更换pip源到国内镜像
pip国内的一些镜像
python国内源
阿里云 https://mirrors.aliyun.com/pypi/simple/
中国科技大学 https://pypi.mirrors.ustc.edu.cn/simple/
豆瓣(douban) http://pypi.douban.com/simple/
清华大学 https://pypi.tuna.tsinghua.edu.cn/simple/
修改源
临时使用：
可以在使用pip的时候在后面加上-i参数，指定pip源为国内源
|
1 |
eg: pip install scrapy -i https://pypi.tuna.tsinghua.edu.cn/simple |
永久修改：
Linux/macOS:
修改 ~/.pip/pip.conf (没有就创建一个)， 内容如下：
|
1 2 |
[global] index-url = https://pypi.tuna.tsinghua.edu.cn/simple |
对于当前用户不是root，而有时候需要使用sudo安装的情况，上述设置是无效的。需要在root用户目录也设置一下，一般情况下配置文件默认是/root/.pip/pip.conf。
Windows:
直接在user目录中创建一个pip目录，如：C:\Users\xx\pip，新建文件pip.ini，内容如下
|
1 2 |
[global] index-url = https://pypi.tuna.tsinghua.edu.cn/simple |
参考链接
解决macOS 10.15.2下Python的tkinter库Entry无法输入中文
问题描述
在 macOS 10.15.2下使用 tkinter 库 Entry 输入框，无法输入中文，不管怎么样都只能输入英文，只能复制进去中文，交互体验很不好ಥ_ಥ。
我的版本是 macOS 10.15.2，python 3.7.5。
ubuntu 18.04下安装CUDA 10.1以及cuDNN的安装
安装最新版本的 CUDA 10.1：
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 |
# 卸载之前已经安装的cuda $ sudo apt-get remove nvidia-cuda-toolkit $ wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-ubuntu1804.pin $ sudo mv cuda-ubuntu1804.pin /etc/apt/preferences.d/cuda-repository-pin-600 $ wget http://developer.download.nvidia.com/compute/cuda/10.1/Prod/local_installers/cuda-repo-ubuntu1804-10-1-local-10.1.243-418.87.00_1.0-1_amd64.deb $ sudo dpkg -i cuda-repo-ubuntu1804-10-1-local-10.1.243-418.87.00_1.0-1_amd64.deb $ sudo apt-key add /var/cuda-repo-10-1-local-10.1.243-418.87.00/7fa2af80.pub $ sudo apt-get update $ sudo apt-get -y install cuda # 部分驱动可能会更新，需要执行更新，否则可能依旧不正常 $ sudo apt-get dist-upgrade $ sudo apt-get autoremove # 可能需要删除一下XWindow的配置文件，否则驱动可能不能正常加载 $ sudo rm -rf ~/.Xauthority # 如果出现如下错误 # ubuntu 18.04 "nvidia-340 导致 /usr/lib/x86_64-linux-gnu/libGL.so.1 # 转移到 /usr/lib/x86_64-linux-gnu/libGL.so.1.distrib" # 参考 http://www.mobibrw.com/?p=21739 # 删除安装源，可以节约几个GB的磁盘，安装完成后这部分已经用不上了 $ sudo apt-get remove --purge cuda-repo-ubuntu1804-10-1-local-10.1.243-418.87.00 |
安装对应版本的cuDNN:
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 |
$ wget https://www.mobibrw.com/wp-content/uploads/2019/11/libcudnn7_7.6.5.32-1cuda10.1_amd64.deb_.zip $ wget https://www.mobibrw.com/wp-content/uploads/2019/11/libcudnn7-dev_7.6.5.32-1cuda10.1_amd64.deb_.zip $ wget https://www.mobibrw.com/wp-content/uploads/2019/11/libcudnn7-doc_7.6.5.32-1cuda10.1_amd64.deb_.zip # 解压缩 $ unzip libcudnn7_7.6.5.32-1cuda10.1_amd64.deb_.zip $ unzip libcudnn7-dev_7.6.5.32-1cuda10.1_amd64.deb_.zip $ unzip libcudnn7-doc_7.6.5.32-1cuda10.1_amd64.deb_.zip # 按照顺序安装 $ sudo dpkg -i libcudnn7_7.6.5.32-1+cuda10.1_amd64.deb $ sudo dpkg -i libcudnn7-dev_7.6.5.32-1+cuda10.1_amd64.deb $ sudo dpkg -i libcudnn7-doc_7.6.5.32-1+cuda10.1_amd64.deb |
检测是否成功安装：
|
1 2 3 4 5 6 7 |
$ cp -r /usr/src/cudnn_samples_v7/ ~/ $ cd ~/cudnn_samples_v7/mnistCUDNN $ make clean && make $ ./mnistCUDNN |
执行输出如下结果代表成功安装：
Visual Studio Code 1.41.0调试python代码时添加参数(args)
前提：代码中设置了arg paser，需要手动设置，VS code的debug没有简洁的添加参数的方式。
解决方式如下：
打开Debug->Open Configurations/ Add Configurations
在用BeautifulSoup解析HTML前对其中以JavaScript渲染部分的处理
在命令行运行：
|
1 2 3 4 5 |
# Python2 $ pip install requests-html # Python3 # pip3 install requests-html |
即可安装该模块。
例子如下：
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 |
#coding=utf-8 from bs4 import BeautifulSoup import requests from requests_html import HTMLSession #使用requests抓取页面内容，并将响应赋值给page变量 html = requests.get('https://xcx.xzlzq.net/#/liftDetail?registerCode=31103301042010020002') session = HTMLSession() first_page = session.get('https://xcx.xzlzq.net/#/liftDetail?registerCode=31103301042010020002') first_page.html.render(sleep=5) #使用content属性获取页面的源页面 #使用BeautifulSoap解析，内容传递到BeautifulSoap类 soup = BeautifulSoup(first_page.html.html,'lxml') links = soup.find_all('div',class_='content') #link的内容就是div，我们取它的span内容就是我们需要段子的内容 for link in links: print(link.span.get_text()) |
参考链接
Samsung Magician 6.0升级固件失败“在选定的驱动器上无法更新固件。”("Failed to update firmware on the selected drive.")
最近的三星固态硬盘提示固件有升级，但是固件升级失败，提示“在选定的驱动器上无法更新固件。”或者"Failed to update firmware on the selected drive."
如下图所示：
继续阅读Samsung Magician 6.0升级固件失败“在选定的驱动器上无法更新固件。”("Failed to update firmware on the selected drive.")
ubuntu 18.04安装Vulkan依赖的工具链SPIR-V Toolchain(glslangValidator)
开发 Vulkan 的时候，需要使用 glslangValidator 编译 Vulkan 代码。
如果是 ubuntu 19.10 版本，可以直接执行:
|
1 2 3 |
$ sudo apt-get install vulkan-tools # 目前ubuntu 20.04 最新需要使用如下命令 sudo apt-get install glslang-tools |
如果低于这个版本，则只能从源代码编译安装了，如下：
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 |
# 编译安装 glslang $ git clone https://github.com/KhronosGroup/glslang.git # 也可本站下载一份拷贝 wget https://www.mobibrw.com/wp-content/uploads/2018/12/glslang.zip $ cd glslang # 目前（2018.12.17）的正式版，最稳定的版本，试过最新的版本，编译部分代码存在问题 $ git checkout 7.10.2984 # 下载 spirv-tools 部分的功能代码 $ python update_glslang_sources.py $ mkdir build $ cd build $ cmake .. $ make $ sudo make install |
Python tkinter实现图片标注
安装依赖
|
1 2 3 4 5 6 7 |
$ python -m pip install --upgrade pip --user $ python -m pip install Pillow --user $ python -m pip install Numpy --user $ python -m pip install OpenCV-Python --user |
源代码：
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 |
#!/usr/bin/python # -*- coding: UTF-8 -*- import os import sys if sys.version_info < (3, 0): import Tkinter as tk # 导入 Tkinter 库 from tkFileDialog import askopenfilename, askdirectory else : import tkinter as tk # 导入 Tkinter 库 from tkinter.filedialog import askopenfilename, askdirectory from PIL import Image, ImageTk, ImageDraw from time import sleep import numpy as np import cv2 as cv import collections DEF_WIDTH = 1080 DEF_HEIGHT = 720 IMAGE_HEIGHT = 720 FRAME_LEFT_WIDTH = 360 # 太小的选定区域我们需要丢弃，防止误操作 MINI_RECT_AREA = 20 class RawImageEditor: def __init__(self, win, img, rects): #变量X和Y用来记录鼠标左键按下的位置 self.X = tk.IntVar(value=0) self.Y = tk.IntVar(value=0) self.sel = False self.lastDraw = None self.lastDraws = [] self.imageScale = 1.0 self.dispWidth = DEF_WIDTH # 图片显示区域的最大高度，宽度 self.dispHeight = DEF_HEIGHT self.rawImage = img self.calcImageScale(self.rawImage) self.dispWidth = int(self.imageScale * self.rawImage.width) self.dispHeight = int(self.imageScale * self.rawImage.height) # 图片缩放 self.dispImage = self.rawImage.resize((self.dispWidth, self.dispHeight)) # 选择区域 self.selPositions = [] for r in rects : self.selPositions.append((r[0] * self.imageScale, r[1] * self.imageScale, r[2] * self.imageScale, r[3] * self.imageScale)) #创建顶级组件容器 self.top = tk.Toplevel(win, width=self.dispWidth, height=self.dispHeight) #不显示最大化、最小化按钮 self.top.overrideredirect(True) #center window on desktop nScreenWid, nScreenHei = self.top.maxsize() winPos = '%sx%s+%s+%s' % (int(self.top.winfo_reqwidth()), int(self.top.winfo_reqheight()), int((nScreenWid - self.top.winfo_reqwidth())/2), int((nScreenHei - self.top.winfo_reqheight())/2)) self.top.geometry(winPos) # Make topLevelWindow remain on top until destroyed, or attribute changes. self.top.attributes('-topmost', 'true') self.canvas = tk.Canvas(self.top, bg='white', width=self.dispWidth, height=self.dispHeight) self.tkImage = ImageTk.PhotoImage(self.dispImage) self.canvas.create_image(self.dispWidth//2, self.dispHeight//2, image=self.tkImage) for r in self.selPositions : draw = self.canvas.create_rectangle(r[0], r[1], r[2], r[3], outline='green') self.lastDraws.append(draw) #鼠标左键按下的位置 def onLeftButtonDown(event): self.X.set(event.x) self.Y.set(event.y) #开始截图 self.sel = True #重新绘制已经选择的区域 for draw in self.lastDraws : self.canvas.delete(draw) self.lastDraws = [] for r in self.selPositions : draw = self.canvas.create_rectangle(r[0], r[1], r[2], r[3], outline='green') self.lastDraws.append(draw) self.canvas.bind('<Button-1>', onLeftButtonDown) #鼠标左键移动，显示选取的区域 def onLeftButtonMove(event): if not self.sel: return try: #删除刚画完的图形，要不然鼠标移动的时候是黑乎乎的一片矩形 self.canvas.delete(self.lastDraw) except Exception as e: pass self.lastDraw = self.canvas.create_rectangle(self.X.get(), self.Y.get(), event.x, event.y, outline='green') self.canvas.bind('<B1-Motion>', onLeftButtonMove) #获取鼠标左键抬起的位置，保存区域截图 def onLeftButtonUp(event): if not self.sel: return self.sel = False sleep(0.1) #考虑鼠标左键从右下方按下而从左上方抬起的截图 left, right = sorted([self.X.get(), event.x]) top, bottom = sorted([self.Y.get(), event.y]) if (right - left) * (bottom - top) > MINI_RECT_AREA : self.selPositions.append((left,top,right,bottom)) #self.top.destroy() #鼠标右键按下 def onRightButtonDown(event): self.sel = False self.top.destroy() self.canvas.bind('<Button-2>', onRightButtonDown) self.canvas.bind('<ButtonRelease-1>', onLeftButtonUp) self.canvas.pack(fill=tk.BOTH, expand=tk.YES) def calcImageScale(self, image) : w = image.width h = image.height self.imageScale = 1.0 # 计算最小的缩放比例，保证原始宽高比 if w > self.dispWidth and h > self.dispHeight : ws = self.dispWidth * 1.0 / w hs = self.dispHeight * 1.0 / h if ws < hs : self.imageScale = ws else : self.imageScale = hs elif w > self.dispWidth and h < self.dispHeight : self.imageScale = self.dispWidth * 1.0 / w elif w < self.dispWidth and h > self.dispHeight : self.imageScale = self.dispHeight * 1.0 / h def waitForWindow(self, win) : win.wait_window(self.top) def selectedPositions(self) : # 转换为原始像素位置 realPos = [] for r in self.selPositions : realPos.append((r[0] / self.imageScale, r[1] / self.imageScale, r[2] / self.imageScale, r[3] / self.imageScale)) return realPos class MainWin(tk.Tk): def __init__(self): if sys.version_info >= (3, 0): super().__init__() else : tk.Tk.__init__(self) self.title('图像处理工具') self.geometry('{}x{}'.format(DEF_WIDTH, DEF_HEIGHT)) self.rawImagePath = '' self.rawImage = None # self.rawImage 原始图像，未经过缩放处理 self.transRawImage = None # self.transRawImage 经过转换处理之后的原始图像，没有经过缩放处理 self.dispImage = None # self.dispImage 显示图像，可能经过缩放处理 self.imageScale = 1.0 # 图片缩放比例，根据缩放比例进行显示的时候的缩放处理，后期选择区域的时候，需要进行缩放还原 self.leftFrameWidth = FRAME_LEFT_WIDTH self.frameDispHeight = DEF_HEIGHT # 整个窗口的高度 self.labelTextHeight = 18 # 文本标签的高度 self.btnHeight = 30 # 按钮的高度 self.brightnessScale = tk.StringVar() # 亮度比 self.brightnessScale.set('1.0') self.defSavePath = '' # 默认保存路径 self.imageDispWidth = IMAGE_HEIGHT # 图片显示区域的最大高度，宽度 self.imageDispHeight = self.frameDispHeight / 2 - self.labelTextHeight * 2 self.liRect = collections.OrderedDict() # 选择区域 self.rawImageEditor = None self.currentListBoxSelIdx = None #当前选择的项目 self.setupUI() def scaleDisplayImage(self, image) : w = image.width h = image.height self.imageScale = 1.0 # 计算最小的缩放比例，保证原始宽高比 if w > self.imageDispWidth and h > self.imageDispHeight : ws = self.imageDispWidth * 1.0 / w hs = self.imageDispHeight * 1.0 / h if ws < hs : self.imageScale = ws else : self.imageScale = hs elif w > self.imageDispWidth and h < self.imageDispHeight : self.imageScale = self.imageDispWidth * 1.0 / w elif w < self.imageDispWidth and h > self.imageDispHeight : self.imageScale = self.imageDispHeight * 1.0 / h # 图片缩放 return image.resize((int(self.imageScale * w), int(self.imageScale * h))) def loadImageCfgFile(self, imf): (path, name) = os.path.split(imf) cfgname = imf + '.txt' if (not os.path.exists(cfgname)) or (not os.path.isfile(cfgname)) : cfgname = os.path.join(path, 'mask', name) + '.txt' if (not os.path.exists(cfgname)) or (not os.path.isfile(cfgname)) : cfgname = os.path.join(path, 'cfg', name) + '.txt' self.liRect.clear() if os.path.exists(cfgname) and os.path.isfile(cfgname) : with open(cfgname, "r") as f: lines = f.readlines() for line in lines: rs = line.split(',') r = (float(rs[0].strip()), float(rs[1].strip()), float(rs[2].strip()), float(rs[3].strip())) if len(rs) > 4 : self.liRect[r] = rs[4].strip() else : self.liRect[r] = '' # 打开图片时使用，传值（图）给展示函数 def openAndDisplayImage(self): imF = self.selectImageFile() if '' != imF : self.rawImagePath = imF self.loadImageCfgFile(self.rawImagePath) self.drawListBox() self.rawImage = Image.open(self.rawImagePath) self.rawImage = self.rawImage.convert('RGBA') self.drawRawImageDisp() self.image_l_trans.image = None self.transRawImage = None def drawListBox(self): self.l_box.delete(0, tk.END) for r in self.liRect.keys(): r = '{},{},{},{}'.format(round(r[0],1), round(r[1],1), round(r[2],1), round(r[3],1)) self.l_box.insert(0, r) def drawRawImageDisp(self, selItems=[]): self.dispImage = self.scaleDisplayImage(self.rawImage) self.dispImage = self.dispImage.convert('RGB') draw = ImageDraw.Draw(self.dispImage) rs = list(self.liRect.keys()) for i in range(len(rs)) : r = rs[i] if i in selItems : draw.rectangle((r[0] * self.imageScale, r[1] * self.imageScale, r[2] * self.imageScale, r[3] * self.imageScale), outline = "red") else : draw.rectangle((r[0] * self.imageScale, r[1] * self.imageScale, r[2] * self.imageScale, r[3] * self.imageScale), outline = "green") img = ImageTk.PhotoImage(self.dispImage) self.image_l_raw.config(image=img) self.image_l_raw.image = img def deleteSelectedItemFromListBox(self): #print(self.l_box.get(self.l_box.curselection())) idx = self.l_box.curselection() if len(idx) > 0 and (None == self.rawImageEditor) : ro = collections.OrderedDict() rs = self.liRect.keys() for i in range(len(rs)) : if i not in idx : r = rs[i] ro[r] = self.liRect[r] self.liRect = ro self.drawListBox() self.drawRawImageDisp() # 打开图片时使用，获得地址 def selectImageFile(self): path = tk.StringVar() file_entry = tk.Entry(self, state='readonly', text=path) path_ = askopenfilename() path.set(path_) return file_entry.get() def rawImageLabelClicked(self, event): if (None != self.rawImage) and (None == self.rawImageEditor) : self.rawImageEditor = RawImageEditor(self, self.rawImage, self.liRect.keys()) self.rawImageEditor.waitForWindow(self.image_l_raw) rs = self.rawImageEditor.selectedPositions() trs = collections.OrderedDict() for k in rs : if k in self.liRect : trs[k] = self.liRect[k] else: trs[k] = '' self.liRect = trs self.rawImageEditor = None self.drawListBox() self.drawRawImageDisp() def onRectListboxSelect(self, event): idx = self.l_box.curselection() if len(idx) > 0 and (self.currentListBoxSelIdx != idx) : self.currentListBoxSelIdx = idx ctx = '' rs = list(self.liRect.keys()) for i in range(len(self.currentListBoxSelIdx)) : v = self.currentListBoxSelIdx[i] k = rs[v] ctx = ctx + self.liRect[k] if i < len(self.currentListBoxSelIdx)-1 : ctx = ctx + ',' self.edtListBoxSel.delete(0, tk.END) self.edtListBoxSel.insert(0, ctx) self.drawRawImageDisp(idx) def onEdtListBoxComplete(self, event): idx = self.l_box.curselection() if (len(idx) > 0) and (self.currentListBoxSelIdx == idx) : ctx = self.edtListBoxSel.get().strip() ctx = ctx.split(',') while len(idx) > len(ctx) : ctx.append('') rs = list(self.liRect.keys()) n = 0 for i in idx : r = rs[i] self.liRect[r] = ctx[n] n = n + 1 def drawTransImageDisp(self): transImage = self.scaleDisplayImage(self.transRawImage) transImage = transImage.convert('L') img = ImageTk.PhotoImage(transImage) self.image_l_trans.config(image=img) self.image_l_trans.image = img def doTransRawImage(self): self.transRawImage = Image.new('L', (self.rawImage.width, self.rawImage.height)) rs = self.liRect.keys() for r in rs : im = self.rawImage.crop(r) cv_im = cv.cvtColor(np.asarray(im), cv.COLOR_RGB2BGR) hsv = cv.cvtColor(cv_im, cv.COLOR_BGR2HSV) _, _, v = cv.split(hsv) avg = np.average(v.flatten()) pixels = im.load() scale = float(self.brightnessScale.get()) for j in range(im.height) : for i in range(im.width) : hv = v[j,i] if hv < avg * scale: #im.putpixel((i, j), 0) # pixels[i, j] = 0 '''else : im.putpixel((i, j), (255, 255, 255, 255))''' self.transRawImage.paste(im, (int(r[0]),int(r[1])), mask = None) self.drawTransImageDisp() def onTransRawImageBtnClicked(self): if None != self.rawImage : self.doTransRawImage() def saveTransCfg(self, pth): (path,name) = os.path.split(self.rawImagePath) cfg = os.path.join(pth, 'cfg') if os.path.exists(cfg) and os.path.isdir(cfg) : cfgname = os.path.join(cfg, name) + '.txt' else : cfgname = os.path.join(pth, name) + '.txt' with open(cfgname, "w") as f: for r in self.liRect.keys() : v = self.liRect[r] line = '{},{},{},{},{}\n'.format(r[0], r[1], r[2], r[3], v) f.write(line) def saveCombinedImage(self, pth): im_A = cv.cvtColor(np.array(self.rawImage), cv.COLOR_RGB2BGR) im_B = cv.cvtColor(np.array(self.transRawImage), cv.COLOR_RGB2BGR) im_AB = np.concatenate([im_A, im_B], 1) ext = os.path.splitext(self.rawImagePath)[-1] ext = ext.lower() name = os.path.basename(self.rawImagePath) sp = name.split('.')[:-1] comb = os.path.join(pth, 'comb') if os.path.exists(comb) and os.path.isdir(comb) : name = sp[0] + ext pth = comb else : name = sp[0] + '_comb' + ext path_AB = os.path.join(pth, name) cv.imwrite(path_AB, im_AB) def saveTransImage(self, path): ext = os.path.splitext(self.rawImagePath)[-1] ext = ext.lower() (path,name) = os.path.split(self.rawImagePath) mask = os.path.join(path, 'mask') fn = name.split('.')[:-1] if os.path.exists(mask) and os.path.isdir(mask) : fn = fn[0] + ext else : fn = fn[0] + '_mask' + ext im = os.path.join(mask, fn) if not im.endswith(ext) : im = im + ext self.transRawImage.save(im) def saveTransInformation(self, path): if '' != path : self.saveTransImage(path) self.saveTransCfg(path) self.saveCombinedImage(path) def onSaveTransRawImageBtnClicked(self): if None != self.transRawImage : (path,name) = os.path.split(self.rawImagePath) dirname = askdirectory(initialdir = os.path.expanduser(path), title = '保存结果') if '' != dirname : self.defSavePath = dirname self.saveTransInformation(self.defSavePath) def onDefaultSaveTransRawImageBtnClicked(self) : if (None != self.transRawImage) and ('' != self.defSavePath) : self.saveTransInformation(self.defSavePath) def setupUI(self): # 左边菜单栏 left_f = tk.Frame(self, height=self.frameDispHeight, width=self.leftFrameWidth) left_f.pack(side=tk.LEFT) # 各种功能按钮名称及位置 btnOpen = tk.Button(left_f, text='打开图像', command=self.openAndDisplayImage) btnOpen.place(y=25, x=30, width=300, height=self.btnHeight) btnTrans = tk.Button(left_f, text='处理图像', command=self.onTransRawImageBtnClicked) btnTrans.place(y=75, x=30, width=300, height=self.btnHeight) l_selRect = tk.Label(left_f, text = '鼠标选定区域') l_selRect.place(x=0, y=125, width=self.leftFrameWidth, height=self.labelTextHeight) '''列表''' self.l_box = tk.Listbox(left_f, exportselection=False) # 创建两个列表组件 self.l_box.place(x=1, y=125+self.labelTextHeight, width=self.leftFrameWidth-2, height=270) self.l_box.bind('<<ListboxSelect>>', self.onRectListboxSelect) self.drawListBox() # 选中列表注释 ledt = tk.Label(left_f, text = '选中区域内容注释（回车结束）：', anchor = 'w') ledt.place(x=1, y=420, width=self.leftFrameWidth-2) self.edtListBoxSel = tk.Entry(left_f) self.edtListBoxSel.place(x=1, y=440, width=self.leftFrameWidth-2) self.edtListBoxSel.bind('<Return>', self.onEdtListBoxComplete) # 亮度比例调节 tkScale = tk.Scale(left_f, from_=0.2, to=1.8, orient=tk.HORIZONTAL, resolution=0.1, length=self.leftFrameWidth-2, variable=self.brightnessScale, label = '亮度比:') tkScale.place(x=1, y=480) # 删除选定项 btnDel = tk.Button(left_f, text='删除选定项', command=self.deleteSelectedItemFromListBox) btnDel.place(y=550, x=30, width=300, height=self.btnHeight) btnSave = tk.Button(left_f, text='保存结果', command=self.onSaveTransRawImageBtnClicked) btnSave.place(y=600, x=30, width=300, height=self.btnHeight) btnDefSave = tk.Button(left_f, text='默认或上次位置保存', command=self.onDefaultSaveTransRawImageBtnClicked) btnDefSave.place(y=650, x=30, width=300, height=self.btnHeight) # 右侧图像显示栏 right_f = tk.Frame(self, height=self.frameDispHeight, width=self.imageDispWidth) right_f.pack(side=tk.RIGHT) l_rawT = tk.Label(right_f, text = '原始图片') l_rawT.place(x=0, y=0, width=self.imageDispWidth, height=self.labelTextHeight) self.image_l_raw = tk.Label(right_f, relief='ridge') self.image_l_raw.place(x=0, y=self.labelTextHeight, width=self.imageDispWidth, height=self.imageDispHeight) self.image_l_raw.bind("<Button-1>",self.rawImageLabelClicked) l_transT = tk.Label(right_f, text = '处理后图片') l_transT.place(x=0, y=self.labelTextHeight + self.imageDispHeight, width=self.imageDispWidth, height=self.labelTextHeight) self.image_l_trans = tk.Label(right_f, relief='ridge') self.image_l_trans.place(x=0, y=self.labelTextHeight + self.imageDispHeight + self.labelTextHeight, width=self.imageDispWidth, height=self.imageDispHeight) if __name__ == '__main__' : win = MainWin() # 进入消息循环 win.mainloop() |
参考链接
- python，一个入门级极简单的图像处理工具（一）
- python，一个入门级极简单的图像处理工具（二）
- Python tkinter 实现本地打开图片进行标注
- Python tkinter Canvas画布完全攻略（超级详细）
- 用python实现选择截图区域
- Python GUI编程(Tkinter)
- Python OpenCV实现鼠标画框
- 用python实现选择截图区域
- OpenCV-Python 选择ROI
- python3.6+opencv3.4实现鼠标交互查看图片像素
- Tkinter Toplevel always in front
- TypeError: __str__ returned non-string (type instance)
- Python Canvas.create_image Examples
- Tkinter tkFileDialog module
- python 调用super()初始化报错“TypeError: super() takes at least 1 argument”
- python – 使用PIL中的Image.point()方法来操作像素数据
- Using the Image.point() method in PIL to manipulate pixel data
- Pixel to pixel edit using PIL and Image.point
- PixelAccess Class
- python tkinter界面居中显示的方法
- MAC 系统中，Tkinter 无法用 中文输入法 输入中文
- macOS下Python的tkinter库Entry输入框无法输入中文问题的原因及解决办法
- IDLE and tkinter with Tcl/Tk on macOS
在ubuntu 18.04(GeForce GTX 760 4GB显存)使用Pytorch Pix2PixGAN(CUDA-10.1)
1. 参照 pytorch 1.0.1在ubuntu 18.04(GeForce GTX 760)编译(CUDA-10.1) 建立 pytorch 1.0.1 的编译环境，并解决编译时遇到的问题。
2. 依旧是推荐在 Anaconda 上建立独立的编译环境，然后执行编译：
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 |
$ sudo apt-get install git # conda remove -n Pix2Pix --all $ conda create -n Pix2Pix -y python=3.6.8 pip $ source activate Pix2Pix $ conda install numpy pyyaml mkl=2019.1 mkl-include=2019.1 setuptools cmake cffi typing pybind11 $ conda install ninja # magma-cuda90 magma-cuda91 magma-cuda92 会编译失败 $ conda install -c pytorch magma-cuda101 $ git clone https://github.com/pytorch/pytorch $ cd pytorch # pytorch 1.0.1 版本支持“Compute Capability” 低于3.0版本的硬件，pytorch 1.2.0需要至少3.5版本的硬件才可以正常运行 # https://github.com/pytorch/pytorch/blob/v1.3.0/torch/utils/cpp_extension.py $ git checkout v1.0.1 -b v1.0.1 $ git submodule sync $ git submodule update --init --recursive $ export CMAKE_PREFIX_PATH=${CONDA_PREFIX:-"$(dirname $(which conda))/../"} # 如果不需要使用cuda的话，这里还要加上一句：export NO_CUDA=1 $ python setup.py clean # 卸载以前安装的pytorch $ conda uninstall pytorch # 从Nvidia开发网站查询到自己硬件对应的“Compute Capability” # 比如 “GeForce GTX 760” 对应 “3.0” 计算能力，能力不正确会导致运行异常 # RuntimeError: cuda runtime error (48) : no kernel image is available for execution on the device $ python setup.py install # 对于开发者模式，可以使用 # python setup.py build develop # 一定要退出 pytorch 的编译目录，在pytorch代码目录下执行命令会出现异常 $ cd .. # 退出环境 $ conda deactivate |
编译出错信息，参考 pytorch 1.0.1在ubuntu 18.04(GeForce GTX 760)编译(CUDA-10.1) 里面的介绍解决。
3. 编译安装 TorchVision
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 |
$ sudo apt-get install git # 进入运行环境 $ source activate Pix2Pix $ git clone https://github.com/pytorch/vision.git # 也可本站下载一份拷贝 wget https://www.mobibrw.com/wp-content/uploads/2019/11/vision.zip $ cd vision $ git checkout v0.2.1 -b v0.2.1 $ python setup.py install # 退出环境 $ conda deactivate |
4. 检出 CycleGAN and pix2pix in PyTorch 的代码，并安装依赖
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 |
# 进入运行环境 $ source activate Pix2Pix $ git clone https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix.git # 也可本站下载 wget https://www.mobibrw.com/wp-content/uploads/2019/12/pytorch-CycleGAN-and-pix2pix.zip $ cd pytorch-CycleGAN-and-pix2pix # 下载人脸替换部分的数据集 $ bash datasets/download_pix2pix_dataset.sh facades # 也可本站下载然后自己参照脚本解压缩到指定目录 https://www.mobibrw.com/wp-content/uploads/2019/12/facades.tar.gz # 安装依赖 $ pip install pillow==6.2.1 $ pip install dominate==2.4.0 $ pip install visdom==0.1.8.9 # 修正错误 models/networks.py # TypeError: cuda() got an unexpected keyword argument 'device_id' $ sed -i "s/netG\.cuda(device_id=gpu_ids\[0\])/netG.cuda(gpu_ids[0])/g" models/networks.py $ sed -i "s/netD\.cuda(device_id=gpu_ids\[0\])/netD.cuda(gpu_ids[0])/g" models/networks.py $ sed -i "s/network\.cuda(device_id=gpu_ids\[0\])/network.cuda(gpu_ids[0])/g" models/base_model.py # 开启WEB服务,主要是第一次运行需要下载部分辅助软件包， # 训练之前需要执行，否则下面训练的时候会报错 $ python -m visdom.server & # 等待屏幕上出现 “You can navigate to http://localhost:8097” 代表服务启动成功 # 执行训练 $ bash scripts/train_pix2pix.sh |
执行训练的时候，如果出现如下错误：
|
1 2 3 4 5 6 |
Traceback (most recent call last): File "train.py", line 47, in <module> errors = model.get_current_errors() File "~/pytorch-CycleGAN-and-pix2pix/models/pix2pix_model.py", line 122, in get_current_errors return OrderedDict([('G_GAN', self.loss_G_GAN.data[0]), IndexError: invalid index of a 0-dim tensor. Use tensor.item() to convert a 0-dim tensor to a Python number |
这个原因是由于 PyTorch 版本差异造成的,（作者在 Pytorch 0.4.1 版本上测试，我们在 Pytorch 1.0.1 版本上测试），执行如下命令修复：
|
1 2 3 4 5 6 7 8 9 |
#loss_G_GAN.data[0] 替换为 loss_G_GAN.item() $ sed -i "s/self\.loss_G_GAN\.data\[0]/self.loss_G_GAN.item()/g" models/pix2pix_model.py $ sed -i "s/self\.loss_G_L1\.data\[0]/self.loss_G_L1.item()/g" models/pix2pix_model.py $ sed -i "s/self\.loss_D_real\.data\[0]/self.loss_D_real.item()/g" models/pix2pix_model.py $ sed -i "s/self\.loss_D_fake\.data\[0]/self.loss_D_fake.item()/g" models/pix2pix_model.py |
5. 测试训练结果
|
1 2 3 |
$ bash scripts/test_pix2pix.sh # 观察结果需要打开 ./results/facades_pix2pix/test_latest/index.html |