README.md

---
title: README
emoji: 🏆
colorFrom: yellow
colorTo: pink
sdk: static
pinned: false
---
<div style="text-align: center;">
  <img style="margin: 0 auto 0 auto;" src="https://huggingface.co/spaces/RedRocket/README/resolve/main/RedRocket.png">
  <span style="font-family:Times New Roman; font-size: 1.7em; font-weight: bold;">🐾 A VNGVLIS ET ALIS NOSTRIS PRO OMNIBVS CREAMVS 🐉</span>
</div>

**FluffyRock Unbound v1.1** - https://huggingface.co/RedRocket/Fluffyrock-Unbound<br>
Overhauled upgrade of the well-regarded FluffyRock Unleashed SD 1.5 model line, featuring a curated dataset and new training techniques from EDM2.<br>

**Joint Tagger Project (JTP PILOT)** - https://huggingface.co/RedRocket/JointTaggerProject &bull; https://huggingface.co/spaces/RedRocket/JointTaggerProject-Inference-Beta<br>
The undispited best-in-class content classifier for furry artwork.<br>

**FARAMIR** - https://github.com/redhottensors/FARAMIR<br>
<u>F</u>antasy <u>A</u>esthetic <u>R</u>ealism <u>A</u>ssessment <u>M</u>odel with <u>I</u>ntegrated <u>R</u>elabeling.<br>

**FARAD** - https://huggingface.co/datasets/RedRocket/FARAD<br>
Furry Aesthetic Realism Annotated Dataset. 18k Artstation realism images classified with FARAMIR and JTP-PILOT2.

**Furception VAE** - https://huggingface.co/RedRocket/furception_vae<br>
Generally regarded as the highest quality SD 1.5 VAE for artwork. Bunded with FluffyRock Unbound/Unleashed and some versions of Fluffusion.<br>

**ComfyUI-Prediction** - https://github.com/redhottensors/ComfyUI-Prediction<br>
Fully customizable Classifier Free Guidance for ComfyUI.<br>

**ComfyUI-ODE** - https://github.com/redhottensors/ComfyUI-ODE<br>
Adaptive ODE Solvers for ComfyUI.<br>

<hr>

<div style="text-align: center; font-weight: bold;">Legacy Releases</div>

**FluffyRock Unleashed v1.0** - https://huggingface.co/RedRocket/Fluffyrock-Unleashed<br>
Continuation of the excellent [FluffyRock](https://huggingface.co/lodestones/furryrock-model-safetensors) foundation model by [Lodestone Rock](https://huggingface.co/lodestones) with v-pred, zero-terminal-SNR, fixed timestep weighting, and a refiner model.