test/dcu/code/train.py

from torchvision.transforms import transforms
from torchvision.datasets import CIFAR10
from torch.utils.data import DataLoader
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Adam
from torch.autograd import Variable
import argparse
import time
import shutil

from c2net.context import prepare,upload_output

parser = argparse.ArgumentParser(description='忽略超参数不存在的报错问题')
#添加自定义参数
parser.add_argument("--output")
parser.add_argument('--epoch', type=int, default=1)
parser.add_argument('--card', type=str, default='cuda:0')

args = parser.parse_args()
args, unknown = parser.parse_known_args()

#初始化导入数据集和预训练模型到容器内
c2net_context = prepare()

# codePath = c2net_context.code_path
# test = codePath + '/pytorch-cnn-cifar10-dcu' + '/test.py'

#获取数据集路径
cifar_10_python_path = c2net_context.dataset_path+"/"+"jo-161M"

#输出结果必须保存在该目录
outputPath = c2net_context.output_path

#回传结果到openi，只有训练任务才能回传
upload_output()

class Network(nn.Module):
    def __init__(self):
        super(Network, self).__init__()

        self.conv1 = nn.Conv2d(in_channels=3, out_channels=12, kernel_size=5, stride=1, padding=1)
        self.bn1 = nn.BatchNorm2d(12)
        self.conv2 = nn.Conv2d(in_channels=12, out_channels=12, kernel_size=5, stride=1, padding=1)
        self.bn2 = nn.BatchNorm2d(12)
        self.pool = nn.MaxPool2d(2, 2)
        self.conv4 = nn.Conv2d(in_channels=12, out_channels=24, kernel_size=5, stride=1, padding=1)
        self.bn4 = nn.BatchNorm2d(24)
        self.conv5 = nn.Conv2d(in_channels=24, out_channels=24, kernel_size=5, stride=1, padding=1)
        self.bn5 = nn.BatchNorm2d(24)
        self.fc1 = nn.Linear(24 * 10 * 10, 10)

    def forward(self, input):
        output = F.relu(self.bn1(self.conv1(input)))
        output = F.relu(self.bn2(self.conv2(output)))
        output = self.pool(output)
        output = F.relu(self.bn4(self.conv4(output)))
        output = F.relu(self.bn5(self.conv5(output)))
        output = output.view(-1, 24 * 10 * 10)
        output = self.fc1(output)

        return output

def saveModel():
    path = outputPath + '/' + 'test.pth'
    torch.save(model.state_dict(), path)
    zipfileName = outputPath + '/' + 'test_database'
    save_zipfile(zipfileName, outputPath)


def testAccuracy(card):
    model.eval()
    accuracy = 0.0
    total = 0.0

    # device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
    device = torch.device(card)
    model.to(device)

    test_dataloader = Get_dataloader(False)
    with torch.no_grad():
        for data in test_dataloader:
            images, labels = data
            images = Variable(images.to(device))
            labels = Variable(labels.to(device))
            outputs = model(images)
            _, predicted = torch.max(outputs.data, 1)
            total += labels.size(0)
            accuracy += (predicted == labels).sum().item()

    accuracy = (100 * accuracy / total)
    return (accuracy)


def train(num_epochs, card):
    # best_accuracy = 0.0

    # device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

    device = torch.device(card)

    model.to(device)

    train_dataloader = Get_dataloader(True)
    for epoch in range(num_epochs):
        running_loss = 0.0
        running_acc = 0.0

        for i, (images, labels) in enumerate(train_dataloader, 0):

            images = Variable(images.to(device))
            labels = Variable(labels.to(device))
            optimizer.zero_grad()
            outputs = model(images)
            loss = loss_fn(outputs, labels)
            loss.backward()
            optimizer.step()

            running_loss += loss.item()
            if i % 1000 == 999:
                print('[%d, %5d] loss: %.3f' %
                      (epoch + 1, i + 1, running_loss / 1000))
                running_loss = 0.0

        accuracy = testAccuracy(card)
        print('For epoch', epoch + 1, 'the test accuracy over the whole test set is %d %%' % (accuracy))

    # saveModel()


def Get_dataloader(train):
    transform_fn = transforms.Compose([
        transforms.ToTensor(),
        transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
    ])
    dataset = CIFAR10(root=DATA_ROOT, train=train, transform=transform_fn)
    data_loader = DataLoader(dataset=dataset, batch_size=batch_size, shuffle=True, num_workers=0)
    return data_loader

def save_zipfile(filename, dest):
    shutil.make_archive(filename, 'zip', dest)


DATA_ROOT = cifar_10_python_path
classes = ('plane', 'car', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck')

batch_size = 10
number_of_labels = 10

model = Network()

loss_fn = nn.CrossEntropyLoss()
optimizer = Adam(model.parameters(), lr=0.001, weight_decay=0.0001)

if __name__ == "__main__":
    accuracy = testAccuracy(args.card)
    print('before training, accuracy for test data is: ', accuracy)

    start = time.perf_counter()
    train(args.epoch, args.card)
    end = time.perf_counter()
    print(f"training completed in {end - start:0.4f} seconds")
    saveModel()