(标签-conda|关键词-File)

Question

#报错内容

C:\Users\ASUS\.conda\envs\mindspore\python.exe C:\PycharmProjects\pythonProject1\venv\CIFAR_10_jiangjie\main.py 
Traceback (most recent call last):
  File "C:\PycharmProjects\pythonProject1\venv\CIFAR_10_jiangjie\main.py", line 1, in <module>
    import torchvision
  File "C:\Users\ASUS\.conda\envs\mindspore\lib\site-packages\torchvision\__init__.py", line 5, in <module>
    from torchvision import datasets
  File "C:\Users\ASUS\.conda\envs\mindspore\lib\site-packages\torchvision\datasets\__init__.py", line 1, in <module>
    from ._optical_flow import KittiFlow, Sintel, FlyingChairs, FlyingThings3D, HD1K
  File "C:\Users\ASUS\.conda\envs\mindspore\lib\site-packages\torchvision\datasets\_optical_flow.py", line 12, in <module>
    from ..io.image import _read_png_16
  File "C:\Users\ASUS\.conda\envs\mindspore\lib\site-packages\torchvision\io\__init__.py", line 22, in <module>
    from .image import (
  File "C:\Users\ASUS\.conda\envs\mindspore\lib\site-packages\torchvision\io\image.py", line 59, in <module>
    module.load_state_dict(torch.load(r'./param/cifar10_cnn.pkl'))
  File "C:\Users\ASUS\.conda\envs\mindspore\lib\site-packages\torch\serialization.py", line 594, in load
    with _open_file_like(f, 'rb') as opened_file:
  File "C:\Users\ASUS\.conda\envs\mindspore\lib\site-packages\torch\serialization.py", line 230, in _open_file_like
    return _open_file(name_or_buffer, mode)
  File "C:\Users\ASUS\.conda\envs\mindspore\lib\site-packages\torch\serialization.py", line 211, in __init__
    super(_open_file, self).__init__(open(name, mode))
FileNotFoundError: [Errno 2] No such file or directory: './param/cifar10_cnn.pkl'

这个是我操作台的设置，目前出现的问题就是不管运行的是什么样的代码出现的都是FileNotFoundError: [Errno 2] No such file or directory: './param/cifar10_cnn.pkl'的错误，这个能有什么解决方法嘛？

#main.py的代码如下

import torchvision
from torch import nn
import torch
from torch.utils.data import DataLoader
from matplotlib import pyplot as plt
from matplotlib import font_manager

font = font_manager.FontProperties(fname=r'C:\Windows\Fonts\STSONG.TTF')

device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

train_loss = list()
test_loss = list()

def plot(train_loss,test_loss):
    plt.figure(figsize=(5,5))
    plt.plot(train_loss,label='train_loss',alpha = 0.5)
    plt.plot(test_loss,label='test_loss',alpha=0.5)
    plt.title('使用非线性函数和图像数据增强手段处理CIFAR-10数据集',fontproperties=font)
    plt.xlabel('训练/测试次数',fontproperties=font)
    plt.ylabel('损失',fontproperties=font)
    plt.legend()
    plt.show()

#图像数据增强手段
train_data_set = torchvision.datasets.CIFAR10(root='dataset',train=True,transform=torchvision.transforms.Compose([
    torchvision.transforms.RandomCrop(size=(32,32),padding=4),
    torchvision.transforms.RandomHorizontalFlip(),
    torchvision.transforms.ToTensor(),
    torchvision.transforms.Normalize(mean=[0.5,0.5,0.5],std=[0.5,0.5,0.5])
]),download=True)

test_data_set = torchvision.datasets.CIFAR10(root='dataset',train=False,transform=torchvision.transforms.Compose([
    torchvision.transforms.RandomCrop(size=(32,32),padding=4),
    torchvision.transforms.RandomHorizontalFlip(),
    torchvision.transforms.ToTensor(),
    torchvision.transforms.Normalize(mean=[0.5,0.5,0.5],std=[0.5,0.5,0.5])
]),download=True)

train_data_load = DataLoader(dataset=train_data_set,batch_size=64,shuffle=True,drop_last=True)
test_data_load = DataLoader(dataset=test_data_set,batch_size=64,shuffle=True,drop_last=True)

train_data_size = len(train_data_set)
test_data_size = len(test_data_set)

print(f'训练集的大小为{train_data_size}')
print(f'测试集的大小为{test_data_size}')

class MyNet(nn.Module):
    def __init__(self):
        super().__init__()
        self.main = nn.Sequential(
            nn.Conv2d(in_channels=3,out_channels=32,kernel_size=3,padding=1),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2,stride=2),
            nn.BatchNorm2d(num_features=32),

            nn.Conv2d(in_channels=32,out_channels=64,kernel_size=3,padding=1),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2,stride=2),
            nn.BatchNorm2d(num_features=64),

            nn.Conv2d(in_channels=64,out_channels=128,kernel_size=3,padding=1),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2,stride=2),
            nn.BatchNorm2d(128),
        )

        self.fc = nn.Sequential(
            nn.Flatten(),

            nn.Linear(128*4*4,1024),
            nn.ReLU(inplace=True),
            nn.Dropout(),

            nn.Linear(1024,256),
            nn.ReLU(inplace=True),
            nn.Dropout(),

            nn.Linear(256,10)
        )

    def forward(self,x):
        return self.fc(self.main(x))

mynet = MyNet()
mynet = mynet.to(device)
print(mynet)

loss_fn = nn.CrossEntropyLoss()
loss_fn = loss_fn.to(device)

learning_rate = 1e-3
optim = torch.optim.Adam(mynet.parameters(),lr=learning_rate)

train_step = 0
test_step = 0

epoch = 5

if __name__ == '__main__':
    for i in range(epoch) :
        print(f'----------第{i+1}轮训练开始----------')
        mynet.train()
        for j , (imgs,targets) in enumerate(train_data_load):

            imgs = imgs.to(device)
            targets = targets.to(device)

            outputs = mynet(imgs)
            loss = loss_fn(outputs,targets)
            optim.zero_grad()
            loss.backward()
            optim.step()

            train_step += 1
            if train_step % 100 == 0:
                print(f'训练第{train_step}次，loss={loss}')
                train_loss.append(loss)

        mynet.eval()
        accuracy = 0
        accuracy_total = 0
        with torch.no_grad():
            for j,(imgs,targets) in enumerate(test_data_load):

                imgs = imgs.to(device)
                targets = targets.to(device)

                outputs = mynet(imgs)
                loss = loss_fn(outputs,targets)

                accuracy = (outputs.argmax(axis=1) == targets).sum()
                accuracy_total += accuracy
                test_step += 1

                if test_step % 100 == 0:
                    test_loss.append(loss)

            print(f'第{i+1}轮训练结束，准确率{accuracy_total/test_data_size}')
            torch.save(mynet,f'CIAFR_10_{i+1}_acc_{accuracy_total/test_data_size}.pth')
    plot(train_loss,test_loss)

#test.py的代码如下

import os
import torch
from PIL import Image
import torchvision
from torch import nn

class MyNet(nn.Module):
    def __init__(self):
        super().__init__()
        self.main = nn.Sequential(
            nn.Conv2d(in_channels=3,out_channels=32,kernel_size=3,padding=1),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2,stride=2),
            nn.BatchNorm2d(num_features=32),

            nn.Conv2d(in_channels=32,out_channels=64,kernel_size=3,padding=1),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2,stride=2),
            nn.BatchNorm2d(num_features=64),

            nn.Conv2d(in_channels=64,out_channels=128,kernel_size=3,padding=1),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2,stride=2),
            nn.BatchNorm2d(128),
        )

        self.fc = nn.Sequential(
            nn.Flatten(),

            nn.Linear(128*4*4,1024),
            nn.ReLU(inplace=True),
            nn.Dropout(),

            nn.Linear(1024,256),
            nn.ReLU(inplace=True),
            nn.Dropout(),

            nn.Linear(256,10)
        )

    def forward(self,x):
        return self.fc(self.main(x))

targets_idx = {
    0:'飞机',
    1:'汽车',
    2:'鸟',
    3:'猫',
    4:'鹿',
    5:'狗',
    6:'青蛙',
    7:'马',
    8:'船',
    9:'卡车'
}

root_dir = 'test_CIFAR_10'
obj_dir = 'test2.png'

img_dir = os.path.join(root_dir,obj_dir)
img = Image.open(img_dir)

tran_pose = torchvision.transforms.Compose([
    torchvision.transforms.Resize(size=(32,32)),
    torchvision.transforms.ToTensor()
])

mynet = torch.load('CIAFR_10_20_acc_0.8222000002861023.pth',map_location=torch.device('cpu'))
img = tran_pose(img)
img = torch.reshape(img,(1,3,32,32))
print(img.shape)
te
output = mynet(img)
print(targets_idx[output.argmax(axis = 1).item()])

Answer 1

【以下回答由 GPT 生成】

1. 首先，确保你的代码中使用的路径是正确的。错误信息显示文件'cifar10_cnn.pkl'位于'./param/'目录下，这意味着代码在运行时应该能够找到该目录。

2. 确保'cifar10_cnn.pkl'文件确实存在于'./param/'目录下。在你的项目文件夹中查找这个文件，如果不存在，可能是因为你没有将该文件添加到项目中或者文件被意外删除了。在这种情况下，你需要找到原始的'cifar10_cnn.pkl'文件并将其放置到正确的目录下。

3. 如果文件确实存在于'./param/'目录下，那么可能是因为文件读取权限的问题导致程序找不到文件。你可以尝试使用绝对路径来加载文件，例如使用`torch.load(r'C:\your_project_path\param\cifar10_cnn.pkl')`来替代相对路径。

4. 最后，如果以上步骤都没有解决问题，那么可能是因为某些库的环境配置问题。你可以尝试重新安装相关的库，例如通过conda重新安装torchvision和torch库。

如果以上解决方案都没有解决问题，那么可能是有其他更深层次的问题导致找不到文件，这可能需要更深入的排查和调试。

如果你已经解决了该问题, 非常希望你能够分享一下解决方案, 写成博客, 将相关链接放在评论区, 以帮助更多的人 ^-^