4. ResNet18

ResNet18是一种卷积神经网络，它有18层深度，其中包括带有权重的卷积层和全连接层。 它是ResNet系列网络的一个变体，使用了残差连接（residual connection）来解决深度网络的退化问题。

本章将简单介绍下PyTorch以及安装环境，然后简单分析下一个ResNet神经网络以及PyTorch的源码实现， 最后我们使用PyTorch简单构建一个ResNet18网络对Cifar-10进行分类，并部署到鲁班猫板卡上。 本章测试的例程源码请参考下 这里。

提示

测试环境：鲁班猫板卡使用Debian10，PC是WSL2（ubuntu20.04），PyTorch是1.10.1 CPU版，rknn-Toolkit2版本1.4.0。

4.1. PyTorch

PyTorch是一个开源的深度学习框架，该框架由Facebook人工智能研究院的Torch7团队开发， 它的底层基于Torch，但实现与运用全部是由python来完成。该框架主要用于人工智能领域的科学研究与应用开发。

4.1.1. PyTorch安装

PyTorch需要根据自己的环境安装，进入PyTorch官网，查看详细的安装教程。 下面示例在ubuntu20.04上简单安装：

先到 PyTorch官网 ，选择你对应环境的，比如下面是选中linux系统，python语言，cpu版本的PyTorch。

broken

该页面上面默认是最新稳定版的PyTorch，安装以前版本可以点击该页面下面的 Previous version of PyTorch 或者直接点击 这里 。

# 使用pytorch，需要先安装python3和pip基础环境，这些可以自行搜索下即可。
# CPU版：
pip3 install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cpu

# GPU版（CUDA 11.6）：
pip3 install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu116

安装GPU版本（英伟达GPU），需要先根据自己的显卡安装或者更新 显卡驱动 ， 然后安装 CUDA工具包 ， cuDNN 看自己情况安装， 最后点击 这里 查看Pytorch和CUDA的版本对应，

验证是否安装成功：

# 测试安装，终端输入python，
>>> import torch
>>> torch.__version__
'1.10.1+cu102'

# 如果安装的是CUDA版本的PyTorch，命令检测PyTorch的安装版本以及绑定的CUDA版本等
>>> torch.version.cuda
'10.2'
>>> torch.cuda.is_available()
True
# PyTorch的安装根据自己的实际环境和需求，也可以使用docker等环境。

关于程序编辑工具，你可以使用Sublime Text，PyCharm，Vim等，这里测试环境是使用WSL2（ubuntu20.04）， 编辑工具使用Jupyter Notebook，安装教程可以参考下 这里 ， 在Linux系统上使用是类似的。

4.1.2. ResNet18结构简介

ResNet（Residual Neural Network）由微软研究院的Kaiming He等人在2015年提出，ResNet的结构可以极快的加速神经网络的训练，模型的准确率也有比较大的提升。

ResNet是一种残差网络，可以把它理解为一个子网络，这个子网络经过堆叠可以构成一个很深的网络。ResNet系列有多种变体，如ResNet18，ResNet34，ResNet50，ResNet101和ResNet152， 其网络结构如下（参考 论文 ）：

broken

这里我们主要看下ResNet18，ResNet18基本含义是网络的基本架构是ResNet，网络的深度是18层，是带有权重的18层，不包括BN层，池化层。 ResNet18使用的基本残差单元，每个单元由两个3x3卷积层组成，中间有一个BN层和一个ReLU激活函数。

4.1.3. PyTorch中的ResNet18实现

PyTorch中的ResNet18源码实现：https://github.com/pytorch/vision/blob/main/torchvision/models/resnet.py

4.2. ResNet18实现

4.2.1. 数据集准备和数据预处理

接下来我们将自定义一个ResNet18网络结构，并使用CIFAR-10数据集进行简单测试。 CIFAR-10数据集由10个类别的60000张32x32彩色图像组成，每个类别有6000张图像，总共分为50000张训练图像和10000张测试图像。

resnet18.py(截取部分，参考配套例程)
1
2
3
4
5
6
7
8
9
# 导入下载的数据集，使用torchvision加载训练集和测试集，
# 其中参数download=True表示从互联网下载数据，存放到./data目录下，也可以自己下载放到指定目录下。
train_dataset = torchvision.datasets.CIFAR10('./data', download=True, train=True, transform=transform_train)
test_dataset = torchvision.datasets.CIFAR10('./data', download=True, train=False, transform=transform_test)

train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=batch_size, shuffle=True)
test_loader = torch.utils.data.DataLoader(test_dataset, batch_size=batch_size, shuffle=False)

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

对划分的数据集进行预处理

resnet18.py(截取部分，参考配套例程)
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
# 预处理
transform_train=torchvision.transforms.Compose([
        torchvision.transforms.Pad(4),
        torchvision.transforms.RandomHorizontalFlip(), #图像一半的概率翻转，一半的概率不翻转
        torchvision.transforms.RandomCrop(32), #图像随机裁剪成32*32
        torchvision.transforms.ToTensor(), #转为Tensor 把灰度范围从0-255变换到0-1，归一化
        #torchvision.transforms.Normalize((0.4914, 0.4822, 0.4465),(0.2023, 0.1994, 0.2010))
        torchvision.transforms.Normalize([0.5,0.5,0.5], [0.5,0.5,0.5]) #归一化用到的均值和方差
])
transform_test=torchvision.transforms.Compose([
        torchvision.transforms.ToTensor(),
        #torchvision.transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010))
        torchvision.transforms.Normalize([0.5,0.5,0.5], [0.5,0.5,0.5])#归一化用到的均值和方差
])

4.2.2. 构建模型

resnet18.py(截取部分，参考配套例程)
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
# 残差块实现
class ResidualBlock(nn.Module):
def __init__(self, in_channels, out_channels, stride=1, downsample=None):
    super(ResidualBlock, self).__init__()
    self.conv1 = conv3x3(in_channels, out_channels, stride)
    self.bn1 = nn.BatchNorm2d(out_channels)
    self.relu = nn.ReLU(inplace=True)                # 原地替换 节省内存开销
    self.conv2 = conv3x3(out_channels, out_channels)
    self.bn2 = nn.BatchNorm2d(out_channels)
    self.downsample = downsample                     # shortcut
def forward(self, x):
    residual=x
    out = self.conv1(x)
    out = self.bn1(out)
    out = self.relu(out)
    out = self.conv2(out)
    out = self.bn2(out)
    if(self.downsample):
        residual = self.downsample(x)
    out += residual
    out = self.relu(out)
    return out

# 自定义一个神经网络，使用nn.model，，通过__init__初始化每一层神经网络。
# 使用forward连接数据
class ResNet(torch.nn.Module):
    def __init__(self, block, layers, num_classes):
        super(ResNet, self).__init__()
        self.in_channels = 16
        self.conv = conv3x3(3, 16)
        self.bn = torch.nn.BatchNorm2d(16)
        self.relu = torch.nn.ReLU(inplace=True)
        self.layer1 = self._make_layers(block, 16, layers[0])
        self.layer2 = self._make_layers(block, 32, layers[1], 2)
        self.layer3 = self._make_layers(block, 64, layers[2], 2)
        self.layer4 = self._make_layers(block, 128, layers[3], 2)
        self.avg_pool = torch.nn.AdaptiveAvgPool2d((1, 1))
        self.fc = torch.nn.Linear(128, num_classes)

    # _make_layers函数重复残差块，以及shortcut部分
    def _make_layers(self, block, out_channels, blocks, stride=1):
        downsample = None
        if (stride != 1) or (self.in_channels != out_channels):      # 卷积核为1 进行升降维
            downsample = torch.nn.Sequential(                        # stride==2的时候 也就是每次输出信道升维的时候
                conv3x3(self.in_channels, out_channels, stride=stride),
                torch.nn.BatchNorm2d(out_channels)
            )
        layers = []
        layers.append(block(self.in_channels, out_channels, stride, downsample))
        self.in_channels = out_channels
        for i in range(1, blocks):
            layers.append(block(out_channels, out_channels))
        return torch.nn.Sequential(*layers)

    def forward(self, x):
        out = self.conv(x)
        out = self.bn(out)
        out = self.relu(out)
        out = self.layer1(out)
        out = self.layer2(out)
        out = self.layer3(out)
        out = self.layer4(out)
        out = self.avg_pool(out)
        out = out.view(out.size(0), -1)
        out = self.fc(out)
        return out

# Make model，使用cpu
model=ResNet(ResidualBlock, [2,2,2,2], num_classes).to(device=device)

# 打印model结构
print(f"Model structure: {model}\n\n")

# Loss and optimizer
criterion = nn.CrossEntropyLoss()  #交叉熵损失函数
optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate) #优化器随机梯度下降

测试时打印模型结构：

Model structure: ResNet(
    (conv): Conv2d(3, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
    (bn): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    (relu): ReLU(inplace=True)
    (layer1): Sequential(
        (0): ResidualBlock(
        (conv1): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (relu): ReLU(inplace=True)
        (conv2): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (bn2): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
        (1): ResidualBlock(
        (conv1): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (relu): ReLU(inplace=True)
        (conv2): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (bn2): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
    )
    (layer2): Sequential(
        (0): ResidualBlock(
        (conv1): Conv2d(16, 32, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
        (bn1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (relu): ReLU(inplace=True)
        (conv2): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (bn2): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (downsample): Sequential(
            (0): Conv2d(16, 32, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
            (1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
        )
        (1): ResidualBlock(
        (conv1): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (bn1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (relu): ReLU(inplace=True)
        (conv2): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (bn2): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
    )
    (layer3): Sequential(
        (0): ResidualBlock(
        (conv1): Conv2d(32, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
        (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (relu): ReLU(inplace=True)
        (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (downsample): Sequential(
            (0): Conv2d(32, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
            (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
        )
        (1): ResidualBlock(
        (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (relu): ReLU(inplace=True)
        (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
    )
    (layer4): Sequential(
        (0): ResidualBlock(
        (conv1): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
        (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (relu): ReLU(inplace=True)
        (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (downsample): Sequential(
            (0): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
            (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
        )
        (1): ResidualBlock(
        (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        (relu): ReLU(inplace=True)
        (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
        (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
    )
    (avg_pool): AdaptiveAvgPool2d(output_size=(1, 1))
    (fc): Linear(in_features=128, out_features=10, bias=True)
)

4.2.3. 训练和测试模型

resnet18.py(截取部分，参考配套例程)
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
if __name__ == "__main__":
    # 训练模型
    total_step = len(train_loader)
    for epoch in range(0,num_epoches):
        for i, (images, labels) in enumerate(train_loader):
            images = images.to(device=device)
            labels = labels.to(device=device)
            outputs = model(images)
            loss = criterion(outputs, labels)
            # 梯度清零
            optimizer.zero_grad()
            # 反向传播
            loss.backward()
            # 更新参数
            optimizer.step()
            if (i+1) % total_step == 0:
                print('Epoch [{}/{}], Step [{}/{}], Loss: {:.4f}'
                    .format(epoch+1, num_epoches, i+1, total_step, loss.item()))
    print("Finished Tranining")

print('\nTest the model')
# 转换到`eval`模式
model.eval()  # eval mode (batchnorm uses moving mean/variance instead of mini-batch mean/variance)
with torch.no_grad():

    correct = 0
    total = 0

    for images, labels in test_loader:

        images = images.to(device=device)
        labels = labels.to(device=device)
        outputs = model(images)

        _, predicted = torch.max(outputs.data, 1)

        total += labels.size(0)
        correct += (predicted == labels).sum().item()

    print('Accuracy of the network on the 10000 test images: {} %'.format(100 * correct / total))

训练模型，后再测试集上测试准确率达89.8400%：

Epoch [88/100], Step [391/391], Loss: 0.0820
Epoch [89/100], Step [391/391], Loss: 0.0185
Epoch [90/100], Step [391/391], Loss: 0.0166
Epoch [91/100], Step [391/391], Loss: 0.0334
Epoch [92/100], Step [391/391], Loss: 0.0641
Epoch [93/100], Step [391/391], Loss: 0.0359
Epoch [94/100], Step [391/391], Loss: 0.0994
Epoch [95/100], Step [391/391], Loss: 0.0069
Epoch [96/100], Step [391/391], Loss: 0.0722
Epoch [97/100], Step [391/391], Loss: 0.0182
Epoch [98/100], Step [391/391], Loss: 0.2182
Epoch [99/100], Step [391/391], Loss: 0.0657
Epoch [100/100], Step [391/391], Loss: 0.0501
Finished Tranining

Test the model
在10000张测试集图片上的准确率:89.8400 %

4.2.4. 保存为onnx模型

这里我们使用torch.onnx.export保存模型为onnx模型（也可以导出pt模型等）：

resnet18.py(截取部分，参考配套例程)
1
2
3
# export onnx (rknn-toolkit2 only support to opset_version=12)
x = torch.randn((1, 3, 32, 32))
torch.onnx.export(model, x, './resnet18_pytorch_100.onnx', opset_version=12, input_names=['input'], output_names=['output'])

4.2.5. 导出RKNN模型和模拟测试

resnet18.py(截取部分，参考配套例程)
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
def show_perfs(perfs):
    perfs = 'perfs: {}\n'.format(perfs)
    print(perfs)

def softmax(x):
    return np.exp(x)/sum(np.exp(x))

if __name__ == '__main__':

    MODEL = './resnet18_pytorch.onnx'

    # 创建RKNN
    # 如果测试遇到问题，请开启verbose=True，查看调试信息。
    # rknn = RKNN(verbose=True)
    rknn = RKNN()

    # 配置模型，预处理
    print('--> Config model')
    rknn.config(mean_values=[125.307, 122.961, 113.8575], std_values=[51.5865, 50.847, 51.255], target_platform='rk3568')
    print('done')

    # 加载模型
    print('--> Loading model')
    #ret = rknn.load_pytorch(model=model, input_size_list=input_size_list)
    ret = rknn.load_onnx(model=MODEL)
    if ret != 0:
        print('Load model failed!')
        exit(ret)
    print('done')

    # 构建模型
    print('--> Building model')
    ret = rknn.build(do_quantization=False)
    if ret != 0:
        print('Build model failed!')
        exit(ret)
    print('done')

    # 导出rknn模型
    print('--> Export rknn model')
    ret = rknn.export_rknn('./resnet_18_100.rknn')
    if ret != 0:
        print('Export rknn model failed!')
        exit(ret)
    print('done')

    # 输入图片处理
    img = cv2.imread('./0_125.jpg')
    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
    img = cv2.resize(img,(32,32))
    img = np.expand_dims(img, 0)

    # 初始化运行环境
    print('--> Init runtime environment')
    ret = rknn.init_runtime()
    if ret != 0:
        print('Init runtime environment failed!')
        exit(ret)
    print('done')

    # 模拟推理
    print('--> Running model')
    outputs = rknn.inference(inputs=[img])
    np.save('./pytorch_resnet18_qat_0.npy', outputs[0])
    #show_outputs(softmax(np.array(outputs[0][0])))
    print(outputs)
    print('done')

    rknn.release()

4.2.6. 板端部署测试

4.2.6.1. 简单测试

rknnlite_inference0.py
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
IMG_PATH = '0_125.jpg'
RKNN_MODEL = './resnet_18_100.rknn'
img_height = 32
img_width = 32
class_names = ["plane","car","bird","cat","deer","dog","frog","horse","ship","truck"]

# Create RKNN object
rknn_lite = RKNNLite()

# load RKNN model
print('--> Load RKNN model')
ret = rknn_lite.load_rknn(RKNN_MODEL)
if ret != 0:
    print('Load RKNN model failed')
    exit(ret)
print('done')

# Init runtime environment
print('--> Init runtime environment')
ret = rknn_lite.init_runtime()
if ret != 0:
    print('Init runtime environment failed!')
    exit(ret)
print('done')

# load image
img = cv2.imread(IMG_PATH)
img = cv2.resize(img,(32,32))
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
img = np.expand_dims(img, 0)

# runing model
print('--> Running model')
outputs = rknn_lite.inference(inputs=[img])
print("result: ", outputs)
print(
    "This image most likely belongs to {}."
    .format(class_names[np.argmax(outputs)])
)
rknn_lite.release()

测试结果：

--> Load RKNN model
done
--> Init runtime environment
I RKNN: [16:02:15.992] RKNN Runtime Information: librknnrt version: 1.4.0 (a10f100eb@2022-09-09T09:07:14)
I RKNN: [16:02:15.992] RKNN Driver Information: version: 0.7.2
I RKNN: [16:02:15.992] RKNN Model Information: version: 1, toolkit version: 1.4.0-22dcfef4(compiler version: 1.4.0 (3b4520e4f@2022-09-05T20:52:35)), target: RKNPU lite, target platform: rk3568, framework name: ONNX, framework layout: NCHW
done
--> Running model
result:  [array([[ -2.0566406, -15.234375 ,   6.6835938,  -6.828125 ,  -9.9921875,
        -6.5390625,  -5.671875 , -15.8515625, -17.96875  , -11.90625  ]],
    dtype=float32)]
This image most likely belongs to bird.

4.2.6.2. 使用测试集图像测试

先转换测试集为.jpg格式图片，然后传输到板卡进行部署测试：

cifar10_to_jpg.py
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
# CIFAR-10数据集所在的绝对路径，根据具体路径修改
base_dir = "/mnt/e/Users/Administrator/Desktop/wsl_user/pytorch/"
data_dir = os.path.join(base_dir, "data", "cifar-10-batches-py")
test_o_dir = os.path.join( base_dir, "Data", "cifar-10-png", "raw_test")

# 解压缩
def unpickle(file):
    with open(file, 'rb') as fo:
        dict_ = pickle.load(fo, encoding='bytes')
    return dict_

# 生成测试集图片
if __name__ == '__main__':
    print("start...")
    test_data_path = os.path.join(data_dir, "test_batch")
    test_data = unpickle(test_data_path)
    for i in range(0, 10000):
        img = np.reshape(test_data[b'data'][i], (3, 32, 32))
        img = img.transpose(1, 2, 0)

        label_num = str(test_data[b'labels'][i])
        o_dir = os.path.join(test_o_dir, label_num)
        if not os.path.isdir(o_dir):
            os.makedirs(o_dir)

        img_name = label_num + '_' + str(i) + '.jpg'
        img_path = os.path.join(o_dir, img_name)
        imwrite(img_path, img)
    print("done.")

然后修改板端测试文件添加：

rknnlite_inference1.py
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
def rknn_inference(root):
    total=0
    correct=0
    for path in os.listdir(root):
        image_filenames = os.listdir(root + '/' + path)
        for image_filename in image_filenames:
            img = cv2.imread(root + '/' + path + '/' + image_filename)
            img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
            outputs = rknn_lite.inference(inputs=[img])
            total += 1
            if np.argmax(outputs) == int(path[:1]) :
                correct += 1
    print("corrorect={}, total={}".format(correct,total))
    print('在{}张测试集图片上的准确率:{:.2f} %'.format(total,100 * correct / total))

简单测试结果：

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
--> Load RKNN model
done
--> Init runtime environment
I RKNN: [10:23:15.384] RKNN Runtime Information: librknnrt version: 1.4.0 (a10f100eb@2022-09-09T09:07:14)
I RKNN: [10:23:15.385] RKNN Driver Information: version: 0.7.2
I RKNN: [10:23:15.385] RKNN Model Information: version: 1, toolkit version:
1.4.0-22dcfef4(compiler version: 1.4.0 (3b4520e4f@2022-09-05T20:52:35)), target:
RKNPU lite, target platform: rk3568, framework name: ONNX, framework layout: NCHW
done
--> Running model
在10000张测试集图片上的准确率:71.21 %
done