devbox
羊村懒王
这个屌丝很懒，什么也没留下！
关注作者
热门标签
热门文章
当前位置:   article > 正文

基于PaddleClas的PP-LCNet模型的动物图像识别与分类

作者：羊村懒王 | 2024-04-08 11:48:48

赞

踩

基于PaddleClas的PP-LCNet模型的动物图像识别与分类

转自AI Studio，原文链接：基于PaddleClas的PP-LCNet模型的动物图像识别与分类 - 飞桨AI Studio

一、基于PP-LCNet的动物图像识别与分类

比赛地址和鲸社区 - Heywhale.com

1.数据集介绍

数据集是一个用于多分类任务的动物图像数据集，包含10种不同动物的图像。数据集来源由Google上的真实图像通过爬虫得到，图片尺寸大小以及格式不固定（包含jpg、jpeg以及png三种图像格式），另外对敏感信息进行了脱敏处理。

1.1训练集

训练集文件夹名为train_data，共有17803张图像，文件夹中包含10个子文件夹，文件名分别是butterfly、cat、chicken、cow、dog、elephant、horse、ragno、sheep、squirrel，文件名为对应文件夹下图像的类别，选手需自行读取标签信息。每个子文件夹下包含若干图像文件，数量约为1000-5000。

1.2测试集

测试集文件夹名为test_data，文件夹中包含8150张图像，选手需根据训练集建立模型，对测试集文件进行预测分类。

2.PP-LCNet介绍

在工业界真实落地的场景中，推理速度才是考量模型好坏的重要指标，然而，推理速度和准确性很难兼得。考虑到工业界有很多基于 Intel CPU 的应用，所以我们本次的工作旨在使骨干网络更好的适应 Intel CPU，从而得到一个速度更快、准确率更高的轻量级骨干网络，与此同时，目标检测、语义分割等下游视觉任务的性能也同样得到提升。针对 Intel CPU 设备以及其加速库 MKLDNN 设计了特定的骨干网络 PP-LCNet，比起其他的轻量级的 SOTA 模型，该骨干网络可以在不增加推理时间的情况下，进一步提升模型的性能，最终大幅度超越现有的 SOTA 模型。

二、数据准备

1.解压缩数据

In [1]

  1. !unzip -qoa data/data140388/traindata.zip -d data/
  2. !unzip -qoa data/data140388/testdata.zip -d data/

In [2]

!mv data/input/animal7479/* data/

2.生成数据列表

In [3]

  1. # paddlex安装
  2. !pip install paddlex >log.log

In [4]

!paddlex --split_dataset --format ImageNet --dataset_dir data/train_data/train_data --val_value 0.2
[04-21 09:55:30 MainThread @logger.py:242] Argv: /opt/conda/envs/python35-paddle120-env/bin/paddlex --split_dataset --format ImageNet --dataset_dir data/train_data/train_data --val_value 0.2
[04-21 09:55:30 MainThread @utils.py:79] WRN paddlepaddle version: 2.2.2. The dynamic graph version of PARL is under development, not fully tested and supported
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/parl/remote/communication.py:38: DeprecationWarning: 'pyarrow.default_serialization_context' is deprecated as of 2.0.0 and will be removed in a future version. Use pickle or the pyarrow IPC functionality instead.
  context = pyarrow.default_serialization_context()
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/matplotlib/__init__.py:107: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
  from collections import MutableMapping
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/matplotlib/rcsetup.py:20: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
  from collections import Iterable, Mapping
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/matplotlib/colors.py:53: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
  from collections import Sized
2022-04-21 09:55:34 [INFO]	Dataset split starts...
2022-04-21 09:55:34 [INFO]	Dataset split done.
2022-04-21 09:55:34 [INFO]	Train samples: 14246
2022-04-21 09:55:34 [INFO]	Eval samples: 3557
2022-04-21 09:55:34 [INFO]	Test samples: 0
2022-04-21 09:55:34 [INFO]	Split files saved in data/train_data/train_data

In [11]

  1. with open('data/train_data/train_data/labels.txt','r') as f:
  2.     lines=f.readlines()
  3.     print(lines)
  4. f_list=open('label_list.txt','w')    
  5. print(len(lines))
  6. for i in range(len(lines)):
  7.     f_list.write(str(i)+' '+ lines[i])
  8. f_list.close()
['butterfly\n', 'cat\n', 'chicken\n', 'cow\n', 'dog\n', 'elephant\n', 'horse\n', 'ragno\n', 'sheep\n', 'squirrel\n']
10

0.8的训练集，0.2的测试集。Train数量为： 14246、Eval 数量为： 3557

  1. 2022-04-20 01:02:48 [INFO]	Train samples: 14246
  2. 2022-04-20 01:02:48 [INFO]	Eval samples: 3557
  3. 2022-04-20 01:02:48 [INFO]	Test samples: 0

二、环境准备

PaddleClas下载,计划使用PaddleClas中的 PP-LCNet 进行训练

In [5]

# !git clone https://gitee.com/paddlepaddle/PaddleClas.git --depth=1

三、修改代码

1.修改配置

以 PaddleClas/ppcls/configs/ImageNet/PPLCNet/PPLCNet_x0_25.yaml 为基础进行配置

  1. # global configs
  2. Global:
  3.   checkpoints: null
  4.   pretrained_model: null
  5.   output_dir: ./output/
  6.   device: gpu
  7.   save_interval: 1
  8.   eval_during_train: True
  9.   eval_interval: 1
  10.   epochs: 300
  11.   print_batch_step: 10
  12.   use_visualdl: False
  13.   # used for static mode and model export
  14.   image_shape: [3, 224, 224]
  15.   save_inference_dir: ./inference
  16. # model architecture
  17. Arch:
  18.   name: PPLCNet_x0_25
  19.   class_num: 10
  20.  
  21. # loss function config for traing/eval process
  22. Loss:
  23.   Train:
  24.     - CELoss:
  25.         weight: 1.0
  26.         epsilon: 0.1
  27.   Eval:
  28.     - CELoss:
  29.         weight: 1.0
  30.  
  31.  
  32. Optimizer:
  33.   name: Momentum
  34.   momentum: 0.9
  35.   lr:
  36.     name: Cosine
  37.     learning_rate: 0.1
  38.     warmup_epoch: 10
  39.   regularizer:
  40.     name: 'L2'
  41.     coeff: 0.0001
  42.  
  43.  
  44. # data loader for train and eval
  45. DataLoader:
  46.   Train:
  47.     dataset:
  48.       name: ImageNetDataset
  49.       image_root: /home/aistudio/data/train_data/train_data/
  50.       cls_label_path: train_list.txt
  51.       transform_ops:
  52.         - DecodeImage:
  53.             to_rgb: True
  54.             channel_first: False
  55.         - RandCropImage:
  56.             size: 224
  57.         - RandFlipImage:
  58.             flip_code: 1
  59.         - AutoAugment:
  60.         - NormalizeImage:
  61.             scale: 1.0/255.0
  62.             mean: [0.485, 0.456, 0.406]
  63.             std: [0.229, 0.224, 0.225]
  64.             order: ''
  65.     batch_transform_ops:
  66.     - CutmixOperator:
  67.         alpha: 0.2
  68.  
  69.     sampler:
  70.       name: DistributedBatchSampler
  71.       batch_size: 2048
  72.       drop_last: False
  73.       shuffle: True
  74.     loader:
  75.       num_workers: 4
  76.       use_shared_memory: False
  77.  
  78.   Eval:
  79.     dataset: 
  80.       name: ImageNetDataset
  81.       image_root: /home/aistudio/data/train_data/train_data/
  82.       cls_label_path: val_list.txt
  83.       transform_ops:
  84.         - DecodeImage:
  85.             to_rgb: True
  86.             channel_first: False
  87.         - ResizeImage:
  88.             resize_short: 256
  89.         - CropImage:
  90.             size: 224
  91.         - NormalizeImage:
  92.             scale: 1.0/255.0
  93.             mean: [0.485, 0.456, 0.406]
  94.             std: [0.229, 0.224, 0.225]
  95.             order: ''
  96.     sampler:
  97.       name: DistributedBatchSampler
  98.       batch_size: 1024
  99.       drop_last: False
  100.       shuffle: False
  101.     loader:
  102.       num_workers: 4
  103.       use_shared_memory: False
  104.  
  105. Infer:
  106.   infer_imgs: docs/images/inference_deployment/whl_demo.jpg
  107.   batch_size: 10
  108.   transforms:
  109.     - DecodeImage:
  110.         to_rgb: True
  111.         channel_first: False
  112.     - ResizeImage:
  113.         resize_short: 256
  114.     - CropImage:
  115.         size: 224
  116.     - NormalizeImage:
  117.         scale: 1.0/255.0
  118.         mean: [0.485, 0.456, 0.406]
  119.         std: [0.229, 0.224, 0.225]
  120.         order: ''
  121.     - ToCHWImage:
  122.   PostProcess:
  123.     name: Topk
  124.     topk: 5
  125.     class_id_map_file: ../label_list.txt
  126.  
  127. Metric:
  128.   Train:
  129.     - TopkAcc:
  130.         topk: [1, 5]
  131.   Eval:
  132.     - TopkAcc:
  133.         topk: [1, 5]
  134.  
  135.

2.修改代码

修改self._cls_path为os.path.join(self._img_root,self._cls_path)

  1.  
  2. from __future__ import print_function
  3.  
  4. import numpy as np
  5. import os
  6.  
  7. from .common_dataset import CommonDataset
  8.  
  9.  
  10. class ImageNetDataset(CommonDataset):
  11.     def _load_anno(self, seed=None):
  12.         # print(self._cls_path)
  13.         # print(self._img_root)
  14.         # 修改
  15.         self._cls_path=os.path.join(self._img_root,self._cls_path)
  16.         assert os.path.exists(self._cls_path)
  17.         assert os.path.exists(self._img_root)
  18.         self.images = []
  19.         self.labels = []
  20.  
  21.         with open(self._cls_path) as fd:
  22.             lines = fd.readlines()
  23.             if seed is not None:
  24.                 np.random.RandomState(seed).shuffle(lines)
  25.             for l in lines:
  26.                 l = l.strip().split(" ")
  27.                 self.images.append(os.path.join(self._img_root, l[0]))
  28.                 self.labels.append(np.int64(l[1]))
  29.                 assert os.path.exists(self.images[-1])

四、模型训练

1.训练模型

没啥说的，配置文件都写好了，跑就完事了。当然配置文件主要做以下工作：

  • 一是数据集地址更改
  • 二是训练轮次、batch size更改
  • 三是数据增强配置

2.注意事项

  • 一是使用预训练模型，傻子才从头开始训练，懂得都懂，掌声响起来
  • 二是不要使用aistudio中复制完整路径，因为复制了是错的

例如：/home/aistudio/PPLCNet_x0_25.yaml路径，会变为/home/aistudio/.jupyter/lab/workspaces/PPLCNet_x0_25.yaml，简直让你防不胜防。

  1.  
  2. !python3 tools/train.py \
  3.     -c /home/aistudio/.jupyter/lab/workspaces/PPLCNet_x0_25.yaml \
  4.     -o Arch.pretrained=False \
  5.     -o Global.device=gpu
  6.     
  7.     
  8. /home/aistudio/PaddleClas
  9. Traceback (most recent call last):
  10.   File "tools/train.py", line 29, in <module>
  11.     args.config, overrides=args.override, show=False)
  12.   File "/home/aistudio/PaddleClas/ppcls/utils/config.py", line 179, in get_config
  13.     'config file({}) is not exist'.format(fname))
  14. AssertionError: config file(/home/aistudio/.jupyter/lab/workspaces/PPLCNet_x0_25.yaml) is not exist

In [ ]

  1. %cd ~/PaddleClas/
  2. # 复制修改后的imagenet_dataset.py到原位
  3. !cp ../imagenet_dataset.py ./ppcls/data/dataloader/imagenet_dataset.py
  4. # 开始训练
  5. !python3 tools/train.py \
  6.     -c ../PPLCNet_x0_25.yaml \
  7.     -o Arch.pretrained=True \
  8.     -o Global.device=gpu

74个epoch可达到91%的准确率，如时间宽裕，可继续提升准确率

  1. [2022/04/21 02:48:33] root INFO: [Train][Epoch 74/300][Iter: 0/7]lr: 0.08875, CELoss: 0.86730, loss: 0.86730, batch_cost: 14.51762s, reader_cost: 12.59892, ips: 141.06996 images/sec, eta: 6:24:28
  2. [2022/04/21 02:48:51] root INFO: [Train][Epoch 74/300][Avg]CELoss: 0.88821, loss: 0.88821
  3. [2022/04/21 02:49:06] root INFO: [Eval][Epoch 74][Iter: 0/4]CELoss: 0.43258, loss: 0.43258, top1: 0.88574, top5: 0.98633, batch_cost: 14.79859s, reader_cost: 13.25228, ips: 69.19578 images/sec
  4. [2022/04/21 02:49:07] root INFO: [Eval][Epoch 74][Avg]CELoss: 0.45654, loss: 0.45654, top1: 0.87855, top5: 0.99044
  5. [2022/04/21 02:49:07] root INFO: [Eval][Epoch 74][best metric: 0.914815859962857]
  6. [2022/04/21 02:49:07] root INFO: Already save model in ./output/PPLCNet_x0_25/epoch_74
  7. [2022/04/21 02:49:07] root INFO: Already save model in ./output/PPLCNet_x0_25/latest

五、模型预测

1.模型导出

在上述模型导出命令中，所使用的配置文件需要与该模型的训练文件相同，在配置文件中有以下字段用于配置模型导出参数：

  • Global.image_shape：用于指定模型的输入数据尺寸，该尺寸不包含 batch 维度；

  • Global.save_inference_dir：用于指定导出的 inference 模型的保存位置；

  • Global.pretrained_model：用于指定训练过程中保存的模型权重文件路径，该路径无需包含模型权重文件后缀名 .pdparams。 上述命令将生成以下三个文件：

  • inference.pdmodel：用于存储网络结构信息；

  • inference.pdiparams：用于存储网络权重信息；

  • inference.pdiparams.info：用于存储模型的参数信息，在分类模型和识别模型中可忽略。

In [4]

  1. %cd ~/PaddleClas/
  2. !python tools/export_model.py \
  3.     -c ../PPLCNet_x0_25.yaml \
  4.     -o Global.pretrained_model=.//output/PPLCNet_x0_25/best_model \
  5.     -o Global.save_inference_dir=./deploy/models/class_PPLCNet_x0_25_ImageNet_infer
/home/aistudio/PaddleClas
[2022/04/21 12:39:48] root INFO: 
===========================================================
==        PaddleClas is powered by PaddlePaddle !        ==
===========================================================
==                                                       ==
==   For more info please go to the following website.   ==
==                                                       ==
==       https://github.com/PaddlePaddle/PaddleClas      ==
===========================================================

[2022/04/21 12:39:48] root INFO: Arch : 
[2022/04/21 12:39:48] root INFO:     class_num : 10
[2022/04/21 12:39:48] root INFO:     name : PPLCNet_x0_25
[2022/04/21 12:39:48] root INFO: DataLoader : 
[2022/04/21 12:39:48] root INFO:     Eval : 
[2022/04/21 12:39:48] root INFO:         dataset : 
[2022/04/21 12:39:48] root INFO:             cls_label_path : val_list.txt
[2022/04/21 12:39:48] root INFO:             image_root : /home/aistudio/data/train_data/train_data/
[2022/04/21 12:39:48] root INFO:             name : ImageNetDataset
[2022/04/21 12:39:48] root INFO:             transform_ops : 
[2022/04/21 12:39:48] root INFO:                 DecodeImage : 
[2022/04/21 12:39:48] root INFO:                     channel_first : False
[2022/04/21 12:39:48] root INFO:                     to_rgb : True
[2022/04/21 12:39:48] root INFO:                 ResizeImage : 
[2022/04/21 12:39:48] root INFO:                     resize_short : 256
[2022/04/21 12:39:48] root INFO:                 CropImage : 
[2022/04/21 12:39:48] root INFO:                     size : 224
[2022/04/21 12:39:48] root INFO:                 NormalizeImage : 
[2022/04/21 12:39:48] root INFO:                     mean : [0.485, 0.456, 0.406]
[2022/04/21 12:39:48] root INFO:                     order : 
[2022/04/21 12:39:48] root INFO:                     scale : 1.0/255.0
[2022/04/21 12:39:48] root INFO:                     std : [0.229, 0.224, 0.225]
[2022/04/21 12:39:48] root INFO:         loader : 
[2022/04/21 12:39:48] root INFO:             num_workers : 4
[2022/04/21 12:39:48] root INFO:             use_shared_memory : False
[2022/04/21 12:39:48] root INFO:         sampler : 
[2022/04/21 12:39:48] root INFO:             batch_size : 1024
[2022/04/21 12:39:48] root INFO:             drop_last : False
[2022/04/21 12:39:48] root INFO:             name : DistributedBatchSampler
[2022/04/21 12:39:48] root INFO:             shuffle : False
[2022/04/21 12:39:48] root INFO:     Train : 
[2022/04/21 12:39:48] root INFO:         batch_transform_ops : 
[2022/04/21 12:39:48] root INFO:             CutmixOperator : 
[2022/04/21 12:39:48] root INFO:                 alpha : 0.2
[2022/04/21 12:39:48] root INFO:         dataset : 
[2022/04/21 12:39:48] root INFO:             cls_label_path : train_list.txt
[2022/04/21 12:39:48] root INFO:             image_root : /home/aistudio/data/train_data/train_data/
[2022/04/21 12:39:48] root INFO:             name : ImageNetDataset
[2022/04/21 12:39:48] root INFO:             transform_ops : 
[2022/04/21 12:39:48] root INFO:                 DecodeImage : 
[2022/04/21 12:39:48] root INFO:                     channel_first : False
[2022/04/21 12:39:48] root INFO:                     to_rgb : True
[2022/04/21 12:39:48] root INFO:                 RandCropImage : 
[2022/04/21 12:39:48] root INFO:                     size : 224
[2022/04/21 12:39:48] root INFO:                 RandFlipImage : 
[2022/04/21 12:39:48] root INFO:                     flip_code : 1
[2022/04/21 12:39:48] root INFO:                 AutoAugment : None
[2022/04/21 12:39:48] root INFO:                 NormalizeImage : 
[2022/04/21 12:39:48] root INFO:                     mean : [0.485, 0.456, 0.406]
[2022/04/21 12:39:48] root INFO:                     order : 
[2022/04/21 12:39:48] root INFO:                     scale : 1.0/255.0
[2022/04/21 12:39:48] root INFO:                     std : [0.229, 0.224, 0.225]
[2022/04/21 12:39:48] root INFO:         loader : 
[2022/04/21 12:39:48] root INFO:             num_workers : 4
[2022/04/21 12:39:48] root INFO:             use_shared_memory : False
[2022/04/21 12:39:48] root INFO:         sampler : 
[2022/04/21 12:39:48] root INFO:             batch_size : 2048
[2022/04/21 12:39:48] root INFO:             drop_last : False
[2022/04/21 12:39:48] root INFO:             name : DistributedBatchSampler
[2022/04/21 12:39:48] root INFO:             shuffle : True
[2022/04/21 12:39:48] root INFO: Global : 
[2022/04/21 12:39:48] root INFO:     checkpoints : None
[2022/04/21 12:39:48] root INFO:     device : gpu
[2022/04/21 12:39:48] root INFO:     epochs : 300
[2022/04/21 12:39:48] root INFO:     eval_during_train : True
[2022/04/21 12:39:48] root INFO:     eval_interval : 1
[2022/04/21 12:39:48] root INFO:     image_shape : [3, 224, 224]
[2022/04/21 12:39:48] root INFO:     output_dir : ./output/
[2022/04/21 12:39:48] root INFO:     pretrained_model : .//output/PPLCNet_x0_25/best_model
[2022/04/21 12:39:48] root INFO:     print_batch_step : 10
[2022/04/21 12:39:48] root INFO:     save_inference_dir : ./deploy/models/class_PPLCNet_x0_25_ImageNet_infer
[2022/04/21 12:39:48] root INFO:     save_interval : 1
[2022/04/21 12:39:48] root INFO:     use_visualdl : False
[2022/04/21 12:39:48] root INFO: Infer : 
[2022/04/21 12:39:48] root INFO:     PostProcess : 
[2022/04/21 12:39:48] root INFO:         class_id_map_file : ../label_list.txt
[2022/04/21 12:39:48] root INFO:         name : Topk
[2022/04/21 12:39:48] root INFO:         topk : 5
[2022/04/21 12:39:48] root INFO:     batch_size : 10
[2022/04/21 12:39:48] root INFO:     infer_imgs : docs/images/inference_deployment/whl_demo.jpg
[2022/04/21 12:39:48] root INFO:     transforms : 
[2022/04/21 12:39:48] root INFO:         DecodeImage : 
[2022/04/21 12:39:48] root INFO:             channel_first : False
[2022/04/21 12:39:48] root INFO:             to_rgb : True
[2022/04/21 12:39:48] root INFO:         ResizeImage : 
[2022/04/21 12:39:48] root INFO:             resize_short : 256
[2022/04/21 12:39:48] root INFO:         CropImage : 
[2022/04/21 12:39:48] root INFO:             size : 224
[2022/04/21 12:39:48] root INFO:         NormalizeImage : 
[2022/04/21 12:39:48] root INFO:             mean : [0.485, 0.456, 0.406]
[2022/04/21 12:39:48] root INFO:             order : 
[2022/04/21 12:39:48] root INFO:             scale : 1.0/255.0
[2022/04/21 12:39:48] root INFO:             std : [0.229, 0.224, 0.225]
[2022/04/21 12:39:48] root INFO:         ToCHWImage : None
[2022/04/21 12:39:48] root INFO: Loss : 
[2022/04/21 12:39:48] root INFO:     Eval : 
[2022/04/21 12:39:48] root INFO:         CELoss : 
[2022/04/21 12:39:48] root INFO:             weight : 1.0
[2022/04/21 12:39:48] root INFO:     Train : 
[2022/04/21 12:39:48] root INFO:         CELoss : 
[2022/04/21 12:39:48] root INFO:             epsilon : 0.1
[2022/04/21 12:39:48] root INFO:             weight : 1.0
[2022/04/21 12:39:48] root INFO: Metric : 
[2022/04/21 12:39:48] root INFO:     Eval : 
[2022/04/21 12:39:48] root INFO:         TopkAcc : 
[2022/04/21 12:39:48] root INFO:             topk : [1, 5]
[2022/04/21 12:39:48] root INFO:     Train : 
[2022/04/21 12:39:48] root INFO:         TopkAcc : 
[2022/04/21 12:39:48] root INFO:             topk : [1, 5]
[2022/04/21 12:39:48] root INFO: Optimizer : 
[2022/04/21 12:39:48] root INFO:     lr : 
[2022/04/21 12:39:48] root INFO:         learning_rate : 0.1
[2022/04/21 12:39:48] root INFO:         name : Cosine
[2022/04/21 12:39:48] root INFO:         warmup_epoch : 10
[2022/04/21 12:39:48] root INFO:     momentum : 0.9
[2022/04/21 12:39:48] root INFO:     name : Momentum
[2022/04/21 12:39:48] root INFO:     regularizer : 
[2022/04/21 12:39:48] root INFO:         coeff : 0.0001
[2022/04/21 12:39:48] root INFO:         name : L2
[2022/04/21 12:39:48] root INFO: train with paddle 2.2.2 and device CUDAPlace(0)
W0421 12:39:48.412415   405 device_context.cc:447] Please NOTE: device: 0, GPU Compute Capability: 7.0, Driver API Version: 11.0, Runtime API Version: 10.1
W0421 12:39:48.417142   405 device_context.cc:465] device: 0, cuDNN Version: 7.6.
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/layers/utils.py:77: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
  return (isinstance(seq, collections.Sequence) and

2.模型预测配置

在配置文件 configs/inference_cls.yaml 中有以下字段用于配置预测参数：

  • Global.infer_imgs：待预测的图片文件路径；
  • Global.inference_model_dir：inference 模型文件所在目录，该目录下需要有文件 inference.pdmodel 和 inference.pdiparams 两个文件；
  • Global.use_tensorrt：是否使用 TesorRT 预测引擎，默认为 False；
  • Global.use_gpu：是否使用 GPU 预测，默认为 True；
  • Global.enable_mkldnn：是否启用 MKL-DNN 加速库，默认为 False。注意 enable_mkldnn 与 use_gpu 同时为 True 时，将忽略 enable_mkldnn，而使用 GPU 预测；
  • Global.use_fp16：是否启用 FP16，默认为 False；
  • PreProcess：用于数据预处理配置；
  • PostProcess：由于后处理配置；
  • PostProcess.Topk.class_id_map_file：数据集 label 的映射文件，默认为 ./utils/imagenet1k_label_list.txt，该文件为 PaddleClas 所使用的 ImageNet 数据集 label 映射文件。

注意:

如果使用 VisionTransformer 系列模型，如 DeiT_384, ViT_384 等，请注意模型的输入数据尺寸，部分模型需要修改参数： PreProcess.resize_short=384, PreProcess.resize=384。

预测文件配置如下：

  1. Global:
  2.   infer_imgs: "../data/testdata/testdata"
  3.   inference_model_dir: "./deploy/models/class_PPLCNet_x0_25_ImageNet_infer"
  4.   batch_size: 1
  5.   use_gpu: True
  6.   enable_mkldnn: True
  7.   cpu_num_threads: 10
  8.   enable_benchmark: True
  9.   use_fp16: False
  10.   ir_optim: True
  11.   use_tensorrt: False
  12.   gpu_mem: 8000
  13.   enable_profile: False
  14.  
  15. PreProcess:
  16.   transform_ops:
  17.     - ResizeImage:
  18.         resize_short: 256
  19.     - CropImage:
  20.         size: 224
  21.     - NormalizeImage:
  22.         scale: 0.00392157
  23.         mean: [0.485, 0.456, 0.406]
  24.         std: [0.229, 0.224, 0.225]
  25.         order: ''
  26.         channel_num: 3
  27.     - ToCHWImage:
  28.  
  29. PostProcess:
  30.   main_indicator: Topk
  31.   Topk:
  32.     topk: 5
  33.     class_id_map_file: "../label_list.txt"
  34.   SavePreLabel:
  35.     save_dir: ./pre_label/Global:
  36.   infer_imgs: "../data/testdata/testdata"
  37.   inference_model_dir: "./deploy/models/class_PPLCNet_x0_25_ImageNet_infer"
  38.   batch_size: 1
  39.   use_gpu: True
  40.   enable_mkldnn: True
  41.   cpu_num_threads: 10
  42.   enable_benchmark: True
  43.   use_fp16: False
  44.   ir_optim: True
  45.   use_tensorrt: False
  46.   gpu_mem: 8000
  47.   enable_profile: False
  48.  
  49. PreProcess:
  50.   transform_ops:
  51.     - ResizeImage:
  52.         resize_short: 256
  53.     - CropImage:
  54.         size: 224
  55.     - NormalizeImage:
  56.         scale: 0.00392157
  57.         mean: [0.485, 0.456, 0.406]
  58.         std: [0.229, 0.224, 0.225]
  59.         order: ''
  60.         channel_num: 3
  61.     - ToCHWImage:
  62.  
  63. PostProcess:
  64.   main_indicator: Topk
  65.   Topk:
  66.     topk: 5
  67.     class_id_map_file: "../label_list.txt"
  68.   SavePreLabel:
  69.     save_dir: ./pre_label/Global:
  70.   infer_imgs: "../data/testdata/testdata"
  71.   inference_model_dir: "./deploy/models/class_PPLCNet_x0_25_ImageNet_infer"
  72.   batch_size: 1
  73.   use_gpu: True
  74.   enable_mkldnn: True
  75.   cpu_num_threads: 10
  76.   enable_benchmark: True
  77.   use_fp16: False
  78.   ir_optim: True
  79.   use_tensorrt: False
  80.   gpu_mem: 8000
  81.   enable_profile: False
  82.  
  83. PreProcess:
  84.   transform_ops:
  85.     - ResizeImage:
  86.         resize_short: 256
  87.     - CropImage:
  88.         size: 224
  89.     - NormalizeImage:
  90.         scale: 0.00392157
  91.         mean: [0.485, 0.456, 0.406]
  92.         std: [0.229, 0.224, 0.225]
  93.         order: ''
  94.         channel_num: 3
  95.     - ToCHWImage:
  96.  
  97. PostProcess:
  98.   main_indicator: Topk
  99.   Topk:
  100.     topk: 5
  101.     class_id_map_file: "../label_list.txt"
  102.   SavePreLabel:
  103.     save_dir: ./pre_label/
  104.

3.预测保存修改

直接修改预测脚本PaddleClas/deploy/python/predict_cls.py的main函数即可

  1. def main(config):
  2.     cls_predictor = ClsPredictor(config)
  3.     image_list = get_image_list(config["Global"]["infer_imgs"])
  4.  
  5.     batch_imgs = []
  6.     batch_names = []
  7.     cnt = 0
  8.     # 写入文件
  9.     f=open('result.csv','w')
  10.     f.write('name,label\n')
  11.     for idx, img_path in enumerate(image_list):
  12.         img = cv2.imread(img_path)
  13.         if img is None:
  14.             logger.warning(
  15.                 "Image file failed to read and has been skipped. The path: {}".
  16.                 format(img_path))
  17.         else:
  18.             img = img[:, :, ::-1]
  19.             batch_imgs.append(img)
  20.             img_name = os.path.basename(img_path)
  21.             batch_names.append(img_name)
  22.             cnt += 1
  23.  
  24.         if cnt % config["Global"]["batch_size"] == 0 or (idx + 1
  25.                                                          ) == len(image_list):
  26.             if len(batch_imgs) == 0:
  27.                 continue
  28.             batch_results = cls_predictor.predict(batch_imgs)
  29.             for number, result_dict in enumerate(batch_results):
  30.                 filename = batch_names[number]
  31.                 clas_ids = result_dict["class_ids"]
  32.                 scores_str = "[{}]".format(", ".join("{:.2f}".format(
  33.                     r) for r in result_dict["scores"]))
  34.                 label_names = result_dict["label_names"]
  35.                 print("{}:\tclass id(s): {}, score(s): {}, label_name(s): {}".
  36.                       format(filename, clas_ids, scores_str, label_names))
  37.                 # 保存预测
  38.                 f.write(filename+','+label_names[0]+'\n')
  39.             batch_imgs = []
  40.             batch_names = []
  41.     if cls_predictor.benchmark:
  42.         cls_predictor.auto_logger.report()
  43.     return

In [ ]

  1. # 覆盖原预测脚本
  2. !cp ~/predict_cls.py ~/PaddleClas/deploy/python/
  3. %cd ~/PaddleClas/
  4. # 开始预测
  5. !python ./deploy/python/predict_cls.py -c ../inference_cls.yaml

预测日志

  1. 0.jpeg:	class id(s): [1, 2, 9, 0, 3], score(s): [0.71, 0.07, 0.06, 0.04, 0.03], label_name(s): ['cat', 'chicken', 'squirrel', 'butterfly', 'cow']
  2. 1.jpeg:	class id(s): [1, 4, 9, 2, 6], score(s): [0.78, 0.14, 0.02, 0.02, 0.01], label_name(s): ['cat', 'dog', 'squirrel', 'chicken', 'horse']
  3. 10.jpeg:	class id(s): [2, 3, 4, 5, 6], score(s): [0.53, 0.15, 0.15, 0.06, 0.05], label_name(s): ['chicken', 'cow', 'dog', 'elephant', 'horse']
  4. 100.jpeg:	class id(s): [1, 4, 9, 2, 0], score(s): [0.61, 0.16, 0.06, 0.04, 0.03], label_name(s): ['cat', 'dog', 'squirrel', 'chicken', 'butterfly']
  5. 1000.jpg:	class id(s): [8, 2, 5, 0, 1], score(s): [0.74, 0.11, 0.03, 0.02, 0.02], label_name(s): ['sheep', 'chicken', 'elephant', 'butterfly', 'cat']

六、提交&总结

1.提交

下载PaddlleClas下的result.csv并提交

In [1]

  1. %cd ~
  2. !head ~/PaddleClas/result.csv
/home/aistudio
name,label
0.jpeg,cat
1.jpeg,cat
10.jpeg,chicken
100.jpeg,cat
1000.jpg,sheep
1001.jpg,sheep
1002.jpg,sheep
1003.jpg,sheep
1004.jpg,sheep

2.总结

  • 一是选择训练模型，其中原数据脚本需要修改
  • 二是预测，需修改预测脚本用于保存结果
  • 三是可持续训练，以获取更好的成绩*
声明：本文内容由网友自发贡献，不代表【wpsshop博客】立场，版权归原作者所有，本站不承担相应法律责任。如您发现有侵权的内容，请联系我们。转载请注明出处：https://www.wpsshop.cn/w/羊村懒王/article/detail/385277
推荐阅读
相关标签

Copyright © 2003-2013 www.wpsshop.cn 版权所有，并保留所有权利。

  

闽ICP备14008679号