Viggle-Animate-pruned-GGUF

DISCLAIMER - TESTING

GGUF quantized versions of the pruned Viggle-Animate model, derived from the original Viggle-Animate release.

Quantizations

File	Quantization	Size
Viggle-Animate-pruned-Q3_K_M.gguf	Q3_K_M	8.91 GB
Viggle-Animate-pruned-Q4_K_M.gguf	Q4_K_M	11.6 GB
Viggle-Animate-pruned-Q5_K_M.gguf	Q5_K_M	14.1 GB
Viggle-Animate-pruned-Q6_K.gguf	Q6_K	16.7 GB

About

This repository provides GGUF quantized versions of the pruned Viggle-Animate model for use with GGUF-compatible inference environments.

The original Viggle-Animate model is available here:

Viggle/Viggle-Animate

Viggle-Animate is designed for video-to-video character replacement, allowing the appearance of a character to be transferred into a driving video while preserving the motion and structure of the shot.

The original model is built on MiniMaxAI/MiniMax-H3.

For the complete methodology, examples, inference details, and limitations, please refer to the original model card.

Quantization Options

Q3_K_M

The smallest version in this repository, intended for systems where memory usage is the primary constraint.

Q4_K_M

A balance between model size and numerical precision.

Q5_K_M

A higher-precision option with a larger memory footprint.

Q6_K

The highest-precision quantization available in this repository.

ComfyUI

These GGUF files are intended for use with GGUF-compatible ComfyUI workflows.

Official Viggle-Animate H3 ComfyUI nodes:

ComfyUI-Viggle-Animate-H3

Original Project

Original model: https://huggingface.co/Viggle/Viggle-Animate

Base model: https://huggingface.co/MiniMaxAI/MiniMax-H3

Disclaimer

These files are derived from the original Viggle-Animate model.

The underlying model is derived from MiniMax-H3. Users should review and comply with the applicable MiniMax H3 Community License and the terms of the original Viggle-Animate release before using, redistributing, or incorporating these weights into products or services.

Citation

@misc{viggle2026animate,
  title  = {Viggle-Animate: Character Replacement in Video from a Single Repainted Frame},
  author = {Viggle Research},
  year   = {2026},
  url    = {https://huggingface.co/Viggle/Viggle-Animate}
}