【AI workshop】增强的PicoDet有多猛?来,跑个RM数据集试试!

P粉084495128
发布: 2025-07-31 14:46:59
原创
324人浏览过
今天我们用PP-PicoDet完成RoboMaster人工智能挑战赛数据集的训练,并在单卡V100上测试效果(没错就是AiStudio的环境) 本次你将:用不到4个A100 hour跑完整个训练过程,然后拿着单卡V100测试。

☞☞☞AI 智能聊天, 问答助手, AI 智能搜索, 免费无限量使用 DeepSeek R1 模型☜☜☜

【ai workshop】增强的picodet有多猛?来,跑个rm数据集试试! - php中文网

增强的PicoDet有多猛?用了它,妈妈再也不用担心我的FPS和mAP啦

项目背景

目标检测

开门见山

今天我们用PP-PicoDet完成RoboMaster人工智能挑战赛数据集的训练,并在单卡V100上测试效果(没错就是AiStudio的环境) 本次你将:用不到4个A100 hour跑完整个训练过程,然后拿着单卡V100测试。

本项目提供: 完整VOC数据集地址, PaddleDetection PicoDet相关配置文件(l_640_lcnet以及s_416_lcnet),voc/train.txt, valid.txt, test.txt,尽可能全面的操作流程

项目版本提供了模型的静态图文件(xs_416和l_640),以及测试用的数据PaddleDetection/dataset/voc/test,可以在安装后直接测试

感兴趣的同学可以去PicoDet详细页面去了解具体的模型概要

单张预测结果:

除了框大了点其他也不戳

【AI workshop】增强的PicoDet有多猛?来,跑个RM数据集试试! - php中文网        

15.47效果一览(注意val和test时的batchsize是1)

CPU(with mkl) mAP(VAL) FPS
picodet_l_640_coco_lcnet 97.43% 20.11307455077152
picodet_xs_416_coco_lcnet 96.30% 27.07027027027027

可以看到在batchsize为1的情况下,作为full版本的picodet_l在精度上比picodet_xs高1.13%,但是picodet_xs的FPS比其高了不少。此测试结果为在V100上导出静态图模型后CPU 4thread 开MKL加速预测。如果使用GPU预测:

GPU mAP FPS
picodet_l_640_coco_lcnet / 15.47
picodet_xs_416_coco_lcnet / 21.21

会发现GPU版本比CPU+MKL差了不少。开了mkl加速的CPU预测比GPU快这么多,这背后究竟是人性的扭曲还是...那么请往后看,谜底就在每一步所耗费的时间上

算法介绍

PP-PicoDet模型有如下特点:

  •  更高的mAP: 第一个在1M参数量之内mAP(0.5:0.95)超越30+(输入416像素时)。
  •  更快的预测速度: 网络预测在ARM CPU下可达150FPS。
  •  部署友好: 支持PaddleLite/MNN/NCNN/OpenVINO等预测库,支持转出ONNX,提供了C++/Python/Android的demo。
  •  先进的算法: 我们在现有SOTA算法中进行了创新, 包括:ESNet, CSP-PAN, SimOTA等等。

感兴趣的同学可以直接阅读论文进行详细学习:

PP-PicoDet

【AI workshop】增强的PicoDet有多猛?来,跑个RM数据集试试! - php中文网    

1.Architecture

【AI workshop】增强的PicoDet有多猛?来,跑个RM数据集试试! - php中文网        

2.backbone:引入了Enhanced ShuffleNet (ESNet)

【AI workshop】增强的PicoDet有多猛?来,跑个RM数据集试试! - php中文网        

顾名思义为ShuffleNetV2改进而来。细节使用了pointwise, deepwise卷积来减少参数,同时解决了ShuffleNetV2中channel shuffle带来的融合特征的丢失

3.neck:CSP-PAN

CSP neck——常见于Yolov4和YoloX。PicoDet中则先使用PAN结构(PAN简单理解就是FPN多了一条Bottom-up path augmentation)提取 multi-level的特征,再用CSP做相邻通道feature map的融合(concatenate,特指通道数的合并)。可见Architecture图的右半部分。

为了减少计算量,引入1x1卷积,1x1的卷积大家都知道可以升维,降维,增加模型非线性。这里则把所有的feature的通道数都等于最小的那个——96,减少了通道数==减少了模型参数量。

除了1x1卷积之外,CSP-PAN中所有的卷积都变成了空洞卷积,增强了感受野,提高了acc。

~剩下的一些策略可以自行去论文PP-PicoDet学习。~

观察数据集

以为数据集是青年挑战赛的,没想到是ICRA的数据集 可以看到分为7个Part,估计是作者分7次去拍摄(不管了 其标签格式并不是传统的coco或者voc

- 00002.txt中的内容:0 0.63125 0.5703125 0.059374999999999956 0.053124999999999981 0.62265625 0.4765625 0.1796875 0.321875
登录后复制
       

众所周知,前面的0,1代表类别,即红色装甲板和红色机器人

red_armor
red_robotblue_robotdead_robotblue_armor
登录后复制
       

第一个数字知道是什么意思了,后面的呢?猜猜会是下面表达方式中的哪一种?

表达方式 说明
x1,y1,x2,y2 (x1,y1)为左上角坐标,(x2,y2)为右下角坐标
x1,y1,w,h (x1,y1)为左上角坐标,w为目标区域宽度,h为目标区域高度
xc,yc,w,h (xc,yc)为目标区域中心坐标,w为目标区域宽度,h为目标区域高度

幸运的是我们从官方repo下面获取到了label信息...是xywh型的 知道了label信息后,开始转换工作吧,写代码!

【AI workshop】增强的PicoDet有多猛?来,跑个RM数据集试试! - php中文网        

注意标注的文件数不等于label数

In [2]
# 测试自己做的coco数据集是否OKimport sysimport jsonfrom pycocotools.coco import COCO

ann_file_train = "PaddleDetection/dataset/coco/annotations/instances_train.json"# json文件的绝对路径ann_file_val = "PaddleDetection/dataset/coco/annotations/instances_val.json"# json文件的绝对路径coco_train = COCO(annotation_file=ann_file_train)
coco_val = COCO(annotation_file=ann_file_val)print("coco_train\nimages.size [%05d]\tannotations.size [%05d]\t category.size [%05d]\ndone!"
      %(len(coco_train.imgs),len(coco_train.anns),len(coco_train.cats)))print("coco_val\nimages.size [%05d]\tannotations.size [%05d]\t category.size [%05d]\ndone!"
      %(len(coco_val.imgs),len(coco_val.anns),len(coco_val.cats)))# 看到输出,就知道已经完成啦
登录后复制
       
loading annotations into memory...
Done (t=0.72s)
creating index...
index created!
loading annotations into memory...
Done (t=0.34s)
creating index...
index created!
coco_train
images.size [13914]	annotations.size [49263]	 category.size [00005]
done!
coco_val
images.size [05969]	annotations.size [21295]	 category.size [00005]
done!
登录后复制
       

用PP-PicoDet完成训练吧

Pico之小,可以放机器人上也。

In [2]
# 进入PaddleDetection目录%cd PaddleDetection/
登录后复制
       
/home/aistudio/PaddleDetection
登录后复制
       
In [3]
# 安装...!pip install -r requirements.txt
登录后复制
       
Looking in indexes: https://pypi.tuna.tsinghua.edu.cn/simple
Requirement already satisfied: tqdm in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 1)) (4.36.1)
Collecting typeguard
  Downloading https://pypi.tuna.tsinghua.edu.cn/packages/9a/bb/d43e5c75054e53efce310e79d63df0ac3f25e34c926be5dffb7d283fb2a8/typeguard-2.13.3-py3-none-any.whl (17 kB)
Requirement already satisfied: visualdl>=2.1.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 3)) (2.2.0)
Requirement already satisfied: opencv-python in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 4)) (4.1.1.26)
Requirement already satisfied: PyYAML in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 5)) (5.1.2)
Collecting shapely
  Downloading https://pypi.tuna.tsinghua.edu.cn/packages/7b/b6/580f795a835f7b93d6e5c4fb125ba4fa81eb0e8489c5ac126ac2669b9521/Shapely-1.8.5.post1-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.whl (2.0 MB)     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 2.0/2.0 MB 1.9 MB/s eta 0:00:0000:0100:01
Requirement already satisfied: scipy in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 7)) (1.6.3)
Collecting terminaltables
  Downloading https://pypi.tuna.tsinghua.edu.cn/packages/c4/fb/ea621e0a19733e01fe4005d46087d383693c0f4a8f824b47d8d4122c87e0/terminaltables-3.1.10-py2.py3-none-any.whl (15 kB)
Requirement already satisfied: Cython in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 9)) (0.29)
Collecting pycocotools
  Downloading https://pypi.tuna.tsinghua.edu.cn/packages/ef/c6/90220be3b39fbc4cbd203775ca47dd8dc97fae06fbd2b500637395621b7c/pycocotools-2.0.6.tar.gz (24 kB)
  Installing build dependencies ... done
  Getting requirements to build wheel ... done
  Preparing metadata (pyproject.toml) ... done
Requirement already satisfied: setuptools>=42.0.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 12)) (56.2.0)
Collecting lap
  Downloading https://pypi.tuna.tsinghua.edu.cn/packages/bf/64/d9fb6a75b15e783952b2fec6970f033462e67db32dc43dfbb404c14e91c2/lap-0.4.0.tar.gz (1.5 MB)     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 1.5/1.5 MB 2.0 MB/s eta 0:00:0000:0100:01
  Preparing metadata (setup.py) ... done
Requirement already satisfied: sklearn in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 14)) (0.0)
Collecting motmetrics
  Downloading https://pypi.tuna.tsinghua.edu.cn/packages/45/41/b019fe934eb811b9aba9b335f852305b804b9c66f098d7e35c2bdb09d1c8/motmetrics-1.2.5-py3-none-any.whl (161 kB)     ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 161.1/161.1 kB 2.5 MB/s eta 0:00:00a 0:00:01Requirement already satisfied: openpyxl in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 16)) (3.0.5)
Requirement already satisfied: Flask-Babel>=1.0.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.1.0->-r requirements.txt (line 3)) (1.0.0)
Requirement already satisfied: pandas in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.1.0->-r requirements.txt (line 3)) (1.1.5)
Requirement already satisfied: protobuf>=3.11.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.1.0->-r requirements.txt (line 3)) (3.20.1)
Requirement already satisfied: six>=1.14.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.1.0->-r requirements.txt (line 3)) (1.16.0)
Requirement already satisfied: Pillow>=7.0.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.1.0->-r requirements.txt (line 3)) (7.1.2)
Requirement already satisfied: matplotlib in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.1.0->-r requirements.txt (line 3)) (2.2.3)
Requirement already satisfied: numpy in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.1.0->-r requirements.txt (line 3)) (1.20.3)
Requirement already satisfied: pre-commit in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.1.0->-r requirements.txt (line 3)) (1.21.0)
Requirement already satisfied: flask>=1.1.1 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.1.0->-r requirements.txt (line 3)) (1.1.1)
Requirement already satisfied: requests in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.1.0->-r requirements.txt (line 3)) (2.22.0)
Requirement already satisfied: bce-python-sdk in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.1.0->-r requirements.txt (line 3)) (0.8.53)
Requirement already satisfied: flake8>=3.7.9 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.1.0->-r requirements.txt (line 3)) (4.0.1)
Requirement already satisfied: shellcheck-py in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.1.0->-r requirements.txt (line 3)) (0.7.1.1)
Requirement already satisfied: scikit-learn in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from sklearn->-r requirements.txt (line 14)) (0.24.2)
Collecting xmltodict>=0.12.0
  Downloading https://pypi.tuna.tsinghua.edu.cn/packages/94/db/fd0326e331726f07ff7f40675cd86aa804bfd2e5016c727fa761c934990e/xmltodict-0.13.0-py2.py3-none-any.whl (10.0 kB)
Requirement already satisfied: et-xmlfile in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from openpyxl->-r requirements.txt (line 16)) (1.0.1)
Requirement already satisfied: jdcal in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from openpyxl->-r requirements.txt (line 16)) (1.4.1)
Requirement already satisfied: importlib-metadata<4.3 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from flake8>=3.7.9->visualdl>=2.1.0->-r requirements.txt (line 3)) (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>=2.1.0->-r requirements.txt (line 3)) (2.4.0)
Requirement already satisfied: pycodestyle<2.9.0,>=2.8.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from flake8>=3.7.9->visualdl>=2.1.0->-r requirements.txt (line 3)) (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>=2.1.0->-r requirements.txt (line 3)) (0.6.1)
Requirement already satisfied: Jinja2>=2.10.1 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from flask>=1.1.1->visualdl>=2.1.0->-r requirements.txt (line 3)) (3.0.0)
Requirement already satisfied: Werkzeug>=0.15 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from flask>=1.1.1->visualdl>=2.1.0->-r requirements.txt (line 3)) (0.16.0)
Requirement already satisfied: itsdangerous>=0.24 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from flask>=1.1.1->visualdl>=2.1.0->-r requirements.txt (line 3)) (1.1.0)
Requirement already satisfied: click>=5.1 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from flask>=1.1.1->visualdl>=2.1.0->-r requirements.txt (line 3)) (7.0)
Requirement already satisfied: pytz in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from Flask-Babel>=1.0.0->visualdl>=2.1.0->-r requirements.txt (line 3)) (2019.3)
Requirement already satisfied: Babel>=2.3 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from Flask-Babel>=1.0.0->visualdl>=2.1.0->-r requirements.txt (line 3)) (2.8.0)
Requirement already satisfied: python-dateutil>=2.1 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from matplotlib->visualdl>=2.1.0->-r requirements.txt (line 3)) (2.8.2)
Requirement already satisfied: cycler>=0.10 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from matplotlib->visualdl>=2.1.0->-r requirements.txt (line 3)) (0.10.0)
Requirement already satisfied: kiwisolver>=1.0.1 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from matplotlib->visualdl>=2.1.0->-r requirements.txt (line 3)) (1.1.0)
Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from matplotlib->visualdl>=2.1.0->-r requirements.txt (line 3)) (3.0.9)
Requirement already satisfied: pycryptodome>=3.8.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from bce-python-sdk->visualdl>=2.1.0->-r requirements.txt (line 3)) (3.9.9)
Requirement already satisfied: future>=0.6.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from bce-python-sdk->visualdl>=2.1.0->-r requirements.txt (line 3)) (0.18.0)
Requirement already satisfied: aspy.yaml in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from pre-commit->visualdl>=2.1.0->-r requirements.txt (line 3)) (1.3.0)
Requirement already satisfied: identify>=1.0.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from pre-commit->visualdl>=2.1.0->-r requirements.txt (line 3)) (1.4.10)
Requirement already satisfied: toml in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from pre-commit->visualdl>=2.1.0->-r requirements.txt (line 3)) (0.10.0)
Requirement already satisfied: nodeenv>=0.11.1 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from pre-commit->visualdl>=2.1.0->-r requirements.txt (line 3)) (1.3.4)
Requirement already satisfied: virtualenv>=15.2 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from pre-commit->visualdl>=2.1.0->-r requirements.txt (line 3)) (16.7.9)
Requirement already satisfied: cfgv>=2.0.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from pre-commit->visualdl>=2.1.0->-r requirements.txt (line 3)) (2.0.1)
Requirement already satisfied: certifi>=2017.4.17 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from requests->visualdl>=2.1.0->-r requirements.txt (line 3)) (2019.9.11)
Requirement already satisfied: idna<2.9,>=2.5 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from requests->visualdl>=2.1.0->-r requirements.txt (line 3)) (2.8)
Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from requests->visualdl>=2.1.0->-r requirements.txt (line 3)) (1.25.6)
Requirement already satisfied: chardet<3.1.0,>=3.0.2 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from requests->visualdl>=2.1.0->-r requirements.txt (line 3)) (3.0.4)
Requirement already satisfied: threadpoolctl>=2.0.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from scikit-learn->sklearn->-r requirements.txt (line 14)) (2.1.0)
Requirement already satisfied: joblib>=0.11 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from scikit-learn->sklearn->-r requirements.txt (line 14)) (0.14.1)
Requirement already satisfied: zipp>=0.5 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from importlib-metadata<4.3->flake8>=3.7.9->visualdl>=2.1.0->-r requirements.txt (line 3)) (3.8.1)
Requirement already satisfied: typing-extensions>=3.6.4 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from importlib-metadata<4.3->flake8>=3.7.9->visualdl>=2.1.0->-r requirements.txt (line 3)) (4.3.0)
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>=1.1.1->visualdl>=2.1.0->-r requirements.txt (line 3)) (2.0.1)
Building wheels for collected packages: pycocotools, lap
  Building wheel for pycocotools (pyproject.toml) ... done
  Created wheel for pycocotools: filename=pycocotools-2.0.6-cp37-cp37m-linux_x86_64.whl size=275102 sha256=8ba9b6cad68c8eefe26e4535d132b6fd21cd1e4155bfb15ac995ee61c6555872
  Stored in directory: /home/aistudio/.cache/pip/wheels/f8/94/70/046149e666bd5812b7de6b87a28dcef238f7162f4108e0b3d8
  Building wheel for lap (setup.py) ... done
  Created wheel for lap: filename=lap-0.4.0-cp37-cp37m-linux_x86_64.whl size=1593889 sha256=fdab7872daf44527d32acee2f60c7affbcc001d6bf813403ec39607b382fbc5f
  Stored in directory: /home/aistudio/.cache/pip/wheels/5c/d0/d2/e331d17a999666b1e2eb99743cfa1742629f9d26c55c657001
Successfully built pycocotools lap
Installing collected packages: lap, xmltodict, typeguard, terminaltables, shapely, pycocotools, motmetrics
Successfully installed lap-0.4.0 motmetrics-1.2.5 pycocotools-2.0.6 shapely-1.8.5.post1 terminaltables-3.1.10 typeguard-2.13.3 xmltodict-0.13.0[notice] A new release of pip available: 22.1.2 -> 22.3.1[notice] To update, run: pip install --upgrade pip
登录后复制
       
In [4]
# 安装...!python3 setup.py install
登录后复制
       
running install
running bdist_egg
running egg_info
writing paddledet.egg-info/PKG-INFO
writing dependency_links to paddledet.egg-info/dependency_links.txt
writing requirements to paddledet.egg-info/requires.txt
writing top-level names to paddledet.egg-info/top_level.txt
adding license file 'LICENSE' (matched pattern 'LICEN[CS]E*')
reading manifest file 'paddledet.egg-info/SOURCES.txt'
writing manifest file 'paddledet.egg-info/SOURCES.txt'
installing library code to build/bdist.linux-x86_64/egg
running install_lib
running build_py
copying ppdet/version.py -> build/lib/ppdet
copying ppdet/model_zoo/MODEL_ZOO -> build/lib/ppdet/model_zoo
creating build/bdist.linux-x86_64/egg
creating build/bdist.linux-x86_64/egg/ppdet
creating build/bdist.linux-x86_64/egg/ppdet/model_zoo
creating build/bdist.linux-x86_64/egg/ppdet/model_zoo/tests
copying build/lib/ppdet/model_zoo/tests/test_list_model.py -> build/bdist.linux-x86_64/egg/ppdet/model_zoo/tests
copying build/lib/ppdet/model_zoo/tests/test_get_model.py -> build/bdist.linux-x86_64/egg/ppdet/model_zoo/tests
copying build/lib/ppdet/model_zoo/tests/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/model_zoo/tests
copying build/lib/ppdet/model_zoo/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/model_zoo
copying build/lib/ppdet/model_zoo/MODEL_ZOO -> build/bdist.linux-x86_64/egg/ppdet/model_zoo
copying build/lib/ppdet/model_zoo/model_zoo.py -> build/bdist.linux-x86_64/egg/ppdet/model_zoo
creating build/bdist.linux-x86_64/egg/ppdet/modeling
creating build/bdist.linux-x86_64/egg/ppdet/modeling/necks
copying build/lib/ppdet/modeling/necks/bifpn.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/necks
copying build/lib/ppdet/modeling/necks/csp_pan.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/necks
copying build/lib/ppdet/modeling/necks/fpn.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/necks
copying build/lib/ppdet/modeling/necks/lc_pan.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/necks
copying build/lib/ppdet/modeling/necks/yolo_fpn.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/necks
copying build/lib/ppdet/modeling/necks/custom_pan.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/necks
copying build/lib/ppdet/modeling/necks/centernet_fpn.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/necks
copying build/lib/ppdet/modeling/necks/hrfpn.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/necks
copying build/lib/ppdet/modeling/necks/ttf_fpn.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/necks
copying build/lib/ppdet/modeling/necks/blazeface_fpn.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/necks
copying build/lib/ppdet/modeling/necks/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/necks
copying build/lib/ppdet/modeling/necks/es_pan.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/necks
copying build/lib/ppdet/modeling/shape_spec.py -> build/bdist.linux-x86_64/egg/ppdet/modeling
creating build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/yolo_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/pico_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/fcos_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/detr_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/ssd_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/mask_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/bbox_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/roi_extractor.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/face_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/centernet_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/tood_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/simota_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/ttf_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/cascade_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/retina_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/solov2_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/ppyoloe_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/sparsercnn_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/keypoint_hrhrnet_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/s2anet_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
copying build/lib/ppdet/modeling/heads/gfl_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/heads
creating build/bdist.linux-x86_64/egg/ppdet/modeling/coders
copying build/lib/ppdet/modeling/coders/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/coders
copying build/lib/ppdet/modeling/coders/delta_bbox_coder.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/coders
copying build/lib/ppdet/modeling/ops.py -> build/bdist.linux-x86_64/egg/ppdet/modeling
copying build/lib/ppdet/modeling/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/modeling
copying build/lib/ppdet/modeling/post_process.py -> build/bdist.linux-x86_64/egg/ppdet/modeling
creating build/bdist.linux-x86_64/egg/ppdet/modeling/tests
copying build/lib/ppdet/modeling/tests/test_ops.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/tests
copying build/lib/ppdet/modeling/tests/test_architectures.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/tests
copying build/lib/ppdet/modeling/tests/test_mstest.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/tests
copying build/lib/ppdet/modeling/tests/test_base.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/tests
copying build/lib/ppdet/modeling/tests/test_yolov3_loss.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/tests
copying build/lib/ppdet/modeling/tests/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/tests
creating build/bdist.linux-x86_64/egg/ppdet/modeling/transformers
copying build/lib/ppdet/modeling/transformers/deformable_transformer.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/transformers
copying build/lib/ppdet/modeling/transformers/position_encoding.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/transformers
copying build/lib/ppdet/modeling/transformers/detr_transformer.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/transformers
copying build/lib/ppdet/modeling/transformers/utils.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/transformers
copying build/lib/ppdet/modeling/transformers/matchers.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/transformers
copying build/lib/ppdet/modeling/transformers/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/transformers
creating build/bdist.linux-x86_64/egg/ppdet/modeling/assigners
copying build/lib/ppdet/modeling/assigners/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/assigners
copying build/lib/ppdet/modeling/assigners/simota_assigner.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/assigners
copying build/lib/ppdet/modeling/assigners/max_iou_assigner.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/assigners
copying build/lib/ppdet/modeling/assigners/atss_assigner.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/assigners
copying build/lib/ppdet/modeling/assigners/task_aligned_assigner.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/assigners
copying build/lib/ppdet/modeling/assigners/utils.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/assigners
creating build/bdist.linux-x86_64/egg/ppdet/modeling/reid
copying build/lib/ppdet/modeling/reid/pyramidal_embedding.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/reid
copying build/lib/ppdet/modeling/reid/jde_embedding_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/reid
copying build/lib/ppdet/modeling/reid/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/reid
copying build/lib/ppdet/modeling/reid/resnet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/reid
copying build/lib/ppdet/modeling/reid/pplcnet_embedding.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/reid
copying build/lib/ppdet/modeling/reid/resnet_embedding.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/reid
copying build/lib/ppdet/modeling/reid/fairmot_embedding_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/reid
creating build/bdist.linux-x86_64/egg/ppdet/modeling/mot
creating build/bdist.linux-x86_64/egg/ppdet/modeling/mot/matching
copying build/lib/ppdet/modeling/mot/matching/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/mot/matching
copying build/lib/ppdet/modeling/mot/matching/deepsort_matching.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/mot/matching
copying build/lib/ppdet/modeling/mot/matching/jde_matching.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/mot/matching
copying build/lib/ppdet/modeling/mot/visualization.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/mot
creating build/bdist.linux-x86_64/egg/ppdet/modeling/mot/tracker
copying build/lib/ppdet/modeling/mot/tracker/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/mot/tracker
copying build/lib/ppdet/modeling/mot/tracker/base_sde_tracker.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/mot/tracker
copying build/lib/ppdet/modeling/mot/tracker/deepsort_tracker.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/mot/tracker
copying build/lib/ppdet/modeling/mot/tracker/jde_tracker.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/mot/tracker
copying build/lib/ppdet/modeling/mot/tracker/base_jde_tracker.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/mot/tracker
creating build/bdist.linux-x86_64/egg/ppdet/modeling/mot/motion
copying build/lib/ppdet/modeling/mot/motion/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/mot/motion
copying build/lib/ppdet/modeling/mot/motion/kalman_filter.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/mot/motion
copying build/lib/ppdet/modeling/mot/utils.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/mot
copying build/lib/ppdet/modeling/mot/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/mot
copying build/lib/ppdet/modeling/initializer.py -> build/bdist.linux-x86_64/egg/ppdet/modeling
creating build/bdist.linux-x86_64/egg/ppdet/modeling/proposal_generator
copying build/lib/ppdet/modeling/proposal_generator/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/proposal_generator
copying build/lib/ppdet/modeling/proposal_generator/anchor_generator.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/proposal_generator
copying build/lib/ppdet/modeling/proposal_generator/rpn_head.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/proposal_generator
copying build/lib/ppdet/modeling/proposal_generator/proposal_generator.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/proposal_generator
copying build/lib/ppdet/modeling/proposal_generator/target_layer.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/proposal_generator
copying build/lib/ppdet/modeling/proposal_generator/target.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/proposal_generator
creating build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/losses/sparsercnn_loss.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/losses/fcos_loss.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/losses/jde_loss.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/losses/iou_aware_loss.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/losses/focal_loss.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/losses/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/losses/ssd_loss.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/losses/varifocal_loss.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/losses/keypoint_loss.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/losses/gfocal_loss.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/losses/detr_loss.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/losses/fairmot_loss.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/losses/smooth_l1_loss.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/losses/iou_loss.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/losses/solov2_loss.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/losses/ctfocal_loss.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/losses/yolo_loss.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/losses
copying build/lib/ppdet/modeling/bbox_utils.py -> build/bdist.linux-x86_64/egg/ppdet/modeling
creating build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/detr.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/yolo.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/ssd.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/retinanet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/faster_rcnn.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/ttfnet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/yolox.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/bytetrack.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/gfl.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/keypoint_hrhrnet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/blazeface.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/fairmot.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/deepsort.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/picodet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/mask_rcnn.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/meta_arch.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/tood.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/jde.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/sparse_rcnn.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/keypoint_hrnet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/cascade_rcnn.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/s2anet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/fcos.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/solov2.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/architectures/centernet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/architectures
copying build/lib/ppdet/modeling/keypoint_utils.py -> build/bdist.linux-x86_64/egg/ppdet/modeling
copying build/lib/ppdet/modeling/layers.py -> build/bdist.linux-x86_64/egg/ppdet/modeling
creating build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/mobilenet_v3.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/lite_hrnet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/vgg.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/shufflenet_v2.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/mobilenet_v1.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/resnet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/blazenet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/cspresnet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/senet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/darknet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/hardnet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/csp_darknet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/ghostnet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/esnet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/swin_transformer.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/lcnet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/hrnet.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/name_adapter.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/dla.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
copying build/lib/ppdet/modeling/backbones/res2net.py -> build/bdist.linux-x86_64/egg/ppdet/modeling/backbones
creating build/bdist.linux-x86_64/egg/ppdet/metrics
copying build/lib/ppdet/metrics/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/metrics
copying build/lib/ppdet/metrics/mcmot_metrics.py -> build/bdist.linux-x86_64/egg/ppdet/metrics
copying build/lib/ppdet/metrics/coco_utils.py -> build/bdist.linux-x86_64/egg/ppdet/metrics
copying build/lib/ppdet/metrics/json_results.py -> build/bdist.linux-x86_64/egg/ppdet/metrics
copying build/lib/ppdet/metrics/munkres.py -> build/bdist.linux-x86_64/egg/ppdet/metrics
copying build/lib/ppdet/metrics/widerface_utils.py -> build/bdist.linux-x86_64/egg/ppdet/metrics
copying build/lib/ppdet/metrics/keypoint_metrics.py -> build/bdist.linux-x86_64/egg/ppdet/metrics
copying build/lib/ppdet/metrics/metrics.py -> build/bdist.linux-x86_64/egg/ppdet/metrics
copying build/lib/ppdet/metrics/map_utils.py -> build/bdist.linux-x86_64/egg/ppdet/metrics
copying build/lib/ppdet/metrics/mot_metrics.py -> build/bdist.linux-x86_64/egg/ppdet/metrics
creating build/bdist.linux-x86_64/egg/ppdet/slim
copying build/lib/ppdet/slim/ofa.py -> build/bdist.linux-x86_64/egg/ppdet/slim
copying build/lib/ppdet/slim/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/slim
copying build/lib/ppdet/slim/distill.py -> build/bdist.linux-x86_64/egg/ppdet/slim
copying build/lib/ppdet/slim/quant.py -> build/bdist.linux-x86_64/egg/ppdet/slim
copying build/lib/ppdet/slim/prune.py -> build/bdist.linux-x86_64/egg/ppdet/slim
copying build/lib/ppdet/slim/unstructured_prune.py -> build/bdist.linux-x86_64/egg/ppdet/slim
creating build/bdist.linux-x86_64/egg/ppdet/engine
copying build/lib/ppdet/engine/trainer.py -> build/bdist.linux-x86_64/egg/ppdet/engine
copying build/lib/ppdet/engine/tracker.py -> build/bdist.linux-x86_64/egg/ppdet/engine
copying build/lib/ppdet/engine/env.py -> build/bdist.linux-x86_64/egg/ppdet/engine
copying build/lib/ppdet/engine/callbacks.py -> build/bdist.linux-x86_64/egg/ppdet/engine
copying build/lib/ppdet/engine/export_utils.py -> build/bdist.linux-x86_64/egg/ppdet/engine
copying build/lib/ppdet/engine/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/engine
creating build/bdist.linux-x86_64/egg/ppdet/utils
copying build/lib/ppdet/utils/checkpoint.py -> build/bdist.linux-x86_64/egg/ppdet/utils
copying build/lib/ppdet/utils/download.py -> build/bdist.linux-x86_64/egg/ppdet/utils
copying build/lib/ppdet/utils/check.py -> build/bdist.linux-x86_64/egg/ppdet/utils
copying build/lib/ppdet/utils/stats.py -> build/bdist.linux-x86_64/egg/ppdet/utils
copying build/lib/ppdet/utils/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/utils
copying build/lib/ppdet/utils/logger.py -> build/bdist.linux-x86_64/egg/ppdet/utils
copying build/lib/ppdet/utils/colormap.py -> build/bdist.linux-x86_64/egg/ppdet/utils
copying build/lib/ppdet/utils/profiler.py -> build/bdist.linux-x86_64/egg/ppdet/utils
copying build/lib/ppdet/utils/voc_utils.py -> build/bdist.linux-x86_64/egg/ppdet/utils
copying build/lib/ppdet/utils/cli.py -> build/bdist.linux-x86_64/egg/ppdet/utils
copying build/lib/ppdet/utils/visualizer.py -> build/bdist.linux-x86_64/egg/ppdet/utils
creating build/bdist.linux-x86_64/egg/ppdet/core
copying build/lib/ppdet/core/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/core
creating build/bdist.linux-x86_64/egg/ppdet/core/config
copying build/lib/ppdet/core/config/schema.py -> build/bdist.linux-x86_64/egg/ppdet/core/config
copying build/lib/ppdet/core/config/yaml_helpers.py -> build/bdist.linux-x86_64/egg/ppdet/core/config
copying build/lib/ppdet/core/config/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/core/config
copying build/lib/ppdet/core/workspace.py -> build/bdist.linux-x86_64/egg/ppdet/core
copying build/lib/ppdet/version.py -> build/bdist.linux-x86_64/egg/ppdet
copying build/lib/ppdet/__init__.py -> build/bdist.linux-x86_64/egg/ppdet
copying build/lib/ppdet/optimizer.py -> build/bdist.linux-x86_64/egg/ppdet
creating build/bdist.linux-x86_64/egg/ppdet/data
copying build/lib/ppdet/data/shm_utils.py -> build/bdist.linux-x86_64/egg/ppdet/data
copying build/lib/ppdet/data/reader.py -> build/bdist.linux-x86_64/egg/ppdet/data
creating build/bdist.linux-x86_64/egg/ppdet/data/transform
copying build/lib/ppdet/data/transform/operators.py -> build/bdist.linux-x86_64/egg/ppdet/data/transform
copying build/lib/ppdet/data/transform/autoaugment_utils.py -> build/bdist.linux-x86_64/egg/ppdet/data/transform
copying build/lib/ppdet/data/transform/batch_operators.py -> build/bdist.linux-x86_64/egg/ppdet/data/transform
copying build/lib/ppdet/data/transform/keypoint_operators.py -> build/bdist.linux-x86_64/egg/ppdet/data/transform
copying build/lib/ppdet/data/transform/gridmask_utils.py -> build/bdist.linux-x86_64/egg/ppdet/data/transform
copying build/lib/ppdet/data/transform/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/data/transform
copying build/lib/ppdet/data/transform/atss_assigner.py -> build/bdist.linux-x86_64/egg/ppdet/data/transform
copying build/lib/ppdet/data/transform/mot_operators.py -> build/bdist.linux-x86_64/egg/ppdet/data/transform
copying build/lib/ppdet/data/transform/op_helper.py -> build/bdist.linux-x86_64/egg/ppdet/data/transform
copying build/lib/ppdet/data/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/data
creating build/bdist.linux-x86_64/egg/ppdet/data/crop_utils
copying build/lib/ppdet/data/crop_utils/annotation_cropper.py -> build/bdist.linux-x86_64/egg/ppdet/data/crop_utils
copying build/lib/ppdet/data/crop_utils/chip_box_utils.py -> build/bdist.linux-x86_64/egg/ppdet/data/crop_utils
copying build/lib/ppdet/data/crop_utils/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/data/crop_utils
creating build/bdist.linux-x86_64/egg/ppdet/data/source
copying build/lib/ppdet/data/source/keypoint_coco.py -> build/bdist.linux-x86_64/egg/ppdet/data/source
copying build/lib/ppdet/data/source/mot.py -> build/bdist.linux-x86_64/egg/ppdet/data/source
copying build/lib/ppdet/data/source/voc.py -> build/bdist.linux-x86_64/egg/ppdet/data/source
copying build/lib/ppdet/data/source/coco.py -> build/bdist.linux-x86_64/egg/ppdet/data/source
copying build/lib/ppdet/data/source/widerface.py -> build/bdist.linux-x86_64/egg/ppdet/data/source
copying build/lib/ppdet/data/source/__init__.py -> build/bdist.linux-x86_64/egg/ppdet/data/source
copying build/lib/ppdet/data/source/category.py -> build/bdist.linux-x86_64/egg/ppdet/data/source
copying build/lib/ppdet/data/source/sniper_coco.py -> build/bdist.linux-x86_64/egg/ppdet/data/source
copying build/lib/ppdet/data/source/dataset.py -> build/bdist.linux-x86_64/egg/ppdet/data/source
byte-compiling build/bdist.linux-x86_64/egg/ppdet/model_zoo/tests/test_list_model.py to test_list_model.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/model_zoo/tests/test_get_model.py to test_get_model.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/model_zoo/tests/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/model_zoo/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/model_zoo/model_zoo.py to model_zoo.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/necks/bifpn.py to bifpn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/necks/csp_pan.py to csp_pan.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/necks/fpn.py to fpn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/necks/lc_pan.py to lc_pan.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/necks/yolo_fpn.py to yolo_fpn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/necks/custom_pan.py to custom_pan.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/necks/centernet_fpn.py to centernet_fpn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/necks/hrfpn.py to hrfpn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/necks/ttf_fpn.py to ttf_fpn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/necks/blazeface_fpn.py to blazeface_fpn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/necks/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/necks/es_pan.py to es_pan.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/shape_spec.py to shape_spec.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/yolo_head.py to yolo_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/pico_head.py to pico_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/fcos_head.py to fcos_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/detr_head.py to detr_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/ssd_head.py to ssd_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/mask_head.py to mask_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/bbox_head.py to bbox_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/roi_extractor.py to roi_extractor.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/face_head.py to face_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/centernet_head.py to centernet_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/tood_head.py to tood_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/simota_head.py to simota_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/ttf_head.py to ttf_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/cascade_head.py to cascade_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/retina_head.py to retina_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/solov2_head.py to solov2_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/ppyoloe_head.py to ppyoloe_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/sparsercnn_head.py to sparsercnn_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/keypoint_hrhrnet_head.py to keypoint_hrhrnet_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/s2anet_head.py to s2anet_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/heads/gfl_head.py to gfl_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/coders/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/coders/delta_bbox_coder.py to delta_bbox_coder.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/ops.py to ops.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/post_process.py to post_process.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/tests/test_ops.py to test_ops.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/tests/test_architectures.py to test_architectures.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/tests/test_mstest.py to test_mstest.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/tests/test_base.py to test_base.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/tests/test_yolov3_loss.py to test_yolov3_loss.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/tests/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/transformers/deformable_transformer.py to deformable_transformer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/transformers/position_encoding.py to position_encoding.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/transformers/detr_transformer.py to detr_transformer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/transformers/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/transformers/matchers.py to matchers.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/transformers/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/assigners/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/assigners/simota_assigner.py to simota_assigner.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/assigners/max_iou_assigner.py to max_iou_assigner.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/assigners/atss_assigner.py to atss_assigner.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/assigners/task_aligned_assigner.py to task_aligned_assigner.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/assigners/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/reid/pyramidal_embedding.py to pyramidal_embedding.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/reid/jde_embedding_head.py to jde_embedding_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/reid/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/reid/resnet.py to resnet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/reid/pplcnet_embedding.py to pplcnet_embedding.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/reid/resnet_embedding.py to resnet_embedding.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/reid/fairmot_embedding_head.py to fairmot_embedding_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/mot/matching/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/mot/matching/deepsort_matching.py to deepsort_matching.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/mot/matching/jde_matching.py to jde_matching.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/mot/visualization.py to visualization.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/mot/tracker/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/mot/tracker/base_sde_tracker.py to base_sde_tracker.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/mot/tracker/deepsort_tracker.py to deepsort_tracker.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/mot/tracker/jde_tracker.py to jde_tracker.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/mot/tracker/base_jde_tracker.py to base_jde_tracker.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/mot/motion/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/mot/motion/kalman_filter.py to kalman_filter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/mot/utils.py to utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/mot/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/initializer.py to initializer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/proposal_generator/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/proposal_generator/anchor_generator.py to anchor_generator.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/proposal_generator/rpn_head.py to rpn_head.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/proposal_generator/proposal_generator.py to proposal_generator.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/proposal_generator/target_layer.py to target_layer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/proposal_generator/target.py to target.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/losses/sparsercnn_loss.py to sparsercnn_loss.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/losses/fcos_loss.py to fcos_loss.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/losses/jde_loss.py to jde_loss.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/losses/iou_aware_loss.py to iou_aware_loss.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/losses/focal_loss.py to focal_loss.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/losses/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/losses/ssd_loss.py to ssd_loss.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/losses/varifocal_loss.py to varifocal_loss.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/losses/keypoint_loss.py to keypoint_loss.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/losses/gfocal_loss.py to gfocal_loss.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/losses/detr_loss.py to detr_loss.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/losses/fairmot_loss.py to fairmot_loss.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/losses/smooth_l1_loss.py to smooth_l1_loss.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/losses/iou_loss.py to iou_loss.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/losses/solov2_loss.py to solov2_loss.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/losses/ctfocal_loss.py to ctfocal_loss.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/losses/yolo_loss.py to yolo_loss.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/bbox_utils.py to bbox_utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/detr.py to detr.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/yolo.py to yolo.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/ssd.py to ssd.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/retinanet.py to retinanet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/faster_rcnn.py to faster_rcnn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/ttfnet.py to ttfnet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/yolox.py to yolox.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/bytetrack.py to bytetrack.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/gfl.py to gfl.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/keypoint_hrhrnet.py to keypoint_hrhrnet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/blazeface.py to blazeface.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/fairmot.py to fairmot.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/deepsort.py to deepsort.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/picodet.py to picodet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/mask_rcnn.py to mask_rcnn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/meta_arch.py to meta_arch.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/tood.py to tood.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/jde.py to jde.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/sparse_rcnn.py to sparse_rcnn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/keypoint_hrnet.py to keypoint_hrnet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/cascade_rcnn.py to cascade_rcnn.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/s2anet.py to s2anet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/fcos.py to fcos.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/solov2.py to solov2.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/architectures/centernet.py to centernet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/keypoint_utils.py to keypoint_utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/layers.py to layers.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/mobilenet_v3.py to mobilenet_v3.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/lite_hrnet.py to lite_hrnet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/vgg.py to vgg.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/shufflenet_v2.py to shufflenet_v2.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/mobilenet_v1.py to mobilenet_v1.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/resnet.py to resnet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/blazenet.py to blazenet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/cspresnet.py to cspresnet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/senet.py to senet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/darknet.py to darknet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/hardnet.py to hardnet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/csp_darknet.py to csp_darknet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/ghostnet.py to ghostnet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/esnet.py to esnet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/swin_transformer.py to swin_transformer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/lcnet.py to lcnet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/hrnet.py to hrnet.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/name_adapter.py to name_adapter.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/dla.py to dla.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/modeling/backbones/res2net.py to res2net.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/metrics/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/metrics/mcmot_metrics.py to mcmot_metrics.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/metrics/coco_utils.py to coco_utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/metrics/json_results.py to json_results.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/metrics/munkres.py to munkres.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/metrics/widerface_utils.py to widerface_utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/metrics/keypoint_metrics.py to keypoint_metrics.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/metrics/metrics.py to metrics.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/metrics/map_utils.py to map_utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/metrics/mot_metrics.py to mot_metrics.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/slim/ofa.py to ofa.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/slim/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/slim/distill.py to distill.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/slim/quant.py to quant.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/slim/prune.py to prune.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/slim/unstructured_prune.py to unstructured_prune.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/engine/trainer.py to trainer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/engine/tracker.py to tracker.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/engine/env.py to env.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/engine/callbacks.py to callbacks.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/engine/export_utils.py to export_utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/engine/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/utils/checkpoint.py to checkpoint.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/utils/download.py to download.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/utils/check.py to check.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/utils/stats.py to stats.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/utils/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/utils/logger.py to logger.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/utils/colormap.py to colormap.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/utils/profiler.py to profiler.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/utils/voc_utils.py to voc_utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/utils/cli.py to cli.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/utils/visualizer.py to visualizer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/core/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/core/config/schema.py to schema.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/core/config/yaml_helpers.py to yaml_helpers.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/core/config/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/core/workspace.py to workspace.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/version.py to version.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/optimizer.py to optimizer.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/shm_utils.py to shm_utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/reader.py to reader.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/transform/operators.py to operators.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/transform/autoaugment_utils.py to autoaugment_utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/transform/batch_operators.py to batch_operators.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/transform/keypoint_operators.py to keypoint_operators.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/transform/gridmask_utils.py to gridmask_utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/transform/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/transform/atss_assigner.py to atss_assigner.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/transform/mot_operators.py to mot_operators.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/transform/op_helper.py to op_helper.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/crop_utils/annotation_cropper.py to annotation_cropper.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/crop_utils/chip_box_utils.py to chip_box_utils.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/crop_utils/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/source/keypoint_coco.py to keypoint_coco.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/source/mot.py to mot.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/source/voc.py to voc.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/source/coco.py to coco.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/source/widerface.py to widerface.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/source/__init__.py to __init__.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/source/category.py to category.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/source/sniper_coco.py to sniper_coco.cpython-37.pyc
byte-compiling build/bdist.linux-x86_64/egg/ppdet/data/source/dataset.py to dataset.cpython-37.pyc
creating build/bdist.linux-x86_64/egg/EGG-INFO
copying paddledet.egg-info/PKG-INFO -> build/bdist.linux-x86_64/egg/EGG-INFO
copying paddledet.egg-info/SOURCES.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying paddledet.egg-info/dependency_links.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying paddledet.egg-info/requires.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
copying paddledet.egg-info/top_level.txt -> build/bdist.linux-x86_64/egg/EGG-INFO
zip_safe flag not set; analyzing archive contents...
ppdet.data.transform.__pycache__.autoaugment_utils.cpython-37: module MAY be using inspect.stack
ppdet.modeling.tests.__pycache__.test_mstest.cpython-37: module references __file__
ppdet.modeling.tests.__pycache__.test_ops.cpython-37: module references __file__
ppdet.modeling.tests.__pycache__.test_yolov3_loss.cpython-37: module references __file__
creating 'dist/paddledet-2.4.0-py3.7.egg' and adding 'build/bdist.linux-x86_64/egg' to it
removing 'build/bdist.linux-x86_64/egg' (and everything under it)
Processing paddledet-2.4.0-py3.7.egg
creating /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddledet-2.4.0-py3.7.egg
Extracting paddledet-2.4.0-py3.7.egg to /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Adding paddledet 2.4.0 to easy-install.pth file

Installed /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddledet-2.4.0-py3.7.egg
Processing dependencies for paddledet==2.4.0
Searching for typeguard==2.13.3
Best match: typeguard 2.13.3
Adding typeguard 2.13.3 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for visualdl==2.2.0
Best match: visualdl 2.2.0
Adding visualdl 2.2.0 to easy-install.pth file
Installing visualdl script to /opt/conda/envs/python35-paddle120-env/bin

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for openpyxl==3.0.5
Best match: openpyxl 3.0.5
Adding openpyxl 3.0.5 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for motmetrics==1.2.5
Best match: motmetrics 1.2.5
Adding motmetrics 1.2.5 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for sklearn==0.0
Best match: sklearn 0.0
Adding sklearn 0.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for lap==0.4.0
Best match: lap 0.4.0
Adding lap 0.4.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for setuptools==56.2.0
Best match: setuptools 56.2.0
Adding setuptools 56.2.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for pycocotools==2.0.6
Best match: pycocotools 2.0.6
Adding pycocotools 2.0.6 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for Cython==0.29
Best match: Cython 0.29
Adding Cython 0.29 to easy-install.pth file
Installing cygdb script to /opt/conda/envs/python35-paddle120-env/bin
Installing cython script to /opt/conda/envs/python35-paddle120-env/bin
Installing cythonize script to /opt/conda/envs/python35-paddle120-env/bin

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for terminaltables==3.1.10
Best match: terminaltables 3.1.10
Adding terminaltables 3.1.10 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for scipy==1.6.3
Best match: scipy 1.6.3
Adding scipy 1.6.3 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for Shapely==1.8.5.post1
Best match: Shapely 1.8.5.post1
Adding Shapely 1.8.5.post1 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for PyYAML==5.1.2
Best match: PyYAML 5.1.2
Adding PyYAML 5.1.2 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for opencv-python==4.1.1.26
Best match: opencv-python 4.1.1.26
Adding opencv-python 4.1.1.26 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for tqdm==4.36.1
Best match: tqdm 4.36.1
Adding tqdm 4.36.1 to easy-install.pth file
Installing tqdm script to /opt/conda/envs/python35-paddle120-env/bin

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for numpy==1.20.3
Best match: numpy 1.20.3
Adding numpy 1.20.3 to easy-install.pth file
Installing f2py script to /opt/conda/envs/python35-paddle120-env/bin
Installing f2py3 script to /opt/conda/envs/python35-paddle120-env/bin
Installing f2py3.7 script to /opt/conda/envs/python35-paddle120-env/bin

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for pandas==1.1.5
Best match: pandas 1.1.5
Adding pandas 1.1.5 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for flake8==4.0.1
Best match: flake8 4.0.1
Adding flake8 4.0.1 to easy-install.pth file
Installing flake8 script to /opt/conda/envs/python35-paddle120-env/bin

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for Pillow==7.1.2
Best match: Pillow 7.1.2
Adding Pillow 7.1.2 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for six==1.16.0
Best match: six 1.16.0
Adding six 1.16.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for Flask==1.1.1
Best match: Flask 1.1.1
Adding Flask 1.1.1 to easy-install.pth file
Installing flask script to /opt/conda/envs/python35-paddle120-env/bin

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for protobuf==3.20.1
Best match: protobuf 3.20.1
Adding protobuf 3.20.1 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for pre-commit==1.21.0
Best match: pre-commit 1.21.0
Adding pre-commit 1.21.0 to easy-install.pth file
Installing pre-commit script to /opt/conda/envs/python35-paddle120-env/bin
Installing pre-commit-validate-config script to /opt/conda/envs/python35-paddle120-env/bin
Installing pre-commit-validate-manifest script to /opt/conda/envs/python35-paddle120-env/bin

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for Flask-Babel==1.0.0
Best match: Flask-Babel 1.0.0
Adding Flask-Babel 1.0.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for matplotlib==2.2.3
Best match: matplotlib 2.2.3
Adding matplotlib 2.2.3 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for requests==2.22.0
Best match: requests 2.22.0
Adding requests 2.22.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for shellcheck-py==0.7.1.1
Best match: shellcheck-py 0.7.1.1
Adding shellcheck-py 0.7.1.1 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for bce-python-sdk==0.8.53
Best match: bce-python-sdk 0.8.53
Adding bce-python-sdk 0.8.53 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for et-xmlfile==1.0.1
Best match: et-xmlfile 1.0.1
Adding et-xmlfile 1.0.1 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for jdcal==1.4.1
Best match: jdcal 1.4.1
Adding jdcal 1.4.1 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for xmltodict==0.13.0
Best match: xmltodict 0.13.0
Adding xmltodict 0.13.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for scikit-learn==0.24.2
Best match: scikit-learn 0.24.2
Adding scikit-learn 0.24.2 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for python-dateutil==2.8.2
Best match: python-dateutil 2.8.2
Adding python-dateutil 2.8.2 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for pytz==2019.3
Best match: pytz 2019.3
Adding pytz 2019.3 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for importlib-metadata==4.2.0
Best match: importlib-metadata 4.2.0
Adding importlib-metadata 4.2.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for pyflakes==2.4.0
Best match: pyflakes 2.4.0
Adding pyflakes 2.4.0 to easy-install.pth file
Installing pyflakes script to /opt/conda/envs/python35-paddle120-env/bin

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for pycodestyle==2.8.0
Best match: pycodestyle 2.8.0
Adding pycodestyle 2.8.0 to easy-install.pth file
Installing pycodestyle script to /opt/conda/envs/python35-paddle120-env/bin

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for mccabe==0.6.1
Best match: mccabe 0.6.1
Adding mccabe 0.6.1 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for Click==7.0
Best match: Click 7.0
Adding Click 7.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for itsdangerous==1.1.0
Best match: itsdangerous 1.1.0
Adding itsdangerous 1.1.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for Werkzeug==0.16.0
Best match: Werkzeug 0.16.0
Adding Werkzeug 0.16.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for Jinja2==3.0.0
Best match: Jinja2 3.0.0
Adding Jinja2 3.0.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for virtualenv==16.7.9
Best match: virtualenv 16.7.9
Adding virtualenv 16.7.9 to easy-install.pth file
Installing virtualenv script to /opt/conda/envs/python35-paddle120-env/bin

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for aspy.yaml==1.3.0
Best match: aspy.yaml 1.3.0
Adding aspy.yaml 1.3.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for identify==1.4.10
Best match: identify 1.4.10
Adding identify 1.4.10 to easy-install.pth file
Installing identify-cli script to /opt/conda/envs/python35-paddle120-env/bin

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for nodeenv==1.3.4
Best match: nodeenv 1.3.4
Adding nodeenv 1.3.4 to easy-install.pth file
Installing nodeenv script to /opt/conda/envs/python35-paddle120-env/bin

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for toml==0.10.0
Best match: toml 0.10.0
Adding toml 0.10.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for cfgv==2.0.1
Best match: cfgv 2.0.1
Adding cfgv 2.0.1 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for Babel==2.8.0
Best match: Babel 2.8.0
Adding Babel 2.8.0 to easy-install.pth file
Installing pybabel script to /opt/conda/envs/python35-paddle120-env/bin

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for pyparsing==3.0.9
Best match: pyparsing 3.0.9
Adding pyparsing 3.0.9 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for kiwisolver==1.1.0
Best match: kiwisolver 1.1.0
Adding kiwisolver 1.1.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for cycler==0.10.0
Best match: cycler 0.10.0
Adding cycler 0.10.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for certifi==2019.9.11
Best match: certifi 2019.9.11
Adding certifi 2019.9.11 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for urllib3==1.25.6
Best match: urllib3 1.25.6
Adding urllib3 1.25.6 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for chardet==3.0.4
Best match: chardet 3.0.4
Adding chardet 3.0.4 to easy-install.pth file
Installing chardetect script to /opt/conda/envs/python35-paddle120-env/bin

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for idna==2.8
Best match: idna 2.8
Adding idna 2.8 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for pycryptodome==3.9.9
Best match: pycryptodome 3.9.9
Adding pycryptodome 3.9.9 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for future==0.18.0
Best match: future 0.18.0
Adding future 0.18.0 to easy-install.pth file
Installing futurize script to /opt/conda/envs/python35-paddle120-env/bin
Installing pasteurize script to /opt/conda/envs/python35-paddle120-env/bin

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for joblib==0.14.1
Best match: joblib 0.14.1
Adding joblib 0.14.1 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for threadpoolctl==2.1.0
Best match: threadpoolctl 2.1.0
Adding threadpoolctl 2.1.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for typing-extensions==4.3.0
Best match: typing-extensions 4.3.0
Adding typing-extensions 4.3.0 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for zipp==3.8.1
Best match: zipp 3.8.1
Adding zipp 3.8.1 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Searching for MarkupSafe==2.0.1
Best match: MarkupSafe 2.0.1
Adding MarkupSafe 2.0.1 to easy-install.pth file

Using /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages
Finished processing dependencies for paddledet==2.4.0
登录后复制
       

这时候可能有人会问为什么安装两次,因为直接运行setup.py的话可能会卡住,他下载不会自动换源。先用pip安装好了之后,再运行就不会卡住了。

多面-AI面试
多面-AI面试

猎聘推出的AI面试平台

多面-AI面试32
查看详情 多面-AI面试
In [ ]
# 一行代码训练,爽# !export CUDA_VISIBLE_DEVICES=0,1,2,3!python3 tools/train.py -c configs/picodet/picodet_xs_416_coco_lcnet.yml\
     --use_vdl True\
     --eval\
     --vdl_log_dir output/vdl_picodet_xs/
登录后复制
   

训练好后测试

这里我们直接用训练好的结果进行预测,多运行几次发现,这个FPS居然可以从20蹦到26,27...不是很靠谱

In [5]
# 预测!python tools/eval.py -c configs/picodet/picodet_xs_416_coco_lcnet.yml \
              -o weights=output/picodet_xs_416_coco_lcnet/best_model.pdparams
登录后复制
       
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/tensor/creation.py:130: DeprecationWarning: `np.object` is a deprecated alias for the builtin `object`. To silence this warning, use `object` by itself. Doing this will not modify any behavior and is safe. 
Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
  if data.dtype == np.object:
W0510 14:21:03.815788  1238 device_context.cc:447] Please NOTE: device: 0, GPU Compute Capability: 7.0, Driver API Version: 10.1, Runtime API Version: 10.1
W0510 14:21:03.821547  1238 device_context.cc:465] device: 0, cuDNN Version: 7.6.
[05/10 14:21:08] ppdet.utils.checkpoint INFO: Finish loading model weights: output/picodet_xs_416_coco_lcnet/best_model.pdparams
[05/10 14:21:08] ppdet.engine INFO: Eval iter: 0
[05/10 14:21:11] ppdet.metrics.metrics INFO: Accumulating evaluatation results...
[05/10 14:21:11] ppdet.metrics.metrics INFO: mAP(0.50, integral) = 94.56%
[05/10 14:21:11] ppdet.engine INFO: Total sample number: 88, averge FPS: 26.069633915698386
登录后复制
       

⭐对于FPS和mAp的讨论

可见动态图下mAp达到了90+

我们的目标可是FPS30+,现在才27,还差3个点。

模型参数量已经很小了,xs_416版本的大小仅有2.7M,这还跑不到35FPS ?

此时我们决定问郑院士怎么办,他建议先导出为静态图看看,我觉得也是...

后来导出静态图后,我们尝试使用trt_int8加速部署,AiStudio上编译trt版本上的好像不支持(主要是没找到相关项目),暂时搁置

【AI workshop】增强的PicoDet有多猛?来,跑个RM数据集试试! - php中文网        

In [ ]
# 导出模型为静态图!python tools/export_model.py -c configs/picodet/picodet_xs_416_coco_lcnet.yml \
              -o weights=output/picodet_xs_416_coco_lcnet/best_model.pdparams \
              --output_dir=output_inference
登录后复制
   

输出,鉴定为静态图

output
  ├── deploy.yaml            # 部署相关的配置文件
  ├── model.pdiparams        # 静态图模型参数
  ├── model.pdiparams.info   # 参数额外信息,一般无需关注
  └── model.pdmodel          # 静态图模型文件
登录后复制
   

我们分了三种方式执行静态图的预测:

  • CPU+4 Thread
  • CPU+MKL+4 Thread
  • GPU

分别用两种模型结果做测试:

  • picodet_l_640_coco_lcnet
  • picodet_xs_416_coco_lcnet

下面三行为测试代码

In [10]
!python deploy/python/infer.py --model_dir=output_inference/picodet_xs_416_coco_lcnet --video_file 1.mp4 --device=CPU --batch_size=1 --cpu_threads=4 --enable_mkldnn=True --threshold=0.64
登录后复制
       
class_id:0, confidence:0.7234, left_top:[537.94,467.07],right_bottom:[661.59,612.36]
class_id:1, confidence:0.8537, left_top:[434.49,51.53],right_bottom:[1288.93,776.08]
class_id:4, confidence:0.6614, left_top:[976.28,482.88],right_bottom:[1138.28,619.30]
detect frame: 363
class_id:2, confidence:0.8559, left_top:[453.89,43.24],right_bottom:[1283.73,778.66]
class_id:4, confidence:0.7588, left_top:[923.57,479.43],right_bottom:[1109.49,630.18]
detect frame: 364
class_id:2, confidence:0.8736, left_top:[484.57,41.56],right_bottom:[1261.76,769.17]
class_id:4, confidence:0.8447, left_top:[864.16,487.46],right_bottom:[1073.51,635.44]
detect frame: 365
class_id:2, confidence:0.8676, left_top:[505.99,48.44],right_bottom:[1242.18,778.10]
class_id:4, confidence:0.7879, left_top:[802.38,497.31],right_bottom:[1022.15,641.81]
detect frame: 366
class_id:2, confidence:0.8714, left_top:[475.93,52.22],right_bottom:[1209.77,781.14]
class_id:4, confidence:0.7459, left_top:[742.96,510.40],right_bottom:[963.97,642.25]
detect frame: 367
class_id:0, confidence:0.7737, left_top:[684.12,509.31],right_bottom:[904.65,649.57]
class_id:1, confidence:0.8269, left_top:[457.69,45.17],right_bottom:[1168.65,786.54]
detect frame: 368
class_id:0, confidence:0.7930, left_top:[623.83,507.65],right_bottom:[850.24,648.36]
class_id:1, confidence:0.8313, left_top:[440.52,45.95],right_bottom:[1146.75,786.18]
detect frame: 369
class_id:2, confidence:0.7943, left_top:[431.26,43.75],right_bottom:[1175.74,788.21]
class_id:4, confidence:0.6610, left_top:[1067.82,446.65],right_bottom:[1132.55,611.89]
detect frame: 370
class_id:0, confidence:0.7595, left_top:[544.45,511.62],right_bottom:[727.46,644.25]
class_id:1, confidence:0.8542, left_top:[405.28,46.02],right_bottom:[1218.14,789.09]
detect frame: 371
class_id:0, confidence:0.8163, left_top:[515.23,506.74],right_bottom:[673.16,638.92]
class_id:0, confidence:0.7437, left_top:[1001.43,453.65],right_bottom:[1124.84,606.06]
class_id:1, confidence:0.8718, left_top:[412.50,55.59],right_bottom:[1267.33,787.53]
detect frame: 372
class_id:0, confidence:0.7881, left_top:[488.46,491.15],right_bottom:[626.15,636.36]
class_id:0, confidence:0.7020, left_top:[966.59,466.03],right_bottom:[1109.25,610.04]
class_id:1, confidence:0.8656, left_top:[407.37,54.01],right_bottom:[1301.85,781.47]
detect frame: 373
class_id:0, confidence:0.7079, left_top:[484.59,484.35],right_bottom:[589.26,637.62]
class_id:1, confidence:0.8466, left_top:[403.27,55.04],right_bottom:[1318.05,780.14]
detect frame: 374
class_id:1, confidence:0.6459, left_top:[417.92,42.50],right_bottom:[1318.36,762.57]
class_id:2, confidence:0.8375, left_top:[414.25,51.29],right_bottom:[1328.15,770.10]
class_id:4, confidence:0.7822, left_top:[871.91,480.46],right_bottom:[1056.86,629.74]
detect frame: 375
class_id:1, confidence:0.6686, left_top:[435.20,47.28],right_bottom:[1307.87,757.52]
class_id:2, confidence:0.7959, left_top:[447.50,50.66],right_bottom:[1305.19,770.54]
class_id:4, confidence:0.7283, left_top:[823.12,486.60],right_bottom:[1015.46,635.79]
detect frame: 376
class_id:1, confidence:0.7167, left_top:[451.56,51.97],right_bottom:[1291.35,773.97]
class_id:2, confidence:0.7003, left_top:[452.67,49.70],right_bottom:[1274.31,775.07]
class_id:4, confidence:0.7641, left_top:[771.40,488.90],right_bottom:[984.43,635.92]
detect frame: 377
class_id:0, confidence:0.6612, left_top:[719.23,491.59],right_bottom:[940.80,635.20]
class_id:1, confidence:0.8560, left_top:[426.17,58.07],right_bottom:[1262.59,775.72]
detect frame: 378
class_id:2, confidence:0.8486, left_top:[410.04,55.50],right_bottom:[1218.59,770.87]
class_id:4, confidence:0.7705, left_top:[670.92,495.73],right_bottom:[894.91,643.01]
detect frame: 379
class_id:2, confidence:0.8722, left_top:[406.94,66.77],right_bottom:[1205.68,772.33]
class_id:4, confidence:0.7798, left_top:[633.79,499.77],right_bottom:[847.05,638.84]
detect frame: 380
class_id:2, confidence:0.8561, left_top:[410.47,72.69],right_bottom:[1245.18,779.02]
class_id:4, confidence:0.7315, left_top:[610.74,510.59],right_bottom:[795.30,630.14]
class_id:4, confidence:0.6828, left_top:[1092.73,457.28],right_bottom:[1174.88,618.27]
detect frame: 381
class_id:2, confidence:0.8759, left_top:[418.20,73.65],right_bottom:[1256.67,787.03]
class_id:4, confidence:0.7660, left_top:[582.69,503.35],right_bottom:[751.46,629.03]
class_id:4, confidence:0.7062, left_top:[1065.70,457.09],right_bottom:[1172.20,623.73]
detect frame: 382
class_id:0, confidence:0.7039, left_top:[562.51,485.60],right_bottom:[713.22,629.59]
class_id:2, confidence:0.8289, left_top:[425.59,69.50],right_bottom:[1277.43,797.34]
detect frame: 383
class_id:2, confidence:0.8610, left_top:[438.34,66.90],right_bottom:[1305.29,799.82]
class_id:4, confidence:0.6893, left_top:[986.62,470.75],right_bottom:[1148.64,630.83]
detect frame: 384
class_id:0, confidence:0.7393, left_top:[531.78,479.24],right_bottom:[648.78,622.96]
class_id:2, confidence:0.8673, left_top:[443.01,64.68],right_bottom:[1305.87,798.33]
class_id:4, confidence:0.7584, left_top:[942.63,478.13],right_bottom:[1121.61,636.15]
detect frame: 385
class_id:2, confidence:0.8762, left_top:[475.03,69.59],right_bottom:[1292.75,794.45]
class_id:4, confidence:0.7801, left_top:[891.48,484.01],right_bottom:[1087.84,640.95]
detect frame: 386
class_id:2, confidence:0.8739, left_top:[500.77,62.65],right_bottom:[1269.87,788.26]
class_id:4, confidence:0.8065, left_top:[836.94,492.39],right_bottom:[1044.85,650.05]
detect frame: 387
class_id:2, confidence:0.8762, left_top:[522.07,67.38],right_bottom:[1238.43,790.86]
class_id:4, confidence:0.8333, left_top:[774.48,503.60],right_bottom:[995.56,650.56]
detect frame: 388
class_id:2, confidence:0.8803, left_top:[480.60,71.52],right_bottom:[1194.81,794.85]
class_id:4, confidence:0.8097, left_top:[731.68,509.81],right_bottom:[945.02,658.52]
detect frame: 389
class_id:2, confidence:0.8725, left_top:[456.92,65.87],right_bottom:[1159.66,796.80]
class_id:4, confidence:0.7294, left_top:[671.99,507.85],right_bottom:[893.47,653.09]
detect frame: 390
class_id:2, confidence:0.8684, left_top:[452.09,66.86],right_bottom:[1155.54,800.14]
class_id:4, confidence:0.7522, left_top:[631.03,514.86],right_bottom:[831.74,648.87]
detect frame: 391
class_id:2, confidence:0.8767, left_top:[447.20,67.53],right_bottom:[1186.18,795.70]
class_id:4, confidence:0.7308, left_top:[593.96,509.74],right_bottom:[784.41,651.03]
class_id:4, confidence:0.6851, left_top:[1066.66,455.35],right_bottom:[1141.32,612.88]
detect frame: 392
class_id:1, confidence:0.7201, left_top:[437.83,68.88],right_bottom:[1208.11,790.21]
class_id:4, confidence:0.7122, left_top:[564.52,516.50],right_bottom:[731.50,643.49]
detect frame: 393
class_id:0, confidence:0.7068, left_top:[537.20,501.41],right_bottom:[684.80,634.42]
class_id:0, confidence:0.6455, left_top:[998.47,471.59],right_bottom:[1123.26,610.38]
class_id:1, confidence:0.8517, left_top:[433.46,69.95],right_bottom:[1266.55,790.43]
detect frame: 394
class_id:2, confidence:0.7295, left_top:[413.76,74.16],right_bottom:[1301.10,794.48]
detect frame: 395
class_id:2, confidence:0.7992, left_top:[417.57,76.96],right_bottom:[1326.24,793.08]
class_id:4, confidence:0.6836, left_top:[929.59,486.10],right_bottom:[1101.96,625.12]
detect frame: 396
class_id:2, confidence:0.8644, left_top:[444.38,81.94],right_bottom:[1316.61,780.00]
class_id:4, confidence:0.7633, left_top:[888.60,492.33],right_bottom:[1078.07,628.87]
detect frame: 397
class_id:2, confidence:0.8753, left_top:[470.87,86.75],right_bottom:[1320.45,777.40]
class_id:4, confidence:0.7915, left_top:[847.17,497.80],right_bottom:[1049.44,634.19]
detect frame: 398
class_id:2, confidence:0.8696, left_top:[455.21,87.31],right_bottom:[1298.91,781.99]
class_id:4, confidence:0.7642, left_top:[805.43,501.95],right_bottom:[1009.39,642.21]
detect frame: 399
class_id:1, confidence:0.7083, left_top:[447.84,79.51],right_bottom:[1281.60,779.87]
class_id:4, confidence:0.6893, left_top:[754.52,501.96],right_bottom:[973.30,642.75]
detect frame: 400
class_id:1, confidence:0.8503, left_top:[432.04,75.26],right_bottom:[1265.56,781.25]
detect frame: 401
class_id:0, confidence:0.6569, left_top:[666.19,505.69],right_bottom:[876.08,645.77]
class_id:1, confidence:0.8317, left_top:[421.65,75.97],right_bottom:[1255.56,780.25]
detect frame: 402
class_id:2, confidence:0.8001, left_top:[411.59,79.40],right_bottom:[1250.29,785.50]
class_id:4, confidence:0.7200, left_top:[644.79,509.81],right_bottom:[833.14,639.69]
detect frame: 403
class_id:1, confidence:0.6504, left_top:[418.27,79.68],right_bottom:[1282.23,795.65]
class_id:2, confidence:0.7000, left_top:[431.48,81.10],right_bottom:[1281.47,797.34]
class_id:4, confidence:0.6999, left_top:[624.32,515.24],right_bottom:[795.67,638.99]
detect frame: 404
class_id:0, confidence:0.7637, left_top:[606.81,492.35],right_bottom:[765.75,630.70]
class_id:0, confidence:0.6594, left_top:[1103.94,475.78],right_bottom:[1206.29,627.88]
class_id:1, confidence:0.8657, left_top:[427.23,73.70],right_bottom:[1296.38,791.18]
detect frame: 405
class_id:1, confidence:0.8525, left_top:[425.18,71.98],right_bottom:[1301.76,789.77]
detect frame: 406
class_id:0, confidence:0.6754, left_top:[1103.85,484.21],right_bottom:[1208.24,631.20]
class_id:1, confidence:0.8401, left_top:[421.69,72.13],right_bottom:[1298.47,800.06]
detect frame: 407
class_id:0, confidence:0.6718, left_top:[609.69,494.78],right_bottom:[760.37,632.80]
class_id:0, confidence:0.6446, left_top:[1110.58,485.63],right_bottom:[1205.62,627.25]
class_id:1, confidence:0.8653, left_top:[426.03,70.42],right_bottom:[1305.05,789.30]
detect frame: 408
class_id:0, confidence:0.7999, left_top:[609.51,503.06],right_bottom:[768.25,638.15]
class_id:1, confidence:0.8625, left_top:[421.58,71.24],right_bottom:[1302.60,801.41]
detect frame: 409
class_id:0, confidence:0.6445, left_top:[1116.68,485.56],right_bottom:[1211.87,630.26]
class_id:1, confidence:0.7984, left_top:[415.91,72.74],right_bottom:[1296.32,801.04]
detect frame: 410
class_id:0, confidence:0.6512, left_top:[620.13,502.08],right_bottom:[790.24,639.38]
class_id:1, confidence:0.7571, left_top:[427.30,76.38],right_bottom:[1295.13,798.19]
detect frame: 411
class_id:1, confidence:0.7940, left_top:[417.25,76.32],right_bottom:[1294.81,800.70]
detect frame: 412
class_id:1, confidence:0.7016, left_top:[417.45,77.44],right_bottom:[1291.82,801.59]
detect frame: 413
class_id:1, confidence:0.7256, left_top:[415.79,77.68],right_bottom:[1291.64,801.00]
detect frame: 414
class_id:1, confidence:0.7999, left_top:[422.07,76.63],right_bottom:[1294.75,803.58]
detect frame: 415
class_id:1, confidence:0.7963, left_top:[422.75,79.05],right_bottom:[1298.31,803.23]
detect frame: 416
class_id:0, confidence:0.6405, left_top:[1133.98,471.90],right_bottom:[1228.00,641.19]
class_id:1, confidence:0.8276, left_top:[420.84,80.25],right_bottom:[1298.68,803.66]
detect frame: 417
class_id:1, confidence:0.8327, left_top:[421.31,80.38],right_bottom:[1297.06,804.00]
detect frame: 418
class_id:1, confidence:0.8291, left_top:[423.16,80.75],right_bottom:[1294.99,804.49]
detect frame: 419
class_id:1, confidence:0.8278, left_top:[423.58,81.29],right_bottom:[1296.24,804.43]
detect frame: 420
class_id:1, confidence:0.8243, left_top:[423.08,82.00],right_bottom:[1295.49,803.66]
detect frame: 421
class_id:1, confidence:0.8279, left_top:[422.64,82.39],right_bottom:[1293.90,803.68]
detect frame: 422
class_id:1, confidence:0.8152, left_top:[423.19,83.09],right_bottom:[1293.32,803.76]
detect frame: 423
class_id:1, confidence:0.8126, left_top:[424.26,83.49],right_bottom:[1293.84,803.86]
detect frame: 424
class_id:1, confidence:0.8073, left_top:[425.78,84.89],right_bottom:[1294.18,804.02]
detect frame: 425
class_id:1, confidence:0.7982, left_top:[424.90,85.21],right_bottom:[1293.81,803.88]
detect frame: 426
class_id:1, confidence:0.7969, left_top:[425.88,84.62],right_bottom:[1293.61,803.71]
detect frame: 427
class_id:1, confidence:0.7505, left_top:[430.32,87.07],right_bottom:[1302.88,801.82]
detect frame: 428
class_id:1, confidence:0.7492, left_top:[425.52,85.88],right_bottom:[1294.59,803.35]
detect frame: 429
class_id:1, confidence:0.7260, left_top:[428.82,87.92],right_bottom:[1301.24,801.30]
detect frame: 430
class_id:1, confidence:0.7207, left_top:[430.81,88.77],right_bottom:[1303.42,802.58]
detect frame: 431
class_id:1, confidence:0.6864, left_top:[432.09,89.36],right_bottom:[1303.23,804.60]
detect frame: 432
class_id:1, confidence:0.7946, left_top:[426.19,85.73],right_bottom:[1306.04,805.64]
detect frame: 433
class_id:0, confidence:0.7366, left_top:[620.40,504.42],right_bottom:[777.35,648.43]
class_id:0, confidence:0.6943, left_top:[1119.69,485.43],right_bottom:[1228.43,639.85]
class_id:1, confidence:0.8590, left_top:[428.01,83.23],right_bottom:[1314.61,805.16]
detect frame: 434
class_id:0, confidence:0.6917, left_top:[1114.37,484.42],right_bottom:[1213.59,638.79]
class_id:0, confidence:0.6850, left_top:[630.21,500.78],right_bottom:[767.51,640.76]
class_id:1, confidence:0.8702, left_top:[442.44,82.04],right_bottom:[1308.72,805.93]
detect frame: 435
class_id:0, confidence:0.6998, left_top:[607.95,504.45],right_bottom:[751.67,637.92]
class_id:1, confidence:0.7667, left_top:[445.64,83.27],right_bottom:[1329.57,802.27]
detect frame: 436
class_id:0, confidence:0.6785, left_top:[606.95,503.05],right_bottom:[733.09,637.98]
class_id:2, confidence:0.8688, left_top:[440.74,84.54],right_bottom:[1331.16,805.70]
detect frame: 437
class_id:2, confidence:0.6641, left_top:[459.10,81.66],right_bottom:[1327.97,803.01]
class_id:4, confidence:0.6674, left_top:[606.77,505.06],right_bottom:[725.43,642.50]
detect frame: 438
class_id:0, confidence:0.7535, left_top:[1040.20,496.26],right_bottom:[1195.59,647.84]
class_id:0, confidence:0.6489, left_top:[604.39,493.90],right_bottom:[717.33,637.56]
class_id:1, confidence:0.8547, left_top:[465.22,75.36],right_bottom:[1332.78,809.29]
detect frame: 439
class_id:0, confidence:0.7350, left_top:[1025.01,498.35],right_bottom:[1179.30,643.71]
class_id:0, confidence:0.6957, left_top:[586.29,489.77],right_bottom:[695.85,638.35]
class_id:1, confidence:0.8232, left_top:[467.44,78.23],right_bottom:[1336.52,807.23]
detect frame: 440
class_id:0, confidence:0.7296, left_top:[575.73,483.25],right_bottom:[684.67,637.47]
class_id:1, confidence:0.7955, left_top:[473.73,78.65],right_bottom:[1327.87,803.54]
detect frame: 441
class_id:0, confidence:0.6710, left_top:[978.18,512.10],right_bottom:[1158.04,656.46]
class_id:1, confidence:0.8435, left_top:[480.52,76.43],right_bottom:[1328.89,808.48]
detect frame: 442
detect frame: 443
class_id:2, confidence:0.6532, left_top:[513.51,85.13],right_bottom:[1328.04,803.38]
detect frame: 444
class_id:0, confidence:0.6823, left_top:[914.00,519.06],right_bottom:[1111.95,667.21]
class_id:1, confidence:0.8428, left_top:[519.65,83.68],right_bottom:[1325.82,809.79]
detect frame: 445
class_id:2, confidence:0.8625, left_top:[540.18,96.18],right_bottom:[1308.71,803.35]
class_id:4, confidence:0.7685, left_top:[880.39,520.73],right_bottom:[1086.38,669.44]
detect frame: 446
class_id:2, confidence:0.8315, left_top:[558.62,105.81],right_bottom:[1308.03,800.23]
class_id:4, confidence:0.7410, left_top:[866.00,515.96],right_bottom:[1068.33,669.00]
detect frame: 447
class_id:0, confidence:0.6881, left_top:[833.26,525.71],right_bottom:[1048.08,669.62]
class_id:1, confidence:0.8583, left_top:[544.21,102.34],right_bottom:[1299.36,803.15]
detect frame: 448
class_id:1, confidence:0.7728, left_top:[531.42,100.54],right_bottom:[1279.67,808.86]
detect frame: 449
class_id:0, confidence:0.7716, left_top:[785.58,529.83],right_bottom:[996.76,671.03]
class_id:1, confidence:0.8382, left_top:[513.40,100.70],right_bottom:[1263.30,808.90]
detect frame: 450
class_id:0, confidence:0.7325, left_top:[763.04,530.59],right_bottom:[972.34,673.34]
class_id:1, confidence:0.8453, left_top:[509.00,90.32],right_bottom:[1249.70,805.18]
detect frame: 451
class_id:0, confidence:0.7919, left_top:[732.61,528.44],right_bottom:[950.05,675.40]
class_id:1, confidence:0.8553, left_top:[491.72,78.90],right_bottom:[1232.93,808.18]
detect frame: 452
class_id:0, confidence:0.8275, left_top:[713.11,533.60],right_bottom:[920.10,672.94]
class_id:1, confidence:0.8672, left_top:[475.03,79.76],right_bottom:[1217.27,811.09]
detect frame: 453
class_id:0, confidence:0.7868, left_top:[691.94,533.22],right_bottom:[900.94,674.01]
class_id:1, confidence:0.8656, left_top:[463.64,82.62],right_bottom:[1202.50,815.24]
detect frame: 454
class_id:0, confidence:0.7894, left_top:[671.77,537.48],right_bottom:[871.11,674.17]
class_id:1, confidence:0.8609, left_top:[459.19,83.49],right_bottom:[1198.57,814.99]
detect frame: 455
class_id:0, confidence:0.8229, left_top:[645.68,535.09],right_bottom:[849.82,671.38]
class_id:1, confidence:0.8684, left_top:[449.99,77.22],right_bottom:[1208.77,815.45]
detect frame: 456
class_id:0, confidence:0.8110, left_top:[630.67,535.26],right_bottom:[828.32,674.14]
class_id:1, confidence:0.8624, left_top:[451.78,77.62],right_bottom:[1214.52,817.50]
detect frame: 457
class_id:0, confidence:0.7821, left_top:[605.17,531.70],right_bottom:[801.14,670.85]
class_id:1, confidence:0.8259, left_top:[444.87,75.12],right_bottom:[1236.44,819.64]
detect frame: 458
class_id:0, confidence:0.8143, left_top:[590.34,529.88],right_bottom:[771.94,668.07]
class_id:0, confidence:0.6902, left_top:[1084.89,480.23],right_bottom:[1159.31,632.27]
class_id:1, confidence:0.8674, left_top:[437.31,78.63],right_bottom:[1258.58,818.78]
detect frame: 459
class_id:0, confidence:0.8148, left_top:[575.46,533.70],right_bottom:[751.81,669.42]
class_id:1, confidence:0.8682, left_top:[439.11,79.58],right_bottom:[1270.13,821.32]
detect frame: 460
class_id:0, confidence:0.7785, left_top:[562.20,533.80],right_bottom:[724.86,667.16]
class_id:0, confidence:0.7301, left_top:[1051.30,481.18],right_bottom:[1156.36,631.40]
class_id:1, confidence:0.8695, left_top:[439.51,91.25],right_bottom:[1280.53,818.27]
detect frame: 461
class_id:0, confidence:0.7966, left_top:[549.15,534.05],right_bottom:[699.58,670.60]
class_id:0, confidence:0.6484, left_top:[1042.56,492.25],right_bottom:[1149.96,631.61]
class_id:1, confidence:0.8695, left_top:[433.12,88.11],right_bottom:[1299.93,815.36]
detect frame: 462
class_id:0, confidence:0.7747, left_top:[538.20,527.57],right_bottom:[683.67,664.84]
class_id:1, confidence:0.8702, left_top:[435.15,83.27],right_bottom:[1319.15,817.06]
detect frame: 463
class_id:0, confidence:0.7199, left_top:[530.32,521.51],right_bottom:[664.53,663.18]
class_id:0, confidence:0.6522, left_top:[1004.22,496.08],right_bottom:[1130.48,638.75]
class_id:1, confidence:0.8712, left_top:[446.14,78.40],right_bottom:[1322.55,813.52]
detect frame: 464
class_id:0, confidence:0.7597, left_top:[530.32,518.52],right_bottom:[648.63,663.94]
class_id:0, confidence:0.7363, left_top:[994.48,486.76],right_bottom:[1134.91,631.11]
class_id:1, confidence:0.8741, left_top:[454.73,81.40],right_bottom:[1330.91,810.76]
detect frame: 465
class_id:0, confidence:0.7840, left_top:[975.20,493.86],right_bottom:[1134.12,640.90]
class_id:0, confidence:0.7135, left_top:[532.56,517.96],right_bottom:[633.26,659.23]
class_id:1, confidence:0.8741, left_top:[453.99,79.83],right_bottom:[1336.01,806.19]
detect frame: 466
class_id:0, confidence:0.7872, left_top:[964.10,492.10],right_bottom:[1125.85,646.63]
class_id:0, confidence:0.7303, left_top:[519.54,509.32],right_bottom:[617.39,658.84]
class_id:1, confidence:0.8669, left_top:[454.72,77.79],right_bottom:[1340.18,806.98]
detect frame: 467
class_id:0, confidence:0.8131, left_top:[935.04,495.47],right_bottom:[1106.70,652.45]
class_id:0, confidence:0.6848, left_top:[519.43,510.18],right_bottom:[601.51,649.18]
class_id:1, confidence:0.8736, left_top:[458.03,83.05],right_bottom:[1342.85,802.25]
detect frame: 468
class_id:0, confidence:0.7914, left_top:[915.16,496.63],right_bottom:[1097.13,653.92]
class_id:1, confidence:0.8681, left_top:[463.42,81.20],right_bottom:[1343.06,798.11]
detect frame: 469
class_id:0, confidence:0.7343, left_top:[896.88,499.94],right_bottom:[1073.23,654.10]
class_id:1, confidence:0.8493, left_top:[472.55,88.71],right_bottom:[1344.77,794.12]
detect frame: 470
class_id:0, confidence:0.7510, left_top:[878.69,499.90],right_bottom:[1062.04,659.61]
class_id:1, confidence:0.8474, left_top:[481.47,92.05],right_bottom:[1351.08,802.35]
detect frame: 471
class_id:0, confidence:0.7428, left_top:[853.13,498.97],right_bottom:[1051.11,659.88]
class_id:1, confidence:0.8639, left_top:[489.11,97.93],right_bottom:[1345.26,798.90]
detect frame: 472
class_id:0, confidence:0.7576, left_top:[837.43,505.38],right_bottom:[1036.82,663.53]
class_id:1, confidence:0.8635, left_top:[495.22,99.17],right_bottom:[1336.32,797.31]
detect frame: 473
class_id:0, confidence:0.7406, left_top:[815.41,508.53],right_bottom:[1020.90,664.90]
class_id:1, confidence:0.8471, left_top:[484.51,94.24],right_bottom:[1336.17,796.63]
detect frame: 474
class_id:0, confidence:0.7378, left_top:[795.67,510.15],right_bottom:[1001.47,666.29]
class_id:1, confidence:0.8510, left_top:[473.96,93.56],right_bottom:[1316.45,795.83]
detect frame: 475
class_id:0, confidence:0.7116, left_top:[772.78,508.06],right_bottom:[984.33,668.94]
class_id:1, confidence:0.8627, left_top:[470.68,90.62],right_bottom:[1302.83,795.91]
detect frame: 476
class_id:0, confidence:0.8004, left_top:[750.06,507.84],right_bottom:[963.14,668.97]
class_id:1, confidence:0.8620, left_top:[457.08,88.95],right_bottom:[1288.90,793.14]
detect frame: 477
class_id:1, confidence:0.7772, left_top:[449.03,88.02],right_bottom:[1280.54,790.32]
detect frame: 478
class_id:0, confidence:0.8279, left_top:[711.55,507.71],right_bottom:[925.60,663.55]
class_id:1, confidence:0.8707, left_top:[440.80,87.75],right_bottom:[1268.97,790.13]
detect frame: 479
class_id:0, confidence:0.8126, left_top:[697.97,507.81],right_bottom:[900.70,664.49]
class_id:1, confidence:0.8695, left_top:[434.81,86.67],right_bottom:[1259.15,787.56]
detect frame: 480
class_id:0, confidence:0.7805, left_top:[673.40,512.29],right_bottom:[878.12,661.30]
class_id:1, confidence:0.8233, left_top:[439.05,91.09],right_bottom:[1250.07,787.68]
detect frame: 481
class_id:2, confidence:0.8564, left_top:[450.61,106.33],right_bottom:[1248.94,787.60]
class_id:4, confidence:0.7184, left_top:[665.97,513.07],right_bottom:[858.70,658.15]
class_id:4, confidence:0.6822, left_top:[1143.51,462.64],right_bottom:[1209.20,642.48]
detect frame: 482
class_id:0, confidence:0.7659, left_top:[649.95,510.18],right_bottom:[841.60,658.00]
class_id:1, confidence:0.8135, left_top:[442.66,92.72],right_bottom:[1264.23,796.18]
detect frame: 483
class_id:0, confidence:0.7302, left_top:[642.79,511.21],right_bottom:[820.50,653.56]
class_id:1, confidence:0.6491, left_top:[440.92,89.18],right_bottom:[1290.87,798.74]
detect frame: 484
class_id:0, confidence:0.6521, left_top:[628.86,511.75],right_bottom:[801.17,653.31]
class_id:2, confidence:0.8519, left_top:[458.27,103.09],right_bottom:[1269.62,804.93]
detect frame: 485
class_id:2, confidence:0.7898, left_top:[455.41,98.48],right_bottom:[1289.87,807.61]
detect frame: 486
class_id:1, confidence:0.6435, left_top:[462.05,89.84],right_bottom:[1314.63,800.51]
class_id:2, confidence:0.7062, left_top:[462.98,90.53],right_bottom:[1307.09,807.23]
detect frame: 487
class_id:0, confidence:0.7643, left_top:[606.27,499.81],right_bottom:[745.58,636.41]
class_id:1, confidence:0.6842, left_top:[463.62,88.23],right_bottom:[1324.49,797.36]
detect frame: 488
class_id:2, confidence:0.8678, left_top:[478.49,91.14],right_bottom:[1336.75,805.97]
detect frame: 489
class_id:2, confidence:0.8426, left_top:[488.79,89.42],right_bottom:[1341.68,808.19]
detect frame: 490
class_id:0, confidence:0.7072, left_top:[591.94,491.58],right_bottom:[704.21,638.65]
class_id:1, confidence:0.8297, left_top:[480.01,87.78],right_bottom:[1342.38,806.52]
detect frame: 491
class_id:0, confidence:0.6548, left_top:[580.57,492.41],right_bottom:[685.17,638.92]
class_id:1, confidence:0.7023, left_top:[484.98,84.34],right_bottom:[1349.74,801.60]
detect frame: 492
class_id:2, confidence:0.7427, left_top:[501.90,88.75],right_bottom:[1338.99,810.39]
detect frame: 493
class_id:2, confidence:0.8473, left_top:[517.21,87.72],right_bottom:[1327.97,812.67]
detect frame: 494
class_id:2, confidence:0.8678, left_top:[534.55,90.30],right_bottom:[1324.66,808.87]
class_id:4, confidence:0.6921, left_top:[926.60,500.16],right_bottom:[1112.62,665.58]
detect frame: 495
class_id:2, confidence:0.8713, left_top:[570.15,94.68],right_bottom:[1307.47,806.06]
class_id:4, confidence:0.6999, left_top:[901.08,498.76],right_bottom:[1098.70,666.78]
detect frame: 496
class_id:2, confidence:0.8673, left_top:[578.17,98.26],right_bottom:[1304.01,801.98]
class_id:4, confidence:0.6511, left_top:[885.31,500.48],right_bottom:[1077.85,668.66]
detect frame: 497
class_id:1, confidence:0.7459, left_top:[578.68,94.28],right_bottom:[1300.92,801.23]
detect frame: 498
class_id:2, confidence:0.7794, left_top:[574.79,98.60],right_bottom:[1276.75,801.44]
detect frame: 499
class_id:2, confidence:0.8796, left_top:[558.57,107.34],right_bottom:[1258.46,800.76]
class_id:4, confidence:0.7596, left_top:[811.23,517.39],right_bottom:[1011.62,672.30]
detect frame: 500
class_id:0, confidence:0.7099, left_top:[777.12,517.87],right_bottom:[984.80,672.62]
class_id:1, confidence:0.8069, left_top:[540.92,88.54],right_bottom:[1249.66,804.13]
detect frame: 501
class_id:2, confidence:0.8584, left_top:[518.91,102.50],right_bottom:[1234.88,807.32]
class_id:4, confidence:0.7395, left_top:[754.10,523.43],right_bottom:[967.90,669.04]
detect frame: 502
class_id:0, confidence:0.6598, left_top:[736.93,523.67],right_bottom:[937.47,674.17]
class_id:2, confidence:0.7278, left_top:[509.83,98.62],right_bottom:[1217.28,806.17]
detect frame: 503
class_id:0, confidence:0.7808, left_top:[711.50,526.09],right_bottom:[907.45,667.69]
class_id:1, confidence:0.8359, left_top:[509.91,83.98],right_bottom:[1206.81,806.49]
detect frame: 504
class_id:2, confidence:0.8559, left_top:[501.16,100.27],right_bottom:[1205.16,810.40]
class_id:4, confidence:0.6891, left_top:[699.58,526.18],right_bottom:[884.84,666.00]
detect frame: 505
class_id:0, confidence:0.6872, left_top:[675.17,527.77],right_bottom:[860.53,667.33]
class_id:2, confidence:0.7790, left_top:[480.82,105.76],right_bottom:[1226.79,809.41]
detect frame: 506
class_id:2, confidence:0.8386, left_top:[488.38,102.82],right_bottom:[1240.56,806.44]
class_id:4, confidence:0.7186, left_top:[1117.80,462.07],right_bottom:[1182.63,617.71]
detect frame: 507
class_id:2, confidence:0.8721, left_top:[484.63,105.57],right_bottom:[1246.08,806.03]
class_id:4, confidence:0.7226, left_top:[643.36,522.99],right_bottom:[819.05,663.19]
class_id:4, confidence:0.6774, left_top:[1104.13,469.51],right_bottom:[1183.09,623.68]
detect frame: 508
class_id:0, confidence:0.7612, left_top:[626.34,521.24],right_bottom:[795.78,662.28]
class_id:1, confidence:0.8419, left_top:[471.18,96.31],right_bottom:[1258.95,806.93]
detect frame: 509
class_id:1, confidence:0.8509, left_top:[470.94,102.69],right_bottom:[1266.38,805.31]
detect frame: 510
class_id:0, confidence:0.7313, left_top:[594.24,509.91],right_bottom:[757.46,655.00]
class_id:0, confidence:0.6771, left_top:[1067.70,480.87],right_bottom:[1170.00,625.40]
class_id:1, confidence:0.8690, left_top:[465.46,103.60],right_bottom:[1288.27,806.51]
detect frame: 511
class_id:0, confidence:0.7162, left_top:[1054.03,485.50],right_bottom:[1174.39,622.95]
class_id:0, confidence:0.6899, left_top:[591.40,517.22],right_bottom:[728.80,652.01]
class_id:1, confidence:0.8665, left_top:[458.89,101.28],right_bottom:[1317.49,803.75]
detect frame: 512
class_id:0, confidence:0.7097, left_top:[1035.60,492.41],right_bottom:[1159.91,631.42]
class_id:0, confidence:0.6529, left_top:[580.32,511.86],right_bottom:[716.12,652.42]
class_id:1, confidence:0.8649, left_top:[459.50,102.90],right_bottom:[1329.45,804.02]
detect frame: 513
class_id:0, confidence:0.6889, left_top:[573.05,510.88],right_bottom:[692.19,650.81]
class_id:0, confidence:0.6493, left_top:[1027.19,489.20],right_bottom:[1162.22,628.52]
class_id:1, confidence:0.8605, left_top:[458.70,103.39],right_bottom:[1337.28,804.38]
detect frame: 514
class_id:0, confidence:0.6985, left_top:[1004.54,491.02],right_bottom:[1161.67,634.72]
class_id:1, confidence:0.8545, left_top:[462.31,103.67],right_bottom:[1336.81,805.76]
detect frame: 515
class_id:0, confidence:0.6965, left_top:[990.06,493.85],right_bottom:[1156.60,636.30]
class_id:1, confidence:0.8675, left_top:[468.32,102.65],right_bottom:[1361.21,804.16]
detect frame: 516
class_id:0, confidence:0.7059, left_top:[972.36,494.19],right_bottom:[1151.96,645.57]
class_id:1, confidence:0.8267, left_top:[477.79,102.99],right_bottom:[1376.73,801.43]
detect frame: 517
class_id:0, confidence:0.7730, left_top:[963.96,496.23],right_bottom:[1134.33,642.41]
class_id:1, confidence:0.8619, left_top:[489.21,102.06],right_bottom:[1372.31,796.71]
detect frame: 518
class_id:0, confidence:0.6967, left_top:[931.53,509.08],right_bottom:[1115.48,648.60]
class_id:1, confidence:0.8400, left_top:[506.10,101.67],right_bottom:[1374.77,797.73]
detect frame: 519
class_id:0, confidence:0.7309, left_top:[918.45,507.36],right_bottom:[1104.87,649.93]
class_id:1, confidence:0.8323, left_top:[523.13,105.24],right_bottom:[1357.37,794.81]
detect frame: 520
class_id:2, confidence:0.7315, left_top:[524.39,109.40],right_bottom:[1364.12,793.29]
detect frame: 521
class_id:0, confidence:0.7707, left_top:[883.67,506.70],right_bottom:[1075.22,656.76]
class_id:1, confidence:0.8472, left_top:[524.82,102.03],right_bottom:[1354.32,796.38]
detect frame: 522
class_id:0, confidence:0.7942, left_top:[851.74,509.39],right_bottom:[1055.92,656.53]
class_id:1, confidence:0.8696, left_top:[513.41,100.44],right_bottom:[1340.85,792.16]
detect frame: 523
class_id:0, confidence:0.7818, left_top:[830.40,509.04],right_bottom:[1040.34,660.49]
class_id:1, confidence:0.8670, left_top:[502.07,100.75],right_bottom:[1335.59,796.15]
detect frame: 524
class_id:0, confidence:0.7144, left_top:[808.09,510.33],right_bottom:[1025.03,657.49]
class_id:1, confidence:0.8561, left_top:[495.97,100.80],right_bottom:[1325.63,799.96]
detect frame: 525
class_id:0, confidence:0.6575, left_top:[782.67,511.07],right_bottom:[1002.44,661.98]
class_id:1, confidence:0.8692, left_top:[490.08,99.46],right_bottom:[1318.47,798.76]
detect frame: 526
class_id:0, confidence:0.8002, left_top:[764.48,511.68],right_bottom:[985.96,659.69]
class_id:1, confidence:0.8745, left_top:[474.58,98.74],right_bottom:[1306.49,797.66]
detect frame: 527
class_id:0, confidence:0.7362, left_top:[739.51,514.48],right_bottom:[955.52,662.59]
class_id:1, confidence:0.8667, left_top:[472.00,103.01],right_bottom:[1289.00,799.93]
detect frame: 528
class_id:4, confidence:0.7260, left_top:[732.55,516.78],right_bottom:[933.77,660.19]
detect frame: 529
class_id:0, confidence:0.7792, left_top:[708.79,515.88],right_bottom:[906.94,665.30]
class_id:1, confidence:0.8300, left_top:[462.74,103.94],right_bottom:[1296.42,792.61]
detect frame: 530
class_id:0, confidence:0.7479, left_top:[700.25,516.12],right_bottom:[881.12,659.34]
class_id:1, confidence:0.7007, left_top:[458.46,106.64],right_bottom:[1308.28,794.83]
detect frame: 531
class_id:2, confidence:0.8397, left_top:[458.83,112.63],right_bottom:[1307.73,800.66]
detect frame: 532
class_id:0, confidence:0.7513, left_top:[667.85,516.83],right_bottom:[838.77,655.31]
class_id:1, confidence:0.7739, left_top:[463.58,104.47],right_bottom:[1307.45,807.84]
detect frame: 533
class_id:0, confidence:0.6895, left_top:[661.74,514.68],right_bottom:[821.37,651.36]
class_id:1, confidence:0.8321, left_top:[460.72,99.68],right_bottom:[1320.61,808.52]
detect frame: 534
class_id:0, confidence:0.7010, left_top:[1139.13,492.08],right_bottom:[1247.32,649.08]
class_id:1, confidence:0.7940, left_top:[470.08,95.88],right_bottom:[1336.43,805.88]
class_id:4, confidence:0.6485, left_top:[652.96,518.64],right_bottom:[798.13,651.46]
detect frame: 535
class_id:0, confidence:0.6896, left_top:[1120.61,493.35],right_bottom:[1231.37,649.50]
class_id:1, confidence:0.8134, left_top:[475.33,93.99],right_bottom:[1355.02,804.27]
class_id:4, confidence:0.7039, left_top:[638.34,521.68],right_bottom:[780.27,653.00]
detect frame: 536
class_id:0, confidence:0.6577, left_top:[1098.51,497.48],right_bottom:[1230.52,647.33]
class_id:2, confidence:0.6545, left_top:[473.02,92.74],right_bottom:[1344.40,807.88]
class_id:4, confidence:0.7043, left_top:[637.27,512.68],right_bottom:[759.46,649.82]
detect frame: 537
class_id:0, confidence:0.6476, left_top:[1077.90,498.88],right_bottom:[1227.35,654.75]
class_id:2, confidence:0.6537, left_top:[480.01,90.79],right_bottom:[1356.13,808.00]
detect frame: 538
class_id:0, confidence:0.6771, left_top:[618.46,507.30],right_bottom:[729.38,648.79]
class_id:0, confidence:0.6631, left_top:[1056.26,505.46],right_bottom:[1208.72,653.64]
class_id:1, confidence:0.8512, left_top:[480.51,82.64],right_bottom:[1349.39,811.34]
detect frame: 539
class_id:0, confidence:0.7710, left_top:[1031.32,510.54],right_bottom:[1202.28,657.00]
class_id:0, confidence:0.6561, left_top:[610.23,493.62],right_bottom:[720.12,639.01]
class_id:1, confidence:0.8682, left_top:[492.05,81.84],right_bottom:[1355.75,811.21]
detect frame: 540
class_id:0, confidence:0.6718, left_top:[597.01,499.15],right_bottom:[696.70,645.11]
class_id:1, confidence:0.8140, left_top:[498.69,82.02],right_bottom:[1352.99,810.82]
class_id:4, confidence:0.6510, left_top:[1012.55,517.79],right_bottom:[1187.75,660.81]
detect frame: 541
class_id:4, confidence:0.7099, left_top:[987.56,517.84],right_bottom:[1167.72,669.07]
detect frame: 542
class_id:2, confidence:0.8142, left_top:[526.10,90.58],right_bottom:[1349.06,808.94]
class_id:4, confidence:0.7120, left_top:[959.74,520.05],right_bottom:[1143.71,667.97]
class_id:4, confidence:0.6730, left_top:[602.84,494.83],right_bottom:[674.41,646.44]
detect frame: 543
class_id:2, confidence:0.8486, left_top:[551.04,94.31],right_bottom:[1345.66,808.44]
class_id:4, confidence:0.7524, left_top:[933.36,521.08],right_bottom:[1132.24,672.79]
detect frame: 544
class_id:0, confidence:0.6627, left_top:[909.95,525.81],right_bottom:[1107.46,673.95]
class_id:1, confidence:0.8574, left_top:[573.43,98.56],right_bottom:[1336.63,809.96]
detect frame: 545
class_id:1, confidence:0.7698, left_top:[579.98,92.00],right_bottom:[1311.61,812.95]
detect frame: 546
class_id:0, confidence:0.7220, left_top:[850.40,529.50],right_bottom:[1059.56,676.15]
class_id:1, confidence:0.8656, left_top:[576.76,93.84],right_bottom:[1300.58,813.59]
detect frame: 547
class_id:0, confidence:0.7040, left_top:[826.29,535.46],right_bottom:[1043.62,677.12]
class_id:1, confidence:0.8591, left_top:[561.78,95.05],right_bottom:[1288.83,816.25]
detect frame: 548
class_id:0, confidence:0.7537, left_top:[796.43,531.55],right_bottom:[1014.88,677.91]
class_id:1, confidence:0.8563, left_top:[537.31,87.79],right_bottom:[1272.71,814.43]
detect frame: 549
class_id:0, confidence:0.7704, left_top:[767.70,528.09],right_bottom:[991.14,680.38]
class_id:1, confidence:0.8607, left_top:[524.28,90.91],right_bottom:[1255.89,813.06]
detect frame: 550
class_id:0, confidence:0.7886, left_top:[750.07,525.45],right_bottom:[965.88,679.09]
class_id:1, confidence:0.8595, left_top:[513.56,85.75],right_bottom:[1247.42,815.22]
detect frame: 551
class_id:0, confidence:0.7817, left_top:[729.40,532.24],right_bottom:[934.98,681.70]
class_id:1, confidence:0.8636, left_top:[499.23,86.81],right_bottom:[1225.45,815.37]
detect frame: 552
class_id:0, confidence:0.7962, left_top:[700.10,539.42],right_bottom:[906.12,677.96]
class_id:1, confidence:0.8651, left_top:[484.12,85.22],right_bottom:[1224.42,819.87]
detect frame: 553
class_id:0, confidence:0.8308, left_top:[678.58,539.35],right_bottom:[884.62,676.21]
class_id:1, confidence:0.8676, left_top:[472.30,86.94],right_bottom:[1227.11,822.04]
detect frame: 554
class_id:0, confidence:0.8123, left_top:[662.05,538.30],right_bottom:[853.76,674.80]
class_id:1, confidence:0.8628, left_top:[468.36,79.36],right_bottom:[1231.92,821.42]
detect frame: 555
class_id:0, confidence:0.8269, left_top:[637.52,539.88],right_bottom:[830.59,676.24]
class_id:1, confidence:0.8592, left_top:[459.94,81.88],right_bottom:[1244.86,821.67]
class_id:4, confidence:0.6443, left_top:[1119.98,485.48],right_bottom:[1189.69,647.40]
detect frame: 556
class_id:0, confidence:0.8219, left_top:[622.49,536.42],right_bottom:[807.31,673.26]
class_id:1, confidence:0.8634, left_top:[458.33,80.76],right_bottom:[1258.80,819.09]
detect frame: 557
class_id:0, confidence:0.8209, left_top:[602.84,538.18],right_bottom:[784.52,673.19]
class_id:1, confidence:0.8640, left_top:[455.02,82.29],right_bottom:[1269.70,821.54]
detect frame: 558
class_id:0, confidence:0.7977, left_top:[591.53,535.49],right_bottom:[762.32,670.87]
class_id:1, confidence:0.8716, left_top:[458.94,91.18],right_bottom:[1278.90,822.13]
detect frame: 559
class_id:0, confidence:0.7916, left_top:[582.16,535.91],right_bottom:[737.90,669.77]
class_id:1, confidence:0.8730, left_top:[457.53,92.25],right_bottom:[1292.87,822.30]
detect frame: 560
class_id:0, confidence:0.7021, left_top:[576.75,532.66],right_bottom:[728.29,667.81]
class_id:0, confidence:0.6784, left_top:[1056.45,485.21],right_bottom:[1186.31,636.63]
class_id:1, confidence:0.8737, left_top:[465.34,90.39],right_bottom:[1308.46,816.29]
detect frame: 561
class_id:1, confidence:0.8655, left_top:[462.64,89.53],right_bottom:[1309.05,815.33]
detect frame: 562
class_id:0, confidence:0.6729, left_top:[575.35,531.21],right_bottom:[717.58,666.10]
class_id:1, confidence:0.8585, left_top:[461.03,87.50],right_bottom:[1321.74,814.87]
detect frame: 563
class_id:0, confidence:0.6735, left_top:[577.78,535.36],right_bottom:[713.62,665.40]
class_id:1, confidence:0.8672, left_top:[467.01,92.82],right_bottom:[1324.88,815.61]
detect frame: 564
class_id:0, confidence:0.6821, left_top:[574.62,535.45],right_bottom:[710.38,664.46]
class_id:0, confidence:0.6569, left_top:[1051.02,487.18],right_bottom:[1175.52,634.34]
class_id:1, confidence:0.8673, left_top:[466.00,91.86],right_bottom:[1328.56,814.08]
detect frame: 565
class_id:0, confidence:0.6685, left_top:[1050.59,487.48],right_bottom:[1172.96,636.16]
class_id:0, confidence:0.6659, left_top:[575.38,534.84],right_bottom:[710.16,663.65]
class_id:1, confidence:0.8686, left_top:[469.12,93.26],right_bottom:[1329.57,813.66]
detect frame: 566
class_id:1, confidence:0.8640, left_top:[470.18,95.61],right_bottom:[1329.16,813.50]
detect frame: 567
class_id:1, confidence:0.8648, left_top:[470.67,96.97],right_bottom:[1329.82,814.10]
detect frame: 568
class_id:1, confidence:0.8602, left_top:[470.05,97.69],right_bottom:[1331.50,813.90]
detect frame: 569
class_id:1, confidence:0.8537, left_top:[469.44,96.77],right_bottom:[1331.15,814.40]
detect frame: 570
class_id:1, confidence:0.8407, left_top:[467.30,96.13],right_bottom:[1329.83,813.61]
detect frame: 571
class_id:1, confidence:0.8484, left_top:[467.85,95.92],right_bottom:[1327.15,812.05]
detect frame: 572
class_id:1, confidence:0.8546, left_top:[469.63,93.40],right_bottom:[1326.43,811.49]
detect frame: 573
class_id:1, confidence:0.8628, left_top:[471.23,90.78],right_bottom:[1326.68,813.56]
detect frame: 574
class_id:1, confidence:0.8640, left_top:[470.05,90.19],right_bottom:[1326.41,812.95]
detect frame: 575
class_id:1, confidence:0.8542, left_top:[471.13,90.17],right_bottom:[1325.28,814.19]
detect frame: 576
class_id:1, confidence:0.8584, left_top:[472.33,92.31],right_bottom:[1326.91,812.46]
detect frame: 577
class_id:1, confidence:0.8584, left_top:[472.52,92.44],right_bottom:[1327.98,812.68]
detect frame: 578
class_id:1, confidence:0.8667, left_top:[472.91,93.68],right_bottom:[1333.20,813.08]
detect frame: 579
class_id:0, confidence:0.6509, left_top:[583.47,531.31],right_bottom:[719.35,660.00]
class_id:1, confidence:0.8682, left_top:[475.06,90.68],right_bottom:[1337.80,812.68]
detect frame: 580
class_id:1, confidence:0.8663, left_top:[474.55,89.11],right_bottom:[1339.44,810.96]
detect frame: 581
class_id:0, confidence:0.6457, left_top:[586.35,531.68],right_bottom:[720.88,660.26]
class_id:0, confidence:0.6413, left_top:[1056.45,483.82],right_bottom:[1177.35,641.21]
class_id:1, confidence:0.8702, left_top:[473.42,87.59],right_bottom:[1340.01,813.18]
detect frame: 582
class_id:1, confidence:0.8690, left_top:[474.71,86.68],right_bottom:[1338.90,810.44]
detect frame: 583
class_id:1, confidence:0.8706, left_top:[474.75,85.01],right_bottom:[1341.67,810.49]
detect frame: 584
class_id:1, confidence:0.8663, left_top:[472.36,83.45],right_bottom:[1338.67,812.18]
detect frame: 585
class_id:0, confidence:0.6598, left_top:[599.42,535.71],right_bottom:[733.30,660.16]
class_id:1, confidence:0.8718, left_top:[477.15,80.71],right_bottom:[1335.94,810.08]
detect frame: 586
class_id:0, confidence:0.7260, left_top:[602.66,523.94],right_bottom:[748.79,657.38]
class_id:0, confidence:0.7191, left_top:[1078.92,475.71],right_bottom:[1200.03,626.58]
class_id:1, confidence:0.8751, left_top:[478.06,80.64],right_bottom:[1332.01,807.57]
detect frame: 587
class_id:0, confidence:0.7338, left_top:[603.47,526.35],right_bottom:[761.94,656.60]
class_id:0, confidence:0.7292, left_top:[1089.16,473.24],right_bottom:[1198.43,622.84]
class_id:1, confidence:0.8734, left_top:[476.51,78.92],right_bottom:[1321.01,810.58]
detect frame: 588
class_id:0, confidence:0.7212, left_top:[602.01,523.42],right_bottom:[762.57,655.78]
class_id:0, confidence:0.6892, left_top:[1091.71,467.59],right_bottom:[1199.28,617.00]
class_id:1, confidence:0.8739, left_top:[479.80,83.00],right_bottom:[1317.70,811.51]
detect frame: 589
class_id:0, confidence:0.7245, left_top:[607.63,527.80],right_bottom:[768.06,659.42]
class_id:1, confidence:0.8715, left_top:[479.10,83.12],right_bottom:[1314.51,810.04]
detect frame: 590
class_id:0, confidence:0.7842, left_top:[612.29,526.37],right_bottom:[776.15,657.02]
class_id:1, confidence:0.8704, left_top:[478.15,82.38],right_bottom:[1301.98,809.02]
detect frame: 591
class_id:0, confidence:0.7806, left_top:[622.98,526.89],right_bottom:[792.91,657.11]
class_id:1, confidence:0.8691, left_top:[481.26,84.36],right_bottom:[1297.18,812.84]
detect frame: 592
class_id:0, confidence:0.7690, left_top:[628.70,524.24],right_bottom:[797.01,658.41]
class_id:1, confidence:0.8575, left_top:[482.02,80.62],right_bottom:[1287.56,809.35]
detect frame: 593
class_id:0, confidence:0.7641, left_top:[627.36,526.03],right_bottom:[803.50,660.81]
class_id:1, confidence:0.8101, left_top:[476.41,79.24],right_bottom:[1285.39,807.31]
detect frame: 594
class_id:0, confidence:0.7813, left_top:[631.94,521.41],right_bottom:[808.09,659.28]
class_id:1, confidence:0.8484, left_top:[480.63,76.23],right_bottom:[1282.70,806.18]
detect frame: 595
class_id:0, confidence:0.8213, left_top:[636.37,522.42],right_bottom:[812.97,660.97]
class_id:1, confidence:0.8503, left_top:[482.16,78.14],right_bottom:[1279.24,806.26]
detect frame: 596
class_id:0, confidence:0.8370, left_top:[642.26,523.87],right_bottom:[829.71,662.36]
class_id:1, confidence:0.8461, left_top:[477.47,74.91],right_bottom:[1273.25,805.99]
detect frame: 597
class_id:0, confidence:0.8256, left_top:[648.86,525.22],right_bottom:[836.58,662.61]
class_id:1, confidence:0.8488, left_top:[478.07,78.44],right_bottom:[1261.43,804.75]
detect frame: 598
class_id:0, confidence:0.8431, left_top:[652.35,522.59],right_bottom:[842.09,660.29]
class_id:1, confidence:0.8503, left_top:[482.85,73.08],right_bottom:[1261.63,805.51]
detect frame: 599
class_id:0, confidence:0.7891, left_top:[656.76,524.58],right_bottom:[844.38,660.76]
class_id:1, confidence:0.8455, left_top:[484.49,77.40],right_bottom:[1252.19,803.81]
detect frame: 600
class_id:0, confidence:0.8234, left_top:[662.50,521.26],right_bottom:[854.42,660.68]
class_id:1, confidence:0.8472, left_top:[483.94,74.65],right_bottom:[1250.29,801.33]
detect frame: 601
class_id:0, confidence:0.8002, left_top:[667.65,525.19],right_bottom:[865.82,659.07]
class_id:1, confidence:0.8522, left_top:[487.88,69.26],right_bottom:[1239.37,799.53]
detect frame: 602
class_id:0, confidence:0.8084, left_top:[672.61,527.96],right_bottom:[864.06,661.55]
class_id:1, confidence:0.8513, left_top:[487.80,67.97],right_bottom:[1247.51,802.78]
detect frame: 603
class_id:0, confidence:0.8412, left_top:[681.41,528.48],right_bottom:[877.51,660.51]
class_id:1, confidence:0.8606, left_top:[484.70,69.38],right_bottom:[1244.97,802.13]
detect frame: 604
class_id:0, confidence:0.7791, left_top:[682.48,526.06],right_bottom:[886.61,657.51]
class_id:1, confidence:0.8536, left_top:[489.90,68.18],right_bottom:[1244.56,800.77]
detect frame: 605
class_id:0, confidence:0.7648, left_top:[692.82,528.71],right_bottom:[894.14,657.53]
class_id:1, confidence:0.8517, left_top:[492.06,67.26],right_bottom:[1238.70,800.07]
detect frame: 606
class_id:0, confidence:0.7975, left_top:[700.36,526.74],right_bottom:[895.48,658.59]
class_id:1, confidence:0.8612, left_top:[491.64,65.71],right_bottom:[1239.69,799.07]
detect frame: 607
class_id:0, confidence:0.7981, left_top:[712.68,526.84],right_bottom:[912.33,662.54]
class_id:1, confidence:0.8583, left_top:[502.05,69.03],right_bottom:[1234.40,799.59]
detect frame: 608
class_id:0, confidence:0.8208, left_top:[715.76,521.26],right_bottom:[912.87,661.64]
class_id:1, confidence:0.8587, left_top:[508.09,67.61],right_bottom:[1239.70,797.64]
detect frame: 609
class_id:0, confidence:0.6651, left_top:[712.96,522.06],right_bottom:[926.88,657.68]
class_id:1, confidence:0.6454, left_top:[514.75,68.28],right_bottom:[1233.98,798.36]
class_id:2, confidence:0.7247, left_top:[510.83,72.00],right_bottom:[1236.61,799.74]
detect frame: 610
class_id:0, confidence:0.7245, left_top:[718.03,523.74],right_bottom:[926.87,657.41]
class_id:1, confidence:0.7675, left_top:[514.27,66.69],right_bottom:[1235.96,792.70]
detect frame: 611
class_id:0, confidence:0.7214, left_top:[730.65,518.16],right_bottom:[943.49,661.73]
class_id:1, confidence:0.8366, left_top:[516.79,71.26],right_bottom:[1237.06,792.99]
detect frame: 612
class_id:2, confidence:0.7856, left_top:[519.72,82.48],right_bottom:[1244.15,796.97]
class_id:4, confidence:0.6695, left_top:[735.82,516.15],right_bottom:[942.77,661.75]
detect frame: 613
class_id:1, confidence:0.7627, left_top:[524.06,67.74],right_bottom:[1244.20,795.32]
detect frame: 614
class_id:0, confidence:0.7776, left_top:[742.21,510.31],right_bottom:[953.16,660.68]
class_id:1, confidence:0.8301, left_top:[529.34,68.66],right_bottom:[1248.70,790.42]
detect frame: 615
class_id:0, confidence:0.7423, left_top:[752.70,505.50],right_bottom:[966.35,657.37]
class_id:1, confidence:0.8467, left_top:[522.06,70.43],right_bottom:[1247.60,794.00]
detect frame: 616
class_id:0, confidence:0.7387, left_top:[753.08,509.57],right_bottom:[972.30,658.62]
class_id:1, confidence:0.8249, left_top:[532.51,67.81],right_bottom:[1251.23,791.84]
detect frame: 617
class_id:0, confidence:0.7509, left_top:[760.68,508.20],right_bottom:[974.38,659.19]
class_id:1, confidence:0.8436, left_top:[527.65,71.86],right_bottom:[1247.43,791.14]
detect frame: 618
class_id:0, confidence:0.7979, left_top:[767.84,505.97],right_bottom:[985.91,659.07]
class_id:1, confidence:0.8539, left_top:[532.48,68.67],right_bottom:[1251.14,791.02]
detect frame: 619
class_id:0, confidence:0.7820, left_top:[768.15,507.02],right_bottom:[986.27,660.18]
class_id:1, confidence:0.8498, left_top:[536.04,69.43],right_bottom:[1255.19,790.39]
detect frame: 620
class_id:1, confidence:0.7997, left_top:[542.55,66.48],right_bottom:[1258.44,790.53]
class_id:4, confidence:0.6770, left_top:[777.95,507.58],right_bottom:[993.65,659.49]
detect frame: 621
class_id:0, confidence:0.7242, left_top:[778.31,515.37],right_bottom:[994.43,660.70]
class_id:1, confidence:0.8182, left_top:[544.28,68.25],right_bottom:[1259.93,790.96]
class_id:4, confidence:0.6470, left_top:[783.49,510.55],right_bottom:[995.42,661.91]
detect frame: 622
class_id:0, confidence:0.7601, left_top:[781.67,521.85],right_bottom:[995.52,659.36]
class_id:1, confidence:0.8548, left_top:[548.07,71.34],right_bottom:[1258.86,789.66]
detect frame: 623
class_id:0, confidence:0.7701, left_top:[780.24,522.61],right_bottom:[994.59,659.92]
class_id:1, confidence:0.8590, left_top:[548.73,72.24],right_bottom:[1259.44,789.61]
detect frame: 624
class_id:0, confidence:0.7591, left_top:[779.77,523.71],right_bottom:[996.63,660.25]
class_id:1, confidence:0.8517, left_top:[549.88,74.50],right_bottom:[1260.69,788.50]
detect frame: 625
class_id:0, confidence:0.7836, left_top:[778.36,524.66],right_bottom:[998.99,661.81]
class_id:1, confidence:0.8611, left_top:[549.51,75.73],right_bottom:[1260.76,788.69]
detect frame: 626
class_id:0, confidence:0.7725, left_top:[776.75,523.76],right_bottom:[996.30,662.99]
class_id:1, confidence:0.8562, left_top:[551.48,77.21],right_bottom:[1259.03,789.64]
detect frame: 627
class_id:0, confidence:0.7312, left_top:[779.89,523.16],right_bottom:[999.32,664.24]
class_id:1, confidence:0.8449, left_top:[550.38,77.50],right_bottom:[1258.61,790.91]
class_id:4, confidence:0.6557, left_top:[789.21,519.53],right_bottom:[997.38,664.23]
登录后复制
       
In [ ]
!python deploy/python/infer.py --model_dir=output_inference/picodet_xs_416_coco_lcnet --image_dir dataset/voc/test/ --device=CPU --batch_size=1 --cpu_threads=4
登录后复制
   
In [ ]
!python deploy/python/infer.py --model_dir=output_inference/picodet_xs_416_coco_lcnet --image_dir dataset/voc/test/ --device=CPU --batch_size=1 --cpu_threads=4 --enable_mkldnn=True
登录后复制
   
In [ ]
!python deploy/python/infer.py --model_dir=output_inference/picodet_xs_416_coco_lcnet --image_dir dataset/voc/test/ --device=GPU --batch_size=1
登录后复制
   

总结

测试结果如下:

picodet_l_640_coco_lcnet preprocess_time(ms) inference_time(ms) postprocess_time(ms) average latency time(ms)
CPU 27.60 934.00 0.10 961.69
CPU+MKL 26.80 84.40 0.10 111.32
GPU 51.10 13.30 0.00 64.47
picodet_xs_416_coco_lcnet preprocess_time(ms) inference_time(ms) postprocess_time(ms) average latency time(ms)
CPU 15.60 175.20 0.10 190.87
CPU+MKL 15.30 19.40 0.10 34.75
GPU 49.70 8.10 0.10 57.87

本项目对大小差距比较大的两个模型进行了比较,推理速度上和精度上各有优势。

可以看到,"大模型"picodet_l_640_coco_lcnet 在GPU模式下比CPU模式下快了约43%,即使开了MKL加速,CPU的还是赶不上GPU的算力。

但是picodet_xs_416_coco_lcnet在开了MKL加速后却比GPU模式下快了约40%。可见在选择轻量级模型时,有时使用CPU预测的速度比GPU快一些。

而快的这一部分主要都快在了preprocess_time上,infer的时间却差别不大,预处理的时间CPU比GPU快了不少(猜测是CPUtoGPU耗时太多)

BONUS CHAPTER:使用ONNX和OpenVINO部署的尝试

ONNX转化步骤如下:

  • 转换paddle静态图模型到onnx
  • 写脚本进行onnx推理(逃

非常nice的借鉴项目

✨【PaddlePaddle+OpenVINO】垃圾邮件检测部署

✨【PaddlePaddle+OpenVINO】OpenVINO驾驶人状态检测

In [ ]
# ONNX转化!pip install onnx
!pip install paddle2onnx==0.9.2!pip install onnxruntime
登录后复制
   
In [ ]
# 只转化xs的, 警告仅支持batchsize = 1!paddle2onnx --model_dir output_inference/picodet_xs_416_coco_lcnet/ \
            --model_filename model.pdmodel  \
            --params_filename model.pdiparams \
            --opset_version 11 \
            --save_file picodet_xs_416_coco_lcnet.onnx
登录后复制
   

测试ONNX模型是否正常加载

In [6]
import osimport onnxruntime
model_path = 'picodet_xs_416_coco_lcnet_sim.onnx'session = onnxruntime.InferenceSession(model_path)

input_names = [input.name for input in session.get_inputs()]
output_names = [output.name for output in session.get_outputs()]print(input_names, output_names)
登录后复制
       
['image', 'scale_factor'] ['multiclass_nms3_0.tmp_0', 'multiclass_nms3_0.tmp_2']
登录后复制
       

ONNX加载模型

代码参考https://blog.csdn.net/ouening/article/details/109249925

进行ONNX推理

In [8]
import numpy as np    # we're going to use numpy to process input and output dataimport onnxruntime    # to inference ONNX models, we use the ONNX Runtimeimport onnxfrom onnx import numpy_helperimport urllib.requestimport jsonimport timeimport pandas as pdfrom imageio import imreadimport warnings
warnings.filterwarnings('ignore')# display images in notebookimport matplotlib.pyplot as pltfrom PIL import Image, ImageDraw, ImageFontimport matplotlib.patches as patches# import IPython.display# import cv2picdet_onnx_model = 'inference_model/onnx/picodet_xs_416_coco_lcnet_sim.onnx'img_file = r"dataset/voc/test/IMG_100.jpg"# 手动预处理def preprocess(img_file, w, h):
    input_shape = (1, 3, w, h)
    img = Image.open(img_file)
    img = img.resize((w, h), Image.BILINEAR)    # convert the input data into the float32 input
    img_data = np.array(img)
    img_data = np.transpose(img_data, [2, 0, 1])
    img_data = np.expand_dims(img_data, 0)
    mean_vec = np.array([0.485, 0.456, 0.406])
    stddev_vec = np.array([0.229, 0.224, 0.225])
    norm_img_data = np.zeros(img_data.shape).astype('float32')    for i in range(img_data.shape[1]):
        norm_img_data[:,i,:,:] = (img_data[:,i,:,:]/255 - mean_vec[i]) / stddev_vec[i]    return norm_img_data.astype('float32'), np.array(img)def infer_sim_picodet(picdet_onnx_model:str):
    
    # Run the model on the backend
    session = onnxruntime.InferenceSession(picdet_onnx_model, None)    
    # get the name of the first input of the model
    input_name = session.get_inputs()[0].name  
    output_name = session.get_outputs()[0].name  
    pre_start = time.time()
    input_data, raw_img = preprocess(img_file, 416, 416)
    image_size = np.array([raw_img.shape[1], raw_img.shape[0]], dtype=np.float32).reshape(1, 2)

    pre_end = time.time()    print('preprocess', pre_end-pre_start,'s')
    start = time.time()
    raw_result = session.run([], {input_name: input_data,                                  'scale_factor':image_size})
    end = time.time()    print('推理时间:', end-start,'s')
    yolonms_layer_1 = raw_result[0]
    yolonms_layer_1_1 = raw_result[1]

    fig, ax = plt.subplots(1)
    ax.imshow(raw_img)    for result in raw_result:        for res in result[:-1]:            # print(res)
            
            cls, score,  x_min, y_min, x_max, y_max = res
            bboxes = [x_min, y_min, x_max, y_max]            # cv2.rectangle(raw_img, (x_min, y_min), (x_max, y_max), (0, 255, 255))
            
            if(score > 0.5):                print(bboxes)
                x1 = 1200*bboxes[0]
                y1 = 1200*bboxes[1]
                x2 = 1200*bboxes[2]
                y2 = 1200*bboxes[3]
                rect = patches.Rectangle((x1,y1),x2-x1,y2-y1,linewidth=1,edgecolor='r',fill=False)
                        
                ax.add_patch(rect)
    x1 = 1200*bboxes[0]
    y1 = 1200*bboxes[1]
    x2 = 1200*bboxes[2]
    y2 = 1200*bboxes[3]
    rect = patches.Rectangle((x1,y1),x2-x1,y2-y1,linewidth=1,edgecolor='r',fill=False)
                        
    ax.add_patch(rect)
    plt.show()    # return raw_resultinfer_sim_picodet(picdet_onnx_model)
登录后复制
       
preprocess 0.02070474624633789 s
推理时间: 1.8522729873657227 s
[0.72810787, 0.5058939, 0.7813135, 0.55807817]
[0.681848, 0.2640958, 0.8350814, 0.61556965]
登录后复制
       
<Figure size 432x288 with 1 Axes>
登录后复制
               

太慢了,推理时间1.85 s,而且框不知道为什么可以print出来但是画不出来,怀疑是resize的原因

OpenVINO预测

只能用python3.6进行安装...而且是2021版本的,装不上2022.1版本

1.环境准备

In [ ]
!python3.6 -m pip install --upgrade pip -i http://pypi.douban.com/simple --trusted-host pypi.douban.com
!python3.6 -m pip install openvino -i http://pypi.douban.com/simple --trusted-host pypi.douban.com
!python3.6 -m pip install paddle2onnx -i http://pypi.douban.com/simple --trusted-host pypi.douban.com
!pip install onnx-simplifier
!python3.6 -m pip install --upgrade openvino -i http://pypi.douban.com/simple --trusted-host pypi.douban.com
登录后复制
   
In [8]
# 使用onnxsim优化模型!python -m onnxsim picodet_xs_416_coco_lcnet.onnx picodet_xs_416_coco_lcnet_sim.onnx --dynamic-input-shape --input-shape image:1,3,416,416!python -m onnxsim picodet_xs_416_coco_lcnet.onnx picodet_xs_416_coco_lcnet_sim.onnx --dynamic-input-shape --input-shape image:1,3,416,416
登录后复制
       
Simplifying...
Finish! Here is the difference:
┏━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━┓
┃                    ┃ Original Model ┃ Simplified Model ┃
┡━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━┩
│ Add                │ 99             │ 83               │
│ BatchNormalization │ 82             │ 0                │
│ Cast               │ 7              │ 7                │
│ Clip               │ 78             │ 78               │
│ Concat             │ 10             │ 10               │
│ Constant           │ 803            │ 16               │
│ Conv               │ 98             │ 98               │
│ Div                │ 79             │ 79               │
│ Gather             │ 13             │ 13               │
│ GlobalAveragePool  │ 6              │ 6                │
│ HardSigmoid        │ 2              │ 2                │
│ Identity           │ 4              │ 0                │
│ MatMul             │ 4              │ 4                │
│ Mul                │ 91             │ 91               │
│ Neg                │ 1              │ 1                │
│ NonMaxSuppression  │ 1              │ 1                │
│ NonZero            │ 1              │ 1                │
│ ReduceMin          │ 1              │ 1                │
│ Relu               │ 2              │ 2                │
│ Reshape            │ 30             │ 14               │
│ Resize             │ 2              │ 2                │
│ Shape              │ 3              │ 3                │
│ Sigmoid            │ 12             │ 12               │
│ Softmax            │ 4              │ 4                │
│ Split              │ 2              │ 2                │
│ Sqrt               │ 4              │ 4                │
│ Squeeze            │ 6              │ 5                │
│ Sub                │ 1              │ 1                │
│ TopK               │ 2              │ 2                │
│ Transpose          │ 4              │ 4                │
│ Unsqueeze          │ 3              │ 3                │
│ Model Size         │ 2.8MiB         │ 2.7MiB           │
└────────────────────┴────────────────┴──────────────────┘
登录后复制
       

2.调用官方脚本进行预测

In [ ]
!python3.6 tools/openvino_infer.py\
            --img_path dataset/voc/test/IMG_100.jpg\
            --onnx_path picodet_xs_416_coco_lcnet_sim.onnx\
            --in_shape 416
登录后复制
   

然而一波操作下来报错...,最好还是安装2022.1版本的OpenVINO进行部署。上github上搜issue,发现解决方法都是说安装了2022版本后解决...然而现在版本的V100还不支持安装OpenVINO,希望早日能支持。

Traceback (most recent call last):  File "tools/openvino_infer.py", line 265, in <module>
    compiled_model = ie.load_network(net, 'CPU')  File "ie_api.pyx", line 403, in openvino.inference_engine.ie_api.IECore.load_network  File "ie_api.pyx", line 442, in openvino.inference_engine.ie_api.IECore.load_network
RuntimeError: Ngraph operation Parameter with name image has dynamic output shape on 0 port, but CPU plug-in supports only static shape
登录后复制
   

以上就是【AI workshop】增强的PicoDet有多猛?来,跑个RM数据集试试!的详细内容,更多请关注php中文网其它相关文章!

最佳 Windows 性能的顶级免费优化软件
最佳 Windows 性能的顶级免费优化软件

每个人都需要一台速度更快、更稳定的 PC。随着时间的推移,垃圾文件、旧注册表数据和不必要的后台进程会占用资源并降低性能。幸运的是,许多工具可以让 Windows 保持平稳运行。

下载
来源:php中文网
本文内容由网友自发贡献,版权归原作者所有,本站不承担相应法律责任。如您发现有涉嫌抄袭侵权的内容,请联系admin@php.cn
最新问题
开源免费商场系统广告
热门教程
更多>
最新下载
更多>
网站特效
网站源码
网站素材
前端模板
关于我们 免责申明 意见反馈 讲师合作 广告合作 最新更新 English
php中文网:公益在线php培训,帮助PHP学习者快速成长!
关注服务号 技术交流群
PHP中文网订阅号
每天精选资源文章推送
PHP中文网APP
随时随地碎片化学习
PHP中文网抖音号
发现有趣的

Copyright 2014-2025 https://www.php.cn/ All Rights Reserved | php.cn | 湘ICP备2023035733号