在线海报设计网站,浙江新华建设有限公司官方网站,外链群发软件,国际大新闻最新消息1、介绍
1.1数据集介绍
flower_data├── train│ └── 1-102#xff08;102个文件夹#xff09;│ └── XXX.jpg#xff08;每个文件夹含若干张图像#xff09;├── valid│ └── 1-102#xff08;102个文件夹#xff09;└── ─── └── XXX.jp…1、介绍
1.1数据集介绍
flower_data├── train│ └── 1-102102个文件夹│ └── XXX.jpg每个文件夹含若干张图像├── valid│ └── 1-102102个文件夹└── ─── └── XXX.jpg每个文件夹含若干张图像 cat_to_name.json每一类花朵的名称-编号对应关系1.2 任务介绍
实现102种花朵的分类任务即通过训练train数据集后从valid数据集中选取某一花朵图像能准确判别其属于哪一类花朵
1.3Resnet介绍
在ResNet网络中有如下两个亮点
提出residual结构残差结构并搭建超深的网络结构(突破1000层)使用Batch Normalization加速训练(丢弃dropout)
在ResNet网络提出之前传统的卷积神经网络都是通过将一系列卷积层与下采样层进行堆叠得到的。但是当堆叠到一定网络深度时就会出现两个问题
梯度消失或梯度爆炸退化问题(degradation problem)
2、数据预处理
2.1引入头文件
import os
import matplotlib.pyplot as plt
import numpy as np
import torch
from torch import nn
import torch.optim as optim
import torchvision
from torchvision import transforms,models,datasets
import imageio
import time
import warnings
import random
import sys
import copy
import json
from PIL import Image
2.2数据读取
#数据读取与预处理操作
data_dir ./flower_data/
# 训练集
train_dir data_dir /train
#验证集
valid_ir data_dir /valid
2.3制作数据源
#制作数据源
data_transfroms {train:transforms.Compose([transforms.RandomRotation(45), #随机旋转(-45~45)transforms.CenterCrop(224), #从中心开始裁剪transforms.RandomHorizontalFlip(p 0.5), #随机水平翻转transforms.RandomVerticalFlip(p 0.5), #随机垂直翻转transforms.ColorJitter(brightness0.2,contrast0.1,saturation0.1,hue 0.1),transforms.RandomGrayscale(p 0.025), #概率转换成灰度率3通道就是RGBtransforms.ToTensor(),transforms.Normalize([0.485,0.456,0.406],[0.229,0.224,0.225])]),valid:transforms.Compose([transforms.Resize(256),transforms.CenterCrop(224),transforms.ToTensor(),transforms.Normalize([0.485,0.456,0.406],[0.229,0.224,0.225])]),
}
2.4batch数据制作
#batch数据制作
batch_size 8
image_datasets {x: datasets.ImageFolder(os.path.join(data_dir,x),data_transfroms[x]) for x in [train,valid]}
dataloaders {x: torch.utils.data.DataLoader(image_datasets[x],batch_size batch_size,shuffle True) for x in [train,valid]}
dataset_sizes {x: len(image_datasets[x]) for x in [train,valid]}
class_names image_datasets[train].classes
2.5读取数据标签
#读取标签对应的实际名字
with open(cat_to_name.json,r) as f:cat_to_name json.load(f)
查看cat_to_name.json文件:
{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: colts 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}3、数据展示
3.1图像处理函数
#展示数据
def im_convert(tensor):image tensor.to(cpu).clone().detach()image image.numpy().squeeze()image image.transpose(1,2,0)image image * np.array((0.229,0.224,0.225)) np.array((0.485,0.456,0.406))image image.clip(0.1)return image3.2展示图像
figplt.figure(figsize(20, 12))
columns 4
rows 2dataiter iter(dataloaders[valid])
inputs, classes dataiter.next()for idx in range (columns*rows):ax fig.add_subplot(rows, columns, idx1, xticks[], yticks[])ax.set_title(cat_to_name[str(int(class_names[classes[idx]]))])plt.imshow(im_convert(inputs[idx]))
plt.show()4、进行迁移学习
迁移学习的关键点:
研究可以用哪些知识在不同的领域或者任务中进行迁移学习即不同领域之间有哪些共有知识可以迁移研究在找到了迁移对象之后针对具体问题所采用哪种迁移学习的特定算法即如何设计出合适的算法来提取和迁移共有知识研究什么情况下适合迁移迁移技巧是否适合具体应用其中涉及到负迁移的问题。
4.1训练全连接层
加载models中提供的模型并且直接用训练好的权重当做初始化参数
下载链接:https://download.pytorch.org/models/resnet152-394f9c45.pth 选择resnet网络
model_name resnet #可选的有 [resnet, alexnet, vgg, squeezenet, densenet, inception]#是否用官方训练好的特征来做
feature_extract True 设置用GPU训练
#是否用GPU来训练
train_on_gpu torch.cuda.is_available()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)
屏蔽预训练模型的权重只训练全连接层的权重
def set_parameter_requires_grad(model,feature_extracting):if feature_extracting:for param in model.parameter():param.requires_grad False
选择resnet152网络
model_ft models.resnet152()
设置优化器:
#优化器设置
optimizer_ft optim.Adam(params_to_update,lr 1e-2)
scheduler optim.lr_scheduler.StepLR(optimizer_ft,step_size7,gamma0.1) #学习率每7个epoch衰减成原来的1/10
criterion nn.NLLLoss()
定义训练模块:
# 训练模块
def train_model(model,dataloaders,criterion,optimizer,num_epochs25,is_inceptionFalse,filename filename):since time.time()best_acc 0model.to(device)val_acc_history []train_acc_history []train_losses []valid_losses []LRs [optimizer.param_groups[0][lr]]best_model_wts copy.deepcopy(model.state_dict())for epoch in range(num_epochs):print(Epoch {} / {}.format(epoch,num_epochs - 1))print(- * 10)#训练与验证for phase in [train,valid]:if phase train:model.train() #训练else:model.eval() #验证running_loss 0.0running_corrects 0#把数据取个遍for inputs,labels in dataloaders[phase]:inputs inputs.to(device)labels labels.to(device)#清零optimizer.zero_grad()#只有训练的时候计算与更新梯度with torch.set_grad_enabled(phase train):if is_inception and phase train:outputs,aux_outputs model(inputs)loss1 criterion(outputs,labels)loss2 criterion(aux_outputs,labels)loss loss1 0.4 * loss2else: #resnet执行的是这里outputs model(inputs)loss criterion(outputs,labels)_, preds torch.max(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(dataloaders[phase].dataset)epoch_acc running_corrects.double() / len(dataloaders[phase].dataset)time_elapsed time.time() - sinceprint(Time elapsed {:.0f}m {:.0f}f.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_wts 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_history.append(epoch_acc)valid_losses.append(epoch_loss)scheduler.step(epoch_loss)if phase train:train_acc_history.append(epoch_acc)train_losses.append(epoch_loss)print(Optimizer learning rate : {:.7f}.format(optimizer.param_groups[0][lr]))LRs.append(optimizer.param_groups[0][lr])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))#训练完后用最好的一次当做模型最终的结果model.load_state_dict(best_model_wts)return model,val_acc_history,train_acc_history.valid_losses,train_losses,LRs开始训练:
# 开始训练
model_ft,val_acc_history,train_acc_history,valid_lossea,train_losses,LRs train_model(model_ft,dataloaders,criterion,optimizer_ft,num_epochs20,is_inception(model_name inception))4.2训练所有层
我们从上次训练好最优的那个全连接层的参数开始以此为基础训练所有层设置param.requires_grad True表明接下来训练全部网络之后把学习率调小一点衰减函数为每7次衰减为原来的1/10损失函数不变
再继续训练所有层
for param in model_ft.parameters():param.requires_grad True#再继续训练所有的参数学习率调小一点(lr)
optimizer optim.Adam(params_to_update,lr 1e-4)
#衰减函数(每七次衰减为原来的七分之一)
scheduler optim.lr_scheduler.StepLR(optimizer_ft,step_size7,gamma0.1)#损失函数
criterion nn.NLLLoss() 导入之前的最优结果并开始训练:
#在之前训练得到最好的模型的基础上继续训练
checkpoint torch.load(filename)
best_acc checkpoint[best_acc]
model_ft.load_state_dict(checkpoint[state_dict])
optimizer.load_state_dict(checkpoint[optimizer])model_ft,val_acc_history,train_acc_history,valid_lossea,train_losses,LRs train_model(model_ft,dataloaders,criterion,optimizer_ft,num_epochs10,is_inception(model_name inception))5、测试网络效果
5.1测试数据预处理
首先将新训练好的checkpoint.pth重命名为serious.pth之后加载训练好的模型:
#加载训练好的模型
model_ft,input_size initialize_model(model_name,102,feature_extract,use_pretrainedTrue)#GPU模型
model_ft model_ft.to(device)
#保存文件的名字
filename serious.pth
#加载模型
checkpoint torch.load(filename)
best_acc checkpoint[beat_acc]
model_ft.load_state_dict(checkpoint[state_dict])
定义图像处理函数:
def process_image(image_path):img Image.open(image_path)#Resize,thumbnail方法只能进行缩小所以进行判断if img.size[0] img.size[1]:img.thumbnail((10000,256))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])std np.array([0.229,0.224,0.225])img (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 img
定义图像展示函数:
#展示数据
def imshow(image,ax None,title None):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
展示一个数据:
image_path image_06621.jpg
img process_image(image_path)
imshow(img) 得到一个batch测试数据:
#测试一个batch数据
dataiter iter(dataloaders[valid])
images,labels dataiter.next()model_ft.eval()if train_on_gpu:output model_ft(images.cuda())
else:output model_ft(images)
利用torch.max()函数计算标签值:
#得到属于类别的八个编号
_,preds_tensor torch.ax(output,1)
preds np.squeeze(preds_tensor.numpy()) if not train_on_gpu else np.squeeze(preds_tensor.cpu().numpy())
5.2结果可视化
#展示预测结果
fig plt.figure(figsize(20,20))
columns 4
rows 2for idx in range(columns * rows):ax fig.add_subplot(rows,columns,idx1,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()
结果如下绿色标题代表识别成功红色标题代表识别失败括号里面为真实值括号外为预测值
