README.md

---
license: creativeml-openrail-m
tags:
- coreml
- stable-diffusion
- text-to-image
---
# Core ML Converted Model:

  - This model was converted to [Core ML for use on Apple Silicon devices](https://github.com/apple/ml-stable-diffusion). Conversion instructions can be found [here](https://github.com/godly-devotion/MochiDiffusion/wiki/How-to-convert-ckpt-or-safetensors-files-to-Core-ML).<br>
  - Provide the model to an app such as Mochi Diffusion [Github](https://github.com/godly-devotion/MochiDiffusion) - [Discord](https://discord.gg/x2kartzxGv) to generate images.<br>
  - `split_einsum` version is compatible with all compute unit options including Neural Engine.<br>
  - `original` version is only compatible with CPU & GPU option.<br>
  - Custom resolution versions are tagged accordingly.<br>
  - `vae` tagged files have a vae embedded into the model.<br>
  - Descriptions are posted as-is from original model source. Not all features and/or results may be available in CoreML format.<br>
  - This model was converted with `vae-encoder` for i2i.
  
# Note: Some models do not have the [unet split into chunks](https://github.com/apple/ml-stable-diffusion#-converting-models-to-core-ml).

# Clarity:
Source(s): [Hugging Face](https://huggingface.co/Schisim/Clarity) - [CivitAI](https://civitai.com/models/5062/clarity)

<h1>Clarity 🤠</h1>
<img src="https://huggingface.co/Schisim/Clarity/resolve/main/Images/00615-2392482435.png" width=512/>


VAE NOT REQUIRED BUT RECOMENDED

Model requires VAE - https://huggingface.co/stabilityai/sd-vae-ft-mse-original/tree/main

File Structure for AUTOMATIC1111-webui :

|──sd

|----|──stable-diffusion-webui

|----|----|──models

|----|----|----|──VAE

|----|----|----|----|──Put your VAE file here

Merged Models

A list of merged models can be found bellow in the description of the attached model version.

Capabilities

NSFW Photography

SFW Photography is also possible, see "Trigger Words" bellow.

Photorealistic 3D renders

Emphasis on human anatomy

Limitations

Anything not listed above. This is model was created as a baseline to a general purpose model I'm working on.

Stylized images and object images are possible, but require a little finesse to generate.

Trigger Words

This checkpoint does not contain any trigger words.

However, placing some tags at the beginning of the prompts can heavily influence the generation.

These tags include:

"nsfw", "sfw", "erotica", and "nudity", "3d render", "cartoon"

Note: For SFW generation, try adding sfw to your prompt and nsfw to your negative prompt. For NSFW generation, try adding either nsfw, erotica, or nudity to your prompt and sfw to your negative prompt. In general, this is more useful for generating sfw images. This concept also applies to 3rd render and cartoon. I recommend leaving 3rd render and cartoon both in your negative prompt for generating photographic images.

Basic Prompt Guide

This model heavily revolves around UnstablePhotorealv.5. This means that you can the tagging system for PhotoReal, although I would recommend using a combination of the PhotoReal comma system and more natural language prompting.

Guide to prompting with PhotoReal - https://docs.google.com/document/d/1-DDIHVbsYfynTp_rsKLu4b2tSQgxtO5F6pNsNla12k0/edit#heading=h.3znysh7

Example prompt using commas and natural language:

Positive

A Professional Full Body Photo, of a beautiful young woman, clothed, standing indoors, Caucasian, toned physique, strawberry red hair, neutral expression

Negative

I recommend something simple like, deformed, bad anatomy, disfigured, missing limb, floating limbs, twisted, blurry, fused fingers, long neck, words, logo, text, mutated hands, mutated fingers

Modify as needed. For example, adding 3d render, cartoon to your negative prompt will help generate photographic images.

The prompts for this model are fairly flexible, experiment to find out what works best for you.