laibel

KO | EN

gitlite — search

#javascript #python #hacktoberfest #react #ai #typescript #llm #go #golang #android #machine-learning #rust #deep-learning #linux

(★ 12)

Easy button for AI image labeling ❤️ You like makesense? You love laibel!

laibel 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

<div align="center">
  <h1><strong>lAIbel</strong></h1>
</div>

A lightweight image annotation tool designed for efficient labeling of images to create downstream AI models. lAIbel provides a cross-platform browser-based interface for drawing and managing bounding box annotations, streamlining the data preparation phase for computer vision tasks. Labels can be exported in a variety of formats.

# Usage
[📺 Quickstart Video](https://private-user-images.githubusercontent.com/70864791/457679789-78e0f6f0-536d-475f-89a5-d392f83cede0.mp4?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.U6wl8KkDMERMghGVLKTX-Yqv-v8vkBCrVa6q8Avt9aY)

[🎥 Watch YOLO Tutorial Video](https://github.com/bortpro/laibel/releases/download/v0.1/YOLO.mp4)

[🎥 Watch YOLOE Tutorial Video](https://github.com/bortpro/laibel/releases/download/v0.1/YOLOE.mp4)

**Core Features:**

*   **Bounding Box Annotation:** Draw and resize bounding boxes directly on the canvas.
*   **Multi-Class Labels:** Press **1–9** (or click a label) to set the active class; new boxes use the **last used** label so you can batch-label one class, then switch.
*   **Annotation Export:**
    *   Export annotations for all images in structured **JSON** format.
    *   Export annotations in the standard **YOLO** `.txt` format, compatible with YOLO & Ultralytics.
*   **Keyboard Shortcuts:** Quick navigation and label selection for faster labeling.
*   **AI-Assisted Labeling:** Lightweight background AI model aids in automatic labeling. (YOLOE!)

### Tech Stack

lAIbel is built with a focus on simplicity and client-side performance.
*   **Backend:** A minimal Flask(Python) web server is used primarily to serve the HTML, CSS, and JavaScript files. It's designed to be lightweight and easy to deploy.
*   **Frontend:** The core annotation logic, UI interactions, state management, and rendering are handled entirely using vanilla JavaScript, HTML5 Canvas, and CSS. This choice avoids heavy framework dependencies, aiming for fast loading times and a responsive user experience directly in the browser.

### Project Goal

The primary goal of laibel is to accelerate the often tedious process of image annotation. The goal is to develop a tool that is exceptionally fast and intuitive to use, reducing the time and effort required to prepare high-quality datasets for machine learning. By focusing on a streamlined user experience, efficient processing, and compatibility with standard formats like YOLO, lAIbel strives to be a valuable asset in any computer vision practitioner's toolkit. The long-term vision is to make lAIbel one of the fastest and most user-friendly open-source labeling tools available.

**Upcoming Features:**

*   Active Learning Labeling: Support active learning labeling.
*   Batch Annotation: Automate annotation tasks for large datasets.
*   Performance Optimization: Improve performance for large datasets and complex annotations.
*   Advanced Annotation Tools: Support for other segmentation tasks and integration of custom models, such as versions ofSegment Anything Model (SAM)
*   Integration with AI Models: Train and evaluate AI models directly within the tool.

## Installation

### With uv (recommended)

[uv](https://docs.astral.sh/uv/) manages Python and dependencies in one step. Install uv from the [official guide](https://docs.astral.sh/uv/getting-started/installation/), then:

```bash
git clone https://github.com/bortpro/laibel
cd laibel
uv sync
uv run python app.py
```

For AI-powered labeling with YOLOE assist, install the extra dependencies and download the model weights:

```bash
uv pip install git+https://github.com/THU-MIG/yoloe.git#subdirectory=third_party/CLIP
uv pip install git+https://github.com/THU-MIG/yoloe.git#subdirectory=third_party/ml-mobileclip

wget https://docs-assets.developer.apple.com/ml-research/datasets/mobileclip/mobileclip_blt.pt
wget https://github.com/ultralytics/assets/releases/download/v8.3.0/yoloe-11s-seg.pt

uv run python app.py
```

### With pip

For manual labeling and lightweight installation, you can quickstart with:

```bash
git clone https://github.com/bortpro/laibel
cd laibel
pip install -r requirements.txt
```

To launch the app, simply run the app framework:

```bash
python app.py
```

For AI powered labeling with the YOLOE assist functionality, a virtual environment is recommended.

```bash
pip install git+https://github.com/THU-MIG/yoloe.git#subdirectory=third_party/CLIP
pip install git+https://github.com/THU-MIG/yoloe.git#subdirectory=third_party/ml-mobileclip

wget https://docs-assets.developer.apple.com/ml-research/datasets/mobileclip/mobileclip_blt.pt
wget https://github.com/ultralytics/assets/releases/download/v8.3.0/yoloe-11s-seg.pt
```

To install via Docker, a recommended Dockerfile is provided:

```bash
docker build -t laibel-app .
docker run -p 5000:5000 laibel-app
```

If using Intel Gaudi, please refer to the provided Dockerfile & yoloe_label.py for integration instructions.
If your accelerator can support PyTorch operations, your accelerator can run Laibel.

## 💬 Citation

You can cite Laibel in your publications if this is useful for your research. Here is an example BibTeX entry:

```BibTeX
@MISC{laibel,
   author = {Dipro Chakraborty},
   title = {{lAIbel}},
   howpublished = "\url{https://github.com/bortpro/laibel/}",
   year = {2025},
}
```

laibel

laibel 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

laibel

laibel 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README