X-AnyLabeling v3.0.3

DigitShortcutDemo.mp4

🚀 New Features

• Added Digit Shortcut Manager for quick shape creation using numeric keys (#945)
• Added support for preserving existing annotations in SegmentAnything2Video model

🛠️ Improvements

• Updated FAQ entries for ONNX model IR version issues with resolution steps
• Added FAQ entry on accessing external models like 'Google Gemini' through proxy settings
• Converted BrightnessContrastDialog to singleton to reduce instantiation overhead (#954)

🌟 Contributors

A total of 1 developer contributed to this release.

Thank @CVHub520

Full Changelog: v3.0.2...v3.0.3

X-AnyLabeling v3.0.2

YOLOE.mp4

Text/Visual Prompting and Prompt-free for Detection & Segmentation

🚀 New Features

• Support YOLOE model for text/visual prompting and prompt-free grounding

🐛 Bug Fixes

• Fixed crash issue during keypoint annotation (#952)

🌟 Contributors

A total of 1 developers contributed to this release.

Thank @CVHub520

Full Changelog: v3.0.1...v3.0.2

X-AnyLabeling v3.0.1

🚀 New Features

• Add support for Ultralytics RT-DETR object detection models (#944)
• Enhance label history management by adding removal and addition methods for label history

🐛 Bug Fixes

• Add 'pillow' to requirements for macOS, Windows, and Linux (#942)
• Initialize image_data before base64 encoding in save_auto_labeling_result (#946)

🛠️ Improvements

• Optimize image folder import performance by disabling automatic EXIF processing (#945)
• Update requirements-macos.txt (#939)
• Add FAQ entry for OpenSSL Uplink error during startup (#941)

🌟 Contributors

A total of 8 developers contributed to this release.

Thank @DenDen047, @4399123, @CVHub520

Full Changelog: v3.0.0...v3.0.1

X-AnyLabeling v3.0.0

annotation.mp4

Visual Prompting for Detection

Visual Prompting for Segmentation

Chatbot

🚀 New Features

• Add ffmpeg acceleration and non-ASCII path support (#891)
• Allow downloading models from ModelScope in addition to existing sources
• Enable one-click import and export of labels for VLM-R1-OVD
• Enable the Chatbot to annotate multimodal datasets for Vision-Language Models (VLMs)
• Enable automatic saving of group IDs when grouping or ungrouping shapes using shortcut keys G (group) and U (ungroup). (#855)
• Introduce GroupID filter and improve label filtering functionality (#686)
• Support GeCo zero-shot counting model (#863)
• Support Grounding-DINO-1.6-API open-set object detection model
• Support YOLO12 object detection model
• Support D-FINE object detection model
• Support RF-DETR object detection model

🐛 Bug Fixes

• Fix bug in predict_shapes for Florence-2 model (#913)
• Replace os_sorted with natsorted to avoid potential segfault (#906)
• Merge multi-part segmentations into single instance on export (#910)
• Handle exceptions in model loading by initializing local_model_data to an empty dictionary for improved stability (#901)
• Prevent UI disappearance when ESC key is pressed during AI annotation (#423)
• Fixed the bug about exporting the empty labels (#881)

🛠️ Improvements

• Move image conversion to avoid redundant processing when using cached embeddings (#915)
• Add new FAQs addressing common runtime errors and file loading issues, including solutions and references to related GitHub issues (#869, #906, #907)
• Enhance image processing logic to support dynamic batch handling based on model type, improving efficiency in auto-labeling operations
• Introduce iou_threshold and conf_threshold parameters across various model configurations for enhanced detection accuracy
• Remove imgviz dependency from requirements and update colormap implementation in labeling utilities for improved modularity
• Optimize batch processing with UI/backend separation (#757)
• Enhance shape visibility handling in labeling interface (#669)
• Updated the merge_shapes method to handle both rectangle and polygon shapes, allowing for more versatile shape unions. (#561)

🌟 Contributors

A total of 8 developers contributed to this release.

Thank @Pecako2001, @liutao, @shyhyawJou, @talebolano, @urbaneman, @wangxiang0722, @Little-King2022, @CVHub520

Full Changelog: v2.5.4...v3.0.0

X-AnyLabeling v2.5.4

Released versions

Version	MD5
X-AnyLabeling-CPU.exe	45a12c6784b917e5a9a498c4739ca227
X-Anylabeling-Linux-CPU	c498c2e5e01d4989c23375c94bc4f6ff

Note

For GPU acceleration and the macOS version, please consult the guide at this link.
If you need to use the following advanced features, please refer to the corresponding documentation to install additional dependencies.

1. Video object tracking based on Segment-Anything-2: Installation Guide
2. Object proposal generation based on UPN or open_vision model: Installation Guide
3. Interactive visual-text prompting for generic vision tasks: Installation Guide

Important updates

• 🚀 feat(model_zoo): Add YOLOv8-SAM2.1 instance segmentation model support by @CVHub520 in be503eb
• 🐛 fix(label_widget): Improve canvas painting state management during image processing by @CVHub520 in 323580b
• 🐛 fix(label_converter): reset pose_data for each new image in pose mode to ensure correct annotation handling (#791) by
  @CVHub520 in 29ba84e
• 📝 docs: Add guide for selecting all annotation shapes (#759) by @CVHub520 and @aiyou9
  in 62542ba
• 🗑 chore: Remove YOLOv8-EfficientViT-SAM model support as it has been superseded by YOLOv8-SAM2 by @CVHub520 in 592cacf

Full Changelog: v2.5.3...v2.5.4

X-AnyLabeling v2.5.3

Released versions

Version	MD5
X-AnyLabeling-CPU.exe	5f6d9ee6e074d21c2947746fbca4200d
X-Anylabeling-Linux-CPU	47b757f789a8a739a1b4e44f9507c5b6

Note

For GPU acceleration and the macOS version, please consult the guide at this link.
If you need to use the following advanced features, please refer to the corresponding documentation to install additional dependencies.

1. Video object tracking based on Segment-Anything-2: Installation Guide
2. Object proposal generation based on UPN: Installation Guide
3. Interactive visual-text prompting for generic vision tasks: Installation Guide

Important updates

• ⚡ perf(label_widget): optimize performance for large files (#743) by @CVHub520 in ea546fc
• 🐛 fix(run_all_images): resolve loading mistake in batch processing (#777) by @CVHub520 in 44f2fa3
• 🐛 fix(label_converter): reset pose_data for each new image in pose mode to ensure correct annotation handling (#791) by @CVHub520 in 29ba84e

Full Changelog: v2.5.2...v2.5.3

X-AnyLabeling v2.5.2

Released versions

Version	MD5
X-AnyLabeling-CPU.exe	46358ce4db8728d9760945bc66d45ea3
X-Anylabeling-Linux-CPU	0338e3ffe7d15ee1bb0c50eab7e486e8

Note

For GPU acceleration and the macOS version, please consult the guide at this link.
If you need to use the following advanced features, please refer to the corresponding documentation to install additional dependencies.

1. Video object tracking based on Segment-Anything-2: Installation Guide
2. Object proposal generation based on UPN: Installation Guide
3. Interactive visual-text prompting for generic vision tasks: Installation Guide

Important updates

• 🚀 feat(overview): enhance export functionality with classes.txt and zip output (#775) by @CVHub520 in 2a811cb
• 🐛 fix(run_all_images): resolve image dimension validation errors (#762) by @CVHub520 in 7438c89
• 📚 docs(README): update downloads badges by @CVHub520 in 85f629f

Full Changelog: v2.5.1...v2.5.2

X-AnyLabeling v2.5.1

Released versions

Version	MD5
X-AnyLabeling-CPU.exe	9f45a1b7fb6f9c530b47ff7e46a95989
X-Anylabeling-Linux-CPU	752fb67594b9ab7e0b00f97e37d3c50e

Note

For GPU acceleration and the macOS version, please consult the guide at this link.
If you need to use the following advanced features, please refer to the corresponding documentation to install additional dependencies.

1. Video object tracking based on Segment-Anything-2: Installation Guide
2. Object proposal generation based on UPN: Installation Guide
3. Interactive visual-text prompting for generic vision tasks: Installation Guide

Important updates

• 🚀 feat(run_all_images): optimize inference efficiency during batch task execution by @chevydream in 900602e
• 🚀 feat(model): add support for Hyper-YOLO by @CVHub520 in 51b8cb2
• 🚀 feat(Florence-2): add CPU support and optimize model loading by @CVHub520 in 3d6239a
• 🐛 fix(RAM): GBK codec decoding error in model loading process by @CVHub520 in 8f7840d
• 🐛 fix(RAM): correct dtype casting in RAM model preprocessing by @CVHub520 in 9a6a036
• 🐛 fix(config): correct text_encoder_type path in open_vision.yaml by @CVHub520 in 64db73c
• 🎨 style(MacOS): improve visibility in dark mode by @CVHub520 in 5c71187
• 📚 docs(OpenVision): update README with new installation instructions by @CVHub520 in 4628963

Full Changelog: v2.5.0...v2.5.1

X-AnyLabeling v2.5.0

Released versions

Version	MD5
X-AnyLabeling-CPU.exe	e92cbfc03300c40312f5b4d4aa821200
X-Anylabeling-Linux-CPU	34f765b5cba2cd6a58ce05931045958a

Note

For GPU acceleration and the macOS version, please consult the guide at this link.
If you need to use the following advanced features, please refer to the corresponding documentation to install additional dependencies.

1. Video object tracking based on Segment-Anything-2: Installation Guide
2. Object proposal generation based on UPN: Installation Guide
3. Interactive visual-text prompting for generic vision tasks: Installation Guide

Interactive Visual-Text Prompting for Generic Vision Tasks

Important updates

• 🚀 feat(model): add support interactive visual-text prompting for generic vision tasks by @CVHub520 in 8f5768b
• 🚀 feat(auto-labeling): optimize rectangle mode to use minimum bounding box by @CVHub520 in 4118efd
• 🚀 feat(model): add support for SAM2.1 model by @CVHub520 in 812634d
• 🚀 feat(model): add support for Florence-2 model (#679) by @CVHub520 in 67aac31
• 🚀 feat(model): add support for the upn model to generate proposal boxes by @CVHub520 in 05eea0a
• 🚀 feat(model): add support for yolov5-sahi model by @CVHub520 in 455b354
• 🚀 feat(tool): add range selection for label batch modification (#708) by @CVHub520 in dcb7c51
• 🚀 feat(export): add options dialog with additional export path selection (#702) by @CVHub520 in 398cb6b
• 🚀 feat(import/export): add support for importing/exporting COCO keypoint annotations (#190) by @CVHub520 in 43cd1da
• 🚀 feat(model): Added support for DocLayout-YOLO model
• 🚀 feat(drawer): add option to color bounding boxes by category or instance by @CVHub520 in ba719c5
• 🚀 feat(ui): add action to loop through each label by @julianstirling in 96f0617
• 🐛 fix(import): handle invalid file paths in natural sort (#734) by @CVHub520 in 2b27a90
• 🐛 fix(export): improve mask overlapping handling in custom_to_mask method by @CVHub520 in 2dd1968
• 🐛 fix(tool): fix the path parsing error of the save_crop function by @wpNZC in 3a7c1e2
• 🐛 fix(ui): disable delete action when no shapes present by @CVHub520 in 71d4cf2
• 🐛 fix(model): correct image normalization in Recognize-Anything-Model preprocessing (#657) by @CVHub520 in 0b07dc4
• 📚 docs(installation): add onnxruntime compatibility by @CVHub520 in 37e87a2
• 🎨 refactor(ui): modernize GroupIDModifyDialog with improved styling by @CVHub520 in 7545561

Full Changelog: v2.4.4...v2.5.0

X-AnyLabeling v2.4.4

Released versions

Version	MD5
X-AnyLabeling-CPU.exe	d8d33f16e5afae68b410ad331c98d2da
X-Anylabeling-Linux-CPU	1e6746f82e095dcc22d51b11594adc96

Note

For GPU acceleration and the macOS version, please consult the guide at this link.
To use the video tracking feature with segment-anything-2, refer to the same document.

Important updates

Add support for YOLO11 Det/OBB/Pose/Seg/Track models

• Integrate YOLO11 models from ultralytics v8.3.0
• Enable detection, oriented bounding box, pose estimation, segmentation, and tracking functionalities
• Update relevant documentation and dependencies)

Full Changelog: v2.4.3...v2.4.4