxpool]
size=2
stride=2
[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=swish
[route]
layers=-3
[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=512
size=3
stride=2
pad=1
activation=swish
# 48
[route]
layers = -1,-4
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=swish
[route]
layers=-2
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=swish
[route]
layers = -1,-3,-5,-7
# 57
[convolutional]
batch_normalize=1
filters=1024
size=1
stride=1
pad=1
activation=swish
##################################
### SPPCSP ###
[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=swish
[route]
layers = -2
[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=swish
[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=swish
### SPP ###
[maxpool]
stride=1
size=5
[route]
layers=-2
[maxpool]
stride=1
size=9
[route]
layers=-4
[maxpool]
stride=1
size=13
[route]
layers=-6,-5,-3,-1
### End SPP ###
[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=swish
[route]
layers = -1, -13
# 72
[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=swish
### End SPPCSP ###
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=swish
[upsample]
stride=2
[route]
layers = 42
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=swish
[route]
layers = -1,-3
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=swish
[route]
layers=-2
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=swish
[route]
layers = -1,-2,-3,-4,-5,-7
# 86
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=swish
[upsample]
stride=2
[route]
layers = 27
[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=swish
[route]
layers = -1,-3
[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=swish
[route]
layers=-2
[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=64
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=64
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=64
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=64
size=3
stride=1
pad=1
activation=swish
[route]
layers = -1,-2,-3,-4,-5,-7
# 100
[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=swish
[maxpool]
size=2
stride=2
[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=swish
[route]
layers=-3
[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=128
size=3
stride=2
pad=1
activation=swish
[route]
layers = -1,-4,86
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=swish
[route]
layers=-2
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=swish
[route]
layers = -1,-2,-3,-4,-5,-7
# 115
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=swish
[maxpool]
size=2
stride=2
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=swish
[route]
layers=-3
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=256
size=3
stride=2
pad=1
activation=swish
[route]
layers = -1,-4,72
[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=swish
[route]
layers=-2
[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=swish
[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=swish
[route]
layers = -1,-2,-3,-4,-5,-7
# 130
[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=swish
#############################
# ============ End of Neck ============ #
# ============ Head ============ #
# P3
[route]
layers = 100
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=256
activation=swish
[convolutional]
size=1
stride=1
pad=1
filters=255
#activation=linear
activation=logistic
[yolo]
mask = 0,1,2
anchors = 12,16, 19,36, 40,28, 36,75, 76,55, 72,146, 142,110, 192,243, 459,401
classes=80
num=9
jitter=.1
scale_x_y = 2.0
objectness_smooth=1
ignore_thresh = .7
truth_thresh = 1
#random=1
resize=1.5
iou_thresh=0.2
iou_normalizer=0.05
cls_normalizer=0.5
obj_normalizer=1.0
iou_loss=ciou
nms_kind=diounms
beta_nms=0.6
new_coords=1
max_delta=2
# P4
[route]
layers = 115
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=swish
[convolutional]
size=1
stride=1
pad=1
filters=255
#activation=linear
activation=logistic
[yolo]
mask = 3,4,5
anchors = 12,16, 19,36, 40,28, 36,75, 76,55, 72,146, 142,110, 192,243, 459,401
classes=80
num=9
jitter=.1
scale_x_y = 2.0
objectness_smooth=1
ignore_thresh = .7
truth_thresh = 1
#random=1
resize=1.5
iou_thresh=0.2
iou_normalizer=0.05
cls_normalizer=0.5
obj_normalizer=1.0
iou_loss=ciou
nms_kind=diounms
beta_nms=0.6
new_coords=1
max_delta=2
# P5
[route]
layers = 130
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=1024
activation=swish
[convolutional]
size=1
stride=1
pad=1
filters=255
#activation=linear
activation=logistic
[yolo]
mask = 6,7,8
anchors = 12,16, 19,36, 40,28, 36,75, 76,55, 72,146, 142,110, 192,243, 459,401
classes=80
num=9
jitter=.1
scale_x_y = 2.0
objectness_smooth=1
ignore_thresh = .7
truth_thresh = 1
#random=1
resize=1.5
iou_thresh=0.2
iou_normalizer=0.05
cls_normalizer=0.5
obj_normalizer=1.0
iou_loss=ciou
nms_kind=diounms
beta_nms=0.6
new_coords=1
max_delta=2