Starstruk Logo MarkStarstruk

v4.0.0-beta.4

CVHub520/X-AnyLabelingv4.0.0-beta.4Apr 20, 2026by CVHub520

AI Summary

This feature-rich release introduces PaddleOCR document parsing workspace, TensorRT inference backend, annotation check status workflow, and Japanese UI localization support.

Key Highlights

  • Added PaddleOCR document parsing workspace with API client mode
  • Added TensorRT inference backend support for auto-labeling
  • Added annotation check status workflow for labeling quality review
  • Added YOLO11 and YOLO26 SAHI support for auto-labeling
  • Added Japanese UI localization support

New Features

  • PaddleOCR document parsing workspace
  • TensorRT inference backend
  • Annotation check status workflow
  • Ultralytics training with checked files only
  • Annotation visualization export
  • YOLO11 and YOLO26 SAHI support
  • File list copy actions and recent menu
  • Japanese UI localization

Full Release Notes

https://github.com/user-attachments/assets/0c018b6e-f8e9-4045-bc22-0d388ab4853d

> PyPI: https://pypi.org/project/x-anylabeling-cvhub/
> Baidu Cloud: https://pan.baidu.com/s/1pgaw02inCvbEgOme9ajDJA?pwd=e528

> [!NOTE]
> Due to compatibility issues across different systems, if the precompiled version doesn’t work properly on your machine, you can try building and running it from source instead. For details, check out the official installation guide and user documentation.

### 🚀 New Features

- Add PaddleOCR document parsing workspace with an official API client mode and asynchronous API job parsing.
- Add TensorRT inference backend support for auto-labeling. (#1343)
- Add annotation check status workflow for labeling quality review. (#1340)
- Add option to train Ultralytics models with checked files only. (#1340)
- Add annotation visualization export for images and videos.
- Add YOLO11 and YOLO26 SAHI support for auto-labeling. (#1321)
- Add file list copy actions and a recent menu action for reopening the last directory. (#1338,#1346)
- Add Japanese UI localization support and Japanese language selection in the labeling widget. (#1331,#1332)

### 🐛 Bug Fixes

- Keep standard models visible in the remote server model dropdown.
- Support custom PPOCR recognition dictionaries in auto-labeling. (#1344)
- Preserve LaTeX arrow commands in PaddleOCR formula previews.
- Keep the PaddleOCR export dialog open when canceling on macOS.
- Prevent PyQt6 attribute panel crashes in the label widget. (#1337)
- Preserve provided image filenames when VOC filenames lack extensions. (#1334)
- Force zlib resource compression for Windows translation loading. (#1332)

### 🛠️ Improvements

- Package macOS PyInstaller release archives with checksums. (#1285)
- Align rotated-box documentation and optimize keyboard rotation repaint behavior. (#1333)
- Update setup guides to install beta pre-release packages by default with quoted optional dependency extras for zsh compatibility.
- Add repository line-ending attributes to keep source, docs, and config files on LF while preserving CRLF for Windows scripts.

### 🌟 Contributors

A total of 3 developers contributed to this release.

Thank @CVHub520, @kame3mikan3, @tribf

**Full Changelog**: https://github.com/CVHub520/X-AnyLabeling/compare/v4.0.0-beta.3...v4.0.0-beta.4