ParC-Net：结合ConvNets和Transformer优点的位置感知循环

本文提出了基于纯卷积的轻量化模型ParC-Net，继承了ViT的全局建模能力，同时又保持了卷积的高计算效率。

AI Studio

708人浏览 · 2023-04-28 11:01:16

AI Studio · 2023-04-28 11:01:16 发布

★★★ 本文源自AlStudio社区精品项目，【点击此处】查看更多精品内容 >>>

摘要

最近，视觉Transformer开始显示出令人印象深刻的结果，显著优于基于大卷积的模型。然而，在面向移动或资源受限设备的小模型领域，ConvNet在性能和模型复杂度方面仍有自己的优势。我们提出了PARC-Net，这是一个纯基于ConvNet的主干模型，通过将视觉Transformer的优点融合到ConvNet中来进一步增强这些优点。具体来说，我们提出了位置感知循环卷积(ParC)，这是一种轻量级的卷积运算，它拥有全局感受野，同时产生像局部卷积一样的位置敏感特征。我们将ParCs和压缩激励操作结合起来，形成一个类似于MetaFormer的模型块，进一步具有类似于Transformer的注意力机制。上述模块可以以即插即用的方式用于替换Convnets或Transformer中的相关模块。实验结果表明，在一般的视觉任务和数据集中，该ParC-Net比目前流行的轻量级ConvNets和基于Transformer的模型具有更好的性能，且参数更少，推理速度更快。在ImageNet-1K上，ParC-Net的分类准确率达到78.6%,约500万个参数，节省11%的参数和13%的计算成本，但比MobileViT（基于ARM的RockChip RK3288)的分类准确率高0.2%,推理速度快23%；与DeiT相比，仅使用0.5×个参数，准确率提高2.7%。在MS-COCO目标检测和PASCAL VOC分割任务上，ParC-Net也表现出较好的性能。

1. ParC Block

如下图所示，本文比较了ConvNets与ViT的优劣，并总结了ViT相对于ConvNets的几个优势：

ViT擅长提取全局特征
ViT使用Meta-Former形式
在信息聚合中，ViT是数据驱动的（即与数据相关）

由此，本文提出了ParC Block，既保留ConvNets的高计算效率，又可以继承ViT的全局建模能力。如图2所示，本文提出将输入沿通道分成两部分，一部分在H维进行全局建模，一部分在V维进行全局建模，具体的全局建模可以表示为如下公式：

$\begin{array}{l} p e^{V}=F\left(\widetilde{p e}^{V}\right)=\left[p e_{0}^{V}, p e_{1}^{V}, \cdots, p e_{h-1}^{V}\right]^{T} \\ p e_{e}^{V}=E V\left(p e^{V}, w\right) \\ k^{V}=F(\widetilde{k})=\left[k_{0}^{V}, k_{1}^{V}, \cdots, k_{h-1}^{V}\right] \\ x^{p}=x+p e_{e}^{V} \\ y_{i, j}=\sum_{t \in(0, h-1)} k_{t}^{V} x_{((i+t) \bmod h, j)}^{p} \end{array}$

上式可以理解为先对输入在V维嵌入位置编码，然后将其复制并合并起来，以方便后续的全局建模能力（使用一个等于V维大小的卷积核来实现，如图3所示），这里的位置编码是必不可少的，因为由于循环的操作，卷积已经无法进行位置感知。

ParC Block的优势：

ParC能够从全局空间中提取全局特征和像素间的交互
自注意力的计算与输入大小呈二次关系，用ParC去替换可以大大降低计算成本

2. 代码复现

2.1 下载并导入所需要的包

%matplotlib inline
import paddle
import paddle.fluid as fluid
import numpy as np
import matplotlib.pyplot as plt
from paddle.vision.datasets import Cifar10
from paddle.vision.transforms import Transpose
from paddle.io import Dataset, DataLoader
from paddle import nn
import paddle.nn.functional as F
import paddle.vision.transforms as transforms
import os
import matplotlib.pyplot as plt
from matplotlib.pyplot import figure
from paddle import ParamAttr
from paddle.nn.layer.norm import _BatchNormBase

2.2 创建数据集

train_tfm = transforms.Compose([
    transforms.RandomResizedCrop(32),
    transforms.RandomHorizontalFlip(0.5),
    transforms.ToTensor(),
    transforms.Normalize(mean=(0.5, 0.5, 0.5), std=(0.5, 0.5, 0.5)),
])

test_tfm = transforms.Compose([
    transforms.Resize((32, 32)),
    transforms.ToTensor(),
    transforms.Normalize(mean=(0.5, 0.5, 0.5), std=(0.5, 0.5, 0.5)),
])

paddle.vision.set_image_backend('cv2')
# 使用Cifar100数据集
train_dataset = Cifar10(data_file='data/data152754/cifar-10-python.tar.gz', mode='train', transform = train_tfm)
val_dataset = Cifar10(data_file='data/data152754/cifar-10-python.tar.gz', mode='test',transform = test_tfm)
print("train_dataset: %d" % len(train_dataset))
print("val_dataset: %d" % len(val_dataset))

train_dataset: 50000
val_dataset: 10000

batch_size=256

train_loader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True, drop_last=True, num_workers=4)
val_loader = DataLoader(val_dataset, batch_size=batch_size, shuffle=False, drop_last=False, num_workers=4)

2.3 标签平滑

class LabelSmoothingCrossEntropy(nn.Layer):
    def __init__(self, smoothing=0.1):
        super().__init__()
        self.smoothing = smoothing

    def forward(self, pred, target):

        confidence = 1. - self.smoothing
        log_probs = F.log_softmax(pred, axis=-1)
        idx = paddle.stack([paddle.arange(log_probs.shape[0]), target], axis=1)
        nll_loss = paddle.gather_nd(-log_probs, index=idx)
        smooth_loss = paddle.mean(-log_probs, axis=-1)
        loss = confidence * nll_loss + self.smoothing * smooth_loss

        return loss.mean()

2.4 ResNet-ParC

2.4.1 ParC

class ParC_op(nn.Layer):
    def __init__(self, dim, type, global_kernel_size, use_pe=True, groups=1):
        super().__init__()
        self.type = type
        self.dim = dim
        self.global_kernel_size = global_kernel_size
        self.use_pe = use_pe
        
        self.kernel_size = (1, self.global_kernel_size) if type == 'H' else (self.global_kernel_size, 1)
        self.gcc_conv = nn.Conv2D(dim, dim, kernel_size=self.kernel_size, stride=1, padding=0, groups=groups)

        if use_pe:
            if self.type == 'H':
                self.pe = self.create_parameter((1, dim, 1, self.global_kernel_size))
            else:
                self.pe = self.create_parameter((1, dim, self.global_kernel_size, 1))
        
    def forward(self, x):
        if self.use_pe:
            x = x + self.pe

        x_cat = paddle.concat([x, x[:, :, :, :-1]], axis=-1) if self.type == 'H' else  paddle.concat([x, x[:, :, :-1, :]], axis=-2)

        x = self.gcc_conv(x_cat)

        return x

class ParCBlock(nn.Layer):

    expansion = 1

    def __init__(
        self,
        inplanes,
        planes,
        stride=1,
        downsample=None,
        groups=1,
        base_width=64,
        dilation=1,
        norm_layer=None,
        use_pe = True,
        global_kernel_size = 14
    ):
        super().__init__()
        if norm_layer is None:
            norm_layer = nn.BatchNorm2D
        width = int(planes * (base_width / 64.0)) * groups

        self.conv1 = nn.Conv2D(inplanes, width, 1, bias_attr=False)
        self.bn1 = norm_layer(width)

        self.parc_H = ParC_op(width//2, 'H', global_kernel_size, use_pe, groups = groups)
        self.parc_W = ParC_op(width//2, 'W', global_kernel_size, use_pe, groups = groups)
        self.bn2 = norm_layer(width)

        self.conv3 = nn.Conv2D(
            width, planes * self.expansion, 1, bias_attr=False
        )
        self.bn3 = norm_layer(planes * self.expansion)
        self.relu = nn.ReLU()
        self.downsample = downsample
        self.stride = stride

    def forward(self, x):
        identity = x

        out = self.conv1(x)
        out = self.bn1(out)
        out = self.relu(out)

        outH, outW = paddle.chunk(out, 2, axis=1)
        outH = self.parc_H(outH)
        outW = self.parc_W(outW)
        out = paddle.concat([outH, outW], axis=1)
        out = self.bn2(out)
        out = self.relu(out)

        out = self.conv3(out)
        out = self.bn3(out)

        if self.downsample is not None:
            identity = self.downsample(x)

        out += identity
        out = self.relu(out)

        return out

model = ParCBlock(64, 64)
paddle.summary(model, (1, 64, 14, 14))

W0215 08:32:45.164451  1049 gpu_resources.cc:61] Please NOTE: device: 0, GPU Compute Capability: 7.0, Driver API Version: 11.2, Runtime API Version: 11.2
W0215 08:32:45.169569  1049 gpu_resources.cc:91] device: 0, cuDNN Version: 8.2.


---------------------------------------------------------------------------
 Layer (type)       Input Shape          Output Shape         Param #    
===========================================================================
   Conv2D-1      [[1, 64, 14, 14]]     [1, 64, 14, 14]         4,096     
 BatchNorm2D-1   [[1, 64, 14, 14]]     [1, 64, 14, 14]          256      
    ReLU-1       [[1, 64, 14, 14]]     [1, 64, 14, 14]           0       
   Conv2D-2      [[1, 32, 14, 27]]     [1, 32, 14, 14]        14,368     
   ParC_op-1     [[1, 32, 14, 14]]     [1, 32, 14, 14]          448      
   Conv2D-3      [[1, 32, 27, 14]]     [1, 32, 14, 14]        14,368     
   ParC_op-2     [[1, 32, 14, 14]]     [1, 32, 14, 14]          448      
 BatchNorm2D-2   [[1, 64, 14, 14]]     [1, 64, 14, 14]          256      
   Conv2D-4      [[1, 64, 14, 14]]     [1, 64, 14, 14]         4,096     
 BatchNorm2D-3   [[1, 64, 14, 14]]     [1, 64, 14, 14]          256      
===========================================================================
Total params: 38,592
Trainable params: 38,208
Non-trainable params: 384
---------------------------------------------------------------------------
Input size (MB): 0.05
Forward/backward pass size (MB): 0.77
Params size (MB): 0.15
Estimated Total Size (MB): 0.96
---------------------------------------------------------------------------






{'total_params': 38592, 'trainable_params': 38208}

2.4.2 ResNet-ParC

class BasicBlock(nn.Layer):
    expansion = 1

    def __init__(
        self,
        inplanes,
        planes,
        stride=1,
        downsample=None,
        groups=1,
        base_width=64,
        dilation=1,
        norm_layer=None,
    ):
        super().__init__()
        if norm_layer is None:
            norm_layer = nn.BatchNorm2D

        if dilation > 1:
            raise NotImplementedError(
                "Dilation > 1 not supported in BasicBlock"
            )

        self.conv1 = nn.Conv2D(
            inplanes, planes, 3, padding=1, stride=stride, bias_attr=False
        )
        self.bn1 = norm_layer(planes)
        self.relu = nn.ReLU()
        self.conv2 = nn.Conv2D(planes, planes, 3, padding=1, bias_attr=False)
        self.bn2 = norm_layer(planes)
        self.downsample = downsample
        self.stride = stride

    def forward(self, x):
        identity = x

        out = self.conv1(x)
        out = self.bn1(out)
        out = self.relu(out)

        out = self.conv2(out)
        out = self.bn2(out)

        if self.downsample is not None:
            identity = self.downsample(x)

        out += identity
        out = self.relu(out)

        return out


class BottleneckBlock(nn.Layer):

    expansion = 4

    def __init__(
        self,
        inplanes,
        planes,
        stride=1,
        downsample=None,
        groups=1,
        base_width=64,
        dilation=1,
        norm_layer=None,
    ):
        super().__init__()
        if norm_layer is None:
            norm_layer = nn.BatchNorm2D
        width = int(planes * (base_width / 64.0)) * groups

        self.conv1 = nn.Conv2D(inplanes, width, 1, bias_attr=False)
        self.bn1 = norm_layer(width)

        self.conv2 = nn.Conv2D(
            width,
            width,
            3,
            padding=dilation,
            stride=stride,
            groups=groups,
            dilation=dilation,
            bias_attr=False,
        )
        self.bn2 = norm_layer(width)

        self.conv3 = nn.Conv2D(
            width, planes * self.expansion, 1, bias_attr=False
        )
        self.bn3 = norm_layer(planes * self.expansion)
        self.relu = nn.ReLU()
        self.downsample = downsample
        self.stride = stride

    def forward(self, x):
        identity = x

        out = self.conv1(x)
        out = self.bn1(out)
        out = self.relu(out)

        out = self.conv2(out)
        out = self.bn2(out)
        out = self.relu(out)

        out = self.conv3(out)
        out = self.bn3(out)

        if self.downsample is not None:
            identity = self.downsample(x)

        out += identity
        out = self.relu(out)

        return out


class ResNet(nn.Layer):
    """ResNet model from
    `"Deep Residual Learning for Image Recognition" <https://arxiv.org/pdf/1512.03385.pdf>`_.
    Args:
        Block (BasicBlock|BottleneckBlock): Block module of model.
        depth (int, optional): Layers of ResNet, Default: 50.
        width (int, optional): Base width per convolution group for each convolution block, Default: 64.
        num_classes (int, optional): Output num_features of last fc layer. If num_classes <= 0, last fc layer
                            will not be defined. Default: 1000.
        with_pool (bool, optional): Use pool before the last fc layer or not. Default: True.
        groups (int, optional): Number of groups for each convolution block, Default: 1.
    Returns:
        :ref:`api_paddle_nn_Layer`. An instance of ResNet model.
    Examples:
        .. code-block:: python
            import paddle
            from paddle.vision.models import ResNet
            from paddle.vision.models.resnet import BottleneckBlock, BasicBlock
            # build ResNet with 18 layers
            resnet18 = ResNet(BasicBlock, 18)
            # build ResNet with 50 layers
            resnet50 = ResNet(BottleneckBlock, 50)
            # build Wide ResNet model
            wide_resnet50_2 = ResNet(BottleneckBlock, 50, width=64*2)
            # build ResNeXt model
            resnext50_32x4d = ResNet(BottleneckBlock, 50, width=4, groups=32)
            x = paddle.rand([1, 3, 224, 224])
            out = resnet18(x)
            print(out.shape)
            # [1, 1000]
    """

    def __init__(
        self,
        block,
        parc_insert_locs=[3, 4, 3, 2],
        use_parc = [False, False, True, True],
        global_kernel_size = [32, 16, 8, 4],
        depth=50,
        width=64,
        num_classes=1000,
        with_pool=True,
        groups=1,
    ):
        super().__init__()
        layer_cfg = {
            18: [2, 2, 2, 2],
            34: [3, 4, 6, 3],
            50: [3, 4, 6, 3],
            101: [3, 4, 23, 3],
            152: [3, 8, 36, 3],
        }
        layers = layer_cfg[depth]
        self.groups = groups
        self.base_width = width
        self.num_classes = num_classes
        self.with_pool = with_pool
        self.parc_insert_locs = parc_insert_locs
        self._norm_layer = nn.BatchNorm2D

        self.inplanes = 64
        self.dilation = 1

        self.conv1 = nn.Conv2D(
            3,
            self.inplanes,
            kernel_size=3,
            stride=1,
            padding=1,
            bias_attr=False,
        )
        self.bn1 = self._norm_layer(self.inplanes)
        self.relu = nn.ReLU()
        self.layer1 = self._make_layer(block, 64, layers[0], use_parc=use_parc[0], 
                            parc_insert_locs=parc_insert_locs[0], global_kernel_size=global_kernel_size[0])
        self.layer2 = self._make_layer(block, 128, layers[1], stride=2, use_parc=use_parc[1],
                            parc_insert_locs=parc_insert_locs[1], global_kernel_size=global_kernel_size[1])
        self.layer3 = self._make_layer(block, 256, layers[2], stride=2, use_parc=use_parc[2], 
                            parc_insert_locs=parc_insert_locs[2], global_kernel_size=global_kernel_size[2])
        self.layer4 = self._make_layer(block, 512, layers[3], stride=2, use_parc=use_parc[3], 
                            parc_insert_locs=parc_insert_locs[3], global_kernel_size=global_kernel_size[3])
        if with_pool:
            self.avgpool = nn.AdaptiveAvgPool2D((1, 1))

        if num_classes > 0:
            self.fc = nn.Linear(512 * block.expansion, num_classes)

    def _make_layer(self, block, planes, blocks, stride=1, dilate=False, use_parc=False, parc_insert_locs=-1, global_kernel_size=14):
        norm_layer = self._norm_layer
        downsample = None
        previous_dilation = self.dilation
        if dilate:
            self.dilation *= stride
            stride = 1
        if stride != 1 or self.inplanes != planes * block.expansion:
            downsample = nn.Sequential(
                nn.Conv2D(
                    self.inplanes,
                    planes * block.expansion,
                    1,
                    stride=stride,
                    bias_attr=False,
                ),
                norm_layer(planes * block.expansion),
            )

        layers = []
        layers.append(
            block(
                self.inplanes,
                planes,
                stride,
                downsample,
                self.groups,
                self.base_width,
                previous_dilation,
                norm_layer,
            )
        )
        self.inplanes = planes * block.expansion
        for i in range(1, blocks):
            if use_parc:
                if i >= parc_insert_locs:
                    layers.append(
                        ParCBlock(
                            self.inplanes,
                            planes,
                            groups=self.groups,
                            base_width=self.base_width,
                            norm_layer=norm_layer,
                            use_pe = True,
                            global_kernel_size = global_kernel_size
                        )
                    )
                    continue

            layers.append(
                block(
                    self.inplanes,
                    planes,
                    groups=self.groups,
                    base_width=self.base_width,
                    norm_layer=norm_layer,
                )
            )

        return nn.Sequential(*layers)

    def forward(self, x):
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.relu(x)

        x = self.layer1(x)
        x = self.layer2(x)
        x = self.layer3(x)
        x = self.layer4(x)

        if self.with_pool:
            x = self.avgpool(x)

        if self.num_classes > 0:
            x = paddle.flatten(x, 1)
            x = self.fc(x)

        return x

model = ResNet(BasicBlock, depth=18, parc_insert_locs=[2, 2, 1, 1], num_classes=10)
paddle.summary(model, (1, 3, 32, 32))

2.5 训练

learning_rate = 0.1
n_epochs = 100
paddle.seed(42)
np.random.seed(42)

work_path = 'work/model'

model = ResNet(BasicBlock, depth=18, parc_insert_locs=[2, 2, 1, 1], num_classes=10)

criterion = LabelSmoothingCrossEntropy()

scheduler = paddle.optimizer.lr.MultiStepDecay(learning_rate=learning_rate, milestones=[30, 60, 80], verbose=False)
optimizer = paddle.optimizer.Momentum(parameters=model.parameters(), learning_rate=scheduler, weight_decay=1e-5)

gate = 0.0
threshold = 0.0
best_acc = 0.0
val_acc = 0.0
loss_record = {'train': {'loss': [], 'iter': []}, 'val': {'loss': [], 'iter': []}}   # for recording loss
acc_record = {'train': {'acc': [], 'iter': []}, 'val': {'acc': [], 'iter': []}}      # for recording accuracy

loss_iter = 0
acc_iter = 0

for epoch in range(n_epochs):
    # ---------- Training ----------
    model.train()
    train_num = 0.0
    train_loss = 0.0

    val_num = 0.0
    val_loss = 0.0
    accuracy_manager = paddle.metric.Accuracy()
    val_accuracy_manager = paddle.metric.Accuracy()
    print("#===epoch: {}, lr={:.10f}===#".format(epoch, optimizer.get_lr()))
    for batch_id, data in enumerate(train_loader):
        x_data, y_data = data
        labels = paddle.unsqueeze(y_data, axis=1)

        logits = model(x_data)

        loss = criterion(logits, y_data)

        acc = paddle.metric.accuracy(logits, labels)
        accuracy_manager.update(acc)
        if batch_id % 10 == 0:
            loss_record['train']['loss'].append(loss.numpy())
            loss_record['train']['iter'].append(loss_iter)
            loss_iter += 1

        loss.backward()

        optimizer.step()
        optimizer.clear_grad()
        
        train_loss += loss
        train_num += len(y_data)
    scheduler.step()

    total_train_loss = (train_loss / train_num) * batch_size
    train_acc = accuracy_manager.accumulate()
    acc_record['train']['acc'].append(train_acc)
    acc_record['train']['iter'].append(acc_iter)
    acc_iter += 1
    # Print the information.
    print("#===epoch: {}, train loss is: {}, train acc is: {:2.2f}%===#".format(epoch, total_train_loss.numpy(), train_acc*100))

    # ---------- Validation ----------
    model.eval()

    for batch_id, data in enumerate(val_loader):

        x_data, y_data = data
        labels = paddle.unsqueeze(y_data, axis=1)
        with paddle.no_grad():
          logits = model(x_data)

        loss = criterion(logits, y_data)

        acc = paddle.metric.accuracy(logits, labels)
        val_accuracy_manager.update(acc)

        val_loss += loss
        val_num += len(y_data)

    total_val_loss = (val_loss / val_num) * batch_size
    loss_record['val']['loss'].append(total_val_loss.numpy())
    loss_record['val']['iter'].append(loss_iter)
    val_acc = val_accuracy_manager.accumulate()
    acc_record['val']['acc'].append(val_acc)
    acc_record['val']['iter'].append(acc_iter)
    
    print("#===epoch: {}, val loss is: {}, val acc is: {:2.2f}%===#".format(epoch, total_val_loss.numpy(), val_acc*100))

    # ===================save====================
    if val_acc > best_acc:
        best_acc = val_acc
        paddle.save(model.state_dict(), os.path.join(work_path, 'best_model.pdparams'))
        paddle.save(optimizer.state_dict(), os.path.join(work_path, 'best_optimizer.pdopt'))

print(best_acc)
paddle.save(model.state_dict(), os.path.join(work_path, 'final_model.pdparams'))
paddle.save(optimizer.state_dict(), os.path.join(work_path, 'final_optimizer.pdopt'))

2.6 实验结果

def plot_learning_curve(record, title='loss', ylabel='CE Loss'):
    ''' Plot learning curve of your CNN '''
    maxtrain = max(map(float, record['train'][title]))
    maxval = max(map(float, record['val'][title]))
    ymax = max(maxtrain, maxval) * 1.1
    mintrain = min(map(float, record['train'][title]))
    minval = min(map(float, record['val'][title]))
    ymin = min(mintrain, minval) * 0.9

    total_steps = len(record['train'][title])
    x_1 = list(map(int, record['train']['iter']))
    x_2 = list(map(int, record['val']['iter']))
    figure(figsize=(10, 6))
    plt.plot(x_1, record['train'][title], c='tab:red', label='train')
    plt.plot(x_2, record['val'][title], c='tab:cyan', label='val')
    plt.ylim(ymin, ymax)
    plt.xlabel('Training steps')
    plt.ylabel(ylabel)
    plt.title('Learning curve of {}'.format(title))
    plt.legend()
    plt.show()

plot_learning_curve(loss_record, title='loss', ylabel='CE Loss')

/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/matplotlib/cbook/__init__.py:2349: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
  if isinstance(obj, collections.Iterator):
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/matplotlib/cbook/__init__.py:2366: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
  return list(data) if isinstance(data, collections.MappingView) else data

在这里插入图片描述

plot_learning_curve(acc_record, title='acc', ylabel='Accuracy')

在这里插入图片描述

import time
work_path = 'work/model'
model = ResNet(BasicBlock, depth=18, parc_insert_locs=[2, 2, 1, 1], num_classes=10)
model_state_dict = paddle.load(os.path.join(work_path, 'best_model.pdparams'))
model.set_state_dict(model_state_dict)
model.eval()
aa = time.time()
for batch_id, data in enumerate(val_loader):

    x_data, y_data = data
    labels = paddle.unsqueeze(y_data, axis=1)
    with paddle.no_grad():
        logits = model(x_data)
bb = time.time()
print("Throughout:{}".format(int(len(val_dataset)//(bb - aa))))

Throughout:3806

3. ResNet

3.1 ResNet

model = paddle.vision.models.resnet18(num_classes=10)
model.conv1 = nn.Conv2D(3, 64, 3, padding=1, bias_attr=False)
model.maxpool = nn.Identity()
paddle.summary(model, (1, 3, 128, 128))

3.2 训练

learning_rate = 0.1
n_epochs = 100
paddle.seed(42)
np.random.seed(42)

work_path = 'work/model1'

model = paddle.vision.models.resnet18(num_classes=10)
model.conv1 = nn.Conv2D(3, 64, 3, padding=1, bias_attr=False)
model.maxpool = nn.Identity()

criterion = LabelSmoothingCrossEntropy()

scheduler = paddle.optimizer.lr.MultiStepDecay(learning_rate=learning_rate, milestones=[30, 60, 80], verbose=False)
optimizer = paddle.optimizer.Momentum(parameters=model.parameters(), learning_rate=scheduler, weight_decay=1e-5)

gate = 0.0
threshold = 0.0
best_acc = 0.0
val_acc = 0.0
loss_record1 = {'train': {'loss': [], 'iter': []}, 'val': {'loss': [], 'iter': []}}   # for recording loss
acc_record1 = {'train': {'acc': [], 'iter': []}, 'val': {'acc': [], 'iter': []}}      # for recording accuracy

loss_iter = 0
acc_iter = 0

for epoch in range(n_epochs):
    # ---------- Training ----------
    model.train()
    train_num = 0.0
    train_loss = 0.0

    val_num = 0.0
    val_loss = 0.0
    accuracy_manager = paddle.metric.Accuracy()
    val_accuracy_manager = paddle.metric.Accuracy()
    print("#===epoch: {}, lr={:.10f}===#".format(epoch, optimizer.get_lr()))
    for batch_id, data in enumerate(train_loader):
        x_data, y_data = data
        labels = paddle.unsqueeze(y_data, axis=1)

        logits = model(x_data)

        loss = criterion(logits, y_data)

        acc = paddle.metric.accuracy(logits, labels)
        accuracy_manager.update(acc)
        if batch_id % 10 == 0:
            loss_record1['train']['loss'].append(loss.numpy())
            loss_record1['train']['iter'].append(loss_iter)
            loss_iter += 1

        loss.backward()

        optimizer.step()
        optimizer.clear_grad()
        
        train_loss += loss
        train_num += len(y_data)
    scheduler.step()

    total_train_loss = (train_loss / train_num) * batch_size
    train_acc = accuracy_manager.accumulate()
    acc_record1['train']['acc'].append(train_acc)
    acc_record1['train']['iter'].append(acc_iter)
    acc_iter += 1
    # Print the information.
    print("#===epoch: {}, train loss is: {}, train acc is: {:2.2f}%===#".format(epoch, total_train_loss.numpy(), train_acc*100))

    # ---------- Validation ----------
    model.eval()

    for batch_id, data in enumerate(val_loader):

        x_data, y_data = data
        labels = paddle.unsqueeze(y_data, axis=1)
        with paddle.no_grad():
          logits = model(x_data)

        loss = criterion(logits, y_data)

        acc = paddle.metric.accuracy(logits, labels)
        val_accuracy_manager.update(acc)

        val_loss += loss
        val_num += len(y_data)

    total_val_loss = (val_loss / val_num) * batch_size
    loss_record1['val']['loss'].append(total_val_loss.numpy())
    loss_record1['val']['iter'].append(loss_iter)
    val_acc = val_accuracy_manager.accumulate()
    acc_record1['val']['acc'].append(val_acc)
    acc_record1['val']['iter'].append(acc_iter)
    
    print("#===epoch: {}, val loss is: {}, val acc is: {:2.2f}%===#".format(epoch, total_val_loss.numpy(), val_acc*100))

    # ===================save====================
    if val_acc > best_acc:
        best_acc = val_acc
        paddle.save(model.state_dict(), os.path.join(work_path, 'best_model.pdparams'))
        paddle.save(optimizer.state_dict(), os.path.join(work_path, 'best_optimizer.pdopt'))

print(best_acc)
paddle.save(model.state_dict(), os.path.join(work_path, 'final_model.pdparams'))
paddle.save(optimizer.state_dict(), os.path.join(work_path, 'final_optimizer.pdopt'))

3.3 实验结果

plot_learning_curve(loss_record1, title='loss', ylabel='CE Loss')

在这里插入图片描述

plot_learning_curve(acc_record1, title='acc', ylabel='Accuracy')

在这里插入图片描述

##### import time
work_path = 'work/model1'
model = paddle.vision.models.resnet18(num_classes=10)
model.conv1 = nn.Conv2D(3, 64, 3, padding=1, bias_attr=False)
model.maxpool = nn.Identity()
model_state_dict = paddle.load(os.path.join(work_path, 'best_model.pdparams'))
model.set_state_dict(model_state_dict)
model.eval()
aa = time.time()
for batch_id, data in enumerate(val_loader):

    x_data, y_data = data
    labels = paddle.unsqueeze(y_data, axis=1)
    with paddle.no_grad():
        logits = model(x_data)
bb = time.time()
print("Throughout:{}".format(int(len(val_dataset)//(bb - aa))))