What's Changed

• [cherry-pick] add project url and fix a bug by @GreatV in #13281
• [cherry-pick] fix slice op parameters not being passed correctly (#13319) by @GreatV in #13324
• Fix the dictionary bug in tablerec inference by @Topdu in #13364

Full Changelog: v2.8.0...v2.8.1

终于等到你！PaddleOCR 新版本发布！

What's Changed

• [Cherry-pick] #10515 by @ToddBear in #10537
• [BugFix]compat_pillow by @shiyutang in #10596
• [bug fix] fix none res in recovery by @andyjiang1116 in #10603
• Fix seed passing issue of build_dataloader by @RuohengMa in #10614
• [bug fix]rm invalid params by @andyjiang1116 in #10605
• [Cherry-pick] #10441 #10512 by @moehuster in #10593
• 修改数据增强导致的DSR报错 by @xu-peng-7 in #10662
• onnxruntime support gpu by @WenmuZhou in #10668
• Update VQA to use the updated LayoutLM syntax from PaddleNLP by @sijunhe in #9791
• 实现功能：当--savefile为true时，在--output下以当前图片名称后接“.txt”为文件名保存ocr推理结果，解决了issues： by @WilliamQf-AI in #10628
• Cherrypicking GH-10217 and GH-10216 to PaddlePaddle:dygraph by @UserUnknownFactor in #10654
• fix numpy speed by @wanghuancoder in #10773
• Cherrypicking GH-10251 & GH-10181 to PaddleOCR:dygraph by @itasli in #10710
• rec_r45_abinet.yml add max_length and image_size by @xlg-go in #10744
• ch_PP-OCRv4_rec_distill.yml, fix KeyError: 'NRTRLabelDecode' by @xlg-go in #10761
• 根据推理对三通道的图像需求，以及opencv中imread参数说明IMREAD_COLOR(If set, always convert … by @Gmgge in #10777
• Update algorithm_kie_vi_layoutxlm_en.md by @sagarjgb in #10736
• Add new recognition method "ParseQ" by @ToddBear in #10836
• rm fluid for paddle dev by @tink2123 in #10931
• rec_r45_abinet for export model by @xlg-go in #10892
• fix:修复通道数不匹配造成的PPOCRLabel启动失败问题#10748,根据更新日志发现#10655，由于paddleocr中增加了对… by @Gmgge in #10847
• [New] add rec CPPD model by @Topdu in #10990
• fix cls_x and bbox_x is possibly unbound by @SigureMo in #10991
• add svtr large model by @zhangyubo0722 in #10937
• [WIP]support eval pre epoch by @zhangyubo0722 in #11003
• Update kie_datasets_en.md by @sagarjgb in #10735
• fix import collection for py310 by @tink2123 in #11012
• update ppocrv4_framework by @tink2123 in #11048
• Update how_to_do_kie_en.md by @sagarjgb in #10731
• add cppd u14m train model and doc by @Topdu in #11052
• Fixed bug with "max_text_length" for VisionLAN by @victor30608 in #11025
• Cherrypicking GH-10923 to PaddleOCR:dygraph by @itasli in #11069
• Update quickstart_en.md by @sagarjgb in #10732
• Update README.md by @sagarjgb in #10733
• Update algorithm_overview_en.md by @sagarjgb in #10734
• [Cherry-pick] Cherry-pick from release/2.6 by @shiyutang in #11092
• [TIPC]update tipc scripts by @USTCKAY in #11097
• fix satrn export for paddle2.5 by @tink2123 in #11096
• [BugFix]Fix parseq net by @shiyutang in #11126
• update uygur dict by @hfengzhi in #11125
• Add tipc for "ParseQ" method by @ToddBear in #10843
• fix SAR inference, when batch size>1, norm_img_batch and valid_ratios… by @shiyunalex in #11238
• v4 det cml configs by @sylarwcy in #11258
• 解決recognition的train test分割程式執行後的文檔每行間多出一行空格 by @DingHsun in #11280
• Fix for Ambiguous Boolean Evaluation Error in PaddleOCR with Python 3.11 by @muhammadAgfian96 in #11287
• Dygraph【benchmark】add max_mem_reserved for benchmark by @mmglove in #11284
• Fix bug when running on XPU by @RuohengMa in #11299
• Dygraph by @RuohengMa in #11301
• Dygraph fix max_mem_reserved for benchmark by @mmglove in #11341
• 在check_gpu时增加对当前环境可用设备的检查 by @TracebaK in #11293
• Fixed some bugs that caused PPOCRLabel to crash, added ability to expand checkboxes by @g39088902 in #11236
• fix a bug for rec_postprocess.py by @Ataraxy33 in #11389
• Optimize prediction on long image and deduplicate similar boxes with multiple lables by @marswen in #11366
• doc: add doc for satrn by @wkml in #11397
• Update zeros' comment in rec_abinet_head.py by @YesianRohn in #11374
• Fix QPointF IndexError: list index out of range by @firmament2008 in #11393
• update paddlex of readme by @zhangyubo0722 in #11422
• chore: add notes for docker gpu deploy PP-OCRv4 by @sheiy in #11390
• Fix words by @co63oc in #11448
• [Feature]Complete the ppocrv4_act by @ranchongzhi in #11345
• rm QR code in the document by @tink2123 in #11512
• rm QR code by @tink2123 in #11532
• Fix dead links by @MatKollar in #11520
• cherry-pick for lazy import pymupdf and pre-commit by @tink2123 in #11692
• adapter new type promotion rule for Paddle 2.6 by @zxcd in #11698
• setup a workflow for publishing package to pypi by @jzhang533 in #11804
• update link mentioned at #11763 by @jzhang533 in #11764
• fix AttributeError by @GreatV in #11686
• fix: Correct misuse of try_import from paddle.utils by @neteroster in #11820
• Update quickstart.md for a better python pdf demo by @qwedc001 in #11927
• Update quickstart_en.md by @qwedc001 in #11934
• Enhance the OCR recognition accuracy of PPStructure. by @RussellLuo in #11916
• add u14m results of cppd by @Topdu in #11943
• use tensor.shape bug not paddle.shape(tensor) by @wanghuancoder in #11919
• add pre-commit workflow by @GreatV in #11973
• docs: Update FAQ.md, delete repeated question by @xu8117 in #11972
• Fix the bug where Python scripts fail to execute PDF text recognition… by @guangyunms in #11994
• 【OCR Issue No.9】以可选形式支持Visualdl by @Liyulingyue in #11947
• fix weird version info by @GreatV in #12003
• 【OCR Issue No.9】移除明确不适合放在ppocr依赖中的依赖项 by @Liyulingyue in #11946
• Burmese Language dict and corpus by @1chimaruGin in #12020
• 面版识别添加onnx支持完善 by @heweisheng in #12068
• Update README.md by @dyning in #12086
• fix readme codestyle by @GreatV in #12095
• fix wrong link for 通用OCR in README.txt by @tackhwa in #12100
• move PPOCRLabel to PFCCLab/PPOCRLabel by @GreatV in #12104
• move StyleText to PFCCLab/StyleText by @GreatV in #12121
• openocr compti code by @Topdu in #12033
• table rec code by @invictuszhao in #11999
• Error with pyclipper inhomogeneous expanded array by @zovelsanj in #12108
• 【OCR Issue No.2】修复训练过程中找不到對應模型和训练时计算精度报错 by @mattheliu in https://github.com/PaddlePaddle/Paddle...

fix broken v2.7.4

This release contains the missed commits from v2.7.0 to v2.7.1.
fixed : #11824

What's Changed

fixed #11808

What's Changed

• add finnish language files by @savikko in #10850
• fix cls_x and bbox_x is possibly unbound by @SigureMo in #10973
• update ppocrv4_framework by @tink2123 in #11047
• Update ONNX conversion readme_ch.md by @greyovo in #11030
• [TIPC]update tipc scripts and rm fluid api by @USTCKAY in #11098
• fix a bug for rec_postprocess.py by @Ataraxy33 in #11408
• Modify readme 27 by @zhangyubo0722 in #11424
• fix: layout recovery image:xxx.png,err msg: list index out of range by @santlchogva in #11405
• rm QR code in the document by @tink2123 in #11511
• rm QR code by @tink2123 in #11533
• Update custom.md by @jzhang533 in #11636
• fix AttributeError by @GreatV in #11556
• update pre-commit config by @jzhang533 in #11682
• lazy import PyMuPDF by @jzhang533 in #11685
• setup a workflow for publishing package to pypi, and bump version to … by @jzhang533 in #11800

New Contributors

• @savikko made their first contribution in #10850
• @greyovo made their first contribution in #11030
• @santlchogva made their first contribution in #11405
• @jzhang533 made their first contribution in #11636

Full Changelog: v2.7.0...v2.7.2

New Projects

• Add Parseq recognition model.(#10836)
• Add text recognition function to return single character coordinates.(#10515)

New Features

• Add savefile option to save OCR output results.(#10628)
• Add more data preprocessing options to ppocr.py.(#10217)
• A single damaged image does not affect data set inference. (#10216)
• Compatible with fitz version. (#10181)
• Compatible with Pillow10.0 upgrade. (#10405)
• Add Finnish dictionary file. (#10850)
• Onnxruntime supports GPU. (#10668)
• TIPC supports XPU and NPU. (#10658, #10460)
• Add inference on mlu devices. (#10249)

BugFix

• Fixed cannot find the library error when packaged into exe on windows. (#10502)
• Fixed the bug of recognize page is affected by the maximum number of PDF files when recognize multiple PDF files. (#10290)
• Fixed the problem of PPOCRLabel startup failure caused by mismatch in channel number. (#10847)
• Fix memory leak problem of cpp inference. (#10441)
• Modify DSR error caused by data enhancement. (#10662)
• Fix training seed problem. (#10614)
• Fix table_master tipc error. (#10514)
• Fixed the problem of error reporting when ppocr.py uses wandb. (#10251)
• Fix memory leak in predict_rec.py. (#10688)
• Fixed the issue where dis and iou cannot be calculated correctly due to the index error of structure_boxes in the PaddleStructure::rebuild_table function. (#10810)
• Compatible with paddle 2.5 fluid exit. (#10391)
• Fix the performance problem of Tensor.numpy under stride. (#10773)
• Adapt the size of ABINet during export to the size of ABINetRecResizeImg. (#10892)
• ABINet training error. (#10744)
• Fix KeyError in ch_PP-OCRv4_rec_distill.yml. (#10761)

Documentations Fix

• Fix algorithm_kie_vi_layoutxlm_en.md, kie_datasets_en.md, README.md, algorithm_overview.md, how_to_do_kie_en.md document issues. (#10717)
• Update documentation issues with setup.py. (#10749)
• Add pyyaml ​​library in requirements.txt. (#10653)

New Projects

• 增加Parseq 识别模型。（#10836）
• 增加文字识别返回单字识别坐标功能。（#10515）

New Features

• 增加savefile选项，保存ocr输出结果。（#10628）
• 增加ppocr.py 更多数据预处理选项。（#10217）
• 单张破损图片不影响整体数据集推理。（#10216）
• 兼容fitz版本。（#10181）
• 兼容Pillow10.0升级。（#10405）
• 增加芬兰语字典文件。（#10850）
• Onnxruntime 支持GPU。（#10668）
• TIPC支持XPU、NPU。（#10658，#10460）
• 增加在mlu设备上的推理。（#10249）

BugFix

• 修复windows打包成exe找不到库的问题。（#10502）
• 修复多次识别pdf，受第一次页面最大数量影响的bug。（#10290）
• 修复通道数不匹配造成的PPOCRLabel启动失败问题。（#10847）
• 修复cpp推理的内存泄漏问题。（#10441）
• 修改数据增强导致的DSR报错。(#10662）
• 修复训练seed传递问题。（ #10614）
• 修复 table_master tipc 报错。（#10514）
• 修复ppocr.py使用wandb报错问题。（#10251）
• 修复predict_rec.py中的内存泄漏问题。（#10688）
• 修复PaddleStructure::rebuild_table函数中structure_boxes的索引错误导致dis和iou无法正确计算的问题。（#10810）
• 兼容paddle 2.5 fluid退场。 （#10391）
• 修复stride下Tensor.numpy的性能问题。（#10773）
• ABINet导出时尺寸适应 ABINetRecResizeImg 的尺寸。（#10892）
• ABINet训练报错问题。（#10744）
• 解决ch_PP-OCRv4_rec_distill.yml中的keyError报错问题。（#10761）

Documentations Fix

• 修复 algorithm_kie_vi_layoutxlm_en.md、kie_datasets_en.md 、README.md、algorithm_overview.md、how_to_do_kie_en.md文档问题（#10717）
• 更新setup.py的文档问题（#10749）
• 修复requirements.txt中没有pyyaml库的问题（#10653）

New Contributors

• @RuohengMa made their first contribution in #10614
• @WilliamQf-AI made their first contribution in #10628
• @xlg-go made their first contribution in #10744
• @Gmgge made their first contribution in #10777
• @victor30608 made their first contribution in #11025

Full Changelog: v2.7.0...v2.7.1

Release Note

• Release PP-OCRv4, support mobile version and server version
  • PP-OCRv4-mobile：When the speed is comparable, the effect of the Chinese scene is improved by 4.5% compared with PP-OCRv3, the English scene is improved by 10%, and the average recognition accuracy of the 80-language multilingual model is increased by more than 8%.
  • PP-OCRv4-server：Release the OCR model with the highest accuracy at present, the detection model accuracy increased by 4.9% in the Chinese and English scenes, and the recognition model accuracy increased by 2%
    refer quickstart quick use by one line command, At the same time, the whole process of model training, reasoning, and high-performance deployment can also be completed with few code in the General OCR Industry Solution in PaddleX.
• ReleasePP-ChatOCR, a new scheme for extracting key information of general scenes using PP-OCR model and ERNIE LLM.

Release Note

• Release PP-Structurev2，with functions and performance fully upgraded, adapted to Chinese scenes, and new support for Layout Recovery and one line command to convert PDF to Word;
• Layout Analysis optimization: model storage reduced by 95%, while speed increased by 11 times, and the average CPU time-cost is only 41ms;
• Table Recognition optimization: 3 optimization strategies are designed, and the model accuracy is improved by 6% under comparable time consumption;
• Key Information Extraction optimization：a visual-independent model structure is designed, the accuracy of semantic entity recognition is increased by 2.8%, and the accuracy of relation extraction is increased by 9.1%.

Release Note

• Release PP-OCRv3: With comparable speed, the effect of Chinese scene is further improved by 5% compared with PP-OCRv2, the effect of English scene is improved by 11%, and the average recognition accuracy of 80 language multilingual models is improved by more than 5%.
• Release PPOCRLabelv2: Add the annotation function for table recognition task, key information extraction task and irregular text image.
• Release interactive e-book "Dive into OCR", covers the cutting-edge theory and code practice of OCR full stack technology.