目錄

1? pytorch的版本：

2? 數(shù)據(jù)下載地址：

3? 原始版本代碼下載：

4? 直接上代碼：

---

?

1? pytorch的版本：

2? 數(shù)據(jù)下載地址：

<https://download.pytorch.org/tutorial/hymenoptera_data.zip>

3? 原始版本代碼下載：

https://pytorch.org/tutorials/_downloads/transfer_learning_tutorial.py

?

4? 直接上代碼：

# -*- coding: utf-8 -*- # @File : test4.py # @Blog : https://blog.csdn.net/caomin1haofrom __future__ import print_function, divisionimport torch import torch.nn as nn import torch.optim as optim from torch.optim import lr_scheduler import numpy as np import torchvision from torchvision import datasets, models, transforms import matplotlib.pyplot as plt import time import os import copydevice = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")plt.ion() # interactive mode###################################################################### # 1.定義模型， 2.加載部分預(yù)訓(xùn)練數(shù)據(jù)， 3.凍結(jié)部分層 ###################################### #1.定義模型 model_conv = models.resnet18() num_ftrs = model_conv.fc.in_features model_conv.fc = nn.Linear(num_ftrs, 2)''' #打印模型的結(jié)構(gòu) print('###打印模型model_conv的結(jié)構(gòu)####') print(model_conv) print('\n')print('###打印模型model_conv加載參數(shù)前的初始值####') print(list(model_conv.parameters())) print('\n') '''############################################# #2.加載部分預(yù)訓(xùn)練數(shù)據(jù) pretrained_dict = torch.load('./08 transfer_learning/resnet18-5c106cde.pth') ''' for k,v in pretrained_dict.items():print(k) ''' #刪除預(yù)訓(xùn)練模型跟當(dāng)前模型層名稱(chēng)相同，層結(jié)構(gòu)卻不同的元素;這里有兩個(gè)'fc.weight'、'fc.bias' pretrained_dict.pop('fc.weight') pretrained_dict.pop('fc.bias')#自己的模型參數(shù)變量 model_dict = model_conv.state_dict() #去除一些不需要的參數(shù) pretrained_dict = {k: v for k, v in pretrained_dict.items() if k in model_dict}#參數(shù)更新 model_dict.update(pretrained_dict)# 加載我們真正需要的state_dict model_conv.load_state_dict(model_dict)''' print('###打印模型model_conv加載參數(shù)后的參數(shù)值####') print(list(model_conv.parameters())) print('\n') ''' ############################################# #3.凍結(jié)部分層 #將滿足條件的參數(shù)的 requires_grad 屬性設(shè)置為False for name, value in model_conv.named_parameters():if (name != 'fc.weight') and (name != 'fc.bias'):value.requires_grad = False ''' #打印各層的requires_grad屬性 print('###打印模型model_conv參數(shù)的requires_grad屬性####') for name, param in model_conv.named_parameters():print(name,param.requires_grad) '''# filter 函數(shù)將模型中屬性 requires_grad = True 的參數(shù)選出來(lái) params_conv = filter(lambda p: p.requires_grad, model_conv.parameters()) model_conv = model_conv.to(device)criterion = nn.CrossEntropyLoss()# Observe that only parameters of final layer are being optimized as # opoosed to before. optimizer_conv = optim.SGD(params_conv, lr=0.001, momentum=0.9)# Decay LR by a factor of 0.1 every 7 epochs exp_lr_scheduler = lr_scheduler.StepLR(optimizer_conv, step_size=7, gamma=0.1)###################################################################### # Training the model #編寫(xiě)一個(gè)通用函數(shù)來(lái)訓(xùn)練模型。 # 下面將說(shuō)明： * 調(diào)整學(xué)習(xí)速率 * 保存最好的模型 #下面的參數(shù)scheduler是一個(gè)來(lái)自 torch.optim.lr_scheduler 的學(xué)習(xí)速率調(diào)整類(lèi)的對(duì)象(LR scheduler object)。def train_model(model, criterion, optimizer, scheduler, num_epochs=25):since = time.time()best_model_wts = copy.deepcopy(model.state_dict())best_acc = 0.0for epoch in range(num_epochs):print('Epoch {}/{}'.format(epoch, num_epochs - 1))print('-' * 10)# 每個(gè)epoch都有一個(gè)訓(xùn)練和驗(yàn)證階段for phase in ['train', 'val']:if phase == 'train':scheduler.step()model.train() # Set model to training modeelse:model.eval() # Set model to evaluate moderunning_loss = 0.0running_corrects = 0# 迭代數(shù)據(jù).for inputs, labels in dataloaders[phase]:inputs = inputs.to(device)labels = labels.to(device)# zero the parameter gradientsoptimizer.zero_grad()# forward# track history if only in trainwith torch.set_grad_enabled(phase == 'train'):outputs = model(inputs)_, preds = torch.max(outputs, 1)loss = criterion(outputs, labels)# 后向+僅在訓(xùn)練階段進(jìn)行優(yōu)化if phase == 'train':loss.backward()optimizer.step()# statisticsrunning_loss += loss.item() * inputs.size(0)running_corrects += torch.sum(preds == labels.data)epoch_loss = running_loss / dataset_sizes[phase]epoch_acc = running_corrects.double() / dataset_sizes[phase]print('{} Loss: {:.4f} Acc: {:.4f}'.format(phase, epoch_loss, epoch_acc))# 深度復(fù)制moif phase == 'val' and epoch_acc > best_acc:best_acc = epoch_accbest_model_wts = copy.deepcopy(model.state_dict())print()time_elapsed = time.time() - sinceprint('Training complete in {:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))print('Best val Acc: {:4f}'.format(best_acc))# 加載最佳模型權(quán)重model.load_state_dict(best_model_wts)return model###################################################################### # 可視化部分訓(xùn)練圖像，以便了解數(shù)據(jù)擴(kuò)充。def imshow(inp, title=None):"""Imshow for Tensor."""inp = inp.numpy().transpose((1, 2, 0))mean = np.array([0.485, 0.456, 0.406])std = np.array([0.229, 0.224, 0.225])inp = std * inp + meaninp = np.clip(inp, 0, 1)plt.imshow(inp)if title is not None:plt.title(title)plt.pause(0.001) # pause a bit so that plots are updated###################################################################### # Visualizing the model predictions # 一個(gè)通用的展示少量預(yù)測(cè)圖片的函數(shù)def visualize_model(model, num_images=6):was_training = model.trainingmodel.eval()images_so_far = 0fig = plt.figure()with torch.no_grad():for i, (inputs, labels) in enumerate(dataloaders['val']):inputs = inputs.to(device)labels = labels.to(device)outputs = model(inputs)_, preds = torch.max(outputs, 1)for j in range(inputs.size()[0]):images_so_far += 1ax = plt.subplot(num_images//2, 2, images_so_far)ax.axis('off')ax.set_title('predicted: {}'.format(class_names[preds[j]]))imshow(inputs.cpu().data[j])if images_so_far == num_images:model.train(mode=was_training)returnmodel.train(mode=was_training)###################################################################### #訓(xùn)練集數(shù)據(jù)擴(kuò)充和歸一化 #在驗(yàn)證集上僅需要?dú)w一化 data_transforms = {'train': transforms.Compose([transforms.RandomResizedCrop(224),transforms.RandomHorizontalFlip(),transforms.ToTensor(),transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])]),'val': transforms.Compose([transforms.Resize(256),transforms.CenterCrop(224),transforms.ToTensor(),transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])]), }data_dir = './08 transfer_learning/hymenoptera_data' image_datasets = {x: datasets.ImageFolder(os.path.join(data_dir, x),data_transforms[x])for x in ['train', 'val']} dataloaders = {x: torch.utils.data.DataLoader(image_datasets[x], batch_size=4,shuffle=True, num_workers=4)for x in ['train', 'val']} dataset_sizes = {x: len(image_datasets[x]) for x in ['train', 'val']} class_names = image_datasets['train'].classesdevice = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")if __name__ == '__main__':# Train and evaluate 2# 訓(xùn)練模型 在CPU上，與前一個(gè)場(chǎng)景相比，這將花費(fèi)大約一半的時(shí)間，因?yàn)椴恍枰獮榇蠖鄶?shù)網(wǎng)絡(luò)計(jì)算梯度。但需要計(jì)算轉(zhuǎn)發(fā)。model_conv = train_model(model_conv, criterion, optimizer_conv,exp_lr_scheduler, num_epochs=11)visualize_model(model_conv)plt.ioff()plt.show()

?

部分運(yùn)行結(jié)果：

總結(jié)

以上是生活随笔為你收集整理的Pytorch 加载部分预训练模型并冻结某些层的全部?jī)?nèi)容，希望文章能夠幫你解決所遇到的問(wèn)題。

如果覺(jué)得生活随笔網(wǎng)站內(nèi)容還不錯(cuò)，歡迎將生活随笔推薦給好友。