Tagging Utility for preparing dataset to train Stable Diffusion Models

This tagging utility is designed for stable diffusion models. It provides a user-friendly interface using the Gradio library and offers the following features:

1. Interrogate Image:

   • Generates tags for a single image using the WD Tagger.

2. Generate Tags for All Images in a Directory using WD Tagger:

   • Generates tags for all images available in a specified directory using the WD Tagger.

3. Generate Tags for All Images in a Directory using Deepbooru Tagger:

   • Generates tags for all images available in a specified directory using the Deepbooru Tagger.

4. Remove Duplicate Tags:

   • Removes duplicate tags from the generated tags.

5. Search and Replace Specific Tags from Files:

   • Allows searching for specific tags and replacing them in files.

6. Append/Prepend Tags to Existing Tags:

   • Appends or prepends tags to existing tags in files.

How to Use:

1. Installation:

   • Clone the repository:

     git clone <git@github.com:grim-reapper/sd_tagging_utility.git>
     cd sd_tagging_utility
     

   • Install the required dependencies:

     pip install -r requirements.txt
     

2. Running the Utility:

   • Run the tagging utility script:

     python tagger.py
     

3. Using the Interface:

   • Once the utility is running, you will be presented with a user-friendly interface powered by Gradio.

   • Select the desired feature from the options available.

   • Follow the instructions provided in the interface to input the necessary information.

   • After processing, the utility will display the results and save them to the specified directory or file.

Requirements:

• Python 3.10
• Gradio
• Other dependencies listed in requirements.txt

Contributors:

• [Imran Ali]