Easy Tagger

A simple GUI application for image tagging using the WD-14 model. This tool makes it easy to automatically generate tags for your images with additional customization options.

Features

• Support for multiple AI models (WD-14, Florence 2, JoyCaption)
• Batch processing of multiple images or folders
• Custom output location
• Additional tags support
• Tag blocking/filtering
• Simple and intuitive interface

Installation

1. Clone this repository:

git clone https://github.com/YOUR_USERNAME/easy-tagger.git
cd easy-tagger

2. Create a virtual environment and activate it:

python -m venv .venv
# On Windows:
.venv\Scripts\activate
# On Linux/Mac:
source .venv/bin/activate

3. Install the requirements:

pip install -r requirements.txt

Usage

1. Run the application:

python easy_tagger.py

2. Using the interface:

   • Select Model: Choose the AI model you want to use (WD-14 recommended)

   • Input Selection:

     • Choose "Select Files" to tag individual images
     • Choose "Select Folder" to tag all images in a folder

   • Output Folder: Choose where to save the tagged images and tag files

   • Tag Settings:

     • Additional Tags: Enter any tags you want to add to ALL images

       • Example: anime, digital_art, high_quality
       • Separate tags with commas

     • Banned Tags: Enter tags you want to exclude from results

       • Example: sensitive, questionable, explicit
       • Separate tags with commas

3. Click "Start Tagging" to begin the process

4. Results:

   • Tagged images will be copied to the output folder
   • Each image will have an accompanying _tags.txt file
   • The original images remain unchanged

Tips

• The WD-14 model is optimized for anime/illustration content
• For best results, use images with clear subjects
• You can use Additional Tags to add metadata like artist name or source
• Use Banned Tags to filter out unwanted or incorrect tags

Requirements

• Python 3.8 or higher
• PyQt6
• torch
• Pillow
• numpy
• onnxruntime
• transformers
• huggingface_hub

License

This project is licensed under the MIT License - see the LICENSE file for details.