Kaloscope Artist Style Classifier

Kaloscope Artist Style Classifier

Classifies an image into Danbooru artist tags with the Kaloscope models, in a Gradio page. It suits people who tag or sort artwork, and developers who want to see how these models are run.

Made by DraconicDragon on Hugging Face. These notes are Mindket's, written from its Space page and README; the app's own files and license are on Hugging Face.

About

What it is: This Space is a Gradio app that classifies images by Danbooru artist tag using the Kaloscope family of models. The repository has scripts to run the models through ONNX, PyTorch or DINOv3, plus lsnet and v3 folders. The Space runs on Hugging Face's free CPU hardware.

Who it suits: It suits people who sort or tag artwork and want a guess at the artist style, and developers who want to see how these models are run.

Industry | Who uses it | How they use it | Example
Art tagging | Archivists and tag editors | Get artist tag suggestions for an image | An editor uploads a picture and reads the suggested artist tags.
AI art | Dataset builders | Label training images by artist style | A builder checks a few images to judge the labels.
Software | Developers | Read how one app switches between ONNX, PyTorch and DINOv3 inference | A developer opens the three inference scripts side by side.

If you need a verdict on who made a piece, skip it, because the app only classifies style.

Where to use it:

• A web browser, through the Space's page on Hugging Face
• A local copy on Windows, macOS or Linux, using Python

Use case | Who | What they get back
Tag one image | Anyone | Danbooru artist tag guesses for the picture
Try other model versions | Hobbyists | Results from the model versions the Space names
Run it locally | Developers | The same page on your own computer

How to set it up:

• Python and pip on Windows, macOS or Linux
• The packages listed in requirements.txt

1. Install the huggingface_hub package.
2. Download the Space into a local folder named Kaloscope-artist-style-classifier.
3. Move into that folder.
4. Install gradio at version 6.29.1.
5. Install the packages listed in requirements.txt.
6. Start the app.

pip install -U huggingface_hub
hf download DraconicDragon/Kaloscope-artist-style-classifier --repo-type space --local-dir Kaloscope-artist-style-classifier
cd Kaloscope-artist-style-classifier
pip install gradio==6.29.1
pip install -r requirements.txt
python app.py

How to use it: Open the Space's page, or start the app locally with the command below. Give the app an image and read the artist tags it returns.

python app.py

Tool | What it does
app.py | The main file, which starts the Gradio page
inference_onnx.py | Runs the model through ONNX, going by its name
inference_pytorch.py | Runs the model through PyTorch, going by its name
inference_dinov3.py | Runs a DINOv3 version, going by its name

Example: You upload a drawing and run the classifier. The page returns a list of Danbooru artist tags that the model thinks match the style.

For newcomers and experts: If you are new, open the Space's page and try one picture. A Danbooru artist tag is the name an image board uses to label a creator.

If you know the field, the Space names several Kaloscope versions, including 2.0 and a 3.0 preview, an ONNX export, and two lsnet_b models. Which one the page uses by default is not stated.

Upload an image
A Kaloscope model reads the style
Artist tags come back as text

Good to know:

• The source gives only a one-line description, so scores and tag lists are not documented
• The page does not say which model version it uses by default
• License: apache-2
• Cost to run: none named
• Data it reaches: the images you give it

On Hugging Face, 13 November 2026

• Likes: 61
• Built with: gradio
• License: apache-2.0

Links

• Space: https://huggingface.co/spaces/DraconicDragon/Kaloscope-artist-style-classifier
• Files: https://huggingface.co/spaces/DraconicDragon/Kaloscope-artist-style-classifier/tree/main

Try it

Open the Space in your browser, or copy it to your own Hugging Face account from its page. Nothing in this kit runs by itself.