pip install labelme
然后在桌面搜索框中找到labelme,然后打开,或者直接在命令行中输入labelme进行打开
安装labelme过程中出现的一些问题:
https://blog.csdn.net/qq_44747572/article/details/127584015?spm=1001.2014.3001.5501
标注步骤:
快捷键:
coco数据集目录结构
如下图所示,其中train2017、test2017、val2017文件夹中保存的是用于训练、测试、验证的图片,而annotations文件夹保存的是这些图片对应的标注信息,分别存在instance_test2017、instance_test2017、instance_val2017三个json文件中。
labelme标注的数据转换成coco格式:
# -*- coding:utf-8 -*-
import argparse
import json
import matplotlib.pyplot as plt
import skimage.io as io
# import cv2
from labelme import utils
import numpy as np
import glob
import PIL.Image
class MyEncoder(json.JSONEncoder):
def default(self, obj):
if isinstance(obj, np.integer):
return int(obj)
elif isinstance(obj, np.floating):
return float(obj)
elif isinstance(obj, np.ndarray):
return obj.tolist()
else:
return super(MyEncoder, self).default(obj)
class labelme2coco(object):
def __init__(self, labelme_json=[], save_json_path='./tran.json'):
self.labelme_json = labelme_json
self.save_json_path = save_json_path
self.images = []
self.categories = []
self.annotations = []
# self.data_coco = {}
self.label = []
self.annID = 1
self.height = 0
self.width = 0
self.save_json()
def data_transfer(self):
for num, json_file in enumerate(self.labelme_json):
with open(json_file, 'r') as fp:
data = json.load(fp) # 加载json文件
self.images.append(self.image(data, num))
for shapes in data['shapes']:
label = shapes['label']
if label not in self.label:
self.categories.append(self.categorie(label))
self.label.append(label)
points = shapes['points'] # 这里的point是用rectangle标注得到的,只有两个点,需要转成四个点
points.append([points[0][0], points[1][1]])
points.append([points[1][0], points[0][1]])
self.annotations.append(self.annotation(points, label, num))
self.annID += 1
def image(self, data, num):
image = {}
img = utils.img_b64_to_arr(data['imageData']) # 解析原图片数据
# img=io.imread(data['imagePath']) # 通过图片路径打开图片
# img = cv2.imread(data['imagePath'], 0)
height, width = img.shape[:2]
img = None
image['height'] = height
image['width'] = width
image['id'] = num + 1
image['file_name'] = data['imagePath'].split('/')[-1]
self.height = height
self.width = width
return image
def categorie(self, label):
categorie = {}
categorie['supercategory'] = label
categorie['id'] = len(self.label) + 1 # 0 默认为背景
categorie['name'] = label
return categorie
def annotation(self, points, label, num):
annotation = {}
annotation['segmentation'] = [list(np.asarray(points).flatten())]
annotation['iscrowd'] = 0
annotation['image_id'] = num + 1
# annotation['bbox'] = str(self.getbbox(points)) # 使用list保存json文件时报错(不知道为什么)
# list(map(int,a[1:-1].split(','))) a=annotation['bbox'] 使用该方式转成list
annotation['bbox'] = list(map(float, self.getbbox(points)))
annotation['area'] = annotation['bbox'][2] * annotation['bbox'][3]
# annotation['category_id'] = self.getcatid(label)
annotation['category_id'] = self.getcatid(label) # 注意,源代码默认为1
annotation['id'] = self.annID
return annotation
def getcatid(self, label):
for categorie in self.categories:
if label == categorie['name']:
return categorie['id']
return 1
def getbbox(self, points):
# img = np.zeros([self.height,self.width],np.uint8)
# cv2.polylines(img, [np.asarray(points)], True, 1, lineType=cv2.LINE_AA) # 画边界线
# cv2.fillPoly(img, [np.asarray(points)], 1) # 画多边形 内部像素值为1
polygons = points
mask = self.polygons_to_mask([self.height, self.width], polygons)
return self.mask2box(mask)
def mask2box(self, mask):
'''从mask反算出其边框
mask:[h,w] 0、1组成的图片
1对应对象,只需计算1对应的行列号(左上角行列号,右下角行列号,就可以算出其边框)
'''
# np.where(mask==1)
index = np.argwhere(mask == 1)
rows = index[:, 0]
clos = index[:, 1]
# 解析左上角行列号
left_top_r = np.min(rows) # y
left_top_c = np.min(clos) # x
# 解析右下角行列号
right_bottom_r = np.max(rows)
right_bottom_c = np.max(clos)
# return [(left_top_r,left_top_c),(right_bottom_r,right_bottom_c)]
# return [(left_top_c, left_top_r), (right_bottom_c, right_bottom_r)]
# return [left_top_c, left_top_r, right_bottom_c, right_bottom_r] # [x1,y1,x2,y2]
return [left_top_c, left_top_r, right_bottom_c - left_top_c,
right_bottom_r - left_top_r] # [x1,y1,w,h] 对应COCO的bbox格式
def polygons_to_mask(self, img_shape, polygons):
mask = np.zeros(img_shape, dtype=np.uint8)
mask = PIL.Image.fromarray(mask)
xy = list(map(tuple, polygons))
PIL.ImageDraw.Draw(mask).polygon(xy=xy, outline=1, fill=1)
mask = np.array(mask, dtype=bool)
return mask
def data2coco(self):
data_coco = {}
data_coco['images'] = self.images
data_coco['categories'] = self.categories
data_coco['annotations'] = self.annotations
return data_coco
def save_json(self):
self.data_transfer()
self.data_coco = self.data2coco()
# 保存json文件
json.dump(self.data_coco, open(self.save_json_path, 'w'), indent=4, cls=MyEncoder) # indent=4 更加美观显示
labelme_json = glob.glob(r'D:\Users\80080947\Desktop\yxLocalWork\ObjectDetection\data\Annotations/*.json')
# labelme_json=['./1.json']
labelme2coco(labelme_json, r'D:\Users\80080947\Desktop\yxLocalWork\ObjectDetection\data\Json\instances_train.json')
下载swin-transformer代码
git clone https://github.com/SwinTransformer/Swin-Transformer-Object-Detection.git
cd Swin-Transformer-Object-Detection
pip install -r requirements.txt
python setup.py develop
环境配置(结合后面的看,这个会出现apex安装的问题)
mmcv-full的安装: 要注意版本的对应,可在下面进行版本的选择,进行安装。
# 命令行输入 可以查看torch和cuda的版本
python -c 'import torch;print(torch.__version__);print(torch.version.cuda)'
查看链接: https://mmcv.readthedocs.io/en/latest/get_started/installation.html
#需要注意的是pytorch版本、cuda版本与mmcv版本需搭配,否则会出错。
#我是cuda10.2 pytorch1.7.0 python3.7 mmcv-full 1.3.1
pip install mmcv-full==1.3.1 -f https://download.openmmlab.com/mmcv/dist/cu102/torch1.7/index.html
测试代码:https://blog.csdn.net/qq_44747572/article/details/127604916?spm=1001.2014.3001.5501
测试结果:
tools
train基本不需要改
config
workdir:生成训练结果
修改的地方
类别
权重文件
图片大小(太大会导致难以训练)
数据集路径配置
batch size设置
# /configs/_base_/models/mask_rcnn_swin_fpn.py
#num_classes=80,#类别
num_classes=2, # 训练的类别是2
配置权重信息
# 修改 configs/base/default_runtime.py 中的 interval,loadfrom
# interval:dict(interval=1) # 表示多少个 epoch 验证一次,然后保存一次权重信息
# loadfrom:表示加载哪一个训练好(预训练)的权重,可以直接写绝对路径如:
# load_from = r"/media/yuanxingWorkSpace/studyProject/ObjectDetection/Swin-Transformer-Object-Detection/checkpoints/mask_rcnn_swin_tiny_patch4_window7.pth"
修改训练图片尺寸大小
# 如果显存够的话可以不改(基本都运行不起来),文件位置为:configs/swin/mask_rcnn_swin_tiny_patch4_window7_mstrain_480-800_adamw_3x_coco.py
# 修改所有的 img_scale 为 :img_scale = [(224, 224)] 或者 img_scale = [(256, 256)] 或者 480,512等。
# 同时 configs/base/datasets/coco_instance.py 中的 img_scale 也要改成 img_scale = [(224, 224)] 或者其他值
# 注意:值应该为32的倍数,大小根据显存或者显卡的性能自行调整
配置数据集路径
# configs/base/datasets/coco_instance.py
# 修改data_root文件的最上面指定了数据集的路径,因此在项目下新建 data/coco目录,下面四个子目录 annotations和test2017,train2017,val2017。
修改该文件下的 train val test 的路径为自己新建的路径
configs/base/datasets/coco_instance.py
修改 batch size 和 线程数
路径:configs/base/datasets/coco_instance.py ,根据自己的显存和CPU来设置
修改分类数组
mmdet/datasets/coco.py
# CLASSES中填写自己的分类:
CLASSES = ('LV', 'LA')
修改最大epoch
configs/swin/mask_rcnn_swin_tiny_patch4_window7_mstrain_480-800_adamw_3x_coco.py
修改72行:runner = dict(type=‘EpochBasedRunnerAmp’, max_epochs=36)#最大epochs
在终端输入
python tools/train.py configs\swin\mask_rcnn_swin_tiny_patch4_window7_mstrain_480-800_adamw_3x_coco.py
报错1:ImportError: cannot import name 'OrderedDict' from 'typing' (/home/yuanxing/anaconda3/envs/ObjectDetection/lib/python3.7/typing.py)
原因:是由于python版本为3.7.1
解决:(ObjectDetection) yuanxing@psdz:/media/yuanxingWorkSpace/studyProject/ObjectDetection/Swin-Transformer-Object-Detection$ conda install python=3.7.2
报错2:ImportError: /home/yuanxing/anaconda3/envs/ObjectDetection/lib/python3.7/site-packages/mmcv/_ext.cpython-37m-x86_64-linux-gnu.so: undefined symbol: _ZN6caffe28TypeMeta21_typeMetaDataInstanceIdEEPKNS_6detail12TypeMetaDataEv
原因:可能会在安装 mmcv-full 后升级您的 pytorch 版本
解决:
(ObjectDetection) yuanxing@psdz:/media/yuanxingWorkSpace/studyProject/ObjectDetection$ python -c 'import torch;print(torch.__version__);print(torch.version.cuda)'
1.13.0+cu117
11.7
发现版本不对,卸载torch和torchvision,再次查看版本,发现版本回到了torch1.7.0和cuda10.2
因此根据版本对应原则卸载mmcv-full,然后再下载
pip install mmcv-full==1.3.1 -f https://download.openmmlab.com/mmcv/dist/cu102/torch1.7/index.html
报错3:路径不对:
全部修改成绝对路径
python tools/train.py /media/yuanxingWorkSpace/studyProject/ObjectDetection/Swin-Transformer-Object-Detection/configs/swin/mask_rcnn_swin_tiny_patch4_window7_mstrain_480-800_adamw_3x_coco.py
报错4:NameError: name 'apex' is not defined
安装成功后
AttributeError: module 'torch.distributed' has no attribute '_all_gather_base'
又报错,应该是torch的版本问题
由于NameError: name 'apex' is not defined
没办法解决,打算重新装下环境
# 创建环境
conda create --name ObjectDetection python==3.7.0
# 安装torch 1.8.0的版本
conda install pytorch==1.8.0 torchvision==0.9.0 torchaudio==0.8.0 cudatoolkit=10.2 -c pytorch
# 安装mmdetection
cd Swin-Transformer-Object-Detection-master
pip install -r requirements.txt -i https://pypi.douban.com/simple/
python setup.py develop
# 安装 mmcv (cuda与torch版本号可自行修改)
# 查看相对应版本
# https://mmcv.readthedocs.io/en/latest/get_started/installation.html
python -c 'import torch;print(torch.__version__);print(torch.version.cuda)'
pip install mmcv-full==1.3.1 -f https://download.openmmlab.com/mmcv/dist/cu102/torch1.8/index.html
# 安装apex
git clone https://github.com/NVIDIA/apex.git
cd apex
pip install -r requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple
pip install -v --no-cache-dir --global-option="--cpp_ext" --global-option="--cuda_ext" . # 报错
python setup.py install --cpp_ext # 可行
# 运行报错
pip uninstall apex #成功,但是import apex.amp会报错
# 再进行配置
pip install -v --disable-pip-version-check --no-cache-dir ./
https://blog.csdn.net/u014061630/article/details/88756644
https://blog.csdn.net/qq_45720073/article/details/125772205
https://blog.csdn.net/hasque2019/article/details/121899614
https://blog.csdn.net/weixin_38429450/article/details/112759862
https://blog.csdn.net/ViatorSun/article/details/124562686
https://segmentfault.com/a/1190000041521916
https://blog.csdn.net/qq_41964545/article/details/115868473
https://blog.csdn.net/weixin_42766091/article/details/112157014
https://blog.csdn.net/qq_41888086/article/details/125647024
https://github.com/nvidia/apex#linux