yolov4关键参数：

[net]
batch=64    # 所有的图片分成all_num/batch个批次，每batch个样本(64)更新一次参数，尽量保证一个batch里面各个类别都能取到样本
subdivisions=64   # 决定每次送入显卡的图片数目 batch/subdivisions
width=608   # 图片宽度
height=608   # 图片高度
channels=3   # 图片通道数
momentum=0.949   # 动量参数，表示梯度下降到最优值的速度
decay=0.0005  # 权重衰减正则项，防止过拟合.
angle=0      #  图片角度变化
saturation = 1.5   # 图片饱和度
exposure = 1.5   #  图片曝光量
hue=.1     # 图片色调learning_rate=0.0013  # 学习率，影响权值更新的速度
burn_in=1000   
max_batches = 500500  # 训练最大次数
policy=steps   # 学习率调整的策略
steps=400000,450000  # 学习率调整时间
scales=.1,.1     # 学习率调整倍数#cutmix=1  # 数据增强cutmix
mosaic=0   # 数据增强mosaic#:104x104 54:52x52 85:26x26 104:13x13 for 416[convolutional]   # 利用32个大小为3*3*3的滤波器步长为1，填充值为1进行过滤然后用mish函数进行激活
batch_normalize=1  # 是否做BN
filters=32         # 输出特征图的数量
size=3             # 卷积核的尺寸
stride=1           # 做卷积运算的步长
pad=1              # 是否做pad
activation=mish    # 激活函数类型     #  （608 + 2 * 1 - 3）/ 1 + 1 = 608   # 608 * 608 * 32# Downsample[convolutional]  # 利用64个大小为3*3*3的滤波器步长为2，填充值为1进行过滤然后用mish函数进行激活
batch_normalize=1
filters=64
size=3
stride=2
pad=1
activation=mish     #  （608 + 2 * 1 - 3）/ 2 + 1 = 304   # 304 * 304 * 64[convolutional]  # 利用64个大小为1*1*3的滤波器步长为1，填充值为1进行过滤然后用mish函数进行激活
batch_normalize=1
filters=64
size=1
stride=1
pad=1
activation=mish   # (304 + 2 * 1 -1) / 1 + 1 = 306   # 304 * 304 * 64[route]
layers = -2   # 当属性layers只有一个值时，它会输出由该值索引的网络层的特征图  # 304 * 304 * 64[convolutional]   # 利用64个大小为1*1*3的滤波器步长为1，填充值为1进行过滤然后用mish函数进行激活
batch_normalize=1
filters=64
size=1
stride=1
pad=1
activation=mish    # (304 + 2 * 1 -1) / 1 + 1 = 306   # 304 * 304 * 64[convolutional]   # 利用32个大小为1*1*3的滤波器步长为1，填充值为1进行过滤然后用mish函数进行激活
batch_normalize=1
filters=32
size=1
stride=1
pad=1
activation=mish   # (304 + 2 * 1 -1) / 1 + 1 = 306   # 304 * 304 * 32[convolutional]   # 利用64个大小为3*3*3的滤波器步长为1，填充值为1进行过滤然后用mish函数进行激活
batch_normalize=1
filters=64
size=3
stride=1
pad=1
activation=mish   # (304 + 2 * 1 -3) / 1 + 1 = 304   # 304 * 304 * 64[shortcut]   #shortcut 操作是类似 ResNet 的跨层连接
from=-3  # 参数 from 是 −3，意思是 shortcut 的输出是当前层与先前的倒数第三层相加而得到，通俗来讲就是 add 操作
activation=linear   # 304 * 304 * 64[convolutional]    # 利用64个大小为1*1*3的滤波器步长为1，填充值为1进行过滤然后用mish函数进行激活
batch_normalize=1
filters=64
size=1
stride=1
pad=1
activation=mish   # (304 + 2 * 1 -1) / 1 + 1 = 306   # 304 * 304 * 64[route]    # 304 * 304 * 128
layers = -1,-7   # 当属性layers有两个值，就是将上一层和从当前层倒数第7层进行融合，大于两个值同理[convolutional]       # 利用64个大小为1*1*3的滤波器步长为1，填充值为1进行过滤然后用mish函数进行激活
batch_normalize=1
filters=64
size=1
stride=1
pad=1
activation=mish    # (304 + 2 * 1 -1) / 1 + 1 = 306   # 304 * 304 * 32# Downsample[convolutional]   # 利用128个大小为3*3*3的滤波器步长为2，填充值为1进行过滤然后用mish函数进行激活
batch_normalize=1
filters=128
size=3
stride=2
pad=1
activation=mish   # (304 + 2 * 1 -3) / 2 + 1 = 152   # 152 * 152 * 32[convolutional]
batch_normalize=1
filters=64
size=1
stride=1
pad=1
activation=mish[route]
layers = -2[convolutional]
batch_normalize=1
filters=64
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=64
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=64
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=64
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=64
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=64
size=1
stride=1
pad=1
activation=mish[route]
layers = -1,-10[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish# Downsample[convolutional]
batch_normalize=1
filters=256
size=3
stride=2
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish[route]
layers = -2[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=128
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=mish[route]
layers = -1,-28[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish# Downsample[convolutional]
batch_normalize=1
filters=512
size=3
stride=2
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish[route]
layers = -2[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=256
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=mish[route]
layers = -1,-28[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish# Downsample[convolutional]
batch_normalize=1
filters=1024
size=3
stride=2
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish[route]
layers = -2[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=512
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=512
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=512
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish[convolutional]
batch_normalize=1
filters=512
size=3
stride=1
pad=1
activation=mish[shortcut]
from=-3
activation=linear[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=mish[route]
layers = -1,-16[convolutional]
batch_normalize=1
filters=1024
size=1
stride=1
pad=1
activation=mish##########################[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=leaky[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=1024
activation=leaky[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=leaky### SPP ###  
[maxpool]
stride=1
size=5[route]
layers=-2[maxpool]
stride=1
size=9[route]
layers=-4[maxpool]
stride=1
size=13[route]
layers=-1,-3,-5,-6
### End SPP ###  每个maxpool的padding=size/2[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=leaky[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=1024
activation=leaky[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=leaky[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=leaky[upsample]
stride=2[route]
layers = 85[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=leaky[route]
layers = -1, -3[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=leaky[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=leaky[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=leaky[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=leaky[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=leaky[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=leaky[upsample]
stride=2[route]
layers = 54[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=leaky[route]
layers = -1, -3[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=leaky[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=256
activation=leaky[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=leaky[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=256
activation=leaky[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=leaky##########################[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=256
activation=leaky[convolutional]
size=1
stride=1
pad=1
filters=33
activation=linear[yolo]
mask = 0,1,2  #对应的anchors索引值
anchors =  23, 40,  25, 79,  44, 77,  30,124,  54,121,  42,197,  87,112,  76,193, 126,255
classes=6
num=9
jitter=.3
ignore_thresh = .7
truth_thresh = 1
scale_x_y = 1.2
iou_thresh=0.213
cls_normalizer=1.0
iou_normalizer=0.07
iou_loss=ciou
nms_kind=greedynms
beta_nms=0.6
max_delta=5[route]
layers = -4[convolutional]
batch_normalize=1
size=3
stride=2
pad=1
filters=256
activation=leaky[route]
layers = -1, -16[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=leaky[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=leaky[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=leaky[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=leaky[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=leaky[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=leaky[convolutional]
size=1
stride=1
pad=1
filters=33
activation=linear[yolo]
mask = 3,4,5
#anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
anchors =  23, 40,  25, 79,  44, 77,  30,124,  54,121,  42,197,  87,112,  76,193, 126,255
classes=6
num=9
jitter=.3
ignore_thresh = .7
truth_thresh = 1
scale_x_y = 1.1
iou_thresh=0.213
cls_normalizer=1.0
iou_normalizer=0.07
iou_loss=ciou
nms_kind=greedynms
beta_nms=0.6
max_delta=5[route]
layers = -4[convolutional]
batch_normalize=1
size=3
stride=2
pad=1
filters=512
activation=leaky[route]
layers = -1, -37[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=leaky[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=1024
activation=leaky[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=leaky[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=1024
activation=leaky[convolutional]
batch_normalize=1
filters=512
size=1
stride=1
pad=1
activation=leaky[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=1024
activation=leaky[convolutional]
size=1
stride=1
pad=1
filters=33
activation=linear[yolo]
mask = 6,7,8
#anchors = 12, 16, 19, 36, 40, 28, 36, 75, 76, 55, 72, 146, 142, 110, 192, 243, 459, 401
anchors =  23, 40,  25, 79,  44, 77,  30,124,  54,121,  42,197,  87,112,  76,193, 126,255
classes=6
num=9
jitter=.3
ignore_thresh = .7
truth_thresh = 1
random=0
scale_x_y = 1.05
iou_thresh=0.213
cls_normalizer=1.0
iou_normalizer=0.07
iou_loss=ciou
nms_kind=greedynms
beta_nms=0.6
max_delta=5