WD1.4 Tagger & Editor

App for tagging images and editing those tags

Note: uses about 8-9 GB of storage after all packages and models are downloaded

Usage

1. open images
2. tag images
3. add/edit/remove tags
4. save tags to file

Tags with grayed out text are not included when saving tags to file.

It is also possible to edit the score for each tag by appending :<score> to the tag.
e.g. tomato:0.8 will set the score for tomato to 0.8

Shortcuts

Key	Action
ALT + A	add tag to all images
ALT + SHIFT + A	add tag to selected images
ALT + Q	tag all images
ALT + SHIFT + Q	tag selected images
ALT + S	save tags to txt file
CTRL + F	find tag
ALT + I	include tag

Settings

Settings	Description
Tagger Model	wd1.4 tagger model used for tagging the images (swinv2, convnextv2, convnext)
Use Tensorflow	use tensorflow instead of onnxruntime*
Batch Size	batch size for tensorflow (onnx model only uses batch 1)
Threshold	tags below this threshold are grayed out and not saved when saving tags
Threshold Low	threshold for the tagger model, tags below this threshold won't be displayed at all
Save Tag Scores	save tag scores when saving tags (for training with weighted captions/tags)

*would only recommend tensorflow if you are tagging a large number of images and have a gpu with more than 12GB of vram.

Run Without Image Tagging Functionality

It is possible to use the app as only a tag editor by placing an empty venv folder within the python folder. This will prevent any tagger models or python packages from being downloaded.

• python
  • venv

Manual Setup

The app searches for python/taggers and python/venv folder. It will download and install the necessary files and packages if those folders don't exist.

python venv

You can use the following commands if you wish to setup the python venv manually.

cd python
python -m venv venv
.\venv\Scripts\activate
pip install torch torchvision --index-url https://download.pytorch.org/whl/cu117
pip install -r requirements.txt

tagger models

These three tagger models are used
https://huggingface.co/SmilingWolf/wd-v1-4-swinv2-tagger-v2
https://huggingface.co/SmilingWolf/wd-v1-4-convnextv2-tagger-v2
https://huggingface.co/SmilingWolf/wd-v1-4-convnext-tagger-v2

Folder structure for the tagger models

• python
  • taggers
    • selected_tags.csv
    • <tagger model name>
      • keras_metadata.pb
      • saved_model.pb
      • model.onnx
      • variables
        • variables.data-00000-of-00001
        • variables.index

Developing

Install dependencies

npm install

Start app in dev mode

npm run start

Package app

npm run package