Issues

• paddleocr-mcp 0.8.5: pp_structurev3 tool always fails (unguarded context.request_context); fastmcp/mcp bounds unsatisfiable

  PaddlePaddle/PaddleOCR#18391 · Filter by author Sbakkalii opened on Oct 8, 2026

• pp-ocr v6 竖排文本的标点识别有大问题，会丢失大量标点符号，这一点,远不如PP-OCR v5Mobile。

  PaddlePaddle/PaddleOCR#18387 · Filter by author baibufeixueyan opened on Oct 5, 2026
  3

• matplotlib版本不对，3.9.3.5 3.7 都尝试过

  PaddlePaddle/PaddleOCR#18381 · Filter by author cqray1990 opened on Sep 28, 2026

• BGR/RGB mismatch in PP-OCRv6 Android recog preprocessing

  PaddlePaddle/PaddleOCR#18380 · Filter by author ankandrew opened on Sep 28, 2026

• PP-OCRv6 returnWordBox 框不精准

  PaddlePaddle/PaddleOCR#18378 · Filter by author tenghian opened on Sep 24, 2026

• VL1.6 HPS部署的服务无法识别页眉，如何恢复识别

  PaddlePaddle/PaddleOCR#18376 · Filter by author doraemonlei opened on Sep 22, 2026

• v6文字检测和识别问题

  PaddlePaddle/PaddleOCR#18373 · Filter by author lasifea opened on Sep 21, 2026

• 使用vllm部署paddleocr-vl-1.6模型，使用API调用报 ‘KeyError: 'pixel_values'’

  PaddlePaddle/PaddleOCR#18372 · Filter by author nuocheng opened on Sep 20, 2026

• Text on stderr, but not in markdown output.

  PaddlePaddle/PaddleOCR#18371 · Filter by author jbarth-ubhd opened on Sep 18, 2026
  2

• Error in "concatenate_markdown_pages" example in Usage tutorial for PPStructureV3: TypeError: write() argument must be str, not MarkdownResult

  PaddlePaddle/PaddleOCR#18369 · Filter by author Heavygrass opened on Sep 17, 2026

• Broken documentation link in langchain-paddleocr/README.md

  PaddlePaddle/PaddleOCR#18367 · Filter by author HammadAshrafDev opened on Sep 16, 2026

• PP-OCRv6 训练可以使用图片高度为32的？对识别影响大？

  PaddlePaddle/PaddleOCR#18366 · Filter by author cqray1990 opened on Sep 16, 2026

• PP-OCRv6/v5 unified rec ONNX 模型非 ASCII 输出为 CP1252 渲染的 UTF-8 字节（PP-OCRv6_small_rec_onnx）

  PaddlePaddle/PaddleOCR#18364 · Filter by author youming-ai opened on Sep 16, 2026

• resize_norm_img 直接resize问题，宽高比大于10的数据丢弃代码怎么没找到呢

  PaddlePaddle/PaddleOCR#18358 · Filter by author cqray1990 opened on Sep 14, 2026
  2

• PIR optimization does not yet support the Kunlun P800 server environment.

  PaddlePaddle/PaddleOCR#18353 · Filter by author ly-c opened on Sep 9, 2026

• arabic_PP-OCRv5_mobile_rec: native aspect-ratio width under-allocates CTC steps for full-width lines (36.7% → 12.9% CER)

  PaddlePaddle/PaddleOCR#18349 · Filter by author Ahmed-Sinkeat opened on Sep 7, 2026
  4

• 识别不准确

  PaddlePaddle/PaddleOCR#18347 · Filter by author xiaoxinmiao opened on Sep 5, 2026
  5

• ppocrv6 ">=" not supported between instance of 'list' and float

  PaddlePaddle/PaddleOCR#18343 · Filter by author vapelouisliu-cyber opened on Sep 3, 2026

• 容器内推理时 GPU 推理报错 MMU Fault: ENGINE GRAPHICS GPC2 GPCCLIENT_T1_0， fallback 到 CPU，重启容器解决

  PaddlePaddle/PaddleOCR#18328 · Filter by author luoxzhg opened on Aug 26, 2026

• [PP-OCRv6][Performance] High CPU inference latency and low CPU utilization - need Small model accuracy with lower latency

  PaddlePaddle/PaddleOCR#18320 · Filter by author Nandhu910 opened on Aug 20, 2026
  2

• PaddleOCR's recognition performance on the following documents is still not satisfactory

  PaddlePaddle/PaddleOCR#18318 · Filter by author xiyuan27-rgb opened on Aug 20, 2026

• npu 容器启动失败

  PaddlePaddle/PaddleOCR#18315 · Filter by author AleutianXie opened on Aug 18, 2026

• PaddleOcr-Vl-1.6 智能版面分析 识别表格时，内容缺失

  PaddlePaddle/PaddleOCR#18314 · Filter by author zhguokai opened on Aug 18, 2026

• 字符识别错误或者无法识别

  PaddlePaddle/PaddleOCR#18313 · Filter by author fs888888889 opened on Aug 18, 2026
  1

• save_to_markdown overwrites images with same coordinates on different pages

  PaddlePaddle/PaddleOCR#18305 · Filter by author Baron1014 opened on Aug 12, 2026
  1