1. 环境准备

本次训练的工程代码为PaddleOCR-2.7可自行去官网下载，linux环境，cuda11.6

（1）安装paddle

python -m pip install paddlepaddle-gpu==2.6.2.post116 -i https://www.paddlepaddle.org.cn/packages/stable/cu116/

（2）安装paddleOCR等一系列依赖安装包

python tools/infer/predict_system.py --image_dir ./ppocr_img/imgs/2.png --use_angle_cls True --use_gpu True --det_model_dir models/ch_PP-OCRv3_det_infer --rec_model_dir models/ch_PP-OCRv3_rec_infer --cls_model_dir models/ch_ppocr_mobile_v2.0_cls_infer

通过上述命令，测试是否能用。

2. PPOCRLabel工具使用

此工具用于标注OCR数据，我是在windows上使用，又重新在windows上配了环境，需要安装paddlepaddle、numpy、opencv等包，注意numpy版本，我出现报错，要求numpy版本不准超过2.0，，所以我安装了 numpy=1.24.4

使用方式：

进入PPOCRLabel目录中 cd D:\workspace\code\PaddleOCR2.7\PPOCRLabel

执行 python PPOCRLabel.py 

标注结束后，点击File->Export Label，会生成Label.txt，点击File->Export Recognition Result，会生成crop_img、rec_gt.txt，分别用于文本区域检测训练和文本内容识别训练。

至此，数据准备工作完成。

3. 文本检测训练

在开始训练之前，需要在官网中下载所文本检测、文本识别需要的预训练权重

3.1 修改配置文件

进入PaddleOCR-2.7\configs\det\ch_PP-OCRv3，修改ch_PP-OCRv3_det_student.yml配置文件

Global:
  debug: false
  use_gpu: true
  epoch_num: 300    #训练轮数
  log_smooth_window: 20
  print_batch_step: 10
  save_model_dir: ./output/ch_PP-OCR_V3_det/  #模型保存路径
  save_epoch_step: 100
  eval_batch_step:
  - 0
  - 400
  cal_metric_during_train: true  #创建计算指标，用于保存最好的模型, 默认是flase
  pretrained_model: /home/workspace/code/PaddleOCR2.7/models/ch_PP-OCRv3_det_distill_train/student   #预训练权重路径
  checkpoints: null      #./output/ch_PP-OCR_V3_det/latest                                                 #断点训练，一旦训练发生中断，可继续恢复当前训练
  save_inference_dir: null
  use_visualdl: false
  infer_img: ppocr_img/imgs/2.png  #doc/imgs_en/img_10.jpg                                                 #推理图片
  save_res_path: ./checkpoints/det_db/predicts_db.txt
  distributed: true

Architecture:
  model_type: det
  algorithm: DB
  Transform:
  Backbone:
    name: MobileNetV3
    scale: 0.5
    model_name: large
    disable_se: True
  Neck:
    name: RSEFPN
    out_channels: 96
    shortcut: True
  Head:
    name: DBHead
    k: 50

Loss:
  name: DBLoss
  balance_loss: true
  main_loss_type: DiceLoss
  alpha: 5
  beta: 10
  ohem_ratio: 3
Optimizer:
  name: Adam
  beta1: 0.9
  beta2: 0.999
  lr:
    name: Cosine
    learning_rate: 0.001
    warmup_epoch: 2
  regularizer:
    name: L2
    factor: 5.0e-05
PostProcess:
  name: DBPostProcess
  thresh: 0.3
  box_thresh: 0.6
  max_candidates: 1000
  unclip_ratio: 1.5
Metric:
  name: DetMetric
  main_indicator: hmean
Train:
  dataset:
    name: SimpleDataSet
    data_dir: /home/workspace/code/PaddleOCR2.7/myData/barcode/                  #训练数据集路径
    label_file_list:
      - /home/workspace/code/PaddleOCR2.7/myData/barcode/Label.txt               #训练数据集标签
    ratio_list: [1.0]
    transforms:
    - DecodeImage:
        img_mode: BGR
        channel_first: false
    - DetLabelEncode: null
    - IaaAugment:
        augmenter_args:
        - type: Fliplr
          args:
            p: 0.5
        - type: Affine
          args:
            rotate:
            - -10
            - 10
        - type: Resize
          args:
            size:
            - 0.5
            - 3
    - EastRandomCropData:
        size:
        - 960
        - 960
        max_tries: 50
        keep_ratio: true
    - MakeBorderMap:
        shrink_ratio: 0.4
        thresh_min: 0.3
        thresh_max: 0.7
    - MakeShrinkMap:
        shrink_ratio: 0.4
        min_text_size: 8
    - NormalizeImage:
        scale: 1./255.
        mean:
        - 0.485
        - 0.456
        - 0.406
        std:
        - 0.229
        - 0.224
        - 0.225
        order: hwc
    - ToCHWImage: null
    - KeepKeys:
        keep_keys:
        - image
        - threshold_map
        - threshold_mask
        - shrink_map
        - shrink_mask
  loader:
    shuffle: true
    drop_last: false
    batch_size_per_card: 8
    num_workers: 4
Eval:
  dataset:
    name: SimpleDataSet
    data_dir: /home/riins/workspace/code/PaddleOCR2.7/myData/barcode/        #验证集路径
    label_file_list:
      - /home/riins/workspace/code/PaddleOCR2.7/myData/barcode/Label.txt     #验证集标签
    transforms:
    - DecodeImage:
        img_mode: BGR
        channel_first: false
    - DetLabelEncode: null
    - DetResizeForTest: null
    - NormalizeImage:
        scale: 1./255.
        mean:
        - 0.485
        - 0.456
        - 0.406
        std:
        - 0.229
        - 0.224
        - 0.225
        order: hwc
    - ToCHWImage: null
    - KeepKeys:
        keep_keys:
        - image
        - shape
        - polys
        - ignore_tags
  loader:
    shuffle: false
    drop_last: false
    batch_size_per_card: 1
    num_workers: 2

3.2 执行训练

python tools/train.py -c configs/det/ch_PP-OCRv3/my_ch_PP-OCRv3_det_student.yml 

3.3 模型验证

python tools/eval.py --config configs/det/ch_PP-OCRv3/my_ch_PP-OCRv3_det_student.yml -o Global.checkpoints=./output/ch_PP-OCR_V3_det/latest

3.4 推理测试

python tools/infer_det.py --config configs/det/ch_PP-OCRv3/my_ch_PP-OCRv3_det_student.yml

预训练权重推理结果

训练模型推理结果

3.5 模型导出

python tools/export_model.py -c configs/det/ch_PP-OCRv3/my_ch_PP-OCRv3_det_student.yml -o Global.checkpoints=./output/ch_PP-OCR_V3_det/latest -o Global.save_inference_dir=./inference/ch_PP-OCRv3_det_barcode

4. 文本识别训练

4.1 修改配置文件

进入PaddleOCR-2.7\configs\rec\PP-OCRv3，修改ch_PP-OCRv3_rec.yml配置文件

Global:
  debug: false
  use_gpu: true
  epoch_num: 300
  log_smooth_window: 20
  print_batch_step: 10
  save_model_dir: ./output/rec_ppocr_v4
  save_epoch_step: 100
  eval_batch_step: [0, 2000]
  cal_metric_during_train: true
  pretrained_model: #/home/riins/workspace/code/PaddleOCR2.7/models/ch_PP-OCRv4_rec_train/student   #预训练模型加载路径
  checkpoints: ./output/rec_ppocr_v4/latest
  save_inference_dir:
  use_visualdl: false
  infer_img: ppocr_img/imgs/ocr1_crop_0.jpg
  character_dict_path: ppocr/utils/ppocr_keys_v1.txt
  max_text_length: &max_text_length 25
  infer_mode: false
  use_space_char: true
  distributed: true
  save_res_path: ./output/rec/predicts_ppocrv3.txt


Optimizer:
  name: Adam
  beta1: 0.9
  beta2: 0.999
  lr:
    name: Cosine
    learning_rate: 0.001
    warmup_epoch: 5
  regularizer:
    name: L2
    factor: 3.0e-05


Architecture:
  model_type: rec
  algorithm: SVTR_LCNet
  Transform:
  Backbone:
    name: PPLCNetV3
    scale: 0.95
  Head:
    name: MultiHead
    head_list:
      - CTCHead:
          Neck:
            name: svtr
            dims: 120
            depth: 2
            hidden_dims: 120
            kernel_size: [1, 3]
            use_guide: True
          Head:
            fc_decay: 0.00001
      - NRTRHead:
          nrtr_dim: 384
          max_text_length: *max_text_length

Loss:
  name: MultiLoss
  loss_config_list:
    - CTCLoss:
    - NRTRLoss:

PostProcess:  
  name: CTCLabelDecode

Metric:
  name: RecMetric
  main_indicator: acc

Train:
  dataset:
    name: MultiScaleDataSet
    ds_width: false
    data_dir: /home/riins/workspace/code/PaddleOCR2.7/myData/barcode/ 
    ext_op_transform_idx: 1
    label_file_list:
    - /home/riins/workspace/code/PaddleOCR2.7/myData/barcode/rec_gt.txt    #验证图片标注文件
    transforms:
    - DecodeImage:
        img_mode: BGR
        channel_first: false
    - RecConAug:
        prob: 0.5
        ext_data_num: 2
        image_shape: [48, 320, 3]
        max_text_length: *max_text_length
    - RecAug:
    - MultiLabelEncode:
        gtc_encode: NRTRLabelEncode
    - KeepKeys:
        keep_keys:
        - image
        - label_ctc
        - label_gtc
        - length
        - valid_ratio
  sampler:
    name: MultiScaleSampler
    scales: [[320, 32], [320, 48], [320, 64]]
    first_bs: &bs 192
    fix_bs: false
    divided_factor: [8, 16] # w, h
    is_training: True
  loader:
    shuffle: true
    batch_size_per_card: *bs
    drop_last: true
    num_workers: 8
Eval:
  dataset:
    name: SimpleDataSet
    data_dir: /home/riins/workspace/code/PaddleOCR2.7/myData/barcode/ 
    label_file_list:
    - /home/riins/workspace/code/PaddleOCR2.7/myData/barcode/rec_gt.txt    #验证图片标注文件
    transforms:
    - DecodeImage:
        img_mode: BGR
        channel_first: false
    - MultiLabelEncode:
        gtc_encode: NRTRLabelEncode
    - RecResizeImg:
        image_shape: [3, 48, 320]
    - KeepKeys:
        keep_keys:
        - image
        - label_ctc
        - label_gtc
        - length
        - valid_ratio
  loader:
    shuffle: false
    drop_last: false
    batch_size_per_card: 1
    num_workers: 4

4.2 执行训练

python tools/train.py -c configs/rec/PP-OCRv4/ch_PP-OCRv4_rec.yml

4.3 模型验证

python tools/eval.py -c configs/rec/PP-OCRv4/ch_PP-OCRv4_rec.yml -o Global.checkpoints=./output/rec_ppocr_v4/latest

4.4 推理测试

python tools/infer_rec.py -c configs/rec/PP-OCRv4/ch_PP-OCRv4_rec.yml

4.5 模型导出

python tools/export_model.py -c configs/rec/PP-OCRv4/ch_PP-OCRv4_rec.yml -o Global.checkpoints=./output/rec_ppocr_v4/latest -o Global.save_inference_dir=./inference/ch_PP-OCRv4_rec_barcode