Pytorch的mean和std调查实例

Python  /  管理员 发布于 5年前   651

如下所示：

# coding: utf-8from __future__ import print_functionimport copyimport clickimport cv2import numpy as npimport torchfrom torch.autograd import Variablefrom torchvision import models, transformsimport matplotlib.pyplot as pltimport load_caffemodelimport scipy.io as sio# if model has LSTM# torch.backends.cudnn.enabled = Falseimgpath = 'D:/ck/files_detected_face224/'   imgname = 'S055_002_00000025.png' # angerimage_path = imgpath + imgnamemean_file = [0.485, 0.456, 0.406]std_file = [0.229, 0.224, 0.225]raw_image = cv2.imread(image_path)[..., ::-1]print(raw_image.shape)raw_image = cv2.resize(raw_image, (224, ) * 2)image = transforms.Compose([  transforms.ToTensor(),  transforms.Normalize(    mean=mean_file,    std =std_file,    #mean = mean_file,    #std = std_file,  )])(raw_image).unsqueeze(0)print(image.shape)convert_image1 = image.numpy()convert_image1 = np.squeeze(convert_image1) # 3* 224 *224, C * H * Wconvert_image1 = convert_image1 * np.reshape(std_file,(3,1,1)) + np.reshape(mean_file,(3,1,1))convert_image1 = np.transpose(convert_image1, (1,2,0)) # H * W * Cprint(convert_image1.shape)convert_image1 = convert_image1 * 255diff = raw_image - convert_image1err = np.max(diff)print(err)plt.imshow(np.uint8(convert_image1))plt.show()

结论：

input_image = (raw_image / 255 - mean) ./ std 

下面调查均值文件和方差文件是如何生成的：

mean_file = [0.485, 0.456, 0.406]std_file = [0.229, 0.224, 0.225]

# coding: utf-8import matplotlib.pyplot as pltimport argparseimport osimport numpy as npimport torchvisionimport torchvision.transforms as transformsdataset_names = ('cifar10','cifar100','mnist')parser = argparse.ArgumentParser(description='PyTorchLab')parser.add_argument('-d', '--dataset', metavar='DATA', default='cifar10', choices=dataset_names,          help='dataset to be used: ' + ' | '.join(dataset_names) + ' (default: cifar10)')args = parser.parse_args()data_dir = os.path.join('.', args.dataset)print(args.dataset)args.dataset = 'cifar10'if args.dataset == "cifar10":  train_transform = transforms.Compose([transforms.ToTensor()])  train_set = torchvision.datasets.CIFAR10(root=data_dir, train=True, download=True, transform=train_transform)  #print(vars(train_set))  print(train_set.train_data.shape)  print(train_set.train_data.mean(axis=(0,1,2))/255)  print(train_set.train_data.std(axis=(0,1,2))/255)  # imshow image  train_data = train_set.train_data  ind = 100  img0 = train_data[ind,...]  ## test channel number, in total , the correct channel is : RGB,not like BGR in caffe  # error produce  #b,g,r=cv2.split(img0)  #img0=cv2.merge([r,g,b])  print(img0.shape)  print(type(img0))  plt.imshow(img0)  plt.show() # in ship in sea  #img0 = cv2.resize(img0,(224,224))  #cv2.imshow('img0',img0)  #cv2.waitKey()elif args.dataset == "cifar100":  train_transform = transforms.Compose([transforms.ToTensor()])  train_set = torchvision.datasets.CIFAR100(root=data_dir, train=True, download=True, transform=train_transform)  #print(vars(train_set))  print(train_set.train_data.shape)  print(np.mean(train_set.train_data, axis=(0,1,2))/255)  print(np.std(train_set.train_data, axis=(0,1,2))/255)elif args.dataset == "mnist":  train_transform = transforms.Compose([transforms.ToTensor()])  train_set = torchvision.datasets.MNIST(root=data_dir, train=True, download=True, transform=train_transform)  #print(vars(train_set))  print(list(train_set.train_data.size()))  print(train_set.train_data.float().mean()/255)  print(train_set.train_data.float().std()/255)

结果：

cifar10Files already downloaded and verified(50000, 32, 32, 3)[ 0.49139968 0.48215841 0.44653091][ 0.24703223 0.24348513 0.26158784](32, 32, 3)<class 'numpy.ndarray'>

使用matlab检测是如何计算mean_file和std_file的：

% load cifar10 datasetdata = load('cifar10_train_data.mat');train_data = data.train_data;disp(size(train_data));temp = mean(train_data,1);disp(size(temp));train_data = double(train_data);% compute mean_file mean_val = mean(mean(mean(train_data,1),2),3)/255;% compute std_file temp1 = train_data(:,:,:,1);std_val1 = std(temp1(:))/255;temp2 = train_data(:,:,:,2);std_val2 = std(temp2(:))/255;temp3 = train_data(:,:,:,3);std_val3 = std(temp3(:))/255;mean_val = squeeze(mean_val);std_val = [std_val1, std_val2, std_val3];disp(mean_val);disp(std_val);% result: mean_val: [0.4914, 0.4822, 0.4465]%     std_val: [0.2470, 0.2435, 0.2616]

均值计算的过程也可以遵循标准差的计算过程。为 了简单，例如对于一个矩阵，所有元素的均值，等于两个方向上先后均值。所以会直接采用如下的形式：

mean_val = mean(mean(mean(train_data,1),2),3)/255;

标准差的计算是每一个通道的对所有样本的求标准差。然后再除以255。

以上这篇Pytorch的mean和std调查实例就是小编分享给大家的全部内容了，希望能给大家一个参考，也希望大家多多支持。