Welcome to Taureal Mix!
--
If you love my work you can buy me a coffee : ko-fi.com/tauron you dont have to, but if you do, it helps me stay awake until midnight (which is the only time im able to work on these stuff lol) -- or.. a like or a review wont hurt or takes you more than 1 mins, while it takes me hours, days or even week to make this - it helps the thread alive and afloat.
--
**Update Changelog can be seen at model version details.**
NOTE : Alot of sample images has been removed by CivitAi even though there is no TOS violation.
--
Suitable to those who needs:
All round realistic model with versatile styles.
Low memory cost.
Accurate txt2img results with short/long prompt.
Low model bias = more accurate lycoris/lora representation.
Better integration with my previous / future lora.
Disclaimers:
tested mostly for txt2img, for other functions use at your own discretion.
tested specifically for my previous or future lora. loras from other creators may work or may not work well with my model.
a pruned model is not suitable to be merged. choose the v3.1 or v2.1 instead.
Prerequisite & Resources
These are the resources i use for my standard workflow :
ADetailer
GitHub - Bing-su/adetailer: Auto detecting, masking and inpainting with detection model.
Negative Embeddings (Use 1 only)
Use when needed only, v3.7 pruned without negative ti is observed to produce significantly better result compared to v2.1 / v3.1
Deep Negative V1.x - V1 75T | Stable Diffusion Textual Inversion | Civitai
accurate prompt such as race, skin tone but smoothen the skin texture and has face bias (change result face), also sometimes misinterpret prompt position occassionally.
EasyNegative - EasyNegative | Stable Diffusion Textual Inversion | Civitai
most accurate amongst other negative ti -- prompt such as race, skin tone, face, and texture is very accurate (90%), but sometimes misinterpret prompt position (ex. cum is supposed to be on body, but is on face instead)
most detailed amongst other negative ti -- this embedding (32v or 75v) pushes skin texture and facial detail to the top. however, it changes skin tone, color, and face significantly producing inaccurate result if you are trying to use lyco.
*DO NOT COMBINE NEGATIVE - TI WITH SAME FUNCTIONS IN THE V3.7 MODEL*
Facial Enhancer Embeddings (optional)
yesyeahvh/ulzzang-6500 · Hugging Face
Pure Eros Face - PureErosFace_v1 | Stable Diffusion Textual Inversion | Civitai
Credits to all creators above that helps the model achieve its sample images quality above.
How To prompt this mix
You can follow your own prompting method, this guide only serves as a convention when you are having trouble producing similar results with my lora.
Style —> realistic , realistic details, detailed for realism / anime for anime / masterpiece , best illustration for conceptual art.
Subject —> describe your subject can be anything. 1girl, solo, cute, beautiful, slender, etc. if you use my character lora this should be where you put trigger words of the character
Attire —> describe the outfit of your subject. If you use my attire lora this is where you should put trigger words
Background / Context —> describe the context or the scene
Textual Inversion & LoRa brackets
How to maximize this mix?
Sampler & Steps
SDE Karras Sampler = 16 / 8 Denoise @ 0.32
DDIM Sampler = 16 / 8 Denoise @0.32 (Observed Fast like 2M, but with SDE Quality, DDIM is very good for concise prompt, but is weak at translating complex prompt)
Using negative embeddings is a mustto create sample image quality.Negative Ti is no longer needed to achieve sample image quality
CFG Scale
CFG Scale (creative) 4 ----------- 10 (accurate) -- 12 (Max)
Resolution & Upscaling
Resolution: 512 x 768,
Upscaling = 4x_Ultrasharp lokCX/4x-Ultrasharp · Hugging Face
To get even better results: Use Lightroom to edit your result.
Utilizing ADetailer
Face = Noise 0.3
Hands = Noise 0.25-0.4
Body = Noise 0.3-0.5
The rule of thumb is that there are no constant equation, you need to find one your own that suits what you are aiming, and sometimes you change these value on the fly depending on your goal. its better to know what these parameters do to your results rather than copying values from people's post.
--
As an afterwords, i hope you enjoy using this model, a like or a review definitely helps, or if you find it distasteful, you don't have to use it. there are many models out there that produces similar or superior results.
Description
Updated to v2.1 Change Log
Improved realism quality in general, it moves away from 2.5D feel from the previous version while maintaining fantasy capabilities.
Improved vanilla NSFW prompts
Improved SDE sampler quality - no more weirdly lit faces when using SDE sampler in ver. 1 especially in dark scene (observed faster baking time, or maybe its me getting more patient)
Lower problems with Dynamic CFG threshold when using my lycoris in this version
hands still sometimes messed up at higher CFG - observed with multiple test scene with controlled seed with 3 comparison model vs. my model. with : 1. no neg-Ti, 2. single neg-Ti, or 3. combined neg-Ti. (Ti in this case doesn't really do shit, it did its job reducing the problems - so its mostly problem with my model, i suspect maybe its because i merged with a SD21 model with hentai model, but who knows.)
the VAE 84000 should be already baked in
Do note that your results may vary than my testing. just pray to the god of Ai often.
FAQ
Comments (4)
Would love to see examples that don't use embeddings/loras/etc. You mention "hands still sometimes messed up" but you have 7 different negative embeddings in your prompts, so you really have no idea what the model is actually doing. Or at least I don't by looking at your examples.
And a lot of those could be fighting each other and sometimes make quality of the base model worse.
I've said it elsewhere, give people a place to start with your model that doesn't require downloading other people's assets as well. These images don't reflect the quality of the base model and you'll get a lot of people that wont realize all of those embeddings in your prompts are other unrelated files.
Hi, thanks for showing interest, will try to make the sample with your suggestions. The prompt shown is basically my standard prompts since the very first i started testing, and i came up with that as i used the model and encountered a problem, then i added more negative prompts, you are right with the statement that the negative ti can fight each other and make the quality worse in this case the hands, but sometime in the past i did some testing and the problem is the same with my model whether i put just single negative ti/ multiples, (i tested with the same seed) so i assume that it should be the problem with my model, not with the ti, it helps reduces the problem 80% of the time. and you are also right that i have no idea what the model is doing lol, im not an ai engineer, and im just an enthusiast that really absorbed with this technology, so all these mix are just me doing trial and error until i come up with a result that i like, that can do what i want it to do. Anyway just as context that each of us has our own process in coming up with those negative ti and its not random how it arrives as it is now. Also, some of the loras used are just showing how it works with my lyco which is the whole point of the mix in the first place, but i get your point. Thanks for taking your time to provide your input!