pytorch繪制并顯示loss曲線和acc曲線，LeNet5識(shí)別圖像準(zhǔn)確率

更新時(shí)間：2020年01月02日 15:00:58 作者：Zaria_z

今天小編就為大家分享一篇pytorch繪制并顯示loss曲線和acc曲線，LeNet5識(shí)別圖像準(zhǔn)確率，具有很好的參考價(jià)值，希望對(duì)大家有所幫助。一起跟隨小編過(guò)來(lái)看看吧

我用的是Anaconda3 ，用spyder編寫(xiě)pytorch的代碼，在Anaconda3中新建了一個(gè)pytorch的虛擬環(huán)境（虛擬環(huán)境的名字就叫pytorch)。

以下內(nèi)容僅供參考哦~~

1.首先打開(kāi)Anaconda Prompt，然后輸入activate pytorch，進(jìn)入pytorch.

2.輸入pip install tensorboardX，安裝完成后，輸入python，用from tensorboardX import SummaryWriter檢驗(yàn)是否安裝成功。如下圖所示：

3.安裝完成之后，先給大家看一下我的文件夾，如下圖：

假設(shè)用LeNet5框架識(shí)別圖像的準(zhǔn)確率，LeNet.py代碼如下：

import torch
import torch.nn as nn
from torchsummary import summary
from torch.autograd import Variable
import torch.nn.functional as F
 
class LeNet5(nn.Module): #定義網(wǎng)絡(luò) pytorch定義網(wǎng)絡(luò)有很多方式，推薦以下方式，結(jié)構(gòu)清晰 
 def __init__(self):  
  super(LeNet5,self).__init__()  
  self.conv1 = nn.Conv2d(3, 6, 5)
  self.conv2 = nn.Conv2d(6, 16, 5)
  self.fc1 = nn.Linear(16*5*5, 120)
  self.fc2 = nn.Linear(120, 84)
  self.fc3 = nn.Linear(84, 2)
 def forward(self,x): 
  # print(x.size())
  x = F.max_pool2d(F.relu(self.conv1(x)), (2, 2))
  # print(x.size())
  x = F.max_pool2d(F.relu(self.conv2(x)), 2)
  # print(x.size()) 
  x = x.view(x.size()[0], -1)#全連接層均使用的nn.Linear()線性結(jié)構(gòu)，輸入輸出維度均為一維，故需要把數(shù)據(jù)拉為一維  
  #print(x.size())
  x = F.relu(self.fc1(x)) 
  # print(x.size())
  x = F.relu(self.fc2(x))
  #print(x.size())
  x = self.fc3(x)
  # print(x.size())
  return x
 
net = LeNet5()
data_input = Variable(torch.randn(16,3,32,32))
print(data_input.size())
net(data_input)
print(summary(net,(3,32,32)))

示網(wǎng)絡(luò)結(jié)構(gòu)如下圖：

訓(xùn)練代碼（LeNet_train_test.py）如下：

# -*- coding: utf-8 -*-
"""
Created on Wed Jan 2 15:53:33 2019
@author: Administrator
"""
 
import torch
import torch.nn as nn
import os
import numpy as np
import matplotlib.pyplot as plt
from torchvision import datasets,transforms
import torchvision
import LeNet
from torch import optim
import time
from torch.optim import lr_scheduler
from tensorboardX import SummaryWriter
 
 
writer = SummaryWriter('LeNet5')
data_transforms = {
  'train':transforms.Compose([
    #transforms.Resize(56),
    transforms.RandomResizedCrop(32),#
    transforms.RandomHorizontalFlip(),#已給定的概率隨即水平翻轉(zhuǎn)給定的PIL圖像
    transforms.ToTensor(),#將圖片轉(zhuǎn)換為T(mén)ensor,歸一化至[0,1]
    transforms.Normalize([0.485,0.456,0.406],[0.229, 0.224, 0.225])#用平均值和標(biāo)準(zhǔn)偏差歸一化張量圖像
    ]),
  'val':transforms.Compose([
    #transforms.Resize(56),
    transforms.CenterCrop(32),
    transforms.ToTensor(),
    transforms.Normalize([0.485,0.456,0.406],[0.229,0.224,0.225])
    ]),
  }
data_dir = 'bees vs ants' #樣本文件夾
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 =16,
            shuffle = True,num_workers = 0)
  for x in ['train','val']
  }
dataset_sizes = {x:len(image_datasets[x]) for x in ['train','val']}
class_names = image_datasets['train'].classes
 
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 
def imshow(inp,title = None):
 #print(inp.size())
 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 + mean
 inp = np.clip(inp,0,1)
 plt.imshow(inp)
 if title is not None:
  plt.title(title)
 plt.pause(0.001)#為了讓圖像更新可以暫停一會(huì)
 
#Get a batch of training data
inputs,classes = next(iter(dataloaders['train']))
#print(inputs.size())
#print(inputs.size())
#Make a grid from batch 
out = torchvision.utils.make_grid(inputs)
#print(out.size())
imshow(out,title=[class_names[x] for x in classes])
 
def train_model(model,criterion,optimizer,scheduler,num_epochs = 25):
 since = time.time()
 
 # best_model_wts = copy.deepcopy(model.state_dict())
 best_acc = 0.0
 
 
 for epoch in range(num_epochs):
  print('Epoch {}/{}'.format(epoch,num_epochs - 1))
  print('-' * 10)
  
  #Each epoch has a training and validation phase
  for phase in ['train','val']:
   if phase == 'train':
    scheduler.step()
    model.train() #Set model to training mode
   else:
    model.eval()
   
   running_loss = 0.0
   running_corrects = 0
   
   #Iterate over data
   for inputs,labels in dataloaders[phase]:
    inputs = inputs.to(device)
    # print(inputs.size())
    labels = labels.to(device)
    #print(inputs.size())
    # print(labels.size())
    
    #zero the parameter gradients(參數(shù)梯度為零)
    optimizer.zero_grad()
    
    #forward
    #track history if only in train
    with torch.set_grad_enabled(phase == 'train'):
     outputs = model(inputs)
     _,preds = torch.max(outputs,1)
     loss = criterion(outputs,labels)
     
     #backward + optimize only if in training phase
     if phase == 'train':
      loss.backward()
      optimizer.step()
    
    #statistics
    running_loss += loss.item() * inputs.size(0)
    running_corrects += torch.sum(preds == labels.data)
   
   if phase == 'train':
    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))
    writer.add_scalar('Train/Loss', epoch_loss,epoch)
    writer.add_scalar('Train/Acc',epoch_acc,epoch)
   else:
    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))
    writer.add_scalar('Test/Loss', epoch_loss,epoch)
    writer.add_scalar('Test/Acc',epoch_acc,epoch)
    if epoch_acc > best_acc:
     best_acc = epoch_acc
  
  print()
 writer.close() 
 time_elapsed = time.time() - since
 print('Training complete in {:.0f}m {:.0f}s'.format(
   time_elapsed // 60 , time_elapsed % 60))
 print('Best val Acc: {:4f}'.format(best_acc))
 
 #load best model weights
 #model.load_state_dict()#best_model_wts)
 return model
 
 
def visualize_model(model,num_images = 6):
 was_training = model.training
 model.eval()
 
 
 images_so_far = 0
 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 += 1
    ax = 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)
     return 
    model.train(mode=was_training)
 
 
net = LeNet.LeNet5()
net = net.to(device)
 
criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(net.parameters(),lr = 0.001,momentum = 0.9)
exp_lr_scheduler = lr_scheduler.StepLR(optimizer,step_size = 7,gamma = 0.1)
 
net = train_model(net,criterion,optimizer,exp_lr_scheduler,num_epochs = 25)
 
 
 
#net1 = train_model(net,criterion,optimizer,exp_lr_scheduler,num_epochs = 25)
visualize_model(net)
 
 
plt.ioff()
plt.show()

最終的二分類(lèi)結(jié)果為：

樣本圖像是pytorch官網(wǎng)中介紹遷移學(xué)習(xí)時(shí)用到的，螞蟻與蜜蜂的二分類(lèi)圖像，圖像大小不一。LeNet5 的輸入圖像是32*32，所以進(jìn)行分類(lèi)時(shí)會(huì)損失一定的圖像像素，導(dǎo)致識(shí)別率較低。

下面介紹顯示loss和acc曲線，在以上訓(xùn)練代碼中，writer = SummaryWriter('LeNet5')，表示在訓(xùn)練過(guò)程中會(huì)生成LeNet5文件夾，保存loss曲線和acc曲線的文件，如下圖：

首先解釋一下這個(gè)文件夾為什么是1，因?yàn)槲抑坝?xùn)練了很多次，在LeNet5文件夾下有很多1文件夾中這樣的文件，待會(huì)用Anaconda Prompt來(lái)顯示loss和acc的時(shí)候，它只識(shí)別一個(gè)文件，所以我就重新建了一個(gè)1文件夾，并將剛剛運(yùn)行完畢的文件放到文件夾中。在LeNet_train_test.py中， writer.add_scalar('Train/Loss', epoch_loss,epoch)和

writer.add_scalar('Train/Acc',epoch_acc,epoch)，這兩行代碼就是生成train數(shù)據(jù)集的loss和acc曲線，同理測(cè)試數(shù)據(jù)集亦是如此。

好啦，下面開(kāi)始顯示loss和acc:

1.打開(kāi)Anaconda Prompt，再次進(jìn)入pytorch虛擬環(huán)境，

2.輸入tensorboard --logdir=C:\Users\Administrator\.spyder-py3\python\pytorch\LeNet\LeNet5\1,紅色部分是來(lái)自上圖文件夾的根目錄，按回車(chē)鍵，會(huì)出現(xiàn)tensorboard的版本和一個(gè)網(wǎng)址，總體顯示效果如下圖：