pytorch实现：Resnet模型识别花朵数据集

  • 一、pytorch实现：Resnet模型识别花朵数据集
    • 1.1 训练模型
    • 1.2 图像预测和可视化
    • 1.3 对新来的数据进行处理和展示

一、pytorch实现：Resnet模型识别花朵数据集

1.1 训练模型

# Author: Liuxin
# Time: 2022/4/22
# 导入所需的包
import torch, torchvision
from torchvision import datasets, transforms, models  # torchvision包最重要的3个模块
import os
import matplotlib.pyplot as plt
import numpy as np
from torch import nn
from torch import optim
import imageio
import time
import warningswarnings.filterwarnings('ignore')  # 忽略告警
import random
import sys
import copy
import json
from PIL import Image# 数据读取路径
data_dir = r'./flower_data/'
train_dir = data_dir + 'train'
test_dir = data_dir + 'valid'# 数据预处理（数据增强）
data_transforms = {'train': transforms.Compose([transforms.RandomRotation(45),  # 随机旋转，-45到45度之间随机选一个值进行旋转transforms.CenterCrop(224),  # 从图像中心开始裁剪图像为224*224的尺寸transforms.RandomHorizontalFlip(p=0.5),  # 水平方向依0.5的概率进行图像翻转transforms.RandomVerticalFlip(p=0.5),  # 垂直方向依0.5的概率进行图像翻转transforms.ColorJitter(brightness=0.2, contrast=0.1, saturation=0.1, hue=0.1),# 参数brightness为亮度，参数contrast为对比度，参数saturation为饱和度，参数hue为色调transforms.RandomGrayscale(p=0.025),  # 以0.025的概率转换为灰度图，当原始图像为彩色图时，转换后R=G=Btransforms.ToTensor(),  # 转换为tensortransforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])]),'valid': transforms.Compose([  # 验证集和测试机不需要进行数据增强操作，当尺寸与训练集不一样时，需要进行Resize操作transforms.Resize(256),  # 尺寸变为256*256transforms.CenterCrop(224),transforms.ToTensor(),transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])  # 均值和标准差需要和训练集一致])
}# 数据导入
batch_size = 8
image_datasets = {x: datasets.ImageFolder(root=os.path.join(data_dir, x), transform=data_transforms[x]) for x in['train', 'valid']}  # 读取图像路径并进行transform操作
dataloaders = {x: torch.utils.data.DataLoader(image_datasets[x], batch_size=batch_size, shuffle=True) for x in['train', 'valid']}
dataset_size = {x: len(image_datasets[x]) for x in ['train', 'valid']}
class_names = image_datasets['train'].classes
print(image_datasets)
print(dataloaders)
print(dataset_size)
print(class_names)# 读取标签对应的实际花的名字
with open('./cat_to_name.json', 'r') as f:cat_to_name = json.load(f)
print(cat_to_name)
#结果
"""
{'train': Dataset ImageFolderNumber of datapoints: 6552Root location: ./flower_data/trainStandardTransform
Transform: Compose(RandomRotation(degrees=[-45.0, 45.0], interpolation=nearest, expand=False, fill=0)CenterCrop(size=(224, 224))RandomHorizontalFlip(p=0.5)RandomVerticalFlip(p=0.5)ColorJitter(brightness=[0.8, 1.2], contrast=[0.9, 1.1], saturation=[0.9, 1.1], hue=[-0.1, 0.1])RandomGrayscale(p=0.025)ToTensor()Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])),
'valid': Dataset ImageFolderNumber of datapoints: 818Root location: ./flower_data/validStandardTransform
Transform: Compose(Resize(size=256, interpolation=bilinear, max_size=None, antialias=None)CenterCrop(size=(224, 224))ToTensor()Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]))}{'train': <torch.utils.data.dataloader.DataLoader object at 0x000001BB8E0CD850>,
'valid': <torch.utils.data.dataloader.DataLoader object at 0x000001BB8E0CD820>}{'train': 6552, 'valid': 818}['1', '10', '100', '101', '102', '11', '12', '13', '14', '15', '16', '17', '18', '19', '2', '20', '21', '22',
'23', '24', '25', '26', '27', '28', '29', '3', '30', '31', '32', '33', '34', '35', '36', '37', '38', '39',
'4', '40', '41', '42', '43', '44', '45', '46', '47', '48', '49', '5', '50', '51', '52', '53', '54', '55',
'56', '57', '58', '59', '6', '60', '61', '62', '63', '64', '65', '66', '67', '68', '69', '7', '70', '71',
'72', '73', '74', '75', '76', '77', '78', '79', '8', '80', '81', '82', '83', '84', '85', '86', '87', '88',
'89', '9', '90', '91', '92', '93', '94', '95', '96', '97', '98', '99']{'21': 'fire lily', '3': 'canterbury bells', '45': 'bolero deep blue', '1': 'pink primrose', '34': 'mexican aster','27': 'prince of wales feathers', '7': 'moon orchid', '16': 'globe-flower', '25': 'grape hyacinth', '26': 'corn poppy', '79': 'toad lily', '39': 'siam tulip', '24': 'red ginger', '67': 'spring crocus', '35': 'alpine sea holly', '32': 'garden phlox', '10': 'globe thistle', '6': 'tiger lily', '93': 'ball moss', '33': 'love in the mist', '9': 'monkshood', '102': 'blackberry lily', '14': 'spear thistle', '19': 'balloon flower', '100': 'blanket flower', '13': 'king protea', '49': 'oxeye daisy', '15': 'yellow iris', '61': 'cautleya spicata', '31': 'carnation', '64': 'silverbush', '68': 'bearded iris', '63': 'black-eyed susan', '69': 'windflower', '62': 'japanese anemone', '20': 'giant white arum lily', '38': 'great masterwort', '4': 'sweet pea', '86': 'tree mallow', '101': 'trumpet creeper', '42': 'daffodil', '22': 'pincushion flower', '2': 'hard-leaved pocket orchid', '54': 'sunflower', '66': 'osteospermum', '70': 'tree poppy', '85': 'desert-rose', '99': 'bromelia', '87': 'magnolia', '5': 'english marigold', '92': 'bee balm', '28': 'stemless gentian', '97': 'mallow', '57': 'gaura', '40': 'lenten rose', '47': 'marigold', '59': 'orange dahlia', '48': 'buttercup', '55': 'pelargonium', '36': 'ruby-lipped cattleya', '91': 'hippeastrum', '29': 'artichoke', '71': 'gazania', '90': 'canna lily', '18': 'peruvian lily', '98': 'mexican petunia', '8': 'bird of paradise', '30': 'sweet william', '17': 'purple coneflower', '52': 'wild pansy', '84': 'columbine', '12': "colt's foot", '11': 'snapdragon', '96': 'camellia', '23': 'fritillary', '50': 'common dandelion', '44': 'poinsettia', '53': 'primula', '72': 'azalea', '65': 'californian poppy', '80': 'anthurium', '76': 'morning glory', '37': 'cape flower', '56': 'bishop of llandaff', '60': 'pink-yellow dahlia', '82': 'clematis', '58': 'geranium', '75': 'thorn apple', '41': 'barbeton daisy', '95': 'bougainvillea', '43': 'sword lily', '83': 'hibiscus', '78': 'lotus lotus', '88': 'cyclamen', '94': 'foxglove', '81': 'frangipani', '74': 'rose', '89': 'watercress', '73': 'water lily', '46': 'wallflower', '77': 'passion flower', '51': 'petunia'}
"""

# 展示数据
def im_convert(tensor):image = tensor.to('cpu').clone().detach()"""tensor.clone():返回tensor的拷贝，返回的新tensor和原来的tensor具有同样的大小和数据类型原tensor的requires_grad=True，clone()返回的tensor是中间节点，梯度会流向原tensor，即返回的tensor的梯度会叠加在原tensor上tensor.detach()：从计算图中脱离出来，返回一个新的tensor，新的tensor和原来的tensor共享数据内存，但不涉及梯度计算"""image = image.numpy()  # 转换为数组image = image.transpose(1, 2, 0)  # 原图片是：通道数、高度、宽度，现在需要转换为高度、宽度、通道数image = image * np.array((0.485, 0.456, 0.406)) + np.array((0.229, 0.224, 0.225))  # 图片经过了标准化，现在需要复原image = image.clip(0, 1)  # 限制数组的上下限，下限为0，上限为1return imagefig = plt.figure(figsize=(20, 12))
cols = 4
rows = 2
dataiter = iter(dataloaders['valid'])  # 每次获取一个batch数据
inputs, classes = dataiter.next()
for idx in range(cols * rows):ax = fig.add_subplot(rows, cols, idx + 1, xticks=[], yticks=[])ax.set_title(cat_to_name[str(int(class_names[idx]))])plt.imshow(im_convert(inputs[idx]))
plt.show()


#定义运算设备CPU/GPU
model_name = 'resnet'
feature_extracting = True  # 是否采用训练好的特征
train_on_gpu = torch.cuda.is_available()  # 是否用GPU来训练
if not train_on_gpu:print('CUDA is not available, Training on CPU!')
else:print('CUDA is available, Training on GPU!')
device = torch.device('cuda:0' if torch.cuda.is_available() else "cpu")
#结果
"""
CUDA is not available, Training on CPU!
"""

# 迁移学习，加载models中提供的模型，用训练好的权重当做初始化参数
def set_parameter_require_grad(model, feature_extract):if feature_extract:for param in model.parameters():param.requires_grad = False# 加载模型
model_ft = models.resnet152()
#print(model_ft)# 参考pytorch官网的代码
def initialize_model(model_name, num_classes, feature_extract, use_pretrained=True):# model_ft = None# input_size = 0if model_name == 'resnet':model_ft = models.resnet152(pretrained=use_pretrained)set_parameter_require_grad(model_ft, feature_extract)num_feature_in = model_ft.fc.in_features  # 改变最后一层的输入和输出model_ft.fc = nn.Sequential(nn.Linear(num_feature_in, num_classes),nn.LogSoftmax(dim=1))input_size = 224elif model_name == 'alexnet':model_ft = models.alexnet(pretrained=use_pretrained)set_parameter_require_grad(model_ft, feature_extract)num_feature_in = model_ft.classifier[6].in_features  # classifier[6]是全连接的最后一层Linear层model_ft.classifier[6] = nn.Linear(num_feature_in, num_classes)input_size = 224elif model_name == 'vgg':model_ft = models.vgg16(pretrained=use_pretrained)set_parameter_require_grad(model_ft, feature_extract)num_feature_in = model_ft.classifier[6].in_featuresmodel_ft.classifier[6] = nn.Linear(num_feature_in, num_classes)input_size = 224elif model_name == 'squeezenet':model_ft = models.squeezenet1_0(pretrained=use_pretrained)set_parameter_require_grad(model_ft, feature_extract)model_ft.classifier[1] = nn.Conv2d(512, num_classes, kernel_size=(1, 1), stride=(1, 1))input_size = 224elif model_name == 'densenet':model_ft = models.densenet121(pretrained=use_pretrained)set_parameter_require_grad(model_ft, feature_extract)num_feature_in = model_ft.classifier.in_featuresmodel_ft.classifier = nn.Linear(num_feature_in, num_classes)input_size = 224elif model_name == 'inception':  # 此模型要求输入图像为229*229model_ft = models.inception_v3(pretrained=use_pretrained)set_parameter_require_grad(model_ft, feature_extract)num_feature_in = model_ft.AuxLogits.fc.in_featuresmodel_ft.AuxLogits.fc = nn.Linear(num_feature_in, num_classes)input_size = 229else:print('Invalid model name, exiting...')return model_ft, input_size# 设置哪些层需要训练
model_ft, input_size = initialize_model(model_name, 102, feature_extracting,use_pretrained=True)  # is_inception参数是给inception模型用的
# CPU进行计算
model_ft = model_ft.to(device)
# 模型保存
file_path = './checkpoint.pth' #训练完成后的参数模型路径，模型参数结果以.pth后缀
# 是否训练所有层
params_to_update = model_ft.parameters()
print('Params to learn')
if feature_extracting:params_to_update = []for name, param in model_ft.named_parameters():if param.requires_grad == True:params_to_update.append(param)print('\t', name)
else:for name, param in model_ft.named_parameters():if param.requires_grad == True:print('\t', name)
print(model_ft)  # 打印网络架构
"""
ResNet((conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True)(maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)(layer1): Sequential((0): Bottleneck((conv1): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True)(downsample): Sequential((0): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)))(1): Bottleneck((conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(2): Bottleneck((conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True)))(layer2): Sequential((0): Bottleneck((conv1): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True)(downsample): Sequential((0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)(1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)))(1): Bottleneck((conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(2): Bottleneck((conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(3): Bottleneck((conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(4): Bottleneck((conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(5): Bottleneck((conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(6): Bottleneck((conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(7): Bottleneck((conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True)))(layer3): Sequential((0): Bottleneck((conv1): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True)(downsample): Sequential((0): Conv2d(512, 1024, kernel_size=(1, 1), stride=(2, 2), bias=False)(1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)))(1): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(2): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(3): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(4): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(5): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(6): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(7): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(8): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(9): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(10): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(11): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(12): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(13): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(14): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(15): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(16): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(17): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(18): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(19): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(20): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(21): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(22): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(23): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(24): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(25): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(26): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(27): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(28): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(29): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(30): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(31): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(32): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(33): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(34): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(35): Bottleneck((conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True)))(layer4): Sequential((0): Bottleneck((conv1): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True)(downsample): Sequential((0): Conv2d(1024, 2048, kernel_size=(1, 1), stride=(2, 2), bias=False)(1): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)))(1): Bottleneck((conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True))(2): Bottleneck((conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)(bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)(relu): ReLU(inplace=True)))(avgpool): AdaptiveAvgPool2d(output_size=(1, 1))(fc): Linear(in_features=2048, out_features=1000, bias=True)
)
-----------------------------------------------------------------------------------
Params to learnfc.0.weightfc.0.bias
"""

# 优化器设置
optimizer_ft = optim.Adam(params_to_update, lr=1e-2)
scheduler = optim.lr_scheduler.StepLR(optimizer_ft, step_size=7, gamma=0.1)  # 学习率衰减，每训练7个epoch，学习率衰减为原来的1/10
# 最后一层已经有LogSoftmax()了，所以不能再nn.CrossEntropyLoss()来计算损失了
# nn.CrossEntropyLoss()相当于LogSoftmax()和nn.NLLLoss()整合
criterion = nn.NLLLoss()# 进行训练
def train(model, dataloader, criterion, optimizer, num_epoch=5, filename=file_path):start = time.time()best_acc = 0model.to(device)val_acc = []val_loss = []train_acc = []train_loss = []LRs = [optimizer.param_groups[0]['lr']]  # 取出lrbest_model_weights = copy.deepcopy(model.state_dict())for epoch in range(num_epoch):print('Epoch {}/{}'.format(epoch, num_epoch - 1))print('-' * 10)for phase in ['train', 'valid']:if phase == 'train':model.train()else:model.eval()running_loss = 0.0running_corrects = 0.0# 遍历数据集for inputs, labels in dataloader[phase]:inputs = inputs.to(device)labels = labels.to(device)optimizer.zero_grad()# 只有训练阶段才进行梯度更新with torch.set_grad_enabled(phase == 'train'):  # torch.set_grad_enabled(True)时才更新梯度，False时不更新梯度outputs = model(inputs)loss = criterion(outputs, labels)preds = torch.argmax(outputs, 1)if phase == 'train':loss.backward()optimizer.step()# 计算损失running_loss += loss.item() * inputs.size(0)running_corrects += torch.sum(preds == labels.data)epoch_loss = running_loss / len(dataloader[phase].dataset)epoch_acc = running_corrects / len(dataloader[phase].dataset)time_elapsed = time.time() - startprint('Time elapsed {:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))print('{} Loss: {:.4f} Acc: {:.4f}'.format(phase, epoch_loss, epoch_acc))# 得到最好的模型if phase == 'valid' and epoch_acc > best_acc:best_acc = epoch_accbest_model_weights = copy.deepcopy(model.state_dict())state = {'state_dict': model.state_dict(),'best_acc': best_acc,'optimizer': optimizer.state_dict()}torch.save(state, filename)if phase == 'valid':val_acc.append(epoch_acc)val_loss.append(epoch_loss)scheduler.step(epoch_loss)if phase == 'train':train_acc.append(epoch_acc)train_loss.append(epoch_loss)print('Optimizer learning rate: {:.7f}'.format(optimizer.param_groups[0]['lr']))LRs.append(optimizer.param_groups[0]['lr'])time_elapsed = time.time() - startprint('Training complete in {:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))print('Best val Acc: {:4f}'.format(best_acc))# 训练完保存最好的模型结果model.load_state_dict(best_model_weights)return model, val_acc, train_acc, val_loss, train_loss, LRs# 开始训练
model_ft, val_acc, train_acc, valid_loss, train_loss, LRs = train(model_ft,dataloaders,criterion,optimizer_ft,num_epoch=5)
#结果
"""
Epoch 0/4
----------
Time elapsed 20m 16s
train Loss: 10.0411 Acc: 0.3132
Time elapsed 22m 34s
valid Loss: 12.0470 Acc: 0.3765
Optimizer learning rate: 0.0010000
Epoch 1/4
----------
Time elapsed 42m 49s
train Loss: 2.3768 Acc: 0.6980
Time elapsed 45m 6s
valid Loss: 3.3266 Acc: 0.6357
Optimizer learning rate: 0.0100000
Epoch 2/4
----------
Time elapsed 66m 21s
train Loss: 9.7728 Acc: 0.4770
Time elapsed 68m 46s
valid Loss: 15.3826 Acc: 0.4389
Optimizer learning rate: 0.0001000
Epoch 3/4
----------
Time elapsed 90m 8s
train Loss: 4.7683 Acc: 0.6784
Time elapsed 92m 27s
valid Loss: 6.8775 Acc: 0.6088
Optimizer learning rate: 0.0100000
Epoch 4/4
----------
Time elapsed 110m 31s
train Loss: 9.1733 Acc: 0.5499
Time elapsed 112m 31s
valid Loss: 14.6585 Acc: 0.4707
Optimizer learning rate: 0.0001000
Training complete in 112m 31s
Best val Acc: 0.635697
"""

1.2 图像预测和可视化

# Author: Liuxin
# Time: 2022/4/27
# 进行测试
import torch, torchvision
from torchvision import datasets, transforms, models  # torchvision包最重要的3个模块
import os
import matplotlib.pyplot as plt
import numpy as np
from torch import nn
from torch import optim
import imageio
import time
import warningswarnings.filterwarnings('ignore')  # 忽略告警
import random
import sys
import copy
import json
from PIL import Image# 数据读取路径
data_dir = r'./flower_data/'
train_dir = data_dir + 'train'
test_dir = data_dir + 'valid'# 数据预处理（数据增强）
data_transforms = {'train': transforms.Compose([transforms.RandomRotation(45),  # 随机旋转，-45到45度之间随机选一个值进行旋转transforms.CenterCrop(224),  # 从图像中心开始裁剪图像为224*224的尺寸transforms.RandomHorizontalFlip(p=0.5),  # 水平方向依0.5的概率进行图像翻转transforms.RandomVerticalFlip(p=0.5),  # 垂直方向依0.5的概率进行图像翻转transforms.ColorJitter(brightness=0.2, contrast=0.1, saturation=0.1, hue=0.1),# 参数brightness为亮度，参数contrast为对比度，参数saturation为饱和度，参数hue为色调transforms.RandomGrayscale(p=0.025),  # 以0.025的概率转换为灰度图，当原始图像为彩色图时，转换后R=G=Btransforms.ToTensor(),  # 转换为tensortransforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])]),'valid': transforms.Compose([  # 验证集和测试机不需要进行数据增强操作，当尺寸与训练集不一样时，需要进行Resize操作transforms.Resize(256),  # 尺寸变为256*256transforms.CenterCrop(224),transforms.ToTensor(),transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])  # 均值和标准差需要和训练集一致])
}# 读取标签对应的实际花的名字
with open('./cat_to_name.json', 'r') as f:cat_to_name = json.load(f)# print(cat_to_name)# 展示数据
def im_convert(tensor):image = tensor.to('cpu').clone().detach()"""tensor.clone():返回tensor的拷贝，返回的新tensor和原来的tensor具有同样的大小和数据类型原tensor的requires_grad=True，clone()返回的tensor是中间节点，梯度会流向原tensor，即返回的tensor的梯度会叠加在原tensor上tensor.detach()：从计算图中脱离出来，返回一个新的tensor，新的tensor和原来的tensor共享数据内存，但不涉及梯度计算"""image = image.numpy()  # 转换为数组image = image.transpose(1, 2, 0)  # 原图片是：通道数、高度、宽度，现在需要转换为高度、宽度、通道数image = image * np.array((0.485, 0.456, 0.406)) + np.array((0.229, 0.224, 0.225))  # 图片经过了标准化，现在需要复原image = image.clip(0, 1)  # 限制数组的上下限，下限为0，上限为1return image# 数据导入
batch_size = 8
image_datasets = {x: datasets.ImageFolder(root=os.path.join(data_dir, x), transform=data_transforms[x]) for x in['train', 'valid']}  # 读取图像路径并进行transform操作
dataloaders = {x: torch.utils.data.DataLoader(image_datasets[x], batch_size=batch_size, shuffle=True) for x in['train', 'valid']}
dataset_size = {x: len(image_datasets[x]) for x in ['train', 'valid']}
class_names = image_datasets['train'].classesdef set_parameter_require_grad(model, feature_extract):if feature_extract:for param in model.parameters():param.requires_grad = False# 参考pytorch官网的代码
def initialize_model(model_name, num_classes, feature_extract, use_pretrained=True):  # 选定模型及训练的函数# model_ft = None# input_size = 0if model_name == 'resnet':model_ft = models.resnet152(pretrained=use_pretrained)set_parameter_require_grad(model_ft, feature_extract)num_feature_in = model_ft.fc.in_features  # 改变最后一层的输入和输出model_ft.fc = nn.Sequential(nn.Linear(num_feature_in, num_classes),nn.LogSoftmax(dim=1))input_size = 224elif model_name == 'alexnet':model_ft = models.alexnet(pretrained=use_pretrained)set_parameter_require_grad(model_ft, feature_extract)num_feature_in = model_ft.classifier[6].in_features  # classifier[6]是全连接的最后一层Linear层model_ft.classifier[6] = nn.Linear(num_feature_in, num_classes)input_size = 224elif model_name == 'vgg':model_ft = models.vgg16(pretrained=use_pretrained)set_parameter_require_grad(model_ft, feature_extract)num_feature_in = model_ft.classifier[6].in_featuresmodel_ft.classifier[6] = nn.Linear(num_feature_in, num_classes)input_size = 224elif model_name == 'squeezenet':model_ft = models.squeezenet1_0(pretrained=use_pretrained)set_parameter_require_grad(model_ft, feature_extract)model_ft.classifier[1] = nn.Conv2d(512, num_classes, kernel_size=(1, 1), stride=(1, 1))input_size = 224elif model_name == 'densenet':model_ft = models.densenet121(pretrained=use_pretrained)set_parameter_require_grad(model_ft, feature_extract)num_feature_in = model_ft.classifier.in_featuresmodel_ft.classifier = nn.Linear(num_feature_in, num_classes)input_size = 224elif model_name == 'inception':  # 此模型要求输入图像为229*229model_ft = models.inception_v3(pretrained=use_pretrained)set_parameter_require_grad(model_ft, feature_extract)num_feature_in = model_ft.AuxLogits.fc.in_featuresmodel_ft.AuxLogits.fc = nn.Linear(num_feature_in, num_classes)input_size = 229else:print('Invalid model name, exiting...')return model_ft, input_sizedevice = torch.device('cuda:0' if torch.cuda.is_available() else "cpu")
model_name = 'resnet'
feature_extracting = True  # 是否采用训练好的特征
model_ft, input_size = initialize_model(model_name, 102, feature_extracting, use_pretrained=True)
# GPU模式
model_ft = model_ft.to(device)
# 保存文件的名字
filename = './checkpoint.pth'
# 加载模型
checkpoint = torch.load(filename)
best_acc = checkpoint['best_acc']
model_ft.load_state_dict(checkpoint['state_dict'])# 得到一个batch的测试数据
dataiter = iter(dataloaders['valid'])
images, labels = dataiter.next()model_ft.eval()
train_on_gpu = torch.cuda.is_available()  # 是否用GPU来训练
if train_on_gpu:output = model_ft(images.cuda())
else:output = model_ft(images)print(output.shape)  # output表示对一个batch中每一个数据得到其属于各个类别的可能性preds_tensor = torch.argmax(output, 1)
print(preds_tensor)preds = np.squeeze(preds_tensor.numpy()) if train_on_gpu else np.squeeze(preds_tensor.cpu().numpy()) #np.squeeze函数将多维数组或矩阵框数据转换为纯数据
#preds=list(preds_tensor.numpy()) 和上一行效果一样
print(preds)fig = plt.figure(figsize=(20, 20))
columns = 4
rows = 2for idx in range(columns * rows):ax = fig.add_subplot(rows, columns, idx + 1, xticks=[], yticks=[])plt.imshow(im_convert(images[idx]))ax.set_title("{} ({})".format(cat_to_name[str(preds[idx])], cat_to_name[str(labels[idx].item())]),color=("green" if cat_to_name[str(preds[idx])] == cat_to_name[str(labels[idx].item())] else "red"))
plt.show()#结果
"""
output.shape：torch.Size([8, 102])
preds_tensor：tensor([43, 24, 64, 49, 64, 34, 18, 89])
preds：[43 24 64 49 64 34 18 89]
"""

1.3 对新来的数据进行处理和展示

def process_image(image_path):# 读取测试数据img = Image.open(image_path)# Resize,thumbnail方法只能进行缩小，所以进行了判断if img.size[0] > img.size[1]:img.thumbnail((10000, 256))  # Image.thumbnail()将此图像制作为缩略图。此方法将图像修改为包含其自身的缩略图版本，但不大于给定的大小# 用法：Image.thumbnail(size, resample=3)，参数：size-要求的尺寸，一个元祖。resample-可选的重采样过滤器。返回类型：Image对象。else:img.thumbnail((256, 10000))# Crop操作left_margin = (img.width - 224) / 2bottom_margin = (img.height - 224) / 2right_margin = left_margin + 224top_margin = bottom_margin + 224img = img.crop((left_margin, bottom_margin, right_margin,top_margin))# 相同的预处理方法img = np.array(img) / 255mean = np.array([0.485, 0.456, 0.406])  # provided meanstd = np.array([0.229, 0.224, 0.225])  # provided stdimg = (img - mean) / std# 注意颜色通道应该放在第一个位置img = img.transpose((2, 0, 1))return imgdef imshow(image, ax=None, title=None):"""展示数据"""if ax is None:fig, ax = plt.subplots()# 颜色通道还原image = np.array(image).transpose((1, 2, 0))# 预处理还原mean = np.array([0.485, 0.456, 0.406])std = np.array([0.229, 0.224, 0.225])image = std * image + meanimage = np.clip(image, 0, 1)ax.imshow(image)ax.set_title(title)return ax
image_path = 'image_04248.jpg'
img = process_image(image_path)
imshow(img)
print(img.shape)
#结果：(3, 224, 224)

pytorch实现：Resnet模型识别花朵数据集（参考pytorch官网代码）相关推荐

  1. 轮播图进度条同步切换(参考漫威官网)

    轮播图进度条同步切换(参考漫威官网) <!DOCTYPE html> <html lang="en"><head><meta charse ...

  2. Pytorch DataLoader一次性封装多种数据集（pytorch Data学习六）

    Pytorch的DataLoader不能够直接达到这个效果,必须要借助DataSet来实现 DataSet的用法可以参考:pytorch 构造读取数据的工具类 Dataset 与 DataLoader ...

  3. VScode的mac版本C++配置（参考官网教程）

    VScode的mac版本C++配置 一. 基本情况 PAST:我使用的是Xcode. NOW:转VScode 参考VScode 官网 实现完本文会生成的3个文件 二.配置基本步骤 Step 1:下载 ...

  4. 基于pytorch搭建神经网络的花朵种类识别（深度学习）

    基于pytorch搭建神经网络的花朵种类识别(深度学习) 文章目录 基于pytorch搭建神经网络的花朵种类识别(深度学习) 一．知识点 1.特征提取.神经元逐层判断 2.中间层(隐藏层) 3.学习权 ...

  5. Deep Learning：基于pytorch搭建神经网络的花朵种类识别项目（内涵完整文件和代码）—超详细完整实战教程

    基于pytorch的深度学习花朵种类识别项目完整教程(内涵完整文件和代码) 相关链接:: 超详细--CNN卷积神经网络教程(零基础到实战) 大白话pytorch基本知识点及语法+项目实战 文章目录 基 ...

  6. 动物数据集+动物分类识别训练代码(Pytorch)

    动物数据集+动物分类识别训练代码(Pytorch) 目录 动物数据集+动物分类识别训练代码(Pytorch) 1. 前言 2. Animals-Dataset动物数据集说明 (1)Animals90动 ...

  7. 图像分类篇：pytorch实现ResNet

    一.ResNet详解 ResNet网络是在2015年由微软实验室提出的,斩获当年ImageNet竞赛中分类任务第一名,目标检测第一名,获得COCO数据集中目标检测第一名,图像分类第一名. 在ResNe ...

  8. 使用pytorch版faster-rcnn训练自己数据集

    使用pytorch版faster-rcnn训练自己数据集 引言 faster-rcnn pytorch代码下载 训练自己数据集 接下来工作 参考文献 引言 最近在复现目标检测代码(师兄强烈推荐FPN, ...

  9. 深度之眼Pytorch打卡（九）：Pytorch数据预处理——预处理过程与数据标准化（transforms过程、Normalize原理、常用数据集均值标准差与数据集均值标准差计算）

    前言   前段时间因为一些事情没有时间或者心情学习,现在两个多月过去了,事情结束了,心态也调整好了,所以又来接着学习Pytorch.这篇笔记主要是关于数据预处理过程.数据集标准化与数据集均值标准差计算 ...

最新文章

  1. 5G NGC — 关键技术 — 计算与存储分离
  2. 山西大学计算机科学与技术全国排名,山西这所大学曾是国内排名前五，如今排名下滑，有点走下坡路了...
  3. KDE和GNOME的区别※切换
  4. vim查找关键字_vim常用命令及操作
  5. Tomcat访问日志详细配置
  6. bat(续六)-windows批处理set命令
  7. azw3转换为pdf_怎么合并几个PDF为一个？快用这个PDF转换器！
  8. JS switch 分支语句
  9. github删除文件_Linux环境下Github操作指南
  10. WCF作为SAO技术的核心：契约
  11. Python虚拟机之if控制流（一）
  12. QTP11.5/UFT官方下载与安装
  13. python全栈工程师视频_python全栈工程师视频教程
  14. 俄罗斯方块中方块的旋转变形
  15. 基于5g的交通运输_一种基于5G的智慧交通基础服务平台
  16. easyui datagrid deleteRow(删除行)的BUG或者updateRow值更新了不展示问题
  17. VC++ excel 写入与写出数据
  18. ARM NEON Intrinsics示例
  19. 46道史上最全Redis面试题
  20. AUTOSAR MCAL解析：MCU

热门文章

  1. 《网络、群体和市场》习题答案
  2. 兼职必看的全行业薪酬对照表
  3. 微信公众号免开发——被关注自动回复2条消息
  4. Spark sql HLL函数
  5. JavaWeb实例：西蒙购物网
  6. 国外云服务器哪个好？云服务器选择什么配置合适？
  7. 爱签电子合同助力医院患者信息表跨科室在线签署，实名认证、自动归档
  8. 文章怎么这么多！！！！！(超分辨率相关)
  9. 《隐秘的角落》影评数据分析
  10. android旅游app动态的,Android平台深秋出游必备APP推荐