[toc]

『深度学习 7 日打卡营·快速入门特辑』

零基础解锁深度学习神器飞桨框架高层 API，七天时间助你掌握 CV、NLP 领域最火模型及应用。

  1. 课程地址

传送门：https://aistudio.baidu.com/aistudio/course/introduce/6771

  1. 目标
  • 掌握深度学习常用模型基础知识
  • 熟练掌握一种国产开源深度学习框架
  • 具备独立完成相关深度学习任务的能力
  • 能用所学为 AI 加一份年味

① 问题定义

十二生肖分类的本质是图像分类任务，我们采用 CNN 网络结构进行相关实践。

② 数据准备

2.1 解压缩数据集

我们将网上获取的数据集以压缩包的方式上传到 aistudio 数据集中，并加载到我们的项目内。

在使用之前我们进行数据集压缩包的一个解压。

1
!unzip -q -o data/data68755/signs.zip

2.2 数据标注

我们先看一下解压缩后的数据集长成什么样子。

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
.
├── test
│   ├── dog
│   ├── dragon
│   ├── goat
│   ├── horse
│   ├── monkey
│   ├── ox
│   ├── pig
│   ├── rabbit
│   ├── ratt
│   ├── rooster
│   ├── snake
│   └── tiger
├── train
│   ├── dog
│   ├── dragon
│   ├── goat
│   ├── horse
│   ├── monkey
│   ├── ox
│   ├── pig
│   ├── rabbit
│   ├── ratt
│   ├── rooster
│   ├── snake
│   └── tiger
└── valid
    ├── dog
    ├── dragon
    ├── goat
    ├── horse
    ├── monkey
    ├── ox
    ├── pig
    ├── rabbit
    ├── ratt
    ├── rooster
    ├── snake
    └── tiger

数据集分为 train、valid、test 三个文件夹，每个文件夹内包含 12 个分类文件夹，每个分类文件夹内是具体的样本图片。

我们对这些样本进行一个标注处理，最终生成 train.txt/valid.txt/test.txt 三个数据标注文件。

1
2
# %cd work
!ls
1
1512224.ipynb  config.py  data	dataset.py  __MACOSX  __pycache__  signs  work
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
import io
import os
from PIL import Image
from config import get


# 数据集根目录
DATA_ROOT = 'signs'

# 标签List
LABEL_MAP = get('LABEL_MAP')

# 标注生成函数
def generate_annotation(mode):
    # 建立标注文件
    with open('{}/{}.txt'.format(DATA_ROOT, mode), 'w') as f:
        # 对应每个用途的数据文件夹，train/valid/test
        train_dir = '{}/{}'.format(DATA_ROOT, mode)

        # 遍历文件夹，获取里面的分类文件夹
        for path in os.listdir(train_dir):
            # 标签对应的数字索引，实际标注的时候直接使用数字索引
            label_index = LABEL_MAP.index(path)

            # 图像样本所在的路径
            image_path = '{}/{}'.format(train_dir, path)

            # 遍历所有图像
            for image in os.listdir(image_path):
                # 图像完整路径和名称
                image_file = '{}/{}'.format(image_path, image)

                try:
                    # 验证图片格式是否ok
                    with open(image_file, 'rb') as f_img:
                        image = Image.open(io.BytesIO(f_img.read()))
                        image.load()

                        if image.mode == 'RGB':
                            f.write('{}\t{}\n'.format(image_file, label_index))
                except:
                    continue


generate_annotation('train')  # 生成训练集标注文件
generate_annotation('valid')  # 生成验证集标注文件
generate_annotation('test')   # 生成测试集标注文件

2.3 数据集定义

接下来我们使用标注好的文件进行数据集类的定义，方便后续模型训练使用。

2.3.1 导入相关库

1
2
3
4
5
import paddle
import numpy as np
from config import get

paddle.__version__
1
'2.0.0'

2.3.2 导入数据集的定义实现

我们数据集的代码实现是在 dataset.py 中。

数据增强data_augumentation为：

1
2
3
4
5
6
    self.transforms = T.Compose([
    T.RandomResizedCrop(IMAGE_SIZE),    # 随机裁剪大小
    T.RandomHorizontalFlip(0.5),        # 随机水平翻转
    T.ToTensor(),                       # 数据的格式转换和标准化 HWC => CHW
    T.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])  # 图像归一化
])
1
from dataset import ZodiacDataset

2.3.3 实例化数据集类

根据所使用的数据集需求实例化数据集类，并查看总样本量。

1
2
3
4
train_dataset = ZodiacDataset(mode='train')
valid_dataset = ZodiacDataset(mode='valid')

print('训练数据集：{}张；验证数据集：{}张'.format(len(train_dataset), len(valid_dataset)))
1
训练数据集：7096张；验证数据集：639张

③ 模型选择和开发

3.1 网络构建

本次我们使用 ResNet50 网络来完成我们的案例实践。

1）ResNet 系列网络

2）ResNet50 结构

3）残差区块

4）ResNet 其他版本

1
2
3
# 请补齐模型实例化代码

network = paddle.vision.models.resnet50(num_classes=get('num_classes'), pretrained=True)
1
2
3
4
5
100%|██████████| 151272/151272 [00:03<00:00, 41104.37it/s]
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/dygraph/layers.py:1263: UserWarning: Skip loading for fc.weight. fc.weight receives a shape [2048, 1000], but the expected shape is [2048, 12].
  warnings.warn(("Skip loading for {}. ".format(key) + str(err)))
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/dygraph/layers.py:1263: UserWarning: Skip loading for fc.bias. fc.bias receives a shape [1000], but the expected shape is [12].
  warnings.warn(("Skip loading for {}. ".format(key) + str(err)))

模型可视化

1
2
model = paddle.Model(network)
model.summary((-1, ) + tuple(get('image_shape')))
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
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
-------------------------------------------------------------------------------
   Layer (type)         Input Shape          Output Shape         Param #
===============================================================================
     Conv2D-1        [[1, 3, 224, 224]]   [1, 64, 112, 112]        9,408
   BatchNorm2D-1    [[1, 64, 112, 112]]   [1, 64, 112, 112]         256
      ReLU-1        [[1, 64, 112, 112]]   [1, 64, 112, 112]          0
    MaxPool2D-1     [[1, 64, 112, 112]]    [1, 64, 56, 56]           0
     Conv2D-3        [[1, 64, 56, 56]]     [1, 64, 56, 56]         4,096
   BatchNorm2D-3     [[1, 64, 56, 56]]     [1, 64, 56, 56]          256
      ReLU-2         [[1, 256, 56, 56]]    [1, 256, 56, 56]          0
     Conv2D-4        [[1, 64, 56, 56]]     [1, 64, 56, 56]        36,864
   BatchNorm2D-4     [[1, 64, 56, 56]]     [1, 64, 56, 56]          256
     Conv2D-5        [[1, 64, 56, 56]]     [1, 256, 56, 56]       16,384
   BatchNorm2D-5     [[1, 256, 56, 56]]    [1, 256, 56, 56]        1,024
     Conv2D-2        [[1, 64, 56, 56]]     [1, 256, 56, 56]       16,384
   BatchNorm2D-2     [[1, 256, 56, 56]]    [1, 256, 56, 56]        1,024
 BottleneckBlock-1   [[1, 64, 56, 56]]     [1, 256, 56, 56]          0
     Conv2D-6        [[1, 256, 56, 56]]    [1, 64, 56, 56]        16,384
   BatchNorm2D-6     [[1, 64, 56, 56]]     [1, 64, 56, 56]          256
      ReLU-3         [[1, 256, 56, 56]]    [1, 256, 56, 56]          0
     Conv2D-7        [[1, 64, 56, 56]]     [1, 64, 56, 56]        36,864
   BatchNorm2D-7     [[1, 64, 56, 56]]     [1, 64, 56, 56]          256
     Conv2D-8        [[1, 64, 56, 56]]     [1, 256, 56, 56]       16,384
   BatchNorm2D-8     [[1, 256, 56, 56]]    [1, 256, 56, 56]        1,024
 BottleneckBlock-2   [[1, 256, 56, 56]]    [1, 256, 56, 56]          0
     Conv2D-9        [[1, 256, 56, 56]]    [1, 64, 56, 56]        16,384
   BatchNorm2D-9     [[1, 64, 56, 56]]     [1, 64, 56, 56]          256
      ReLU-4         [[1, 256, 56, 56]]    [1, 256, 56, 56]          0
     Conv2D-10       [[1, 64, 56, 56]]     [1, 64, 56, 56]        36,864
  BatchNorm2D-10     [[1, 64, 56, 56]]     [1, 64, 56, 56]          256
     Conv2D-11       [[1, 64, 56, 56]]     [1, 256, 56, 56]       16,384
  BatchNorm2D-11     [[1, 256, 56, 56]]    [1, 256, 56, 56]        1,024
 BottleneckBlock-3   [[1, 256, 56, 56]]    [1, 256, 56, 56]          0
     Conv2D-13       [[1, 256, 56, 56]]    [1, 128, 56, 56]       32,768
  BatchNorm2D-13     [[1, 128, 56, 56]]    [1, 128, 56, 56]         512
      ReLU-5         [[1, 512, 28, 28]]    [1, 512, 28, 28]          0
     Conv2D-14       [[1, 128, 56, 56]]    [1, 128, 28, 28]       147,456
  BatchNorm2D-14     [[1, 128, 28, 28]]    [1, 128, 28, 28]         512
     Conv2D-15       [[1, 128, 28, 28]]    [1, 512, 28, 28]       65,536
  BatchNorm2D-15     [[1, 512, 28, 28]]    [1, 512, 28, 28]        2,048
     Conv2D-12       [[1, 256, 56, 56]]    [1, 512, 28, 28]       131,072
  BatchNorm2D-12     [[1, 512, 28, 28]]    [1, 512, 28, 28]        2,048
 BottleneckBlock-4   [[1, 256, 56, 56]]    [1, 512, 28, 28]          0
     Conv2D-16       [[1, 512, 28, 28]]    [1, 128, 28, 28]       65,536
  BatchNorm2D-16     [[1, 128, 28, 28]]    [1, 128, 28, 28]         512
      ReLU-6         [[1, 512, 28, 28]]    [1, 512, 28, 28]          0
     Conv2D-17       [[1, 128, 28, 28]]    [1, 128, 28, 28]       147,456
  BatchNorm2D-17     [[1, 128, 28, 28]]    [1, 128, 28, 28]         512
     Conv2D-18       [[1, 128, 28, 28]]    [1, 512, 28, 28]       65,536
  BatchNorm2D-18     [[1, 512, 28, 28]]    [1, 512, 28, 28]        2,048
 BottleneckBlock-5   [[1, 512, 28, 28]]    [1, 512, 28, 28]          0
     Conv2D-19       [[1, 512, 28, 28]]    [1, 128, 28, 28]       65,536
  BatchNorm2D-19     [[1, 128, 28, 28]]    [1, 128, 28, 28]         512
      ReLU-7         [[1, 512, 28, 28]]    [1, 512, 28, 28]          0
     Conv2D-20       [[1, 128, 28, 28]]    [1, 128, 28, 28]       147,456
  BatchNorm2D-20     [[1, 128, 28, 28]]    [1, 128, 28, 28]         512
     Conv2D-21       [[1, 128, 28, 28]]    [1, 512, 28, 28]       65,536
  BatchNorm2D-21     [[1, 512, 28, 28]]    [1, 512, 28, 28]        2,048
 BottleneckBlock-6   [[1, 512, 28, 28]]    [1, 512, 28, 28]          0
     Conv2D-22       [[1, 512, 28, 28]]    [1, 128, 28, 28]       65,536
  BatchNorm2D-22     [[1, 128, 28, 28]]    [1, 128, 28, 28]         512
      ReLU-8         [[1, 512, 28, 28]]    [1, 512, 28, 28]          0
     Conv2D-23       [[1, 128, 28, 28]]    [1, 128, 28, 28]       147,456
  BatchNorm2D-23     [[1, 128, 28, 28]]    [1, 128, 28, 28]         512
     Conv2D-24       [[1, 128, 28, 28]]    [1, 512, 28, 28]       65,536
  BatchNorm2D-24     [[1, 512, 28, 28]]    [1, 512, 28, 28]        2,048
 BottleneckBlock-7   [[1, 512, 28, 28]]    [1, 512, 28, 28]          0
     Conv2D-26       [[1, 512, 28, 28]]    [1, 256, 28, 28]       131,072
  BatchNorm2D-26     [[1, 256, 28, 28]]    [1, 256, 28, 28]        1,024
      ReLU-9        [[1, 1024, 14, 14]]   [1, 1024, 14, 14]          0
     Conv2D-27       [[1, 256, 28, 28]]    [1, 256, 14, 14]       589,824
  BatchNorm2D-27     [[1, 256, 14, 14]]    [1, 256, 14, 14]        1,024
     Conv2D-28       [[1, 256, 14, 14]]   [1, 1024, 14, 14]       262,144
  BatchNorm2D-28    [[1, 1024, 14, 14]]   [1, 1024, 14, 14]        4,096
     Conv2D-25       [[1, 512, 28, 28]]   [1, 1024, 14, 14]       524,288
  BatchNorm2D-25    [[1, 1024, 14, 14]]   [1, 1024, 14, 14]        4,096
 BottleneckBlock-8   [[1, 512, 28, 28]]   [1, 1024, 14, 14]          0
     Conv2D-29      [[1, 1024, 14, 14]]    [1, 256, 14, 14]       262,144
  BatchNorm2D-29     [[1, 256, 14, 14]]    [1, 256, 14, 14]        1,024
      ReLU-10       [[1, 1024, 14, 14]]   [1, 1024, 14, 14]          0
     Conv2D-30       [[1, 256, 14, 14]]    [1, 256, 14, 14]       589,824
  BatchNorm2D-30     [[1, 256, 14, 14]]    [1, 256, 14, 14]        1,024
     Conv2D-31       [[1, 256, 14, 14]]   [1, 1024, 14, 14]       262,144
  BatchNorm2D-31    [[1, 1024, 14, 14]]   [1, 1024, 14, 14]        4,096
 BottleneckBlock-9  [[1, 1024, 14, 14]]   [1, 1024, 14, 14]          0
     Conv2D-32      [[1, 1024, 14, 14]]    [1, 256, 14, 14]       262,144
  BatchNorm2D-32     [[1, 256, 14, 14]]    [1, 256, 14, 14]        1,024
      ReLU-11       [[1, 1024, 14, 14]]   [1, 1024, 14, 14]          0
     Conv2D-33       [[1, 256, 14, 14]]    [1, 256, 14, 14]       589,824
  BatchNorm2D-33     [[1, 256, 14, 14]]    [1, 256, 14, 14]        1,024
     Conv2D-34       [[1, 256, 14, 14]]   [1, 1024, 14, 14]       262,144
  BatchNorm2D-34    [[1, 1024, 14, 14]]   [1, 1024, 14, 14]        4,096
BottleneckBlock-10  [[1, 1024, 14, 14]]   [1, 1024, 14, 14]          0
     Conv2D-35      [[1, 1024, 14, 14]]    [1, 256, 14, 14]       262,144
  BatchNorm2D-35     [[1, 256, 14, 14]]    [1, 256, 14, 14]        1,024
      ReLU-12       [[1, 1024, 14, 14]]   [1, 1024, 14, 14]          0
     Conv2D-36       [[1, 256, 14, 14]]    [1, 256, 14, 14]       589,824
  BatchNorm2D-36     [[1, 256, 14, 14]]    [1, 256, 14, 14]        1,024
     Conv2D-37       [[1, 256, 14, 14]]   [1, 1024, 14, 14]       262,144
  BatchNorm2D-37    [[1, 1024, 14, 14]]   [1, 1024, 14, 14]        4,096
BottleneckBlock-11  [[1, 1024, 14, 14]]   [1, 1024, 14, 14]          0
     Conv2D-38      [[1, 1024, 14, 14]]    [1, 256, 14, 14]       262,144
  BatchNorm2D-38     [[1, 256, 14, 14]]    [1, 256, 14, 14]        1,024
      ReLU-13       [[1, 1024, 14, 14]]   [1, 1024, 14, 14]          0
     Conv2D-39       [[1, 256, 14, 14]]    [1, 256, 14, 14]       589,824
  BatchNorm2D-39     [[1, 256, 14, 14]]    [1, 256, 14, 14]        1,024
     Conv2D-40       [[1, 256, 14, 14]]   [1, 1024, 14, 14]       262,144
  BatchNorm2D-40    [[1, 1024, 14, 14]]   [1, 1024, 14, 14]        4,096
BottleneckBlock-12  [[1, 1024, 14, 14]]   [1, 1024, 14, 14]          0
     Conv2D-41      [[1, 1024, 14, 14]]    [1, 256, 14, 14]       262,144
  BatchNorm2D-41     [[1, 256, 14, 14]]    [1, 256, 14, 14]        1,024
      ReLU-14       [[1, 1024, 14, 14]]   [1, 1024, 14, 14]          0
     Conv2D-42       [[1, 256, 14, 14]]    [1, 256, 14, 14]       589,824
  BatchNorm2D-42     [[1, 256, 14, 14]]    [1, 256, 14, 14]        1,024
     Conv2D-43       [[1, 256, 14, 14]]   [1, 1024, 14, 14]       262,144
  BatchNorm2D-43    [[1, 1024, 14, 14]]   [1, 1024, 14, 14]        4,096
BottleneckBlock-13  [[1, 1024, 14, 14]]   [1, 1024, 14, 14]          0
     Conv2D-45      [[1, 1024, 14, 14]]    [1, 512, 14, 14]       524,288
  BatchNorm2D-45     [[1, 512, 14, 14]]    [1, 512, 14, 14]        2,048
      ReLU-15        [[1, 2048, 7, 7]]     [1, 2048, 7, 7]           0
     Conv2D-46       [[1, 512, 14, 14]]     [1, 512, 7, 7]       2,359,296
  BatchNorm2D-46      [[1, 512, 7, 7]]      [1, 512, 7, 7]         2,048
     Conv2D-47        [[1, 512, 7, 7]]     [1, 2048, 7, 7]       1,048,576
  BatchNorm2D-47     [[1, 2048, 7, 7]]     [1, 2048, 7, 7]         8,192
     Conv2D-44      [[1, 1024, 14, 14]]    [1, 2048, 7, 7]       2,097,152
  BatchNorm2D-44     [[1, 2048, 7, 7]]     [1, 2048, 7, 7]         8,192
BottleneckBlock-14  [[1, 1024, 14, 14]]    [1, 2048, 7, 7]           0
     Conv2D-48       [[1, 2048, 7, 7]]      [1, 512, 7, 7]       1,048,576
  BatchNorm2D-48      [[1, 512, 7, 7]]      [1, 512, 7, 7]         2,048
      ReLU-16        [[1, 2048, 7, 7]]     [1, 2048, 7, 7]           0
     Conv2D-49        [[1, 512, 7, 7]]      [1, 512, 7, 7]       2,359,296
  BatchNorm2D-49      [[1, 512, 7, 7]]      [1, 512, 7, 7]         2,048
     Conv2D-50        [[1, 512, 7, 7]]     [1, 2048, 7, 7]       1,048,576
  BatchNorm2D-50     [[1, 2048, 7, 7]]     [1, 2048, 7, 7]         8,192
BottleneckBlock-15   [[1, 2048, 7, 7]]     [1, 2048, 7, 7]           0
     Conv2D-51       [[1, 2048, 7, 7]]      [1, 512, 7, 7]       1,048,576
  BatchNorm2D-51      [[1, 512, 7, 7]]      [1, 512, 7, 7]         2,048
      ReLU-17        [[1, 2048, 7, 7]]     [1, 2048, 7, 7]           0
     Conv2D-52        [[1, 512, 7, 7]]      [1, 512, 7, 7]       2,359,296
  BatchNorm2D-52      [[1, 512, 7, 7]]      [1, 512, 7, 7]         2,048
     Conv2D-53        [[1, 512, 7, 7]]     [1, 2048, 7, 7]       1,048,576
  BatchNorm2D-53     [[1, 2048, 7, 7]]     [1, 2048, 7, 7]         8,192
BottleneckBlock-16   [[1, 2048, 7, 7]]     [1, 2048, 7, 7]           0
AdaptiveAvgPool2D-1  [[1, 2048, 7, 7]]     [1, 2048, 1, 1]           0
     Linear-1           [[1, 2048]]            [1, 12]            24,588
===============================================================================
Total params: 23,585,740
Trainable params: 23,479,500
Non-trainable params: 106,240
-------------------------------------------------------------------------------
Input size (MB): 0.57
Forward/backward pass size (MB): 261.48
Params size (MB): 89.97
Estimated Total Size (MB): 352.02
-------------------------------------------------------------------------------






{'total_params': 23585740, 'trainable_params': 23479500}

超参数配置

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
CONFIG = {
'model_save_dir': "./output/zodiac",
'num_classes': 12,
'total_images': 7096,
'epochs': 20,
'batch_size': 32,
'image_shape': [3, 224, 224],
'LEARNING_RATE': {
    'params': {
        'lr': 0.00375
    }
},
'OPTIMIZER': {
    'params': {
        'momentum': 0.9
    },
    'regularizer': {
        'function': 'L2',
        'factor': 0.000001
    }
},
'LABEL_MAP': [
    "ratt",
    "ox",
    "tiger",
    "rabbit",
    "dragon",
    "snake",
    "horse",
    "goat",
    "monkey",
    "rooster",
    "dog",
    "pig",
]
}

④ 模型训练和优化

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
EPOCHS = get('epochs')
BATCH_SIZE = get('batch_size')

# 请补齐模型训练过程代码

def create_optim(parameters):
    step_each_epoch = get('total_images') // get('batch_size')
    lr = paddle.optimizer.lr.CosineAnnealingDecay(learning_rate=get('LEARNING_RATE.params.lr'),
                                                  T_max=step_each_epoch * EPOCHS)

    return paddle.optimizer.Momentum(learning_rate=lr,
                                     parameters=parameters,
                                     weight_decay=paddle.regularizer.L2Decay(get('OPTIMIZER.regularizer.factor')))

# 模型训练配置
model.prepare(create_optim(network.parameters()),  # 优化器
              paddle.nn.CrossEntropyLoss(),        # 损失函数
              paddle.metric.Accuracy(topk=(1, 5))) # 评估指标

# 训练可视化VisualDL工具的回调函数
visualdl = paddle.callbacks.VisualDL(log_dir='visualdl_log')

# 启动模型全流程训练
model.fit(train_dataset,            # 训练数据集
          valid_dataset,            # 评估数据集
          epochs=EPOCHS,            # 总的训练轮次
          batch_size=BATCH_SIZE,    # 批次计算的样本量大小
          shuffle=True,             # 是否打乱样本集
          verbose=1,                # 日志展示格式
          save_dir='./chk_points/', # 分阶段的训练模型存储路径
          callbacks=[visualdl])     # 回调函数使用
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
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
The loss value printed in the log is the current step, and the metric is the average value of previous step.
Epoch 1/20


/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/layers/utils.py:77: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
  return (isinstance(seq, collections.Sequence) and
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/nn/layer/norm.py:636: UserWarning: When training, we now always track global mean and variance.
  "When training, we now always track global mean and variance.")


step 222/222 [==============================] - loss: 0.5499 - acc_top1: 0.7851 - acc_top5: 0.9548 - 734ms/step
save checkpoint at /home/aistudio/chk_points/0
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.3935 - acc_top1: 0.9092 - acc_top5: 0.9922 - 812ms/step
Eval samples: 639
Epoch 2/20
step 222/222 [==============================] - loss: 0.3459 - acc_top1: 0.8519 - acc_top5: 0.9790 - 732ms/step
save checkpoint at /home/aistudio/chk_points/1
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.7412 - acc_top1: 0.8983 - acc_top5: 0.9890 - 833ms/step
Eval samples: 639
Epoch 3/20
step 222/222 [==============================] - loss: 0.4244 - acc_top1: 0.8671 - acc_top5: 0.9817 - 728ms/step
save checkpoint at /home/aistudio/chk_points/2
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.5254 - acc_top1: 0.9218 - acc_top5: 0.9984 - 819ms/step
Eval samples: 639
Epoch 4/20
step 222/222 [==============================] - loss: 0.2774 - acc_top1: 0.8878 - acc_top5: 0.9858 - 738ms/step
save checkpoint at /home/aistudio/chk_points/3
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.3364 - acc_top1: 0.9280 - acc_top5: 0.9969 - 823ms/step
Eval samples: 639
Epoch 5/20
step 222/222 [==============================] - loss: 0.2692 - acc_top1: 0.8922 - acc_top5: 0.9884 - 728ms/step
save checkpoint at /home/aistudio/chk_points/4
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.2550 - acc_top1: 0.9311 - acc_top5: 0.9969 - 804ms/step
Eval samples: 639
Epoch 6/20
step 222/222 [==============================] - loss: 0.5775 - acc_top1: 0.9121 - acc_top5: 0.9894 - 727ms/step
save checkpoint at /home/aistudio/chk_points/5
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.5690 - acc_top1: 0.9531 - acc_top5: 0.9969 - 821ms/step
Eval samples: 639
Epoch 7/20
step 222/222 [==============================] - loss: 0.2791 - acc_top1: 0.9136 - acc_top5: 0.9897 - 731ms/step
save checkpoint at /home/aistudio/chk_points/6
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.5091 - acc_top1: 0.9609 - acc_top5: 0.9969 - 820ms/step
Eval samples: 639
Epoch 8/20
step 222/222 [==============================] - loss: 0.1944 - acc_top1: 0.9253 - acc_top5: 0.9920 - 734ms/step
save checkpoint at /home/aistudio/chk_points/7
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.2852 - acc_top1: 0.9531 - acc_top5: 0.9953 - 821ms/step
Eval samples: 639
Epoch 9/20
step 222/222 [==============================] - loss: 0.3202 - acc_top1: 0.9287 - acc_top5: 0.9927 - 735ms/step
save checkpoint at /home/aistudio/chk_points/8
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.4450 - acc_top1: 0.9531 - acc_top5: 0.9969 - 824ms/step
Eval samples: 639
Epoch 10/20
step 222/222 [==============================] - loss: 0.1250 - acc_top1: 0.9359 - acc_top5: 0.9925 - 745ms/step
save checkpoint at /home/aistudio/chk_points/9
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.4062 - acc_top1: 0.9609 - acc_top5: 0.9953 - 844ms/step
Eval samples: 639
Epoch 11/20
step 222/222 [==============================] - loss: 0.0311 - acc_top1: 0.9366 - acc_top5: 0.9917 - 737ms/step
save checkpoint at /home/aistudio/chk_points/10
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.1610 - acc_top1: 0.9577 - acc_top5: 1.0000 - 831ms/step
Eval samples: 639
Epoch 12/20
step 222/222 [==============================] - loss: 0.2705 - acc_top1: 0.9479 - acc_top5: 0.9938 - 726ms/step
save checkpoint at /home/aistudio/chk_points/11
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.1312 - acc_top1: 0.9577 - acc_top5: 1.0000 - 811ms/step
Eval samples: 639
Epoch 13/20
step 222/222 [==============================] - loss: 0.1966 - acc_top1: 0.9548 - acc_top5: 0.9942 - 729ms/step
save checkpoint at /home/aistudio/chk_points/12
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.3287 - acc_top1: 0.9562 - acc_top5: 1.0000 - 833ms/step
Eval samples: 639
Epoch 14/20
step 222/222 [==============================] - loss: 0.2755 - acc_top1: 0.9563 - acc_top5: 0.9956 - 740ms/step
save checkpoint at /home/aistudio/chk_points/13
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.1597 - acc_top1: 0.9546 - acc_top5: 0.9969 - 811ms/step
Eval samples: 639
Epoch 15/20
step 222/222 [==============================] - loss: 0.1340 - acc_top1: 0.9589 - acc_top5: 0.9945 - 741ms/step
save checkpoint at /home/aistudio/chk_points/14
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.3858 - acc_top1: 0.9531 - acc_top5: 0.9937 - 828ms/step
Eval samples: 639
Epoch 16/20
step 222/222 [==============================] - loss: 0.0603 - acc_top1: 0.9600 - acc_top5: 0.9975 - 741ms/step
save checkpoint at /home/aistudio/chk_points/15
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.1964 - acc_top1: 0.9656 - acc_top5: 0.9984 - 826ms/step
Eval samples: 639
Epoch 17/20
step 222/222 [==============================] - loss: 0.0447 - acc_top1: 0.9581 - acc_top5: 0.9955 - 740ms/step
save checkpoint at /home/aistudio/chk_points/16
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.1808 - acc_top1: 0.9703 - acc_top5: 0.9969 - 818ms/step
Eval samples: 639
Epoch 18/20
step 222/222 [==============================] - loss: 0.5067 - acc_top1: 0.9625 - acc_top5: 0.9956 - 741ms/step
save checkpoint at /home/aistudio/chk_points/17
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.2021 - acc_top1: 0.9656 - acc_top5: 0.9984 - 833ms/step
Eval samples: 639
Epoch 19/20
step 222/222 [==============================] - loss: 0.0673 - acc_top1: 0.9593 - acc_top5: 0.9946 - 738ms/step
save checkpoint at /home/aistudio/chk_points/18
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.5409 - acc_top1: 0.9577 - acc_top5: 0.9969 - 836ms/step
Eval samples: 639
Epoch 20/20
step 222/222 [==============================] - loss: 0.1440 - acc_top1: 0.9596 - acc_top5: 0.9965 - 742ms/step
save checkpoint at /home/aistudio/chk_points/19
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.2487 - acc_top1: 0.9640 - acc_top5: 0.9984 - 809ms/step
Eval samples: 639
save checkpoint at /home/aistudio/chk_points/final

VisualDL 训练过程可视化展示

visualdl = paddle.callbacks.VisualDL(log_dir='VisualDL_log')

模型存储

将我们训练得到的模型进行保存，以便后续评估和测试使用。

1
model.save(get('model_save_dir'))

⑤ 模型评估和测试

5.1 批量预测测试

5.1.1 测试数据集

1
2
predict_dataset = ZodiacDataset(mode='test')
print('测试数据集样本量：{}'.format(len(predict_dataset)))
1
测试数据集样本量：646

5.1.2 执行预测

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
from paddle.static import InputSpec

# 请补充网络结构
network = paddle.vision.models.resnet50(num_classes=get('num_classes'))

# 模型封装
model_2 = paddle.Model(network, inputs=[InputSpec(shape=[-1] + get('image_shape'), dtype='float32', name='image')])

# 请补充模型文件加载代码
# 训练好的模型加载
model_2.load(get('model_save_dir'))

# 模型配置
model_2.prepare()

# 执行预测
result = model_2.predict(predict_dataset)
1
2
3
Predict begin...
step 646/646 [==============================] - 33ms/step
Predict samples: 646
1
2
3
4
5
6
7
8
9
10
11
# 样本映射
LABEL_MAP = get('LABEL_MAP')

# 随机取样本展示
indexs = np.random.randint(1, 646, size=20)

for idx in indexs:
    predict_label = np.argmax(result[0][idx])
    real_label = predict_dataset[idx][1]

    print('样本ID：{}, 真实标签：{}, 预测值：{}'.format(idx, LABEL_MAP[real_label], LABEL_MAP[predict_label]))
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
样本ID：393, 真实标签：pig, 预测值：pig
样本ID：107, 真实标签：monkey, 预测值：monkey
样本ID：466, 真实标签：tiger, 预测值：tiger
样本ID：279, 真实标签：horse, 预测值：horse
样本ID：425, 真实标签：pig, 预测值：pig
样本ID：127, 真实标签：monkey, 预测值：monkey
样本ID：432, 真实标签：tiger, 预测值：tiger
样本ID：377, 真实标签：pig, 预测值：pig
样本ID：99, 真实标签：dragon, 预测值：dragon
样本ID：322, 真实标签：dog, 预测值：dog
样本ID：460, 真实标签：tiger, 预测值：tiger
样本ID：554, 真实标签：ox, 预测值：ox
样本ID：77, 真实标签：dragon, 预测值：dragon
样本ID：335, 真实标签：dog, 预测值：dog
样本ID：398, 真实标签：pig, 预测值：pig
样本ID：176, 真实标签：snake, 预测值：snake
样本ID：207, 真实标签：snake, 预测值：snake
样本ID：29, 真实标签：rooster, 预测值：rooster
样本ID：476, 真实标签：tiger, 预测值：tiger
样本ID：424, 真实标签：pig, 预测值：pig

⑥ 模型部署

1
model_2.save('infer/zodiac', training=False)
1
2
3
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/layers/math_op_patch.py:298: UserWarning: /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/vision/models/resnet.py:145
The behavior of expression A + B has been unified with elementwise_add(X, Y, axis=-1) from Paddle 2.0. If your code works well in the older versions but crashes in this version, try to use elementwise_add(X, Y, axis=0) instead of A + B. This transitional warning will be dropped in the future.
  op_type, op_type, EXPRESSION_MAP[method_name]))

MobileNet_V2 测试

由于ResNet模型和计算量都比较大，在 AI_Studio 的 GPU 下跑 20 个 Epoch 还是跑了很久，这里测试一下用MobileNet来进行训练，测试运算效率和准确性。

1
2
3
network = paddle.vision.models.mobilenet_v2(num_classes=get('num_classes'), pretrained=True)
model_lite = paddle.Model(network)
model_lite.summary((-1, ) + tuple(get('image_shape')))
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
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
100%|██████████| 20795/20795 [00:00<00:00, 44614.48it/s]
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/dygraph/layers.py:1263: UserWarning: Skip loading for classifier.1.weight. classifier.1.weight receives a shape [1280, 1000], but the expected shape is [1280, 12].
  warnings.warn(("Skip loading for {}. ".format(key) + str(err)))
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/dygraph/layers.py:1263: UserWarning: Skip loading for classifier.1.bias. classifier.1.bias receives a shape [1000], but the expected shape is [12].
  warnings.warn(("Skip loading for {}. ".format(key) + str(err)))


-------------------------------------------------------------------------------
   Layer (type)         Input Shape          Output Shape         Param #
===============================================================================
    Conv2D-160       [[1, 3, 224, 224]]   [1, 32, 112, 112]         864
  BatchNorm2D-107   [[1, 32, 112, 112]]   [1, 32, 112, 112]         128
      ReLU6-1       [[1, 32, 112, 112]]   [1, 32, 112, 112]          0
    Conv2D-161      [[1, 32, 112, 112]]   [1, 32, 112, 112]         288
  BatchNorm2D-108   [[1, 32, 112, 112]]   [1, 32, 112, 112]         128
      ReLU6-2       [[1, 32, 112, 112]]   [1, 32, 112, 112]          0
    Conv2D-162      [[1, 32, 112, 112]]   [1, 16, 112, 112]         512
  BatchNorm2D-109   [[1, 16, 112, 112]]   [1, 16, 112, 112]         64
InvertedResidual-1  [[1, 32, 112, 112]]   [1, 16, 112, 112]          0
    Conv2D-163      [[1, 16, 112, 112]]   [1, 96, 112, 112]        1,536
  BatchNorm2D-110   [[1, 96, 112, 112]]   [1, 96, 112, 112]         384
      ReLU6-3       [[1, 96, 112, 112]]   [1, 96, 112, 112]          0
    Conv2D-164      [[1, 96, 112, 112]]    [1, 96, 56, 56]          864
  BatchNorm2D-111    [[1, 96, 56, 56]]     [1, 96, 56, 56]          384
      ReLU6-4        [[1, 96, 56, 56]]     [1, 96, 56, 56]           0
    Conv2D-165       [[1, 96, 56, 56]]     [1, 24, 56, 56]         2,304
  BatchNorm2D-112    [[1, 24, 56, 56]]     [1, 24, 56, 56]          96
InvertedResidual-2  [[1, 16, 112, 112]]    [1, 24, 56, 56]           0
    Conv2D-166       [[1, 24, 56, 56]]     [1, 144, 56, 56]        3,456
  BatchNorm2D-113    [[1, 144, 56, 56]]    [1, 144, 56, 56]         576
      ReLU6-5        [[1, 144, 56, 56]]    [1, 144, 56, 56]          0
    Conv2D-167       [[1, 144, 56, 56]]    [1, 144, 56, 56]        1,296
  BatchNorm2D-114    [[1, 144, 56, 56]]    [1, 144, 56, 56]         576
      ReLU6-6        [[1, 144, 56, 56]]    [1, 144, 56, 56]          0
    Conv2D-168       [[1, 144, 56, 56]]    [1, 24, 56, 56]         3,456
  BatchNorm2D-115    [[1, 24, 56, 56]]     [1, 24, 56, 56]          96
InvertedResidual-3   [[1, 24, 56, 56]]     [1, 24, 56, 56]           0
    Conv2D-169       [[1, 24, 56, 56]]     [1, 144, 56, 56]        3,456
  BatchNorm2D-116    [[1, 144, 56, 56]]    [1, 144, 56, 56]         576
      ReLU6-7        [[1, 144, 56, 56]]    [1, 144, 56, 56]          0
    Conv2D-170       [[1, 144, 56, 56]]    [1, 144, 28, 28]        1,296
  BatchNorm2D-117    [[1, 144, 28, 28]]    [1, 144, 28, 28]         576
      ReLU6-8        [[1, 144, 28, 28]]    [1, 144, 28, 28]          0
    Conv2D-171       [[1, 144, 28, 28]]    [1, 32, 28, 28]         4,608
  BatchNorm2D-118    [[1, 32, 28, 28]]     [1, 32, 28, 28]          128
InvertedResidual-4   [[1, 24, 56, 56]]     [1, 32, 28, 28]           0
    Conv2D-172       [[1, 32, 28, 28]]     [1, 192, 28, 28]        6,144
  BatchNorm2D-119    [[1, 192, 28, 28]]    [1, 192, 28, 28]         768
      ReLU6-9        [[1, 192, 28, 28]]    [1, 192, 28, 28]          0
    Conv2D-173       [[1, 192, 28, 28]]    [1, 192, 28, 28]        1,728
  BatchNorm2D-120    [[1, 192, 28, 28]]    [1, 192, 28, 28]         768
     ReLU6-10        [[1, 192, 28, 28]]    [1, 192, 28, 28]          0
    Conv2D-174       [[1, 192, 28, 28]]    [1, 32, 28, 28]         6,144
  BatchNorm2D-121    [[1, 32, 28, 28]]     [1, 32, 28, 28]          128
InvertedResidual-5   [[1, 32, 28, 28]]     [1, 32, 28, 28]           0
    Conv2D-175       [[1, 32, 28, 28]]     [1, 192, 28, 28]        6,144
  BatchNorm2D-122    [[1, 192, 28, 28]]    [1, 192, 28, 28]         768
     ReLU6-11        [[1, 192, 28, 28]]    [1, 192, 28, 28]          0
    Conv2D-176       [[1, 192, 28, 28]]    [1, 192, 28, 28]        1,728
  BatchNorm2D-123    [[1, 192, 28, 28]]    [1, 192, 28, 28]         768
     ReLU6-12        [[1, 192, 28, 28]]    [1, 192, 28, 28]          0
    Conv2D-177       [[1, 192, 28, 28]]    [1, 32, 28, 28]         6,144
  BatchNorm2D-124    [[1, 32, 28, 28]]     [1, 32, 28, 28]          128
InvertedResidual-6   [[1, 32, 28, 28]]     [1, 32, 28, 28]           0
    Conv2D-178       [[1, 32, 28, 28]]     [1, 192, 28, 28]        6,144
  BatchNorm2D-125    [[1, 192, 28, 28]]    [1, 192, 28, 28]         768
     ReLU6-13        [[1, 192, 28, 28]]    [1, 192, 28, 28]          0
    Conv2D-179       [[1, 192, 28, 28]]    [1, 192, 14, 14]        1,728
  BatchNorm2D-126    [[1, 192, 14, 14]]    [1, 192, 14, 14]         768
     ReLU6-14        [[1, 192, 14, 14]]    [1, 192, 14, 14]          0
    Conv2D-180       [[1, 192, 14, 14]]    [1, 64, 14, 14]        12,288
  BatchNorm2D-127    [[1, 64, 14, 14]]     [1, 64, 14, 14]          256
InvertedResidual-7   [[1, 32, 28, 28]]     [1, 64, 14, 14]           0
    Conv2D-181       [[1, 64, 14, 14]]     [1, 384, 14, 14]       24,576
  BatchNorm2D-128    [[1, 384, 14, 14]]    [1, 384, 14, 14]        1,536
     ReLU6-15        [[1, 384, 14, 14]]    [1, 384, 14, 14]          0
    Conv2D-182       [[1, 384, 14, 14]]    [1, 384, 14, 14]        3,456
  BatchNorm2D-129    [[1, 384, 14, 14]]    [1, 384, 14, 14]        1,536
     ReLU6-16        [[1, 384, 14, 14]]    [1, 384, 14, 14]          0
    Conv2D-183       [[1, 384, 14, 14]]    [1, 64, 14, 14]        24,576
  BatchNorm2D-130    [[1, 64, 14, 14]]     [1, 64, 14, 14]          256
InvertedResidual-8   [[1, 64, 14, 14]]     [1, 64, 14, 14]           0
    Conv2D-184       [[1, 64, 14, 14]]     [1, 384, 14, 14]       24,576
  BatchNorm2D-131    [[1, 384, 14, 14]]    [1, 384, 14, 14]        1,536
     ReLU6-17        [[1, 384, 14, 14]]    [1, 384, 14, 14]          0
    Conv2D-185       [[1, 384, 14, 14]]    [1, 384, 14, 14]        3,456
  BatchNorm2D-132    [[1, 384, 14, 14]]    [1, 384, 14, 14]        1,536
     ReLU6-18        [[1, 384, 14, 14]]    [1, 384, 14, 14]          0
    Conv2D-186       [[1, 384, 14, 14]]    [1, 64, 14, 14]        24,576
  BatchNorm2D-133    [[1, 64, 14, 14]]     [1, 64, 14, 14]          256
InvertedResidual-9   [[1, 64, 14, 14]]     [1, 64, 14, 14]           0
    Conv2D-187       [[1, 64, 14, 14]]     [1, 384, 14, 14]       24,576
  BatchNorm2D-134    [[1, 384, 14, 14]]    [1, 384, 14, 14]        1,536
     ReLU6-19        [[1, 384, 14, 14]]    [1, 384, 14, 14]          0
    Conv2D-188       [[1, 384, 14, 14]]    [1, 384, 14, 14]        3,456
  BatchNorm2D-135    [[1, 384, 14, 14]]    [1, 384, 14, 14]        1,536
     ReLU6-20        [[1, 384, 14, 14]]    [1, 384, 14, 14]          0
    Conv2D-189       [[1, 384, 14, 14]]    [1, 64, 14, 14]        24,576
  BatchNorm2D-136    [[1, 64, 14, 14]]     [1, 64, 14, 14]          256
InvertedResidual-10  [[1, 64, 14, 14]]     [1, 64, 14, 14]           0
    Conv2D-190       [[1, 64, 14, 14]]     [1, 384, 14, 14]       24,576
  BatchNorm2D-137    [[1, 384, 14, 14]]    [1, 384, 14, 14]        1,536
     ReLU6-21        [[1, 384, 14, 14]]    [1, 384, 14, 14]          0
    Conv2D-191       [[1, 384, 14, 14]]    [1, 384, 14, 14]        3,456
  BatchNorm2D-138    [[1, 384, 14, 14]]    [1, 384, 14, 14]        1,536
     ReLU6-22        [[1, 384, 14, 14]]    [1, 384, 14, 14]          0
    Conv2D-192       [[1, 384, 14, 14]]    [1, 96, 14, 14]        36,864
  BatchNorm2D-139    [[1, 96, 14, 14]]     [1, 96, 14, 14]          384
InvertedResidual-11  [[1, 64, 14, 14]]     [1, 96, 14, 14]           0
    Conv2D-193       [[1, 96, 14, 14]]     [1, 576, 14, 14]       55,296
  BatchNorm2D-140    [[1, 576, 14, 14]]    [1, 576, 14, 14]        2,304
     ReLU6-23        [[1, 576, 14, 14]]    [1, 576, 14, 14]          0
    Conv2D-194       [[1, 576, 14, 14]]    [1, 576, 14, 14]        5,184
  BatchNorm2D-141    [[1, 576, 14, 14]]    [1, 576, 14, 14]        2,304
     ReLU6-24        [[1, 576, 14, 14]]    [1, 576, 14, 14]          0
    Conv2D-195       [[1, 576, 14, 14]]    [1, 96, 14, 14]        55,296
  BatchNorm2D-142    [[1, 96, 14, 14]]     [1, 96, 14, 14]          384
InvertedResidual-12  [[1, 96, 14, 14]]     [1, 96, 14, 14]           0
    Conv2D-196       [[1, 96, 14, 14]]     [1, 576, 14, 14]       55,296
  BatchNorm2D-143    [[1, 576, 14, 14]]    [1, 576, 14, 14]        2,304
     ReLU6-25        [[1, 576, 14, 14]]    [1, 576, 14, 14]          0
    Conv2D-197       [[1, 576, 14, 14]]    [1, 576, 14, 14]        5,184
  BatchNorm2D-144    [[1, 576, 14, 14]]    [1, 576, 14, 14]        2,304
     ReLU6-26        [[1, 576, 14, 14]]    [1, 576, 14, 14]          0
    Conv2D-198       [[1, 576, 14, 14]]    [1, 96, 14, 14]        55,296
  BatchNorm2D-145    [[1, 96, 14, 14]]     [1, 96, 14, 14]          384
InvertedResidual-13  [[1, 96, 14, 14]]     [1, 96, 14, 14]           0
    Conv2D-199       [[1, 96, 14, 14]]     [1, 576, 14, 14]       55,296
  BatchNorm2D-146    [[1, 576, 14, 14]]    [1, 576, 14, 14]        2,304
     ReLU6-27        [[1, 576, 14, 14]]    [1, 576, 14, 14]          0
    Conv2D-200       [[1, 576, 14, 14]]     [1, 576, 7, 7]         5,184
  BatchNorm2D-147     [[1, 576, 7, 7]]      [1, 576, 7, 7]         2,304
     ReLU6-28         [[1, 576, 7, 7]]      [1, 576, 7, 7]           0
    Conv2D-201        [[1, 576, 7, 7]]      [1, 160, 7, 7]        92,160
  BatchNorm2D-148     [[1, 160, 7, 7]]      [1, 160, 7, 7]          640
InvertedResidual-14  [[1, 96, 14, 14]]      [1, 160, 7, 7]           0
    Conv2D-202        [[1, 160, 7, 7]]      [1, 960, 7, 7]        153,600
  BatchNorm2D-149     [[1, 960, 7, 7]]      [1, 960, 7, 7]         3,840
     ReLU6-29         [[1, 960, 7, 7]]      [1, 960, 7, 7]           0
    Conv2D-203        [[1, 960, 7, 7]]      [1, 960, 7, 7]         8,640
  BatchNorm2D-150     [[1, 960, 7, 7]]      [1, 960, 7, 7]         3,840
     ReLU6-30         [[1, 960, 7, 7]]      [1, 960, 7, 7]           0
    Conv2D-204        [[1, 960, 7, 7]]      [1, 160, 7, 7]        153,600
  BatchNorm2D-151     [[1, 160, 7, 7]]      [1, 160, 7, 7]          640
InvertedResidual-15   [[1, 160, 7, 7]]      [1, 160, 7, 7]           0
    Conv2D-205        [[1, 160, 7, 7]]      [1, 960, 7, 7]        153,600
  BatchNorm2D-152     [[1, 960, 7, 7]]      [1, 960, 7, 7]         3,840
     ReLU6-31         [[1, 960, 7, 7]]      [1, 960, 7, 7]           0
    Conv2D-206        [[1, 960, 7, 7]]      [1, 960, 7, 7]         8,640
  BatchNorm2D-153     [[1, 960, 7, 7]]      [1, 960, 7, 7]         3,840
     ReLU6-32         [[1, 960, 7, 7]]      [1, 960, 7, 7]           0
    Conv2D-207        [[1, 960, 7, 7]]      [1, 160, 7, 7]        153,600
  BatchNorm2D-154     [[1, 160, 7, 7]]      [1, 160, 7, 7]          640
InvertedResidual-16   [[1, 160, 7, 7]]      [1, 160, 7, 7]           0
    Conv2D-208        [[1, 160, 7, 7]]      [1, 960, 7, 7]        153,600
  BatchNorm2D-155     [[1, 960, 7, 7]]      [1, 960, 7, 7]         3,840
     ReLU6-33         [[1, 960, 7, 7]]      [1, 960, 7, 7]           0
    Conv2D-209        [[1, 960, 7, 7]]      [1, 960, 7, 7]         8,640
  BatchNorm2D-156     [[1, 960, 7, 7]]      [1, 960, 7, 7]         3,840
     ReLU6-34         [[1, 960, 7, 7]]      [1, 960, 7, 7]           0
    Conv2D-210        [[1, 960, 7, 7]]      [1, 320, 7, 7]        307,200
  BatchNorm2D-157     [[1, 320, 7, 7]]      [1, 320, 7, 7]         1,280
InvertedResidual-17   [[1, 160, 7, 7]]      [1, 320, 7, 7]           0
    Conv2D-211        [[1, 320, 7, 7]]     [1, 1280, 7, 7]        409,600
  BatchNorm2D-158    [[1, 1280, 7, 7]]     [1, 1280, 7, 7]         5,120
     ReLU6-35        [[1, 1280, 7, 7]]     [1, 1280, 7, 7]           0
AdaptiveAvgPool2D-3  [[1, 1280, 7, 7]]     [1, 1280, 1, 1]           0
     Dropout-1          [[1, 1280]]           [1, 1280]              0
     Linear-4           [[1, 1280]]            [1, 12]            15,372
===============================================================================
Total params: 2,273,356
Trainable params: 2,205,132
Non-trainable params: 68,224
-------------------------------------------------------------------------------
Input size (MB): 0.57
Forward/backward pass size (MB): 152.87
Params size (MB): 8.67
Estimated Total Size (MB): 162.12
-------------------------------------------------------------------------------






{'total_params': 2273356, 'trainable_params': 2205132}
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
EPOCHS = 10
BATCH_SIZE = 32

# 请补齐模型训练过程代码


# 模型训练配置
model_lite.prepare(paddle.optimizer.Adam(learning_rate=0.0001, parameters=model_lite.parameters()),  # 优化器
              paddle.nn.CrossEntropyLoss(),        # 损失函数
              paddle.metric.Accuracy(topk=(1, 5))) # 评估指标

# 训练可视化VisualDL工具的回调函数
visualdl = paddle.callbacks.VisualDL(log_dir='./mobilenet/visualdl_log')

# 启动模型全流程训练
model_lite.fit( train_dataset,            # 训练数据集
                valid_dataset,            # 评估数据集
                epochs=EPOCHS,            # 总的训练轮次
                batch_size=BATCH_SIZE,    # 批次计算的样本量大小
                shuffle=True,             # 是否打乱样本集
                verbose=1,                # 日志展示格式
                save_dir='./mobilenet/chk_points/', # 分阶段的训练模型存储路径
                callbacks=[visualdl])     # 回调函数使用


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
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
The loss value printed in the log is the current step, and the metric is the average value of previous step.
Epoch 1/10


/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/nn/layer/norm.py:636: UserWarning: When training, we now always track global mean and variance.
  "When training, we now always track global mean and variance.")


step 222/222 [==============================] - loss: 0.2848 - acc_top1: 0.8271 - acc_top5: 0.9762 - 690ms/step
save checkpoint at /home/aistudio/mobilenet/chk_points/0
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.5077 - acc_top1: 0.9124 - acc_top5: 0.9906 - 803ms/step
Eval samples: 639
Epoch 2/10
step 222/222 [==============================] - loss: 0.2549 - acc_top1: 0.8541 - acc_top5: 0.9800 - 689ms/step
save checkpoint at /home/aistudio/mobilenet/chk_points/1
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.3955 - acc_top1: 0.9218 - acc_top5: 0.9937 - 813ms/step
Eval samples: 639
Epoch 3/10
step 222/222 [==============================] - loss: 0.3388 - acc_top1: 0.8660 - acc_top5: 0.9842 - 693ms/step
save checkpoint at /home/aistudio/mobilenet/chk_points/2
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.2251 - acc_top1: 0.9390 - acc_top5: 0.9937 - 817ms/step
Eval samples: 639
Epoch 4/10
step 222/222 [==============================] - loss: 0.6369 - acc_top1: 0.8773 - acc_top5: 0.9845 - 688ms/step
save checkpoint at /home/aistudio/mobilenet/chk_points/3
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.3293 - acc_top1: 0.9311 - acc_top5: 0.9937 - 813ms/step
Eval samples: 639
Epoch 5/10
step 222/222 [==============================] - loss: 0.4066 - acc_top1: 0.8908 - acc_top5: 0.9872 - 696ms/step
save checkpoint at /home/aistudio/mobilenet/chk_points/4
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.2782 - acc_top1: 0.9343 - acc_top5: 0.9937 - 808ms/step
Eval samples: 639
Epoch 6/10
step 222/222 [==============================] - loss: 0.3100 - acc_top1: 0.8961 - acc_top5: 0.9876 - 686ms/step
save checkpoint at /home/aistudio/mobilenet/chk_points/5
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.4607 - acc_top1: 0.9390 - acc_top5: 0.9937 - 810ms/step
Eval samples: 639
Epoch 7/10
step 222/222 [==============================] - loss: 0.1829 - acc_top1: 0.9016 - acc_top5: 0.9887 - 688ms/step
save checkpoint at /home/aistudio/mobilenet/chk_points/6
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.3421 - acc_top1: 0.9264 - acc_top5: 0.9969 - 815ms/step
Eval samples: 639
Epoch 8/10
step 222/222 [==============================] - loss: 0.4560 - acc_top1: 0.8932 - acc_top5: 0.9870 - 688ms/step
save checkpoint at /home/aistudio/mobilenet/chk_points/7
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.1464 - acc_top1: 0.9343 - acc_top5: 0.9922 - 830ms/step
Eval samples: 639
Epoch 9/10
step 222/222 [==============================] - loss: 0.4075 - acc_top1: 0.9040 - acc_top5: 0.9873 - 686ms/step
save checkpoint at /home/aistudio/mobilenet/chk_points/8
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.5118 - acc_top1: 0.9468 - acc_top5: 0.9953 - 820ms/step
Eval samples: 639
Epoch 10/10
step 222/222 [==============================] - loss: 0.0882 - acc_top1: 0.9094 - acc_top5: 0.9866 - 686ms/step
save checkpoint at /home/aistudio/mobilenet/chk_points/9
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 20/20 [==============================] - loss: 0.1651 - acc_top1: 0.9468 - acc_top5: 0.9953 - 821ms/step
Eval samples: 639
save checkpoint at /home/aistudio/mobilenet/chk_points/final



---------------------------------------------------------------------------

KeyError                                  Traceback (most recent call last)

<ipython-input-49-266b4a21aaf0> in <module>
     23                 callbacks=[visualdl])     # 回调函数使用
     24
---> 25 model.save(get('./mobilenet/model'))


~/config.py in get(full_path)
     43             config = CONFIG
     44
---> 45         config = config[name]
     46
     47     return config


KeyError: ''
1
model.save('./mobilenet/model')
1
2
3
# 模型评估
val = model_lite.evaluate(valid_dataset, verbose=1)
print(val)
1
2
3
4
5
Eval begin...
The loss value printed in the log is the current batch, and the metric is the average value of previous step.
step 639/639 [==============================] - loss: 1.0871e-04 - acc_top1: 0.9374 - acc_top5: 0.9890 - 46ms/step
Eval samples: 639
{'loss': [0.000108712964], 'acc_top1': 0.9374021909233177, 'acc_top5': 0.9890453834115805}
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
# 执行预测
pred = model_lite.predict(predict_dataset)

# 样本映射
LABEL_MAP = get('LABEL_MAP')

# 随机取样本展示
indexs = np.random.randint(1, 646, size=20)

for idx in indexs:
    predict_label = np.argmax(pred[0][idx])
    real_label = predict_dataset[idx][1]

    real_label = predict_dataset[idx][1]

    print('样本ID：{}, 真实标签：{}, 预测值：{}'.format(idx, LABEL_MAP[real_label], LABEL_MAP[predict_label]))
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Predict begin...
step 646/646 [==============================] - 42ms/step
Predict samples: 646
样本ID：611, 真实标签：ratt, 预测值：dragon
样本ID：42, 真实标签：rooster, 预测值：dragon
样本ID：11, 真实标签：rooster, 预测值：rooster
样本ID：625, 真实标签：ratt, 预测值：ratt
样本ID：263, 真实标签：goat, 预测值：goat
样本ID：635, 真实标签：ratt, 预测值：ratt
样本ID：78, 真实标签：dragon, 预测值：dragon
样本ID：567, 真实标签：ox, 预测值：ox
样本ID：280, 真实标签：horse, 预测值：horse
样本ID：406, 真实标签：pig, 预测值：pig
样本ID：167, 真实标签：snake, 预测值：snake
样本ID：19, 真实标签：rooster, 预测值：rooster
样本ID：41, 真实标签：rooster, 预测值：rooster
样本ID：406, 真实标签：pig, 预测值：pig
样本ID：497, 真实标签：rabbit, 预测值：rabbit
样本ID：71, 真实标签：dragon, 预测值：dragon
样本ID：300, 真实标签：horse, 预测值：horse
样本ID：247, 真实标签：goat, 预测值：goat
样本ID：450, 真实标签：tiger, 预测值：tiger
样本ID：82, 真实标签：dragon, 预测值：dragon