anylabeling

What's Changed

• Add Segment Anything 2 support.
• Remove YOLOv5.
• Fix crash when removing points.
  • PR: #163
• Canvas performance improvements.
  • PR: #165
• Fix for loading model configs.
  • PR: #166

What's Changed

• Add Segment Anything 2 support.
• Remove YOLOv5.
• Fix crash when removing points.
  • PR: #163
• Canvas performance improvements.
  • PR: #165
• Fix for loading model configs.
  • PR: #166

If you find this project useful, please consider sponsoring its development.

What's Changed

• Fixed some issues when inferencing SAM.
• Adjust color for a better look in Dark mode.

If you find this project useful, please consider sponsoring its development.

What's Changed

• Support models from SAM Exporter.
• Use zip format for model downloading.
• Support MobileSAM - super fast auto labeling experience.
• Fix wrong model URL for SAM ViT-L.

If you find this project useful, please consider sponsoring its development.

What's Changed

• Fix building .dmg for macOS.

If you find this project useful, please consider sponsoring its development.

What's Changed

• Handle some issues with model downloading:
  • Wrong status.
  • Timeout when downloading models.

If you find this project useful, please consider sponsoring its development.

What's Changed

• Add the option to load custom models.
• Please download sample custom models from AnyLabeling Assets v0.4.0.

If you find this project useful, please consider sponsoring its development.

What's Changed

• Add Auto Use Last Label to speed up labeling #74 . Documentation: https://anylabeling.com/docs/auto-use-last-label.

If you find this project useful, please consider sponsoring its development.

What's Changed

• Separated two versions: with GPU and without GPU support.

Installing from Pypi

• For CPU version:

pip install -U anylabeling

• For GPU version:

pip install -U anylabeling-gpu

Download binary files

Use versions with -GPU for GPU support.

If you find this project useful, please consider sponsoring its development.

What's Changed

• Fixed random crashing when scrolling. Reported by @yibaoby.
• Ignore SSL verification when downloading models to avoid errors when using a proxy. #52

If you find this project useful, please consider sponsoring its development.

• Restore the app icon.

What's Changed

• Don't specify providers for ONNXRuntime - Let it decide automatically.
• Optimize image size to max 64x64.

If you find this project useful, please consider sponsoring its development.

What's Changed

Support multiple languages for all UI features:

• English (default language).
• Vietnamese.
• Chinese (zh_CN). Thank @yibaoby for the translation.

Now you can select the language from the menu.

If you find this project useful, please consider sponsoring its development.

What's Changed

• Update icons.
• Fixed: Could not open single image file.

If you find this project useful, please consider sponsoring its development.

What's Changed

• Fixed #50: Crashing when running auto labeling for gray images.
• Fixed wrong label color for new objects from Segment Anything.

If you find this project useful, please consider sponsoring its development.

What's Changed

• Add auto release workflow
  • PR: #49

If you find this project useful, please consider sponsoring its development.

What's Changed

• 🐞fix: fix shape.label is none caused by label_dialog return empty s… by @liaozihang in https://github.com/vietanhdev/anylabeling/pull/44

New Contributors

• @liaozihang made their first contribution in https://github.com/vietanhdev/anylabeling/pull/44
• @qqqhhh-any made their first contribution in https://github.com/vietanhdev/anylabeling/pull/47

Full Changelog: https://github.com/vietanhdev/anylabeling/compare/v0.2.8...v0.2.9

What's Changed

• Rectangle output for SAM
• Fix text editing when switching between modes

• Fix crashing due to QPoint type
• Draw all points only when selecting the shape
• Remove small contours
• Added documentation website at https://anylabeling.com.

• Add all Segment Anything models.
• Validate ONNX models on load.
• Run model in separate thread + add cache.
• Build wheels with Github workflow.
• Inference encoder for Segment Anything in the background.