diff --git a/doc/doc_ch/recognition.md b/doc/doc_ch/recognition.md index 7b51ae9788c397249e10e6e374226b003bcbbe00..b657e01e1b5bcc8495bb2c88ab0f255b8071252e 100644 --- a/doc/doc_ch/recognition.md +++ b/doc/doc_ch/recognition.md @@ -375,7 +375,7 @@ python3 -m paddle.distributed.launch --gpus '0' tools/eval.py -c configs/rec/rec 根据配置文件中设置的的 `save_model_dir` 和 `save_epoch_step` 字段，会有以下几种参数被保存下来： ``` -seed_ch/ +output/rec/ ├── best_accuracy.pdopt ├── best_accuracy.pdparams ├── best_accuracy.states diff --git a/doc/doc_ch/training.md b/doc/doc_ch/training.md index 2c588aec6db10c14f8cffe7756712110f10f958f..25bd276c733b1d06cb96713bd07277c44901c8cb 100644 --- a/doc/doc_ch/training.md +++ b/doc/doc_ch/training.md @@ -4,14 +4,14 @@ 同时会简单介绍PaddleOCR模型训练数据的组成部分，以及如何在垂类场景中准备数据finetune模型。 -### 基本概念 +### 1. 基本概念 OCR(Optical Character Recognition,光学字符识别)是指对图像进行分析识别处理，获取文字和版面信息的过程，是典型的计算机视觉任务， 通常由文本检测和文本识别两个子任务构成。 模型调优时需要关注以下参数： -- 学习率 +#### 1.1 学习率 学习率是训练神经网络的重要超参数之一，它代表在每一次迭代中梯度向损失函数最优解移动的步长。 在PaddleOCR中提供了多种学习率更新策略,可以通过配置文件修改，例如： @@ -30,7 +30,7 @@ Piecewise 代表分段常数衰减，在不同的学习阶段指定不同的学 warmup_epoch 代表在前5个epoch中，学习率将逐渐从0增加到base_lr。全部策略可以参考代码[learning_rate.py](../../ppocr/optimizer/learning_rate.py) 。 -- 正则化 +#### 1.2 正则化 正则化可以有效的避免算法过拟合，PaddleOCR中提供了L1、L2正则方法，L1 和 L2 正则化是最常用的正则化方法。L1 正则化向目标函数添加正则化项，以减少参数的绝对值总和；而 L2 正则化中，添加正则化项的目的在于减少参数平方的总和。配置方法如下： @@ -43,7 +43,7 @@ Optimizer: ``` -- 评估指标： +#### 1.3 评估指标： （1）检测阶段：先按照检测框和标注框的IOU评估，IOU大于某个阈值判断为检测准确。这里检测框和标注框不同于一般的通用目标检测框，是采用多边形进行表示。检测准确率：正确的检测框个数在全部检测框的占比，主要是判断检测指标。检测召回率：正确的检测框个数在全部标注框的占比，主要是判断漏检的指标。 @@ -51,7 +51,7 @@ Optimizer: （3）端到端统计： 端对端召回率：准确检测并正确识别文本行在全部标注文本行的占比； 端到端准确率：准确检测并正确识别文本行在 检测到的文本行数量 的占比； 准确检测的标准是检测框与标注框的IOU大于某个阈值，正确识别的的检测框中的文本与标注的文本相同。 -### 常见问题 +### 2. 常见问题 **Q**: 基于深度学习的文字检测方法有哪几种？各有什么优缺点？ @@ -77,9 +77,9 @@ Optimizer: （2）统计训练样本文字数目。最长字符数目的选取考虑满足80%的训练样本。然后中文字符长宽比近似认为是1，英文认为3：1，预估一个最长宽度。 -### 数据与垂类场景 +### 3. 数据与垂类场景 -- 训练数据： +#### 3.1 训练数据： 目前开源的模型，数据集和量级如下： - 检测： @@ -94,12 +94,12 @@ Optimizer: 其中，公开数据集都是开源的，用户可自行搜索下载，也可参考[中文数据集](./datasets.md)，合成数据暂不开源，用户可使用开源合成工具自行合成，可参考的合成工具包括[text_renderer](https://github.com/Sanster/text_renderer) 、[SynthText](https://github.com/ankush-me/SynthText) 、[TextRecognitionDataGenerator](https://github.com/Belval/TextRecognitionDataGenerator) 等。 -- 垂类场景 +#### 3.2 垂类场景 PaddleOCR主要聚焦通用OCR，如果有垂类需求，您可以用PaddleOCR+垂类数据自己训练； 如果缺少带标注的数据，或者不想投入研发成本，建议直接调用开放的API，开放的API覆盖了目前比较常见的一些垂类。 -- 自己构建数据集 +#### 3.3 自己构建数据集 在构建数据集时有几个经验可供参考： diff --git a/doc/doc_en/config_en.md b/doc/doc_en/config_en.md index 5e5847c4b298553b2d376b90196b61b7e0286efe..6780bc8b48c51a83c680dee7cff699f8fd24c274 100644 --- a/doc/doc_en/config_en.md +++ b/doc/doc_en/config_en.md @@ -120,3 +120,86 @@ In ppocr, the network is divided into four stages: Transform, Backbone, Neck and | batch_size_per_card | Single card batch size during training | 256 | \ | | drop_last | Whether to discard the last incomplete mini-batch because the number of samples in the data set cannot be divisible by batch_size | True | \ | | num_workers | The number of sub-processes used to load data, if it is 0, the sub-process is not started, and the data is loaded in the main process | 8 | \ | + + +## MULTILINGUAL CONFIG FILE GENERATION + +PaddleOCR currently supports 80 (except Chinese) language recognition. A multi-language configuration file template is +provided under the path `configs/rec/multi_languages`: [rec_multi_language_lite_train.yml](../../configs/rec/multi_language/rec_multi_language_lite_train.yml)。 + +There are two ways to create the required configuration file:： + +### Automatically generated by script + +[generate_multi_language_configs.py](../../configs/rec/multi_language/generate_multi_language_configs.py) Can help you generate configuration files for multi-language models + +- Take Italian as an example, if your data is prepared in the following format: + ``` + |-train_data + |- it_train.txt # train_set label + |- it_val.txt # val_set label + |- data + |- word_001.jpg + |- word_002.jpg + |- word_003.jpg + | ... + ``` + + You can use the default parameters to generate a configuration file: + + ```bash + # The code needs to be run in the specified directory + cd PaddleOCR/configs/rec/multi_language/ + # Set the configuration file of the language to be generated through the -l or --language parameter. + # This command will write the default parameters into the configuration file + python3 generate_multi_language_configs.py -l it + ``` + +- If your data is placed in another location, or you want to use your own dictionary, you can generate the configuration file by specifying the relevant parameters: + + ```bash + # -l or --language field is required + # --train to modify the training set + # --val to modify the validation set + # --data_dir to modify the data set directory + # --dict to modify the dict path + # -o to modify the corresponding default parameters + cd PaddleOCR/configs/rec/multi_language/ + python3 generate_multi_language_configs.py -l it \ # language + --train {path/of/train_label.txt} \ # path of train_label + --val {path/of/val_label.txt} \ # path of val_label + --data_dir {train_data/path} \ # root directory of training data + --dict {path/of/dict} \ # path of dict + -o Global.use_gpu=False # whether to use gpu + ... + + ``` +Italian is made up of Latin letters, so after executing the command, you will get the rec_latin_lite_train.yml. + +### Manually modify the configuration file + + You can also manually modify the following fields in the template: + + ``` + Global: + use_gpu: True + epoch_num: 500 + ... + character_type: it # language + character_dict_path: {path/of/dict} # path of dict + + Train: + dataset: + name: SimpleDataSet + data_dir: train_data/ # root directory of training data + label_file_list: ["./train_data/train_list.txt"] # train label path + ... + + Eval: + dataset: + name: SimpleDataSet + data_dir: train_data/ # root directory of val data + label_file_list: ["./train_data/val_list.txt"] # val label path + ... + + ``` diff --git a/doc/doc_en/recognition_en.md b/doc/doc_en/recognition_en.md index ff5b802d8dbf54e842c035bce9ed480428eade83..f15eb347e064c97d213728535a53c35ed9db94d8 100644 --- a/doc/doc_en/recognition_en.md +++ b/doc/doc_en/recognition_en.md @@ -8,8 +8,8 @@ - [2 TRAINING](#TRAINING) - [2.1 Data Augmentation](#Data_Augmentation) - - [2.2 Training](#Training) - - [2.3 Multi-language](#Multi_language) + - [2.2 General Training](#Training) + - [2.3 Multi-language Training](#Multi_language) - [3 EVALUATION](#EVALUATION) @@ -21,8 +21,8 @@ PaddleOCR supports two data formats: -- `LMDB` is used to train data sets stored in lmdb format; -- `general data` is used to train data sets stored in text files: +- `LMDB` is used to train data sets stored in lmdb format（LMDBDataSet）; +- `general data` is used to train data sets stored in text files（SimpleDataSet）: Please organize the dataset as follows: @@ -86,9 +86,10 @@ Similar to the training set, the test set also needs to be provided a folder con #### 1.2 Dataset download -If you do not have a dataset locally, you can download it on the official website [icdar2015](http://rrc.cvc.uab.es/?ch=4&com=downloads). Also refer to [DTRB](https://github.com/clovaai/deep-text-recognition-benchmark#download-lmdb-dataset-for-traininig-and-evaluation-from-here) ，download the lmdb format dataset required for benchmark +- ICDAR2015 -If you want to reproduce the paper indicators of SRN, you need to download offline [augmented data](https://pan.baidu.com/s/1-HSZ-ZVdqBF2HaBZ5pRAKA), extraction code: y3ry. The augmented data is obtained by rotation and perturbation of mjsynth and synthtext. Please unzip the data to {your_path}/PaddleOCR/train_data/data_lmdb_Release/training/path. +If you do not have a dataset locally, you can download it on the official website [icdar2015](http://rrc.cvc.uab.es/?ch=4&com=downloads). +Also refer to [DTRB](https://github.com/clovaai/deep-text-recognition-benchmark#download-lmdb-dataset-for-traininig-and-evaluation-from-here) ，download the lmdb format dataset required for benchmark PaddleOCR provides label files for training the icdar2015 dataset, which can be downloaded in the following ways: @@ -99,6 +100,26 @@ wget -P ./train_data/ic15_data https://paddleocr.bj.bcebos.com/dataset/rec_gt_t wget -P ./train_data/ic15_data https://paddleocr.bj.bcebos.com/dataset/rec_gt_test.txt ``` +PaddleOCR also provides a data format conversion script, which can convert ICDAR official website label to a data format +supported by PaddleOCR. The data conversion tool is in `ppocr/utils/gen_label.py`, here is the training set as an example: + +``` +# convert the official gt to rec_gt_label.txt +python gen_label.py --mode="rec" --input_path="{path/of/origin/label}" --output_label="rec_gt_label.txt" +``` + +The data format is as follows, (a) is the original picture, (b) is the Ground Truth text file corresponding to each picture: + +![](../datasets/icdar_rec.png) + + +- Multilingual dataset + +The multi-language model training method is the same as the Chinese model. The training data set is 100w synthetic data. A small amount of fonts and test data can be downloaded using the following two methods. +* [Baidu Netdisk](https://pan.baidu.com/s/1bS_u207Rm7YbY33wOECKDA) ,Extraction code:frgi. +* [Google drive](https://drive.google.com/file/d/18cSWX7wXSy4G0tbKJ0d9PuIaiwRLHpjA/view) + + #### 1.3 Dictionary @@ -154,6 +175,18 @@ If you want to support the recognition of the `space` category, please set the ` ### 2 TRAINING + +#### 2.1 Data Augmentation + +PaddleOCR provides a variety of data augmentation methods. All the augmentation methods are enabled by default. + +The default perturbation methods are: cvtColor, blur, jitter, Gasuss noise, random crop, perspective, color reverse, TIA augmentation. + +Each disturbance method is selected with a 40% probability during the training process. For specific code implementation, please refer to: [rec_img_aug.py](../../ppocr/data/imaug/rec_img_aug.py) + + +#### 2.2 General Training + PaddleOCR provides training scripts, evaluation scripts, and prediction scripts. In this section, the CRNN recognition model will be used as an example: First download the pretrain model, you can download the trained model to finetune on the icdar2015 data: @@ -170,21 +203,15 @@ tar -xf rec_mv3_none_bilstm_ctc_v2.0_train.tar && rm -rf rec_mv3_none_bilstm_ctc Start training: ``` -# GPU training Support single card and multi-card training, specify the card number through --gpus +# GPU training Support single card and multi-card training # Training icdar15 English data and The training log will be automatically saved as train.log under "{save_model_dir}" + +#specify the single card training(Long training time, not recommended) +python3 tools/train.py -c configs/rec/rec_icdar15_train.yml +#specify the card number through --gpus python3 -m paddle.distributed.launch --gpus '0,1,2,3' tools/train.py -c configs/rec/rec_icdar15_train.yml ``` - -#### 2.1 Data Augmentation - -PaddleOCR provides a variety of data augmentation methods. All the augmentation methods are enabled by default. -The default perturbation methods are: cvtColor, blur, jitter, Gasuss noise, random crop, perspective, color reverse, TIA augmentation. - -Each disturbance method is selected with a 40% probability during the training process. For specific code implementation, please refer to: [rec_img_aug.py](../../ppocr/data/imaug/rec_img_aug.py) - - -#### 2.2 Training PaddleOCR supports alternating training and evaluation. You can modify `eval_batch_step` in `configs/rec/rec_icdar15_train.yml` to set the evaluation frequency. By default, it is evaluated every 500 iter and the best acc model is saved under `output/rec_CRNN/best_accuracy` during the evaluation process. @@ -277,87 +304,7 @@ Eval: **Note that the configuration file for prediction/evaluation must be consistent with the training.** -#### 2.3 Multi-language - -PaddleOCR currently supports 80 (except Chinese) language recognition. A multi-language configuration file template is -provided under the path `configs/rec/multi_languages`: [rec_multi_language_lite_train.yml](../../configs/rec/multi_language/rec_multi_language_lite_train.yml)。 - -There are two ways to create the required configuration file:： - -1. Automatically generated by script - -[generate_multi_language_configs.py](../../configs/rec/multi_language/generate_multi_language_configs.py) Can help you generate configuration files for multi-language models - -- Take Italian as an example, if your data is prepared in the following format: - ``` - |-train_data - |- it_train.txt # train_set label - |- it_val.txt # val_set label - |- data - |- word_001.jpg - |- word_002.jpg - |- word_003.jpg - | ... - ``` - - You can use the default parameters to generate a configuration file: - - ```bash - # The code needs to be run in the specified directory - cd PaddleOCR/configs/rec/multi_language/ - # Set the configuration file of the language to be generated through the -l or --language parameter. - # This command will write the default parameters into the configuration file - python3 generate_multi_language_configs.py -l it - ``` - -- If your data is placed in another location, or you want to use your own dictionary, you can generate the configuration file by specifying the relevant parameters: - - ```bash - # -l or --language field is required - # --train to modify the training set - # --val to modify the validation set - # --data_dir to modify the data set directory - # --dict to modify the dict path - # -o to modify the corresponding default parameters - cd PaddleOCR/configs/rec/multi_language/ - python3 generate_multi_language_configs.py -l it \ # language - --train {path/of/train_label.txt} \ # path of train_label - --val {path/of/val_label.txt} \ # path of val_label - --data_dir {train_data/path} \ # root directory of training data - --dict {path/of/dict} \ # path of dict - -o Global.use_gpu=False # whether to use gpu - ... - - ``` -Italian is made up of Latin letters, so after executing the command, you will get the rec_latin_lite_train.yml. - -2. Manually modify the configuration file - - You can also manually modify the following fields in the template: - - ``` - Global: - use_gpu: True - epoch_num: 500 - ... - character_type: it # language - character_dict_path: {path/of/dict} # path of dict - - Train: - dataset: - name: SimpleDataSet - data_dir: train_data/ # root directory of training data - label_file_list: ["./train_data/train_list.txt"] # train label path - ... - - Eval: - dataset: - name: SimpleDataSet - data_dir: train_data/ # root directory of val data - label_file_list: ["./train_data/val_list.txt"] # val label path - ... - - ``` +#### 2.3 Multi-language Training Currently, the multi-language algorithms supported by PaddleOCR are: @@ -376,9 +323,6 @@ Currently, the multi-language algorithms supported by PaddleOCR are: For more supported languages, please refer to : [Multi-language model](https://github.com/PaddlePaddle/PaddleOCR/blob/release/2.1/doc/doc_en/multi_languages_en.md#4-support-languages-and-abbreviations) -The multi-language model training method is the same as the Chinese model. The training data set is 100w synthetic data. A small amount of fonts and test data can be downloaded using the following two methods. -* [Baidu Netdisk](https://pan.baidu.com/s/1bS_u207Rm7YbY33wOECKDA),Extraction code:frgi. -* [Google drive](https://drive.google.com/file/d/18cSWX7wXSy4G0tbKJ0d9PuIaiwRLHpjA/view) If you want to finetune on the basis of the existing model effect, please refer to the following instructions to modify the configuration file: @@ -429,18 +373,37 @@ python3 -m paddle.distributed.launch --gpus '0' tools/eval.py -c configs/rec/rec ### 4 PREDICTION - -#### 4.1 Training engine prediction Using the model trained by paddleocr, you can quickly get prediction through the following script. -The default prediction picture is stored in `infer_img`, and the weight is specified via `-o Global.checkpoints`: +The default prediction picture is stored in `infer_img`, and the trained weight is specified via `-o Global.checkpoints`: + + +According to the `save_model_dir` and `save_epoch_step` fields set in the configuration file, the following parameters will be saved: + +``` +output/rec/ +├── best_accuracy.pdopt +├── best_accuracy.pdparams +├── best_accuracy.states +├── config.yml +├── iter_epoch_3.pdopt +├── iter_epoch_3.pdparams +├── iter_epoch_3.states +├── latest.pdopt +├── latest.pdparams +├── latest.states +└── train.log +``` + +Among them, best_accuracy.* is the best model on the evaluation set; iter_epoch_x.* is the model saved at intervals of `save_epoch_step`; latest.* is the model of the last epoch. ``` # Predict English results python3 tools/infer_rec.py -c configs/rec/ch_ppocr_v2.0/rec_chinese_lite_train_v2.0.yml -o Global.pretrained_model={path/to/weights}/best_accuracy Global.load_static_weights=false Global.infer_img=doc/imgs_words/en/word_1.jpg ``` + Input image: ![](../imgs_words/en/word_1.png) diff --git a/doc/doc_en/training_en.md b/doc/doc_en/training_en.md index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..36eae509784b3a14863ba3d734d296a428d74c6c 100644 --- a/doc/doc_en/training_en.md +++ b/doc/doc_en/training_en.md @@ -0,0 +1,126 @@ +## MODEL TRAINING + +This article will introduce the basic concepts that need to be mastered during model training and the tuning methods during training. + +At the same time, it will briefly introduce the components of the PaddleOCR model training data and how to prepare the data finetune model in the vertical scene. + +### 1. Basic concepts + +OCR (Optical Character Recognition) refers to the process of analyzing and recognizing images to obtain text and layout information. It is a typical computer vision task. +It usually consists of two subtasks: text detection and text recognition. + +The following parameters need to be paid attention to when tuning the model: + +#### 1.1 Learning rate + +The learning rate is one of the important hyperparameters for training neural networks. It represents the step length of the gradient moving to the optimal solution of the loss function in each iteration. +A variety of learning rate update strategies are provided in PaddleOCR, which can be modified through configuration files, for example: + +``` +Optimizer: + ... + lr: + name: Piecewise + decay_epochs : [700, 800] + values : [0.001, 0.0001] + warmup_epoch: 5 +``` + +Piecewise stands for piecewise constant attenuation. Different learning rates are specified in different learning stages, +and the learning rate is the same in each stage. + +warmup_epoch means that in the first 5 epochs, the learning rate will gradually increase from 0 to base_lr. For all strategies, please refer to the code [learning_rate.py](../../ppocr/optimizer/learning_rate.py). + +#### 1.2 Regularization + +Regularization can effectively avoid algorithm overfitting. PaddleOCR provides L1 and L2 regularization methods. +L1 and L2 regularization are the most commonly used regularization methods. +L1 regularization adds a regularization term to the objective function to reduce the sum of absolute values of the parameters; +while in L2 regularization, the purpose of adding a regularization term is to reduce the sum of squared parameters. +The configuration method is as follows: + +``` +Optimizer: + ... + regularizer: + name: L2 + factor: 2.0e-05 +``` + +#### 1.3 Evaluation indicators: + +(1) Detection stage: First, evaluate according to the IOU of the detection frame and the labeled frame. If the IOU is greater than a certain threshold, it is judged that the detection is accurate. Here, the detection frame and the label frame are different from the general general target detection frame, and they are represented by polygons. Detection accuracy: the percentage of the correct detection frame number in all detection frames is mainly used to judge the detection index. Detection recall rate: the percentage of correct detection frames in all marked frames, which is mainly an indicator of missed detection. + +(2) Recognition stage: Character recognition accuracy, that is, the ratio of correctly recognized text lines to the number of marked text lines. Only the entire line of text recognition pairs can be regarded as correct recognition. + +(3) End-to-end statistics: End-to-end recall rate: accurately detect and correctly identify the proportion of text lines in all labeled text lines; End-to-end accuracy rate: accurately detect and correctly identify the number of text lines in the detected text lines The standard for accurate detection is that the IOU of the detection box and the labeled box is greater than a certain threshold, and the text in the correctly identified detection box is the same as the labeled text. + + +### 2. FAQ + +**Q**: What are the text detection methods based on deep learning? What are the advantages and disadvantages of each? + + A: Commonly used deep learning-based text detection methods can generally be divided into two categories: regression-based and segmentation-based, and of course there are some methods that combine the two. + + (1) Methods based on regression are divided into box regression and pixel value regression. a. The methods that use box regression mainly include CTPN, Textbox series and EAST. This type of algorithm has a better effect on regular shape text detection, but it cannot accurately detect irregular shape text. b. The methods of pixel value regression mainly include CRAFT and SA-Text. This type of algorithm can detect curved text and has an excellent effect on small text, but the real-time performance is not enough. + + (2) Algorithms based on segmentation, such as PSENet, are not limited by the shape of the text, and can achieve better results for texts of various shapes, but the post-processing is often more complicated, leading to serious time-consuming. At present, there are also some algorithms that are specifically improved for this problem, such as DB, which approximates the binarization, makes it guideable, and integrates it into training, so as to obtain a more accurate boundary, which greatly reduces the time-consuming post-processing. + + + +**Q**: For Chinese line text recognition, which is better, CTC or Attention? + + A: + (1) From the point of view of effect, the recognition effect of CTC in general OCR scene is better than Attention, because there are more characters in the dictionary with recognition, and the commonly used Chinese characters are more than 3,000 characters. If the training samples are insufficient, for these characters Sequence relationship mining is more difficult. The advantages of the Attention model in the Chinese scene cannot be reflected. Moreover, Attention is suitable for short sentence recognition, and it is relatively poor in recognition of long sentences. + + (2) In terms of training and prediction speed, Attention's serial decoding structure limits the prediction speed, while the CTC network structure is more efficient and has an advantage in prediction speed. + +**Q**: How to choose a suitable network input shape when training CRNN recognition? + + A: The general height is 32, the longest width is selected, there are two methods: + + (1) Calculate the aspect ratio distribution of training sample images. The selection of the maximum aspect ratio considers 80% of the training samples. + + (2) Count the number of texts in training samples. The selection of the longest number of characters considers the training sample that satisfies 80%. Then the aspect ratio of Chinese characters is approximately considered to be 1, and that of English is 3:1, and the longest width is estimated. + + + +### 3. Data and vertical scenes + +#### 3.1 Training data + +The current open source models, data sets and magnitudes are as follows: + +- Detection: + - English data set, ICDAR2015 + - Chinese data set, LSVT street view data set training data 3w pictures + +- Identification: + - English data set, MJSynth and SynthText synthetic data, the data volume is tens of millions. + - Chinese data set, LSVT street view data set crops the image according to the truth value, and performs position calibration, a total of 30w images. In addition, based on the LSVT corpus, 500w of synthesized data. + - Small language data set, using different corpora and fonts, respectively generated 100w synthetic data set, and using ICDAR-MLT as the verification set. + +Among them, the public data sets are all open source, users can search and download by themselves, or refer to [Chinese data set](./datasets.md), synthetic data is not open source, users can use open source synthesis tools to synthesize by themselves. Synthesis tools include [text_renderer](https://github.com/Sanster/text_renderer), [SynthText](https://github.com/ankush-me/SynthText), [TextRecognitionDataGenerator](https://github.com/Belval/TextRecognitionDataGenerator) etc. + + +#### 3.2 Vertical scene + +PaddleOCR mainly focuses on general OCR. If you have vertical requirements, you can use PaddleOCR + vertical data to train yourself; +If there is a lack of labeled data, or if you do not want to invest in research and development costs, it is recommended to directly call the open API, which covers some of the more common vertical categories. + + +#### 3.3 Build your own data set + +There are several experiences for reference when constructing the data set: + +(1) The amount of data in the training set: + + a. The data required for detection is relatively small. For Fine-tune based on the PaddleOCR model, 500 sheets are generally required to achieve good results. + b. Recognition is divided into English and Chinese. Generally, English scenarios require hundreds of thousands of data to achieve good results, while Chinese requires several million or more. + + +(2) When the amount of training data is small, you can try the following three ways to get more data: + + a. Manually collect more training data, the most direct and effective way. + b. Basic image processing or transformation based on PIL and opencv. For example, the three modules of ImageFont, Image, ImageDraw in PIL write text into the background, opencv's rotating affine transformation, Gaussian filtering and so on. + c. Use data generation algorithms to synthesize data, such as algorithms such as pix2pix.