一、首先下载配置Yolov3框架
git clone https://github.com/pjreddie/darknet
cd darknet
make
下载常用的模型文件
wget https://pjreddie.com/media/files/yolov3.weights
wget https://pjreddie.com/media/files/yolov3-tiny.weights
wget https://pjreddie.com/media/files/darknet53.conv.74
如果需要使用cuda进行训练,配置开通cuda
vim Makefile
将默认不开启数字0改为1,然后保存关闭,重新运行一下make
GPU=1
CUDNN=1
二、数据转换
LabelImage标记的数据生成的为xml格式,利用Python转换成txt格式,进行中心归一化。贴出voc转换代码。
import xml.etree.ElementTree as ET
import pickle
import os
from os import listdir, getcwd
from os.path import join
sets=[('2012', 'train'), ('2012', 'val'), ('2007', 'train'), ('2007', 'val'), ('2007', 'test')]
classes = ["aeroplane", "bicycle", "bird", "boat", "bottle", "bus", "car", "cat", "chair", "cow", "diningtable", "dog", "horse", "motorbike", "person", "pottedplant", "sheep", "sofa", "train", "tvmonitor"]
def convert(size, box):
dw = 1./size[0]
dh = 1./size[1]
x = (box[0] + box[1])/2.0
y = (box[2] + box[3])/2.0
w = box[1] - box[0]
h = box[3] - box[2]
x = x*dw
w = w*dw
y = y*dh
h = h*dh
return (x,y,w,h)
def convert_annotation(year, image_id):
in_file = open('VOCdevkit/VOC%s/Annotations/%s.xml'%(year, image_id))
out_file = open('VOCdevkit/VOC%s/labels/%s.txt'%(year, image_id), 'w')
tree=ET.parse(in_file)
root = tree.getroot()
size = root.find('size')
w = int(size.find('width').text)
h = int(size.find('height').text)
for obj in root.iter('object'):
difficult = obj.find('difficult').text
cls = obj.find('name').text
if cls not in classes or int(difficult) == 1:
continue
cls_id = classes.index(cls)
xmlbox = obj.find('bndbox')
b = (float(xmlbox.find('xmin').text), float(xmlbox.find('xmax').text), float(xmlbox.find('ymin').text), float(xmlbox.find('ymax').text))
bb = convert((w,h), b)
out_file.write(str(cls_id) + " " + " ".join([str(a) for a in bb]) + '\n')
wd = getcwd()
for year, image_set in sets:
if not os.path.exists('VOCdevkit/VOC%s/labels/'%(year)):
os.makedirs('VOCdevkit/VOC%s/labels/'%(year))
image_ids = open('VOCdevkit/VOC%s/ImageSets/Main/%s.txt'%(year, image_set)).read().strip().split()
list_file = open('%s_%s.txt'%(year, image_set), 'w')
for image_id in image_ids:
list_file.write('%s/VOCdevkit/VOC%s/JPEGImages/%s.jpg\n'%(wd, year, image_id))
convert_annotation(year, image_id)
list_file.close()
三、将所有的图片生成在一个txt文档中保存
用Python将所有的图片数据的路径保存在一个txt文档中,便于后期训练。
四、配置yourlabel.names、yourtrain.data、youryolov3.cfg
在yourlabel.names里面存放标记的类别名称,在yourtrain.data里面如下
classes= 4 #需要标记的类别,与label数相对应
train = /media/pico/886835D26835C02C/Kevin_ubuntu/darknet/data/myv3.txt #合成的图片训练数据集
names = /media/pico/886835D26835C02C/Kevin_ubuntu/darknet/data/myv3.names #label文件
backup= /media/pico/886835D26835C02C/Kevin_ubuntu/darknet/backup #存放模型路径
在youryolov3.cfg中配置你的网络,可以使用voc模型但是需要修改每一个yolo层的前一层的filters,filters数值=3×(类别数+5),比如你分类标记的类别为4类,则filters=3×(4+5)=27。
yourlabel.names与生成图片路径文档放在data文件夹下,yourtrain.data与youryolov3.cfg放在cfg文件夹下。
五、训练自己的模型
在darknet路径下启动命令窗口,然后执行如下命令,模型会在backup下进行保存。
./darknet detector train cfg/voc.data cfg/yolov3-voc.cfg darknet53.conv.74
六、测试模型
./darknet detect cfg/my_yolov3.cfg backup/my_yolov3_900.weights data/jam7_0002.jpg
七、yolov3参数配置
[net]
# Testing ### 测试模式
# batch=1
# subdivisions=1
# Training ### 训练模式,每次前向的图片数目 = batch/subdivisions
batch=64
subdivisions=16 ### 如果cuda out of memory,需要调小batch_size
width=416 ### 网络的输入宽、高、通道数
height=416
channels=3 ### RGB图使用3通道,灰度图使用1
momentum=0.9 ### 动量
decay=0.0005 ### 权重衰减
angle=0
saturation = 1.5 ### 饱和度
exposure = 1.5 ### 曝光度
hue=.1 ### 色调
learning_rate=0.001 ### 学习率
burn_in=1000 ### 学习率控制的参数
max_batches = 50200 ### 迭代次数
policy=steps ### 学习率策略
steps=40000,45000 ### 学习率变动步长
scales=.1,.1 ### 学习率变动因子
[convolutional]
batch_normalize=1 ### BN
filters=32 ### 卷积核数目
size=3 ### 卷积核尺寸
stride=1 ### 卷积核步长
pad=1 ### pad
activation=leaky ### 激活函数
......
[convolutional]
size=1
stride=1
pad=1
filters=45 #3*(10+4+1)
activation=linear
[yolo]
mask = 6,7,8
anchors = 10,13, 16,30, 33,23, 30,61, 62,45, 59,119, 116,90, 156,198, 373,326
classes=10 #类别
num=9
jitter=.3
ignore_thresh = .5
truth_thresh = 1
random=0 #1,如果显存很小,将random设置为0,关闭多尺度训练;
......
[convolutional]
size=1
stride=1
pad=1
filters=45 #3*(10+4+1)
activation=linear
[yolo]
mask = 3,4,5
anchors = 10,13, 16,30, 33,23, 30,61, 62,45, 59,119, 116,90, 156,198, 373,326
classes=10 #类别
num=9
jitter=.3
ignore_thresh = .5
truth_thresh = 1
random=0 #1,如果显存很小,将random设置为0,关闭多尺度训练;
......
[convolutional]
size=1
stride=1
pad=1
filters=45 #3*(10+4+1)
activation=linear
[yolo]
mask = 0,1,2
anchors = 10,13, 16,30, 33,23, 30,61, 62,45, 59,119, 116,90, 156,198, 373,326
classes=10 #类别
num=9
jitter=.3 # 数据扩充的抖动操作
ignore_thresh = .5 #文章中的阈值1
truth_thresh = 1 #文章中的阈值2
random=0 #1,如果显存很小,将random设置为0,关闭多尺度训练;
八、yolov3中断后继续训练
./darknet detector train cfg/hand.data cfg/hand.cfg backup/hand.backup