Skip to content

Commit ad792f1

Browse files
committed
update docs and remove old docs
1 parent 5f0b90a commit ad792f1

10 files changed

+19
-272
lines changed

doc/doc_ch/algorithm_rec_svtrv2.md

-150
This file was deleted.

doc/doc_ch/algorithm_table_slanet.md

-111
This file was deleted.

docs/algorithm/table_recognition/algorithm_table_slanet.md

+5-5
Original file line numberDiff line numberDiff line change
@@ -6,7 +6,7 @@ comments: true
66

77
## 1. 算法简介
88

9-
PaddleOCR 算法模型挑战赛 - 赛题二:通用表格识别任务排行榜第一算法。核心思路
9+
该算法由来自北京交通大学机器学习与认识计算研究团队的ocr识别队研发,其在PaddleOCR算法模型挑战赛 - 赛题二:通用表格识别任务中排行榜荣获一等奖,排行榜精度相比PP-Structure表格识别模型提升0.8%,推理速度提升3倍。优化思路如下
1010

1111
- 1. 改善推理过程,至EOS停止,速度提升3倍
1212
- 2. 升级Backbone为LCNetV2(SSLD版本)
@@ -16,9 +16,9 @@ PaddleOCR 算法模型挑战赛 - 赛题二:通用表格识别任务排行榜
1616

1717
在PubTabNet表格识别公开数据集上,算法复现效果如下:
1818

19-
|模型|骨干网络|配置文件|acc|
20-
| --- | --- | --- | --- |
21-
|SLANet|LCNetV2|[configs/table/SLANet_lcnetv2.yml](../../configs/table/SLANet_lcnetv2.yml)|76.67%|
19+
|模型|骨干网络|配置文件|acc|下载链接|
20+
| --- | --- | --- | --- | --- |
21+
|SLANet|LCNetV2|[configs/table/SLANet_lcnetv2.yml](../../configs/table/SLANet_lcnetv2.yml)|76.67%| [训练模型](https://paddleocr.bj.bcebos.com/openatom/ch_ppstructure_openatom_SLANetv2_train.tar) /[推理模型](https://paddleocr.bj.bcebos.com/openatom/ch_ppstructure_openatom_SLANetv2_infer.tar) |
2222

2323
## 2. 环境配置
2424

@@ -67,7 +67,7 @@ python3 tools/export_model.py -c configs/table/SLANet_lcnetv2.yml -o Global.pret
6767

6868
```bash linenums="1"
6969
cd ppstructure/
70-
python3.7 table/predict_structure.py --table_model_dir=../inference/slanet_lcnetv2_infer/ --table_char_dict_path=../ppocr/utils/dict/table_structure_dict.txt --image_dir=docs/table/table.jpg --output=../output/table_slanet_lcnetv2 --use_gpu=False --benchmark=True --enable_mkldnn=True
70+
python table/predict_structure.py --table_model_dir=../inference/slanet_lcnetv2_infer/ --table_char_dict_path=../ppocr/utils/dict/table_structure_dict.txt --image_dir=docs/table/table.jpg --output=../output/table_slanet_lcnetv2 --use_gpu=False --benchmark=True --enable_mkldnn=True --table_max_len=512
7171
# 预测文件夹下所有图像时,可修改image_dir为文件夹,如 --image_dir='docs/table'。
7272
```
7373

docs/algorithm/text_recognition/algorithm_rec_svtrv2.md

+11-3
Original file line numberDiff line numberDiff line change
@@ -8,7 +8,15 @@ comments: true
88

99
### SVTRv2算法简介
1010

11-
[PaddleOCR 算法模型挑战赛 - 赛题一:OCR 端到端识别任务](https://aistudio.baidu.com/competition/detail/1131/0/introduction)排行榜第一算法。主要思路:1、检测和识别模型的Backbone升级为RepSVTR;2、识别教师模型升级为SVTRv2,可识别长文本。
11+
🔥 该算法由来自复旦大学视觉与学习实验室([FVL](https://fvl.fudan.edu.cn))的[OpenOCR](https://github.com/Topdu/OpenOCR)团队研发,其在[PaddleOCR算法模型挑战赛 - 赛题一:OCR端到端识别任务](https://aistudio.baidu.com/competition/detail/1131/0/introduction)中荣获一等奖,B榜端到端识别精度相比PP-OCRv4提升2.5%,推理速度持平。主要思路:1、检测和识别模型的Backbone升级为RepSVTR;2、识别教师模型升级为SVTRv2,可识别长文本。
12+
13+
|模型|配置文件|端到端|下载链接|
14+
| --- | --- | --- | --- |
15+
|PP-OCRv4| |A榜 62.77% <br> B榜 62.51%| [Model List](../../doc/doc_ch/models_list.md) |
16+
|SVTRv2(Rec Sever)|[configs/rec/SVTRv2/rec_svtrv2_ch.yml](../../configs/rec/SVTRv2/rec_svtrv2_ch.yml)|A榜 68.81% (使用PP-OCRv4检测模型)| [训练模型](https://paddleocr.bj.bcebos.com/openatom/openatom_rec_svtrv2_ch_train.tar) / [推理模型](https://paddleocr.bj.bcebos.com/openatom/openatom_rec_svtrv2_ch_infer.tar) |
17+
|RepSVTR(Mobile)|[识别](../../configs/rec/SVTRv2/rec_repsvtr_ch.yml) <br> [识别蒸馏](../../configs/rec/SVTRv2/rec_svtrv2_ch_distillation.yml) <br> [检测](../../configs/det/det_repsvtr_db.yml)|B榜 65.07%| 识别: [训练模型](https://paddleocr.bj.bcebos.com/openatom/openatom_rec_repsvtr_ch_train.tar) / [推理模型](https://paddleocr.bj.bcebos.com/openatom/openatom_rec_repsvtr_ch_infer.tar) <br> 识别蒸馏: [训练模型](https://paddleocr.bj.bcebos.com/openatom/openatom_rec_svtrv2_distill_ch_train.tar) / [推理模型](https://paddleocr.bj.bcebos.com/openatom/openatom_rec_svtrv2_distill_ch_infer.tar) <br> 检测: [训练模型](https://paddleocr.bj.bcebos.com/openatom/openatom_det_repsvtr_ch_train.tar) / [推理模型](https://paddleocr.bj.bcebos.com/openatom/openatom_det_repsvtr_ch_infer.tar) |
18+
19+
🚀 快速使用:参考PP-OCR推理[说明文档](../../doc/doc_ch/inference_ppocr.md),将检测和识别模型替换为上表中对应的RepSVTR或SVTRv2推理模型即可使用。
1220

1321
## 2. 环境配置
1422

@@ -96,15 +104,15 @@ Predicts of ./doc/imgs_words_en/word_10.png:('pain', 0.9999998807907104)
96104

97105
### 4.2 C++推理部署
98106

99-
由于C++预处理后处理还未支持SVTRv2
107+
准备好推理模型后,参考[cpp infer](https://github.com/PaddlePaddle/PaddleOCR/tree/main/deploy/cpp_infer)教程进行操作即可。
100108

101109
### 4.3 Serving服务化部署
102110

103111
暂不支持
104112

105113
### 4.4 更多推理部署
106114

107-
暂不支持
115+
- Paddle2ONNX推理:准备好推理模型后,参考[paddle2onnx](https://github.com/PaddlePaddle/PaddleOCR/tree/main/deploy/paddle2onnx)教程操作。
108116

109117
## 5. FAQ
110118

File renamed without changes.
File renamed without changes.

doc/doc_en/inference_args_en.md renamed to docs/ppocr/inference/inference_args.en.md

+2-2
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,6 @@
11
# PaddleOCR Model Inference Parameter Explanation
22

3-
When using PaddleOCR for model inference, you can customize the modification parameters to modify the model, data, preprocessing, postprocessing, etc. (parameter file: [utility.py](../../tools/infer/utility.py)),The detailed parameter explanation is as follows:
3+
When using PaddleOCR for model inference, you can customize the modification parameters to modify the model, data, preprocessing, postprocessing, etc. (parameter file: [utility.py](https://github.com/PaddlePaddle/PaddleOCR/blob/main/tools/infer/utility.py)),The detailed parameter explanation is as follows:
44

55
* Global parameters
66

@@ -88,7 +88,7 @@ The relevant parameters of the PSE algorithm are as follows
8888
| :--: | :--: | :--: | :--: |
8989
| rec_algorithm | str | "CRNN" | Text recognition algorithm name, currently supports `CRNN`, `SRN`, `RARE`, `NETR`, `SAR`, `ViTSTR`, `ABINet`, `VisionLAN`, `SPIN`, `RobustScanner`, `SVTR`, `SVTR_LCNet` |
9090
| rec_model_dir | str | None, it is required if using the recognition model | recognition inference model paths |
91-
| rec_image_shape | str | "3,48,320" ] | Image size at the time of recognition |
91+
| rec_image_shape | str | "3,48,320" | Image size at the time of recognition |
9292
| rec_batch_num | int | 6 | batch size |
9393
| max_text_length | int | 25 | The maximum length of the recognition result, valid in `SRN` |
9494
| rec_char_dict_path | str | "./ppocr/utils/ppocr_keys_v1.txt" | character dictionary file |

0 commit comments

Comments
 (0)