一、环境
Ubuntu 18.04 + CUDA10.0 +CUDNN
下方是我训练时的真实文件,可供大家参考
大家需要我的文件资源的请使用曲奇云盘下载,下面是下载链接:https://quqi.gblhgk.com/s/981990/7C1nyIvTyl7SRFjx
二、修改重要文件
1.yolov3-voc.cfg(darknet/cfg//yolov3-voc.cfg)
[convolutional]
size=1
stride=1
pad=1
filters=21//修改为filters=(classes+5)x3
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=2//修改为需要检测的类别数目
num=9
jitter=.3
ignore_thresh = .5
truth_thresh = 1
random=1//当显卡内存不够时可以修改为0
以上共修改三处均已标出,在整个文件里有三个这样的结构也就是说需要修改九处,建议先拉到文件的最下方,开始修改。
2.voc.names(darknet/data//voc.names)
wheel
goods
按照上述方法写入要检测的类别的名称,其中第一个类别的id为0,依次升序!
3.voc.data(darknet/data//voc.data)
classes= 2//类别数目
train = /home/ustc/Kaixiang/YOLO/darknet/scripts/2007_train.txt//训练集目录
valid = /home/ustc/Kaixiang/YOLO/darknet/scripts/2007_val.txt//验证集目录
names = data/voc.names//类别名称目录
backup = backup//模型的存储目录
4.Makefile(darknet//Makefile)
GPU=1
CUDNN=1
OPENCV=1
OPENMP=0
DEBUG=0
根据硬件是否有GPU,是否安装OPENCV等进行选择,有 1无0
三、生成训练集验证集等的txt文件
1.文件的目录格式
darknet/scripts/如下图
2.generator_voc.py(darknet\scripts\VOCdevkit\VOC2007\\generator_voc.py)
import os
import random
trainval_percent = 0.9
train_percent = 0.8
xmlfilepath = 'Annotations'
txtsavepath = 'ImageSets\Main'
total_xml = os.listdir(xmlfilepath)
num=len(total_xml)
list=range(num)
tv=int(num*trainval_percent)
tr=int(tv*train_percent)
trainval= random.sample(list,tv)
train=random.sample(trainval,tr)
ftrainval = open('ImageSets/Main/trainval.txt', 'w')
ftest = open('ImageSets/Main/test.txt', 'w')
ftrain = open('ImageSets/Main/train.txt', 'w')
fval = open('ImageSets/Main/val.txt', 'w')
for i in list:
name=total_xml[i][:-4]+'\n'
if i in trainval:
ftrainval.write(name)
if i in train:
ftrain.write(name)
else:
fval.write(name)
else:
ftest.write(name)
ftrainval.close()
ftrain.close()
fval.close()
ftest .close()
生成相应的验证集和训练集以及测试集的txt文件.
四、生成labels的txt文件
用labelIamge打标签时获得的是xml格式的文件,而用于YOLOv3进行训练的是txt文件格式,下面是voc_label.py(darknet\scripts)将xml文件的格式转换为txt文件的格式,注意xml文件放在darknet\scripts\VOCdevkit\VOC2007\Annotations,运行成功后,txt文件会存放在darknet\scripts\VOCdevkit\VOC2007\labels文件夹下。下面是voc_label.py的源代码,需要修改classes为自己的训练类别,其余无需改动。
import xml.etree.ElementTree as ET
import pickle
import os
from os import listdir, getcwd
from os.path import join
sets=[ ('2007', 'train'), ('2007', 'val'), ('2007', 'test')]
classes = ["wheel","goods"]
def convert(size, box):
dw = 1./(size[0])
dh = 1./(size[1])
x = (box[0] + box[1])/2.0 - 1
y = (box[2] + box[3])/2.0 - 1
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()
os.system("cat 2007_train.txt 2007_val.txt")
os.system("cat 2007_train.txt 2007_val.txt 2007_test.txt")
五、开始训练
1.下载预训练模型
darknet53.conv.74(darknet/)是预训练模型已经下载好了
2.训练的命令
./darknet detector train cfg/voc.data cfg/yolov3-voc.cfg darknet53.conv.74
在训练一段时间后会在darknet/backup/文件夹下生成一些权重文件即模型,其中.backup是最新的权重文件,如果中途停止了训练,可以在.backup的基础上进行训练,训练的命令为
./darknet detector train cfg/voc.data cfg/yolov3-voc.cfg backup/yolov3-voc.backup
当训练的loss降至0.1以下就可以进行测试了,如果测试满足要求就可以进行使用了。
六、进行测试
测试命令为:
./darknet detector test cfg/voc.data cfg/yolov3-voc.cfg backup/yolov3-voc.backup test/4.jpg
测试成功的话会在darknet下生成predictions.jpg文件