---
thumbnail: "https://media.discordapp.net/attachments/1116849695642374154/1123058242512224256/jem.png"
license: creativeml-openrail-m
tags:
- stable-diffusion
- text-to-image
- safetensors
pipeline_tag: text-to-image
---
----
# GyozanMix
A stylized anime model. This model is the result of attempts to get the textures from [MouseyMix](https://huggingface.co./closertodeath/mouseymix) on characters with more realistic body proportions. Named after 'Gyokai' and 'Zankuro', Gyozan.
GyozanMixs can be considered 'prototype' GyozaMix. GyozaMixV1 coincides with the release of [SD-Silicon](https://huggingface.co./Xynon/SD-Silicon), which I think has the best aesthetic compositional sense out of any SD1.X based model, to afix MouseyMix's other characteristics (mostly proportions) to a more normalized standard. V2 is a further refinement of that idea by using block-weight merging to specifically target the compositional sense, which allow mouseymix overall control over the textures. V3 is a (large-successful) attempt to curtail the tendency of Silicon-29 to head towards 2.5d realism through the introduction of a variety of loras and locons and retain more MouseyMix while not giving up any of Silicon's aesthetic compositional sense.
----
# Table of Contents
- [GyozanMix](#gyozanmix)
- [License](#license)
- [Terms of Use](#terms-of-use)
- [Disclaimer](#disclaimer)
- [Hashes/Downloads](#hashes)
- [GyozanMixs](#gyozanmixs)
- [Store Bought Gyoza餃子Mix](#store-bouught-gyoza餃子mix)
- [Components](#components)
- [Checkpoints](#checkpoints)
- [GyozanMixes](#gyozanmix-checkpoints)
- [GyozaMixV1](#gyozamixv1-checkpoints)
- [FlatandFlatterGyoza](#lingesgyozalines-checkpoints)
- [GyozaMixV2](#gyozamixv2-checkpoints)
- [GyozaMixV3](#gyozamixv3-checkpoints)
- [GyozaMixV3.5](#gyozamixv35-checkpoints)
- [GyozaMixV5](#gyozamixv5-checkpoints)
- [GyozaMixWinterHoliday2023](#gyozanwinterholiday2023-checkpoints)
- [LoRA/LoCons](#loralocons)
- [FlatandFlatterGyoza](#lingesgyozalines-loras)
- [GyozaMixV2](#gyozamixv2-loras)
- [GyozaMixV3](#gyozamixv3-loralocons)
- [GyozaMixV3.5](#gyozamixv35-loralocons)
- [GyozaMixV5](#gyozamixv5-loralocons)
- [GyozaMixWinterHoliday2023](#gyozanwinterholiday2023-loralocons)
- [Models](#Models)
- [GyozanMixes](#gyozanmixes)
- [Recipe](#gyozanmixes-recipes)
- [GyozaMixV1.X](#gyozamixv1)
- [Recipe](#gyozamixv1-recipes)
- [FlatandFlatterGyoza](#lingesgyozalines)
- [Recipe](#lingesgyozalines-recipes)
- [GyozaMixV2](#gyozamixv2)
- [Recipe](#gyozamixv2-recipe)
- [GyozaMixV3.X](#gyozamixv3)
- [Recipe](#gyozamixv3-recipe)
- [GyozaMixV3.5](#gyozamixv35)
- [Recipe](#gyozamixv35-recipe)
- [GyozaMixV5](#gyozamixv5)
- [Recipe](#gyozamixv5-recipe)
- [GyozaMixWinterHoliday2023](#gyozanwinterholiday2023)
- [Recipe](#gyozanwinterholiday2023-recipe)
- [Spinoff Mixes](#spinoff-mixes)
- [SanXianWonton](#sanxianwonton)
- [TangYuan](#tangyuan)
- [Bayesian Mixes](#bayesian-mixes)
- [GyozaSalute!](#gyozasalute)
----
# Licence
This model inherits the Silicon-29 license and is open access and available to all, with a CreativeML OpenRAIL-M license further specifying rights and usage. The CreativeML OpenRAIL License specifies:
1. You can't use the model to deliberately produce nor share illegal or harmful outputs or content
2. The authors claims no rights on the outputs you generate, you are free to use them and are accountable for their use which must not go against the provisions set in the license
3. You may re-distribute the weights and use the model commercially and/or as a service. If you do, please be aware you have to include the same use restrictions as the ones in the license and share a copy of the CreativeML OpenRAIL-M to all your users (please read the license entirely and carefully) Please read the full license here ：https://huggingface.co./spaces/CompVis/stable-diffusion-license
## Terms of Use
- **Clearly indicate where modifications have been made.**
If you used it for merging, please state what steps you took to do so.
## Disclaimer
READ MORE: Disclaimer
The user has complete control over whether or not to generate NSFW content, and the user's decision to enjoy either SFW or NSFW is entirely up to the user.The learning model does not contain any obscene visual content that can be viewed with a single click.The posting of the Learning Model is not intended to display obscene material in a public place.
In publishing examples of the generation of copyrighted characters, I consider the following cases to be exceptional cases in which unauthorised use is permitted.
"when the use is for private use or research purposes; when the work is used as material for merchandising (however, this does not apply when the main use of the work is to be merchandised); when the work is used in criticism, commentary or news reporting; when the work is used as a parody or derivative work to demonstrate originality."
In these cases, use against the will of the copyright holder or use for unjustified gain should still be avoided, and if a complaint is lodged by the copyright holder, it is guaranteed that the publication will be stopped as soon as possible.
I would also like to note that I am aware of the fact that many of the merged models use NAI, which is learned from Danbooru and other sites that could be interpreted as illegal, and whose model data itself is also a leak, and that this should be watched carefully. I believe that the best we can do is to expand the possibilities of GenerativeAI while protecting the works of illustrators and artists.
## GyozanMixes
GyozanMixes are mostly just simple mixes with mouseypastel with various models. They're basically just squishier versions of their mix components. They preserve a some of the mousy-mix feel but only in the sense that they are slightly softer and smaller than they would be otherwise.
More Examples
## GyozaMixV1
GyozaMix was the first to merge mouseypastel (and other associated mousy mixes) with Silicon. This made GyozaMix extremely good at landscapes, which is carried on into future Gyozamixes. GyozaMixV1 comes with a latenight version that has noise offset mixed in since it does not have it by default. GyozaMixV1 was not very successful at its goal, so while it cannot be called a failed model, it cannot be called a successful one either. It resembles Silicon with slightly less realism and more illustrative textures mostly.
Day vs. Night?
Examples
## LingesGyoza/LINES
Linges Gyoza and Gyoza_LINES were attempts to flatten out the 3d texture that Silicon had given to Gyoza and incorporate LoRA mixing. Flat Gyoza is slightly flatter, Linges Gyoza is flatter than that. Reviews indicate that they are flexible models that respond well to prompting.
[LingesGyoza](https://huggingface.co./Jemnite/GyoZanMix/blob/main/factorygyozan/LingesGyoza.safetensors)
[LingesGyozaAfterDark](https://huggingface.co./Jemnite/GyoZanMix/blob/main/factorygyozan/LingesGyozaAfterDark.safetensors)
[GyozaLINES](https://huggingface.co./Jemnite/GyoZanMix/blob/main/factorygyozan/GyozaLINES.safetensors)
Examples
What's the difference?
## GyozaMixV2
GyozaMixV2 is a complete overhaul with the way GyozaMix was done by incorporating more Block Weight Merging. It retains more of the GyozaMix feel in shading in texture but the lines are still quite thin.
[GyozaMixV2](https://huggingface.co./Jemnite/GyoZanMix/blob/main/factorygyozan/GyozaMixV2.safetensors).
Examples
## GyozaMixV3
GyozaMixV3 utilizes loCons as well as LoRAs to finally return to mousymix's texture while retaining GyozaMixV1's ability to do fantastic backgrounds and proper proportions. In this respect it has successfully fulfilled the original goal of GyozaMix.
[GyozaMixV3](https://huggingface.co./Jemnite/GyoZanMix/blob/main/factorygyozan/GyozaMixV3ClipFixed.safetensors)
[GyozaMixV3fp16](https://huggingface.co./Jemnite/GyoZanMix/blob/main/factorygyozan/GyozaMixV3fp16.safetensors)
[GyozaMixV3.2](https://huggingface.co./Jemnite/GyoZanMix/blob/main/factorygyozan/GyozaMixV3.2.safetensors)
[GyozaMixV3.2fp16](https://huggingface.co./Jemnite/GyoZanMix/blob/main/factorygyozan/GyozaMixV3.2fp16.safetensors)
Examples
### GyozaMixV3.5
GyozaMixV3.5D is the long put off, long unavoidable, move from illustrative to 2.25D as a result of numerous and various lighting upgrades. It prompts similarly due to V3.2 due to face that it is only a style enforcement upgrade on V3.2.
Examples
Examples
## GyozanWinterHoliday2023
WinterHoliday2023 is a special edition made especially for Unknownno3 that abandons many of the traditional precepts of Gyozanmix. It trades flexibility for a focus on cinematic shots, hardbakes in lighting choices, and is primarily built on moidel components which are already heavily finetuned via block-merging. Overall it is a very aesthetic model that can create very beautiful pictures but also has many weaknesses. Generally speaking, it should be considered a upgrade to V5, but a flawed one that accepts steps-back in some areas in order to provide a more cinematic output.
Examples
## SanXianWonton
SanXianWonton is a GyozaV2-based mix to affect the style of ink washed paintings (水墨画, lit. water rubbed art) popular in China, Japan, and Korea from the late 10th century until the modern day on anime illustrations. It has its own repository [here](https://huggingface.co./Jemnite/ShuimohuaAnime).
Examples
## TangYuan
TangYuan is a SanXianWontonV2-based mix to convert it into a general use illustration model. It is good with mimicking tradition forms of paintings but is also good at other things too. It has amazing landscapes and does wide angle shots well. It has its own repository [here](https://huggingface.co./Jemnite/SomethingSweetAndSticky).
Examples
## Bayesian Mixes
### GyozaSalute!
Weights
1.0,0.18607648082955117,0.0,0.4650677066890658,0.5615966350132885,0.0,0.8401037755276394,0.332580097221715,1.0,1.0,1.0,0.5930444012946376,0.8652689667099235,0.4050236946871791,0.25976747288269475,0.09796104560611762,0.0,0.1339274177431354,0.0,1.0,1.0,1.0,1.0,0.6328046281809198,0.0