Issues

• HunyuanOCR-1.5 skips one image in a simple two-image OCR request

  Tencent-Hunyuan/HunyuanOCR#133 · Filter by author ASheng1019 opened on Oct 4, 2026

• CHAOS-Bench

  Tencent-Hunyuan/HunyuanOCR#132 · Filter by author caowanxia123 opened on Sep 24, 2026
  1

• Batch inference CLI exits successfully even when images fail

  Tencent-Hunyuan/HunyuanOCR#127 · Filter by author kaluli123123 opened on Sep 14, 2026
  2

• Streaming inference leaves HTTP responses open on errors and interruption

  Tencent-Hunyuan/HunyuanOCR#126 · Filter by author kaluli123123 opened on Sep 14, 2026
  2

• doc_parse normalization deletes coordinate content inside multiline HTML tables

  Tencent-Hunyuan/HunyuanOCR#125 · Filter by author kaluli123123 opened on Sep 14, 2026
  2

• Use merged Draft speculative decoding in llama.cpp

  Tencent-Hunyuan/HunyuanOCR#124 · Filter by author Thomzoy opened on Sep 3, 2026
  2

• 坐标漂移

  Tencent-Hunyuan/HunyuanOCR#123 · Filter by author ProtectedInventory opened on Aug 20, 2026

• Why does the DFlash draft use target layers [1,8,15,22]?

  Tencent-Hunyuan/HunyuanOCR#121 · Filter by author sleepyshep opened on Aug 10, 2026

• DOC_parser幻觉情况很严重呀

  Tencent-Hunyuan/HunyuanOCR#119 · Filter by author LuLu2kANG opened on Aug 4, 2026
  1

• HunyuanOCR-1.5: doc_parse task emits a degenerate coordinate string on a figure-heavy sparse-text page (deterministic)

  Tencent-Hunyuan/HunyuanOCR#118 · Filter by author AIwork4me opened on Jul 31, 2026
  1

• SFT Failed SFT失败

  Tencent-Hunyuan/HunyuanOCR#117 · Filter by author xinuan opened on Jul 30, 2026
  7

• Verify evals on Papers with Code

  Tencent-Hunyuan/HunyuanOCR#116 · Filter by author NielsRogge opened on Jul 22, 2026

• Prompt of Video Subtitle Extraction Task

  Tencent-Hunyuan/HunyuanOCR#115 · Filter by author Breeze-Zero opened on Jul 20, 2026
  1

• Recommended max image resolution / vision-token budget for HunyuanOCR-1.5 stable inference? (high-resolution ViT instability observed on ROCm)

  Tencent-Hunyuan/HunyuanOCR#114 · Filter by author AIwork4me opened on Jul 15, 2026
  5

• Information Extraction for Multipage PDF Documents

  Tencent-Hunyuan/HunyuanOCR#113 · Filter by author Docubytes opened on Jul 15, 2026

• Question about DFlash / AR decoding throughput on A800 80G

  Tencent-Hunyuan/HunyuanOCR#112 · Filter by author wwwjjjj opened on Jul 14, 2026

• Some Questions About Benchmark

  Tencent-Hunyuan/HunyuanOCR#111 · Filter by author cuicheng01 opened on Jul 13, 2026

• 文字的坐标和图片背景的颜色相关，并且差异比较大

  Tencent-Hunyuan/HunyuanOCR#105 · Filter by author lhb2 opened on Mar 16, 2026
  1

• Question regarding commercial license cost for >100M MAU / 关于MAU超过1亿时的商业授权费用问题

  Tencent-Hunyuan/HunyuanOCR#103 · Filter by author frkake opened on Mar 10, 2026

• 推理时会出现重复某段字符串直到最大token

  Tencent-Hunyuan/HunyuanOCR#101 · Filter by author simenh4444-max opened on Feb 27, 2026
  2

• 不同图片分辨率对于识别效果的影响

  Tencent-Hunyuan/HunyuanOCR#100 · Filter by author WZK040104 opened on Feb 11, 2026

• 对于多行文本的value，结构化输出效果不是很好

  Tencent-Hunyuan/HunyuanOCR#99 · Filter by author F0undLinks opened on Feb 10, 2026

• 会出官方api服务吗？

  Tencent-Hunyuan/HunyuanOCR#97 · Filter by author qidianlinjin1990 opened on Jan 27, 2026

• Question: Training HunyuanOCR for Information Extraction with BBox output

  Tencent-Hunyuan/HunyuanOCR#96 · Filter by author Docubytes opened on Jan 25, 2026
  1

• Feedback and Issues with HunyuanOCR Model Inference and Recognition Performance

  Tencent-Hunyuan/HunyuanOCR#95 · Filter by author relic-yuexi opened on Jan 21, 2026
  3