."
},
{
"path": "MANIFEST.in",
"content": "recursive-include . *"
},
{
"path": "README.md",
"content": "\n
\n ![](\"https://github.com/0ssamaak0/DLTA-AI/blob/master/assets/icon.png?raw=true\")
\n
\n\n
\n\n
\nData Labeling, Tracking and Annotation with AI\n
\n\nDLTA-AI is the next generation of annotation tools, integrating the power of Computer Vision SOTA models to
Labelme in a seamless expirence and intuitive workflow to make creating image datasets easier than ever before\n\n\n[](https://0ssamaak0.github.io/DLTA-AI/)\n[](https://www.youtube.com/@DLTA-AI)\n[](https://discord.gg/9Rmwcnc4)\n[](https://pypi.org/project/DLTA-AI/)\n[](https://github.com/0ssamaak0/DLTA-AI/releases)\n[](https://github.com/0ssamaak0/DLTA-AI/issues)\n[](https://github.com/0ssamaak0/DLTA-AI/commits)\n[](https://github.com/0ssamaak0/DLTA-AI/blob/master/LICENSE)\n\n\n\n\n[Installation](#installation-%EF%B8%8F) 🛠️ | [Segment Anything](#Segment-Anything-) 🪄 | [Model Selection](#model-selection-) 🤖 | [Segmentation](#segmentation-) 🎨 | [Object Tracking](#object-tracking-) 🚗 | [Export](#export-) 📤 | [Other Features](#other-features-) 🌟| [Contributing](#contributing-) 🤝| [Acknowledgements](#acknowledgements-)🙏| [Resources](#resources-) 🌐 | [License](#license-) 📜\n\n
\n\n\n
\n\n# Model Selection 🤖\nFor model selection, DLTA-AI provides the **Model Explorer** to utilize the power of the numerous models in [mmdetection](https://github.com/open-mmlab/mmdetection/tree/2.x) and [ultralytics YOLOv8](https://github.com/ultralytics/ultralytics) as well as the models of [SAM](https://github.com/facebookresearch/segment-anything)\n\nthe to give the user the ability to compare, download and select from the library of models\n\n\n\n
\n\n# Segmentation 🎨\nUsing the models from the **Model Explorer**, DLTA-AI provides a seamless expirence to annotate single image or batch of images, with options to select classes, modify threshold, and full control to edit the segmentation results.\n\n\n\n\n
\nand as mentioned before, **SAM** is fully integrated in DLTA-AI to provide zero-shot segmentation for any class, and to improve the quality of segmentation\n\n# Object Tracking 🚗\nBuilt on top of the segmentation and detection models, DLTA-AI provides a complete solution for Object Tracking, with 5 different models for tracking\n\nTo import DLTA-AI have options for video navigation, tracking settings and different visualization options with the ability to export the tracking results to a video file\n\nBeside this, DLTA-AI provides a completely new way to modify the tracking results, including edit and delete propagation across frames\n\n\n\n\n\n
\n\nBeside automatic tracking models, DLTA-AI provides different methods of interpolation and filling gaps between frames to fix occlusions and unpredicted behaviors in a semi-automatic way\n\n\n\n\n
\n\n# Export 📤\nFor Instance Segmentation, DLTA-AI provides to option to export the segmentation to standard COCO format, and the results of tracking to MOT format, and a video file for the tracking results with desired visualization options e.g., show id, bbox, class name, etc.\n\n\n\n\n\n
\n\nDLTA-AI provides also the ability to add user-defined or custom export formats that can be used for any purpose, once the user defines his own format, it will be available in the export menu.\n\n# Other Features 🌟\n\n- Threshold Selection (Confidence and IoU)\n- Select Classes (from 80 COCO classes) with option to save default classes\n- Track assigned objects only\n- Merging models (Run both models and merge the results)\n- Show Runtime Type (CPU/GPU)\n- Show GPU Memory Usage\n- Video Navigation (Frame by Frame, Fast Forward, Fast Backward, Play/Pause)\n- Light / Dark Theme Support (syncs with OS theme)\n- Fully Customizable UI (drag and drop, show/hide)\n- OS Notifications (for long running tasks)\n- using orjson for faster json serialization\n- additional script (external) to evaluate the results of segmentation (COCO)\n- additional script (external) to extract frames from a video file for future use\n- User shortcuts and preferences settings\n\n# Contributing 🤝\nDLTA-AI is an open source project and contributions are very welcome, specially in this early stage of development.\n\nYou can contribute in many ways:\n- Create an [issue](https://github.com/0ssamaak0/DLTA-AI/issues) Reporting bugs 🐞 or suggesting new features 🌟 or just give your feedback 📝\n\n- Create a [pull request](https://github.com/0ssamaak0/DLTA-AI/pulls) to fix bugs or add new features, or just to improve the code quality, optimize performance, documentation, or even just to fix typos\n\n- Review [pull requests](https://github.com/0ssamaak0/DLTA-AI/pulls) and help with the code review process\n\n- Spread the word about DLTA-AI and help us grow the community 🌎, by sharing the project on social media, or just by telling your friends about it\n\n# Acknowledgements 🙏\nThis tool is part of a Graduation Project at [Faculty of Engineering, Ain Shams University](https://eng.asu.edu.eg/) under the supervision of:\n\n- [Dr. Karim Ismail](https://carleton.ca/cee/profile/karim-ismail/)\n- [Dr. Ahmed Osama](ahmed.osama@eng.asu.edu.eg)\n- Dr. Watheq El-Kharashy\n- [Eng. Yousra El-Qattan](https://www.linkedin.com/in/youssra-elqattan/)\n\nwe want also to thank our friends who helped us with testing, feedback and suggestions:\n\n- [Eng. Houssam Siyoufi](https://www.linkedin.com/in/houssam-siyoufi-163627110/)\n- [Amin Mohamed](https://github.com/AminMohamed-3)\n- [Badr Mohamed](https://github.com/Badr-1)\n- [Ahmed Mahmoud](https://github.com/ahmedd-mahmoudd)\n- [Youssef Ashraf](https://github.com/0xNine9)\n- [Chadi Ashraf](https://github.com/Chady00)\n\n\n# Resources 🌐\n- [Labelme](https://github.com/wkentaro/labelme)\n- [Segment Anything (SAM)](https://github.com/facebookresearch/segment-anything)\n- [MMDetection](https://github.com/open-mmlab/mmdetection/tree/2.x)\n- [ultralytics YOLOv8](https://github.com/ultralytics/ultralytics)\n- [mikelbrostrom yolov8_tracking](https://github.com/mikel-brostrom/yolov8_tracking)\n- [orjson](https://github.com/ijl/orjson)\n- [icons8](https://icons8.com/)\n\n# License 📜\nDLTA-AI is released under the [GPLv3 license](https://github.com/0ssamaak0/DLTA-AI/blob/master/LICENSE). \n"
},
{
"path": "additional_scripts/coco_eval.py",
"content": "import sys\nimport os\nimport json\nimport argparse\nfrom pycocotools.coco import COCO\nfrom pycocotools.cocoeval import COCOeval\n\n# Define a class to suppress print statements\n\n\nclass HiddenPrints:\n \"\"\"\n A context manager to suppress print statements.\n \"\"\"\n\n def __enter__(self):\n self._original_stdout = sys.stdout\n sys.stdout = open(os.devnull, 'w')\n\n def __exit__(self, exc_type, exc_val, exc_tb):\n sys.stdout.close()\n sys.stdout = self._original_stdout\n\n\n# Define the function to evaluate coco\ndef evaluate_coco(gt_file: str, pred_file: str, task: str = \"bbox\", evaluation_type: str = \"full\") -> None:\n \"\"\"\n Evaluates the performance of a COCO object detection model.\n\n Args:\n gt_file (str): Path to the ground truth file.\n pred_file (str): Path to the prediction file.\n task (str, optional): The type of task to evaluate (bbox or segm). Defaults to \"bbox\".\n evaluation_type (str, optional): The type of evaluation to perform (full or mAP). Defaults to \"full\".\n \"\"\"\n # Use HiddenPrints to suppress print statements\n with HiddenPrints():\n # Load the ground truth file\n coco_gt = COCO(gt_file)\n\n # Load the prediction file\n with open(pred_file, 'r') as f:\n pred_file = json.load(f)\n pred_file = pred_file[0]['annotations'] # type: ignore\n\n coco_dt = coco_gt.loadRes(pred_file)\n\n # Create a COCO evaluator object\n coco_eval = COCOeval(coco_gt, coco_dt, task) # type: ignore\n\n # Evaluate the model\n coco_eval.evaluate()\n coco_eval.accumulate()\n\n # Compute stats\n coco_eval.summarize()\n\n # Print the results based on the evaluation type\n if evaluation_type == \"full\":\n coco_eval.summarize()\n elif evaluation_type == \"mAP\":\n print(f\"{task} mAP: {coco_eval.stats[0]:.3f}\")\n\n\n# Create an argument parser\nparser = argparse.ArgumentParser()\nparser.add_argument(\"--gt_file\", required=True, help=\"ground truth file\")\nparser.add_argument(\"--pred_file\", required=True, help=\"prediction file\")\nparser.add_argument(\"--task\", default=\"bbox\",\n choices=[\"bbox\", \"segm\"], help=\"task (bbox or segm)\")\nparser.add_argument(\"--evaluation_type\", default=\"full\",\n choices=[\"full\", \"mAP\"], help=\"evaluation type (full or mAP)\")\n\n# Parse the arguments\nargs = parser.parse_args()\n\n# Run the function with the arguments\nevaluate_coco(args.gt_file, args.pred_file, args.task, args.evaluation_type)\n"
},
{
"path": "docs/Installation/executable.md",
"content": "---\nlabel: Executable (CPU Only) \norder: 2\nicon: \":package:\"\n---\nDLTA-AI is available as an executable, however it's CPU only, so it's not recommended for large datasets. It's currently available for windows and linux only\n\nyou can download the [lastest release](https://github.com/0ssamaak0/DLTA-AI/releases) Executable under **Assets** \n\nThe Executable doesn't require any installation, just download and run it from the executable file \n\n"
},
{
"path": "docs/Installation/full installation.md",
"content": "---\nlabel: full installation\nicon: \":inbox_tray:\"\norder: 3\n---\n\n# Full Installation\n## Create a Virtual Environment\nIt is highly recommended to install DTLA-AI in virtual environment using conda. This will ensure a clean and isolated environment for the installation process. use `python=3.8` to avoid any compatibility issues\n```\nconda create -n DLTA-AI python=3.8\nconda activate DLTA-AI\n```\n\n\n## Install Pytorch\n\nFirst, you need to install [pytorch](https://pytorch.org/get-started/locally/) according to your device and your OS, if you have GPU, choose CUDA version, otherwise choose CPU version\n\nExample:\n```\nconda install pytorch torchvision torchaudio .... -c pytorch>\n```\n\n!!!\nDue to inconsistency between the current versions of `pytorch` and `mmcv`, some users may face issues when using `pytorch 2.x`\nif you face any issues, try to install `pytorch 1.13.1` instead\n!!!\n\n## Option 1: Using pip\nInstallation using pip is more easier since it handles all dependencies\n```\npip install DLTA-AI\n```\nthen run it from anywhere using\n```\nDLTA-AI\n```\nnote that first time running DLTA-AI, it will download a required module, it may take some time\n\nyou can also use pip for updating DLTA-AI\n```\npip install DLTA-AI -U\n```\n\n\n\n## Option 2: Manual Installation\nDownload the lastest release from [here](https://github.com/0ssamaak0/DLTA-AI/releases)\n\ninstall requirements\n\n```\npip install -r requirements.txt\nmim install mmcv-full==1.7.0\n```\nthen \nRun the tool from `DLTA_AI_app` directory\n```\ncd DLTA_AI_app\npython __main__.py\n```\n\n\n\n\n\n\n"
},
{
"path": "docs/Installation/index.yml",
"content": "label: Installation\nicon: \":hammer_and_wrench:\"\nexpanded: true\norder: 50"
},
{
"path": "docs/Installation/problems.md",
"content": "---\nlabel: possible problems\norder: 1\nicon: \":interrobang:\"\n---\n\n# Solutions to possible problems\n\n## Qt Platform Plugin Error in OpenCV on Linux Machines 🐧\nsome linux machines may have this problem \n```\nCould not load the Qt platform plugin \"xcb\" in \"/home//miniconda3/envs/test/lib/python3.8/site-packages/cv2/qt/plugins\" even though it was found.\nThis application failed to start because no Qt platform plugin could be initialized. Reinstalling the application may fix this problem.\n\nAvailable platform plugins are: xcb, eglfs, linuxfb, minimal, minimalegl, offscreen, vnc, wayland-egl, wayland, wayland-xcomposite-egl, wayland-xcomposite-glx, webgl.\n```\nit can be solved simply be installing opencv-headless\n```\npip3 install opencv-python-headless\n```\n## Microsoft Visual C++ Build Tools Error When Installing MMDetection on Windows Machines 🪟\nsome windows machines may have this problem when installing **mmdet**\n```\nBuilding wheel for pycocotools (setup.py) ... error\n...\nerror: Microsoft Visual C++ 14.0 or greater is required. Get it with \"Microsoft C++ Build Tools\": https://visualstudio.microsoft.com/visual-cpp-build-tools/\n```\nYou can try\n```\nconda install -c conda-forge pycocotools\n```\nor just use Visual Studio installer to Install `MSVC v143 - VS 2022 C++ x64/x86 build tools (Latest)**`\n## Problem in installing mmcv-full\nyou may often stuck in installing `mmcv-full` with this message\n```\nBuilding wheels for collected packages: mmcv-full\n Building wheel for mmcv-full (setup.py) ...\n```\nyou can try installing [pytorch 1.13.1](https://pytorch.org/get-started/previous-versions/#v1131), instead of the lastest version, you can also refer to [this issue](https://github.com/open-mmlab/mmcv/issues/1386)\n\n## Multiple copies of the OpenMP runtime have been linked into the program\nyou may encounter this problem\n```\nOMP: Error #15: Initializing libiomp5md.dll, but found libiomp5md.dll already initialized.\n\nOMP: Hint This means that multiple copies of the OpenMP runtime have been linked into the program.\nThat is dangerous, since it can degrade performance or cause incorrect results. The best thing to do is to\nensure that only a single OpenMP runtime is linked into the process, e.g. by avoiding static linking of the\nOpenMP runtime in any library. As an unsafe, unsupported, undocumented workaround you can set the environment\nvariable KMP_DUPLICATE_LIB_OK=TRUE to allow the program to continue to execute, but that may cause crashes or silently\nproduce incorrect results. For more information, please see http://www.intel.com/software/products/support/.\n\n```\nyou can solve this by upgrading numpy\n```\npip install numpy==1.23.3\n```\n\nThanks for [mohamedraafat96's issue](https://github.com/0ssamaak0/DLTA-AI/issues/52), you can check this [stackoverflow answer](https://stackoverflow.com/questions/64209238/error-15-initializing-libiomp5md-dll-but-found-libiomp5md-dll-already-initial) for more details"
},
{
"path": "docs/index.md",
"content": "---\nlabel: \"DLTA-AI User Guide\"\nicon: \"assets/icon.png\"\n---\n\n# DLTA-AI User Guide\n\n\n\n\nData Labeling, Tracking and Annotation with AI.\n
\n\nDLTA-AI is the next generation of annotation tools, integrating the power of Computer Vision SOTA models to Labelme in a seamless expirence and intuitive workflow to make creating image datasets easier than ever before\n\n\n## Why DLTA-AI?\nOpen source and customizable annotation tool was created to fill a gap in annotation tools.\nThe customization and giving the user the full control was and will be our priority, from the model selection, input formats and inference parameters, to the export formats and even the User Interface itself. From these options, the goal was to extend the use cases of the concept of annotation tool to other use cases for end users beyond just preparing datasets to train models.\n\n## Features\n-\tEasy and straightforward Installation process, support for all Operating Systems\n-\tUser Guide with detailed tutorials for all the features\n-\tFull Support of Auto Annotation with different models.\n-\tDifferent annotation options and parameters (e.g., Thresholds)\n-\tExport to (literally) any format\n-\tModern and functional User Interface\n-\tDedicated Video Mode\n- Object Tracking Support \n- Completely free and open-source, and will always be.\n\n\n\n## Contributing\nDLTA-AI is an open source project and contributions are very welcome\n\nYou can contribute in many ways:\n\n- Create an [issue](https://github.com/0ssamaak0/DLTA-AI/issues) Reporting bugs 🐞 or suggesting new features 🌟 or just give your feedback 📝\n\n- Create a [pull request](https://github.com/0ssamaak0/DLTA-AI/pulls) to fix bugs or add new features, or just to improve the code quality, optimize performance, documentation, or even just to fix typos\n\n- Review [pull requests](https://github.com/0ssamaak0/DLTA-AI/pulls) and help with the code review process\n\n- Spread the word about DLTA-AI and help us grow the community 🌎, by sharing the project on social media, or just by telling your friends about it\n\n## Resources\n- [Labelme](https://github.com/wkentaro/labelme)\n- [Segment Anything (SAM)](https://github.com/facebookresearch/segment-anything)\n- [MMDetection](https://github.com/open-mmlab/mmdetection/tree/2.x)\n- [ultralytics YOLOv8](https://github.com/ultralytics/ultralytics)\n- [mikelbrostrom yolov8_tracking](https://github.com/mikel-brostrom/yolov8_tracking)\n- [orjson](https://github.com/ijl/orjson)\n- [icons8](https://icons8.com/)\n"
},
{
"path": "docs/main_features/Export.md",
"content": "---\nicon: \":outbox_tray:\"\norder: 3\n---\n\n# Export\nThis page is under construction 🚧, please check back later."
},
{
"path": "docs/main_features/SAM.md",
"content": "---\nicon: https://github.com/0ssamaak0/DLTA-AI/blob/master/DLTA_AI_app/labelme/icons/SAM.png?raw=true\norder: 10\n---\n\n# Segment Anything (SAM)\nMETA AI model [Segment Anything](https://segment-anything.com/) or `SAM` is integrated in DLTA-AI in many ways to increase the accuracy of the Annotation process, in a very native user expirience with almost zero effort to install.\n\n## Installation\nLike all other models, The [Model Explorer](\"../model_selection/model_explorer.md\") can be used to install the checkpoints directly with just a single click\n\n[!embed](https://youtu.be/8g15M9bE1uA?t=5)\n\n## Segmentation\nSegment Anything can be used to make Zero-Shot Segmentation of any object.\nDLTA-AI provides an expiernce similar to the [Original Demo](https://segment-anything.com/demo#) with the simple SAM toolbar that supports user-customized shortcuts, and runs locally on the user machine on any image or video.\n\n[!embed](https://youtu.be/8g15M9bE1uA?t=41)\n\n## Enhance Polygons\nBeside the usual functionality of Zero-Shot Segmentation, Segment Anything can be used to enhance the accuracy of any polygon, weather it was created by the user or by any other model, by simply selecting the polygon(s) and enhancing them from the toolbar or the context menu.\n\n\n[!embed](https://youtu.be/8g15M9bE1uA?t=84)\n\n## Interpolation Tracking\nDLTA-AI utilizes the power of Segment Anything to provide a very accurate interpolation tracking, that can be used to track any object in a video, and can be used to track multiple objects at the same time.\n\n[!embed](https://youtu.be/8g15M9bE1uA?t=116)"
},
{
"path": "docs/main_features/index.yml",
"content": "label: Main Features\nicon: \":star2:\"\nexpanded: true\norder: 40"
},
{
"path": "docs/main_features/inputs.md",
"content": "---\nicon: \":clapper:\"\norder: 8\n---\n\n# Input Modes\nDLTA-AI provides different options for inputs\n\n## Image Mode\nimage mode is very simple, just open an imgae and start annotating\n\n[!embed](https://youtu.be/zkm_GhX8OtM?t=6)\n## Directory Mode\nDirectory mode is used to annotate a directory of images, it's very useful when you have a dataset of images and you want to annotate them all at once.\n\nNote that it shows all images within the directory and all subdirectories.\n\n[!embed](https://youtu.be/zkm_GhX8OtM?t=17)\n## Video Mode\nVideo mode is used to annotate a video, and provides an integrated video player that allows you to naivgate, play, pause, forward, backward, and jump to a specific frame.\n\n[!embed](https://youtu.be/zkm_GhX8OtM?t=41)\n\n## Video as Frames\nyou can open a video as a directory of frames, this is useful when you want to just annotate some frames of a video. you have the option to set start and end frame, and also the sampling rate i.e., the step between frames.\n\n[!embed](https://youtu.be/zkm_GhX8OtM?t=71)\n"
},
{
"path": "docs/main_features/segmentation.md",
"content": "---\nicon: \":art:\"\norder: 9\n---\n\n# Segmentation\nInstance Segmentation is one of the major features in DLTA-AI, from the huge library to the different options and pramaters, to the ability to apply manual edits to the results, DLTA-AI proivdes a fully customizable and easy to use segmentation experience.\n## Model Selection\nThe model Selection can be done directly by selecting a segmentation model from the menu, or by selecting a model from the huge library of models in the [Model Explorer](../model_selection/model_explorer.md)\n\n[!embed](https://youtu.be/bYjy82Ug2wU?t=10)\n## Inferencing\nThe model can be run on the current image only (works in all [input modes](inputs.md)) or on all images (directory mode only) \n\n[!embed](https://youtu.be/bYjy82Ug2wU?t=28)\n## Visualization Options\nyou can select the visualization options from the menu, such as showing the segmentation mask, or just bounding box, and the class name and the cofidence scor as well\n\n[!embed](https://youtu.be/bYjy82Ug2wU?t=60)\n## Select Classes\nyou can select some classes among the 80 of [COCO classes](https://cocodataset.org/) you can select for just this use (forgotten when you close DLTA-AI) or set them as default classes (saved when you close DLTA-AI)\n\n\n[!embed](https://youtu.be/bYjy82Ug2wU?t=84)\n## Thresholds\nTo give the annotator the full control and the ability to choose the optimum point in the precision/recall tradeoff, DLTA-AI provides 2 thresholding options\n\n### Confidence Threshold\nConfidence threshold is very simple, by just typing the threshold value of setting it through the slider, all predictions with confidence less than the threshold will be ignored\n\n[!embed](https://youtu.be/bYjy82Ug2wU?t=112)\n\n### IOU Threshold (For Non Maximum Suppression)\nDLTA-AI internally applies Non Maximum Suppression (NMS) to the predictions, to remove the overlapping predictions, the IOU threshold is the threshold used in NMS.\n\n[!embed](https://youtu.be/bYjy82Ug2wU?t=150)"
},
{
"path": "docs/main_features/tracking/index.yml",
"content": "label: Tracking\nicon: https://github.com/0ssamaak0/DLTA-AI/blob/master/DLTA_AI_app/labelme/icons/tracking.png?raw=true\nexpanded: true\norder: 4"
},
{
"path": "docs/main_features/tracking/interpolation.md",
"content": "---\nicon: \norder: 1\n---\n\n# Interpolation Tracking\n\n## Interpolation Method\n\n### Linear Interpolation\n\n### SAM Interpolation\n\n## Interpolation Between\n\n### Selected Keyframes\n\n### Detected Frames\n\n"
},
{
"path": "docs/main_features/tracking/tracking.md",
"content": "---\nicon: \norder: 2\n---\n\n# Tracking\n\n## Model Selection\n\n## Tracking Options\n\n## Visualization Options\n\n## Edit propagation\n\n## Delete Options"
},
{
"path": "docs/model_selection/index.yml",
"content": "label: Model Selection\nicon: \":robot_face:\"\nexpanded: true\norder: 30"
},
{
"path": "docs/model_selection/merge.md",
"content": "---\nicon: https://github.com/0ssamaak0/DLTA-AI/blob/master/DLTA_AI_app/labelme/icons/merge.png?raw=true\norder: 5\n\n---\n\n# Merge Models \nThis page is under construction 🚧, please check back later."
},
{
"path": "docs/model_selection/model_explorer.md",
"content": "---\nicon: https://github.com/0ssamaak0/DLTA-AI/blob/master/DLTA_AI_app/labelme/icons/checklist.png?raw=true\norder: 10\n\n---\n\n# Model Explorer\nThis page is under construction 🚧, please check back later."
},
{
"path": "docs/retype.yml",
"content": "input: .\noutput: .retype\n\nurl: # Add your website address here\n\nbranding:\n title: DLTA-AI User Guide\n logo: assets/icon.png\n colors:\n label:\n text: \"#ffffff\"\n background: \"#ff0000\"\n\nfavicon: assets/icon.png\n\nlinks:\n- text: GitHub\n link: https://github.com/0ssamaak0/DLTA-AI\n icon: mark-github\n\n- text: Issues\n link: https://github.com/0ssamaak0/DLTA-AI/issues\n icon: bug\n\n- text: Release Notes\n link: https://github.com/0ssamaak0/DLTA-AI/releases\n icon: note\n\n- text: Youtube Channel\n link: https://www.youtube.com/@DLTA-AI\n icon: video\n\n\nfooter:\n copyright: \"DLTA-AI is licensed under GPLv3 License\" \n links: \n - text: License \n link: https://github.com/0ssamaak0/DLTA-AI/blob/master/LICENSE\n icon: law\n\n"
},
{
"path": "docs/user_interface.md",
"content": "---\nlabel: User Interface\nicon: https://github.com/0ssamaak0/DLTA-AI/blob/master/DLTA_AI_app/labelme/icons/UI.png?raw=true\norder: 1\n---\nThis page is under construction 🚧, please check back later."
},
{
"path": "releasenotes.md",
"content": "# New Features 🌟\n-\n\n# Bug Fixes 🐞\n- Closing the save dialog without saving annotations no more enables the export button #45"
},
{
"path": "requirements.txt",
"content": "PyQt6==6.6.0\r\nQtPy==2.3.1\r\ntermcolor==2.2.0\r\nimgviz==1.7.2\r\nopencv-python==4.7.0.72\r\npyqtdarktheme==2.1.0\r\nsupervision==0.3.2\r\ngdown==4.7.1\r\nultralytics==8.0.61\r\nonemetric==0.1.2\r\neasydict==1.10\r\nopenmim==0.3.2\r\nmmdet==2.25.2\r\nscikit-image==0.20.0\r\nfilterpy==1.4.5\r\nsegment-anything==1.0\r\nlap==0.4.0\r\norjson==3.8.12\r\nnotify-py==0.3.42\r\npsutil==5.9.4\r\nshapely==2.0.1\r\nscreeninfo==0.8.1\r\n"
},
{
"path": "setup.py",
"content": "import setuptools\n\nwith open(\"README.md\", \"r\", encoding=\"utf8\") as fh:\n long_description = fh.read()\n\nwith open(\"requirements.txt\", \"r\") as f:\n requirements = f.read().splitlines()\n \n__version__ = \"0.0.0\"\nwith open(\"DLTA_AI_app/labelme/__init__.py\", \"r\") as f:\n for line in f.readlines():\n if line.startswith(\"__version__\"):\n __version__ = line.split(\"=\")[1].strip().strip('\"')\n break\n\nsetuptools.setup(\n name=\"DLTA-AI\",\n version=f\"{__version__}\",\n author=\"0ssamaak0\",\n author_email=\"0ssamaak0@gmail.com\",\n description=\"DLTA-AI is the next generation of annotation tools, integrating the power of Computer Vision SOTA models to Labelme in a seamless expirence and intuitive workflow to make creating image datasets easier than ever before\",\n long_description=long_description,\n long_description_content_type=\"text/markdown\",\n url=\"https://github.com/0ssamaak0/DLTA-AI\",\n package_dir={\"DLTA_AI_app\": \"DLTA_AI_app\"},\n python_requires='>=3.8',\n install_requires=requirements,\n package_data={\"\": [\"*\"]},\n license=\"GPLv3\",\n entry_points={\n \"console_scripts\": [\n \"DLTA-AI=DLTA_AI_app.__main__:main\"\n ]\n }\n)\n\n"
},
{
"path": "yolo training commands.txt",
"content": "yolo task=segment mode=train model=yolov8n-seg.pt epochs= 1 imgsz = 320 workers =2 batch = 4 data=datasets\\thermalseg.v6i.yolov8\\data.yaml"
}
]