第八章：精通深度学习的高级内容 - 飞桨场景应用开发套件-PaddleX
文档简介：
飞桨场景应用开发套件-PaddleX：
PaddleX是飞桨场景应用开发套件，它集成飞桨智能视觉领域图像分类、目标检测、语义分割、实例分割任务能力，将深度学习开发全流程从数据准备、模型训练与优化到多端部署端到端打通，并提供统一任务API接口及图形化开发界面Demo。开发者无需分别安装不同套件，以低代码的形式即可快速完成飞桨全流程开发。
咨询热线：400-826-7010，为您提供专业的售前咨询，让您快速了解云产品，助您轻松上云！
微信咨询
免费试用、价格特惠
飞桨场景应用开发套件-PaddleX
PaddleX是飞桨场景应用开发套件，它集成飞桨智能视觉领域图像分类、目标检测、语义分割、实例分割任务能力，将深度学习开发全流程从数据准备、模型训练与优化到多端部署端到端打通，并提供统一任务API接口及图形化开发界面Demo。开发者无需分别安装不同套件，以低代码的形式即可快速完成飞桨全流程开发。
PaddleX 经过质检、安防、巡检、遥感、零售、医疗等十多个行业实际应用场景验证，沉淀产业实际经验，并提供丰富的案例实践教程，全程助力开发者产业实践落地。
PaddleX提供可视化界面GUI和Python API两种使用模式，通过可视化界面，即使是不熟悉编码的用户也可以快速上手飞桨的使用。通过简明易懂的Python API，方便用户根据实际生产需求直接调用或二次开发，为开发者提供飞桨全流程开发的最佳实践。同时，PaddleX提供的Restful API可以让用户通过简单集成即可生产所在行业的专属AI工具。
PaddleX具备以下特点：
- 全流程打通：打通从数据接入到推理部署的深度学习开发全流程，简化各环节串联工作，大幅提升开发效率。
- 开源技术内核：集成飞桨领先的视觉算法和工具组件，提供简明易懂的Python API，完全开源开放，易于集成和二次开发。
- 产业深度兼容：沉淀质检、安防、遥感、零售、医疗等十多个行业实际应用场景的验证，有丰富的预置案例，并具备符合产业应用需要的诸多特点，如兼容Windows、Mac、Linux系统，支持GPU加速模型训练，支持本地开发以确保数据安全等。
- 完善的教程与服务：丰富的全流程开发文档，高效的技术服务支持，提供多种方式方便用户与技术团队直接交流。
PaddleX 快速使用方法
下面以MobileNetV3_ssld完成化妆品分类为例，介绍PaddleX训练模型方式。
该案例是创建一个识别化妆品的模型，输入时日常可见的各种化妆品图片，输出是化妆品的类别名称。如果读者是男性用户，该案例可以让大家在女朋友面前露一把脸。如果读者是女性用户，可以制作一个真懂化妆品的AI取悦自己。
MobileNetV3_ssld是通过SSLD（简单的半监督标签知识蒸馏）方式得到的新模型。相对比原有的MobileNetV3预训练模型，在参数量不变的情况下，MobileNetV3_ssld预训练模型在ImageNet数据集上的精度提升3%。
在下文中，我们将先后介绍API和可视化界面Demo两种使用PaddleX的方法。
PaddleX API的使用方法
1. 安装PaddleX
! pip install paddlex==1.3.11 -i https://mirror.baidu.com/pypi/simple
Looking in indexes: https://mirror.baidu.com/pypi/simple Collecting paddlex==1.3.11 Downloading https://mirror.baidu.com/pypi/packages/d6/a2/07435f4aa1e51fe22bdf06c
95d03bf1b78b7bc6625adbb51e35dc0804cc7/paddlex-1.3.11-py3-none-any.whl (516 kB) ━
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 517.0/
517.0 KB 15.8 MB/s eta 0:00:00 Requirement already satisfied: tqdm in /opt/conda/e
nvs/python35-paddle120-env/lib/python3.7/site-packages (from paddlex==1.3.11) (4.27.0) Requirement already satisfied: psutil in /opt/conda/envs/python35-paddle120-env/lib
/python3.7/site-packages (from paddlex==1.3.11) (5.7.2) Collecting paddleslim==1.1.1 Downloading https://mirror.baidu.com/pypi/packages/d1/77/e257227bed9a70ff0d35a4a3c
4e70ac2d2362c803834c4c52018f7c4b762/paddleslim-1.1.1-py2.py3-none-any.whl (145 kB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 145.3/145.3
KB 41.1 MB/s eta 0:00:00 Requirement already satisfied: colorama in /opt/conda/envs/p
ython35-paddle120-env/lib/python3.7/site-packages (from paddlex==1.3.11) (0.4.4) Collecting paddlehub==2.1.0
Downloading https://mirror.baidu.com/pypi/packages/7a/29/3bd0ca43c787181e9c22fe44b9
44b64d7fcb14ce66d3bf4602d9ad2ac76c/paddlehub-2.1.0-py3-none-any.whl (211 kB) ━━━━
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 211.4/211.4 KB 43
.8 MB/s eta 0:00:00 Collecting shapely>=1.7.0 Downloading https://mirror.baidu.com/pypi/packages/d1/ec/3038263d69a0065d3ab6944ae839
f5f00896efd29b13ae62d73c00345b95/Shapely-1.8.2-cp37-cp37m-manylinux_2_12_x86_64.manylinu
x2010_x86_64.whl (2.0 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
━━━━━━━━━ 2.0/2.0 MB 10.9 MB/s eta 0:00:0000:0100:01 Requirement already satisfied: opencv-python in /opt/conda/envs/python35-paddle120-env/l
ib/python3.7/site-packages (from paddlex==1.3.11) (4.1.1.26) Requirement already satisfied: flask-cors in /opt/conda/envs/python35-paddle120-env/lib/p
ython3.7/site-packages (from paddlex==1.3.11) (3.0.8) Collecting pycocotools Downloading https://mirror.baidu.com/pypi/packages/75/5c/ac61ea715d7a89ecc31c090753bde2
8810238225ca8b71778dfe3e6a68bc/pycocotools-2.0.4.tar.gz (106 kB) ━━━━━━━━━━━━
━━━━━━━━━━━━━━━━━━━━━━━━━━ 106.6/106.6 KB 12.4 MB/
s eta 0:00:00 Installing build dependencies ... done Getting requirements to build wheel ... done Preparing metadata (pyproject.toml) ... done Requirement already satisfied: pyyaml in /opt/conda/envs/python35-paddle120-env/lib
/python3.7/site-packages (from paddlex==1.3.11) (5.1.2) Requirement already satisfied: sklearn in /opt/conda/envs/python35-paddle120-env/lib/
python3.7/site-packages (from paddlex==1.3.11) (0.0) Requirement already satisfied: visualdl>=2.0.0 in /opt/conda/envs/python35-paddle120-
env/lib/python3.7/site-packages (from paddlex==1.3.11) (2.2.3) Collecting xlwt Downloading https://mirror.baidu.com/pypi/packages/44/48/def306413b25c3d01753603b1
a222a011b8621aed27cd7f89cbc27e6b0f4/xlwt-1.3.0-py2.py3-none-any.whl (99 kB) ━━━━
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 100.0/100.0 K
B 12.3 MB/s eta 0:00:00 Requirement already satisfied: paddlenlp>=2.0.0rc5 in /opt/
conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from paddlehub==2.1.0-
>paddlex==1.3.11) (2.1.1) Collecting paddle2onnx>=0.5.1 Downloading https://mirror.baidu.com/pypi/packages/cf/40/7ebb5e820e80b94dbd132164f6108
2df67f9588118580b93670543d6f7ad/paddle2onnx-0.9.5-cp37-cp37m-manylinux_2_17_x86_64.manyli
nux2014_x86_64.whl (2.7 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
━━━━━━━━━━ 2.7/2.7 MB 10.8 MB/s eta 0:00:0000:0100:01 Requirement already satisfied: easydict in /opt/conda/envs/python35-paddle120-env/lib/pyt
hon3.7/site-packages (from paddlehub==2.1.0->paddlex==1.3.11) (1.9) Requirement already satisfied: pyzmq in /opt/conda/envs/python35-paddle120-env/lib/python3
.7/site-packages (from paddlehub==2.1.0->paddlex==1.3.11) (22.3.0) Requirement already satisfied: gunicorn>=19.10.0 in /opt/conda/envs/python35-paddle120-env
/lib/python3.7/site-packages (from paddlehub==2.1.0->paddlex==1.3.11) (20.0.4) Requirement already satisfied: numpy in /opt/conda/envs/python35-paddle120-env/lib/python3
.7/site-packages (from paddlehub==2.1.0->paddlex==1.3.11) (1.19.5) Requirement already satisfied: matplotlib in /opt/conda/envs/python35-paddle120-env/lib/py
thon3.7/site-packages (from paddlehub==2.1.0->paddlex==1.3.11) (2.2.3) Requirement already satisfied: rarfile in /opt/conda/envs/python35-paddle120-env/lib/pyth
on3.7/site-packages (from paddlehub==2.1.0->paddlex==1.3.11) (3.1) Requirement already satisfied: packaging in /opt/conda/envs/python35-paddle120-env/lib/py
thon3.7/site-packages (from paddlehub==2.1.0->paddlex==1.3.11) (21.3) Requirement already satisfied: filelock in /opt/conda/envs/python35-paddle120-env/lib/pyt
hon3.7/site-packages (from paddlehub==2.1.0->paddlex==1.3.11) (3.0.12) Requirement already satisfied: gitpython in /opt/conda/envs/python35-paddle120-env/lib/p
ython3.7/site-packages (from paddlehub==2.1.0->paddlex==1.3.11) (3.1.14) Requirement already satisfied: Pillow in /opt/conda/envs/python35-paddle120-env/lib/pytho
n3.7/site-packages (from paddlehub==2.1.0->paddlex==1.3.11) (8.2.0) Requirement already satisfied: colorlog in /opt/conda/envs/python35-paddle120-env/lib/pyt
hon3.7/site-packages (from paddlehub==2.1.0->paddlex==1.3.11) (4.1.0) Requirement already satisfied: flask>=1.1.0 in /opt/conda/envs/python35-paddle120-env/lib/
python3.7/site-packages (from paddlehub==2.1.0->paddlex==1.3.11) (1.1.1) Requirement already satisfied: pre-commit in /opt/conda/envs/python35-paddle120-env/lib/p
ython3.7/site-packages (from visualdl>=2.0.0->paddlex==1.3.11) (1.21.0) Requirement already satisfied: protobuf>=3.11.0 in /opt/conda/envs/python35-paddle120-env/
lib/python3.7/site-packages (from visualdl>=2.0.0->paddlex==1.3.11) (3.14.0) Requirement already satisfied: Flask-Babel>=1.0.0 in /opt/conda/envs/python35-paddle120-en
v/lib/python3.7/site-packages (from visualdl>=2.0.0->paddlex==1.3.11) (1.0.0) Requirement already satisfied: bce-python-sdk in /opt/conda/envs/python35-paddle120-env/l
ib/python3.7/site-packages (from visualdl>=2.0.0->paddlex==1.3.11) (0.8.53) Requirement already satisfied: six>=1.14.0 in /opt/conda/envs/python35-paddle120-env/lib/p
ython3.7/site-packages (from visualdl>=2.0.0->paddlex==1.3.11) (1.16.0) Requirement already satisfied: requests in /opt/conda/envs/python35-paddle120-env/lib/pyt
hon3.7/site-packages (from visualdl>=2.0.0->paddlex==1.3.11) (2.24.0) Requirement already satisfied: shellcheck-py in /opt/conda/envs/python35-paddle120-env/li
b/python3.7/site-packages (from visualdl>=2.0.0->paddlex==1.3.11) (0.7.1.1) Requirement already satisfied: pandas in /opt/conda/envs/python35-paddle120-env/lib/pyth
on3.7/site-packages (from visualdl>=2.0.0->paddlex==1.3.11) (1.1.5) Requirement already satisfied: flake8>=3.7.9 in /opt/conda/envs/python35-paddle120-env/lib
/python3.7/site-packages (from visualdl>=2.0.0->paddlex==1.3.11) (4.0.1) Requirement already satisfied: scikit-learn in /opt/conda/envs/python35-paddle120-env/lib
/python3.7/site-packages (from sklearn->paddlex==1.3.11) (0.24.2) Requirement already satisfied: pycodestyle<2.9.0,>=2.8.0 in /opt/conda/envs/python35-pad
dle120-env/lib/python3.7/site-packages (from flake8>=3.7.9->visualdl>=2.0.0->paddlex==1.3.11) (2.8.0) Requirement already satisfied: mccabe<0.7.0,>=0.6.0 in /opt/conda/envs/python35-paddle120-
env/lib/python3.7/site-packages (from flake8>=3.7.9->visualdl>=2.0.0->paddlex==1.3.11) (0.6.1) Requirement already satisfied: importlib-metadata<4.3 in /opt/conda/envs/python35-paddle120-en
v/lib/python3.7/site-packages (from flake8>=3.7.9->visualdl>=2.0.0->paddlex==1.3.11) (4.2.0) Requirement already satisfied: pyflakes<2.5.0,>=2.4.0 in /opt/conda/envs/python35-paddle120-
env/lib/python3.7/site-packages (from flake8>=3.7.9->visualdl>=2.0.0->paddlex==1.3.11) (2.4.0) Requirement already satisfied: Werkzeug>=0.15 in /opt/conda/envs/python35-paddle120-env/lib/p
ython3.7/site-packages (from flask>=1.1.0->paddlehub==2.1.0->paddlex==1.3.11) (0.16.0) Requirement already satisfied: itsdangerous>=0.24 in /opt/conda/envs/python35-paddle120-env/l
ib/python3.7/site-packages (from flask>=1.1.0->paddlehub==2.1.0->paddlex==1.3.11) (1.1.0) Requirement already satisfied: click>=5.1 in /opt/conda/envs/python35-paddle120-env/lib/pytho
n3.7/site-packages (from flask>=1.1.0->paddlehub==2.1.0->paddlex==1.3.11) (7.0) Requirement already satisfied: Jinja2>=2.10.1 in /opt/conda/envs/python35-paddle120-env/lib/p
ython3.7/site-packages (from flask>=1.1.0->paddlehub==2.1.0->paddlex==1.3.11) (3.0.0) Requirement already satisfied: Babel>=2.3 in /opt/conda/envs/python35-paddle120-env/lib/pytho
n3.7/site-packages (from Flask-Babel>=1.0.0->visualdl>=2.0.0->paddlex==1.3.11) (2.8.0) Requirement already satisfied: pytz in /opt/conda/envs/python35-paddle120-env/lib/python3.7/
site-packages (from Flask-Babel>=1.0.0->visualdl>=2.0.0->paddlex==1.3.11) (2019.3) Requirement already satisfied: setuptools>=3.0 in /opt/conda/envs/python35-paddle120-env/lib/
python3.7/site-packages (from gunicorn>=19.10.0->paddlehub==2.1.0->paddlex==1.3.11) (56.2.0) Requirement already satisfied: python-dateutil>=2.1 in /opt/conda/envs/python35-paddle120-env
/lib/python3.7/site-packages (from matplotlib->paddlehub==2.1.0->paddlex==1.3.11) (2.8.2) Requirement already satisfied: kiwisolver>=1.0.1 in /opt/conda/envs/python35-paddle120-env/
lib/python3.7/site-packages (from matplotlib->paddlehub==2.1.0->paddlex==1.3.11) (1.1.0) Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /opt/conda/envs/p
ython35-paddle120-env/lib/python3.7/site-packages (from matplotlib->paddlehub==2.1.0->paddlex==1.3.11) (3.0.8) Requirement already satisfied: cycler>=0.10 in /opt/conda/envs/python35-paddle120-env/lib/pyt
hon3.7/site-packages (from matplotlib->paddlehub==2.1.0->paddlex==1.3.11) (0.10.0) Collecting onnx<=1.9.0 Downloading https://mirror.baidu.com/pypi/packages/3f/9b/54c950d3256e27
f970a83cd0504efb183a24312702deed0179453316dbd0/onnx-1.9.0-cp37-cp37m-manylinux2010_x86_64.whl
(12.2 MB) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 1
2.2/12.2 MB 10.0 MB/s eta 0:00:0000:0100:01 Requirement already satisfied: seqeval in /opt/conda/envs/python35-paddle120-env/lib/python3.
7/site-packages (from paddlenlp>=2.0.0rc5->paddlehub==2.1.0->paddlex==1.3.11) (1.2.2) Requirement already satisfied: paddlefsl==1.0.0 in /opt/conda/envs/python35-paddle120-env/li
b/python3.7/site-packages (from paddlenlp>=2.0.0rc5->paddlehub==2.1.0->paddlex==1.3.11) (1.0.0) Requirement already satisfied: multiprocess in /opt/conda/envs/python35-paddle120-env/lib/pytho
n3.7/site-packages (from paddlenlp>=2.0.0rc5->paddlehub==2.1.0->paddlex==1.3.11) (0.70.11.1) Requirement already satisfied: jieba in /opt/conda/envs/python35-paddle120-env/lib/python3.7/s
ite-packages (from paddlenlp>=2.0.0rc5->paddlehub==2.1.0->paddlex==1.3.11) (0.42.1) Requirement already satisfied: h5py in /opt/conda/envs/python35-paddle120-env/lib/python3.7/si
te-packages (from paddlenlp>=2.0.0rc5->paddlehub==2.1.0->paddlex==1.3.11) (2.9.0) Requirement already satisfied: idna<3,>=2.5 in /opt/conda/envs/python35-paddle120-env/lib/pytho
n3.7/site-packages (from requests->visualdl>=2.0.0->paddlex==1.3.11) (2.8) Requirement already satisfied: certifi>=2017.4.17 in /opt/conda/envs/python35-paddle120-env/li
b/python3.7/site-packages (from requests->visualdl>=2.0.0->paddlex==1.3.11) (2019.9.11) Requirement already satisfied: chardet<4,>=3.0.2 in /opt/conda/envs/python35-paddle120-env/li
b/python3.7/site-packages (from requests->visualdl>=2.0.0->paddlex==1.3.11) (3.0.4) Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /opt/conda/envs/py
thon35-paddle120-env/lib/python3.7/site-packages (from requests->visualdl>=2.0.0->paddlex==1.3.11) (1.25.6) Requirement already satisfied: future>=0.6.0 in /opt/conda/envs/python35-paddle120-env/lib/py
thon3.7/site-packages (from bce-python-sdk->visualdl>=2.0.0->paddlex==1.3.11) (0.18.0) Requirement already satisfied: pycryptodome>=3.8.0 in /opt/conda/envs/python35-paddle120-env/
lib/python3.7/site-packages (from bce-python-sdk->visualdl>=2.0.0->paddlex==1.3.11) (3.9.9) Requirement already satisfied: gitdb<5,>=4.0.1 in /opt/conda/envs/python35-paddle120-env/lib
/python3.7/site-packages (from gitpython->paddlehub==2.1.0->paddlex==1.3.11) (4.0.5) Requirement already satisfied: cfgv>=2.0.0 in /opt/conda/envs/python35-paddle120-env/lib/p
ython3.7/site-packages (from pre-commit->visualdl>=2.0.0->paddlex==1.3.11) (2.0.1) Requirement already satisfied: toml in /opt/conda/envs/python35-paddle120-env/lib/python3
.7/site-packages (from pre-commit->visualdl>=2.0.0->paddlex==1.3.11) (0.10.0) Requirement already satisfied: virtualenv>=15.2 in /opt/conda/envs/python35-paddle120-e
nv/lib/python3.7/site-packages (from pre-commit->visualdl>=2.0.0->paddlex==1.3.11) (16.7.9) Requirement already satisfied: aspy.yaml in /opt/conda/envs/python35-paddle120-env/lib/py
thon3.7/site-packages (from pre-commit->visualdl>=2.0.0->paddlex==1.3.11) (1.3.0) Requirement already satisfied: identify>=1.0.0 in /opt/conda/envs/python35-paddle120-env/l
ib/python3.7/site-packages (from pre-commit->visualdl>=2.0.0->paddlex==1.3.11) (1.4.10) Requirement already satisfied: nodeenv>=0.11.1 in /opt/conda/envs/python35-paddle120-env/li
/python3.7/site-packages (from pre-commit->visualdl>=2.0.0->paddlex==1.3.11) (1.3.4) Requirement already satisfied: threadpoolctl>=2.0.0 in /opt/conda/envs/python35-paddle120-en
v/lib/python3.7/site-packages (from scikit-learn->sklearn->paddlex==1.3.11) (2.1.0) Requirement already satisfied: joblib>=0.11 in /opt/conda/envs/python35-paddle120-env/lib/py
thon3.7/site-packages (from scikit-learn->sklearn->paddlex==1.3.11) (0.14.1) Requirement already satisfied: scipy>=0.19.1 in /opt/conda/envs/python35-paddle120-env/lib/py
thon3.7/site-packages (from scikit-learn->sklearn->paddlex==1.3.11) (1.6.3) Requirement already satisfied: smmap<4,>=3.0.1 in /opt/conda/envs/python35-paddle120-env/lib/
python3.7/site-packages (from gitdb<5,>=4.0.1->gitpython->paddlehub==2.1.0->paddlex==1.3.11) (3.0.5) Requirement already satisfied: typing-extensions>=3.6.4 in /opt/conda/envs/python35-paddle12
0-env/lib/python3.7/site-packages (from importlib-metadata<4.3->flake8>=3.7.9->visualdl>=2.0
.0->paddlex==1.3.11) (4.2.0) Requirement already satisfied: zipp>=0.5 in /opt/conda/envs/python35-paddle120-env/lib/pyth
on3.7/site-packages (from importlib-metadata<4.3->flake8>=3.7.9->visualdl>=2.0.0->paddlex==1.3.11) (3.8.0) Requirement already satisfied: MarkupSafe>=2.0.0rc2 in /opt/conda/envs/python35-paddle120-
env/lib/python3.7/site-packages (from Jinja2>=2.10.1->flask>=1.1.0->paddlehub==2.1.0->paddlex==1.3.11) (2.0.1) Requirement already satisfied: dill>=0.3.3 in /opt/conda/envs/python35-paddle120-env/lib/
python3.7/site-packages (from multiprocess->paddlenlp>=2.0.0rc5->paddlehub==2.1.0->paddlex==1.3.11) (0.3.3) Building wheels for collected packages: pycocotools Building wheel for pycocotools (pyproject.toml) ... done Created wheel for pycocotools: filename=pycocotools-2.0.4-cp37-cp37m-linux_x86_64.whl
size=273789 sha256=bafea6d86df5bf63e6dc2e6183a323192674c734375255b9c330cc05479f701b Stored in directory: /home/aistudio/.cache/pip/wheels/d0/74/13/98b11419a029
f3c25590419747f1ec26f5494beae1d457560b
Successfully built pycocotools
Installing collected packages: xlwt, shapely, paddleslim, onnx, pycocotools, paddle2onnx, paddlehub, paddlex
Attempting uninstall: paddlehub
Found existing installation: paddlehub 2.0.4
Uninstalling paddlehub-2.0.4:
Successfully uninstalled paddlehub-2.0.4
Successfully installed onnx-1.9.0 paddle2onnx-0.9.5 paddlehub-2.1.0 paddleslim-
1.1.1 paddlex-1.3.11 pycocotools-2.0.4 shapely-1.8.2 xlwt-1.3.0
# jupyter中使用paddlex需要设置matplotlib import matplotlib
matplotlib.use('Agg') # 设置使用0号GPU卡（如无GPU，执行此代码后仍然会使用CPU训练模型） import os
os.environ['CUDA_VISIBLE_DEVICES'] = '0' import paddlex as pdx
from paddlex.cls import transforms
train_transforms = transforms.Compose([
transforms.RandomCrop(crop_size=224),
transforms.RandomHorizontalFlip(),
transforms.Normalize()
])
eval_transforms = transforms.Compose([
transforms.ResizeByShort(short_size=256),
transforms.CenterCrop(crop_size=224),
transforms.Normalize()
])
train_dataset = pdx.datasets.ImageNet(
data_dir='makeup',
file_list='makeup/train_list.txt',
label_list='makeup/labels.txt',
transforms=train_transforms,
shuffle=True)
eval_dataset = pdx.datasets.ImageNet(
data_dir='makeup',
file_list='makeup/val_list.txt',
label_list='makeup/labels.txt',
transforms=eval_transforms)
2022-05-06 17:06:50 [INFO] Starting to read file list from dataset... 2022-05-06 17:06:50 [INFO] 710 samples in file makeup/train_list.txt 2022-05-06 17:06:50 [INFO] Starting to read file list from dataset... 2022-05-06 17:06:50 [INFO] 132 samples in file makeup/val_list.txt
3.4 开始训练模型
在定义好数据集后，即可选择分类模型（这里使用了MobileNetV3_large_ssld模型）开始训练。MobileNetV3_large是面向移动端应用场景的模型，而MobileNetV3_large_ssld是百度通过SSLD蒸馏策略所得的模型，具有更高的精度表现。
关于分类模型训练，更多参数介绍可参见文档paddlex.cls.MobileNetV3_large_ssld。在如下代码中，模型训练过程每间隔save_interval_epochs轮次，会保存一次模型在save_dir目录下，同时在保存的过程中也会在验证数据集上计算相关指标，模型训练过程中相关日志的含义可参见文档。
注：本数据集在P40 GPU上训练MobileNetV3_large_ssld，模型的训练过程预估为10分钟左右；如无GPU，则预估为30分钟左右。
num_classes = len(train_dataset.labels)
model = pdx.cls.MobileNetV3_large_ssld(num_classes=num_classes)
model.train(num_epochs=10,
train_dataset=train_dataset,
train_batch_size=32,
eval_dataset=eval_dataset,
lr_decay_epochs=[4, 6, 8],
save_interval_epochs=1,
learning_rate=0.025,
save_dir='output/mobilenetv3_large_ssld')
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/layers
/math_op_patch.py:341: UserWarning: /opt/conda/envs/python35-paddle120-env/lib/python3.7
/site-packages/paddlex/cv/nets/mobilenet_v3.py:231 The behavior of expression A * B has been unified with elementwise_mul(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_mul(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]))
2022-05-06 17:06:52 [INFO] Decompressing output/mobilenetv3_large_ssld/pr
etrain/MobileNetV3_large_x1_0_ssld_pretrained.tar...
W0506 17:06:52.906878 498 gpu_context.cc:244] Please NOTE: device: 0, GPU Compute
Capability: 7.0, Driver API Version: 10.1, Runtime API Version: 10.1 W0506 17:06:52.913384 498 gpu_context.cc:272] device: 0, cuDNN Version: 7.6.
result = model.predict('makeup/mascara/27.jpeg', topk=1)
print("Predict Result:", result)
当然，PaddleX的功能不止这么简单，这是一个极简功能的展示案例。实际上，PaddleX可以和PaddleSeg、PaddleClas、PaddleDetection等开发套件一样，实现非常丰富的模型和训练配置。如果读者对这些功能感兴趣，欢迎继续查阅PaddleX的文档，或者可以通过PaddleX官方提供的图形化界面Demo了解。该Demo完整展示了基于PaddleX的API可以完成的功能，而且PaddleX的界面读者可以根据自己的需要重新设计。
5.项目开发
（1）选择数据：项目创建完成后，需要选择已载入客户端并校验后的数据集，并点击“下一步”，进入参数配置页面。
（2）配置参数：主要分为模型参数、训练参数、优化策略三部分。可根据实际需求选择模型结构及对应的训练参数、优化策略，使得任务效果最佳。
另外，可以在客户端中选择不同的数据增强方式：
参数配置完成后，点击“启动训练”，模型开始训练并进行效果评估。
（3）训练可视化：在训练过程中，可通过VisualDL查看模型训练过程时的参数变化、日志详情，及当前最优的训练集和验证集训练指标。模型在训练过程中通过点击“终止训练”随时终止训练过程。
PaddleX集成了飞桨可视化分析工具VisualDL，可以很方便地查看训练过程的指标数据。
模型训练结束后，点击”下一步“，从客户端中，也可以看到训练的完成进度和验证集精度。
（4）模型发布：当模型效果满意后，可根据实际的生产环境需求，将模型发布为需要的版本。
注意：这个带可视化界面的AI研发软件仅仅是基于PaddleX API做出来的一个Demo。受此启发，欢迎读者使用PaddleX API研发一款适合自己所在企业或行业使用的AI研发工具，图形化界面的软件功能可以根据场景的需要来灵活定制。
往届优秀学员作品展示
PaddleX-API扩展预览transform的效果
- 项目背景
我们在做深度学习计算机视觉项目的时候，经常会用到图像增广数据增强的操作，可是操作后的结果是什么样子呢？此项目带你揭开Transform的神秘面纱。
- 项目内容
使用tkinter的操作界面，可视化显示图像增强之后的图像形式。
- 实现方案
使用PaddleX的API扩展，预览Transform的效果。
- 实现结果
- 项目点评
不仅应用PaddleX GUI实现了应用自有数据对模型进行训练，增加了对卸妆水/卸妆油品类的识别，又自主基于PaddleX API完成了对数据增强Transform的可视化呈现工程。
- 项目链接 https://aistudio.baidu.com/aistudio/projectdetail/545565
