If you want to use more my checkpoint online generation, please visit here.
https://tensor.art/u/762555264535746522
You can run WAI-ANI-NSFW-PONYXL and use its API on SinkIn:
V14 released.
Increased composition diversity and improved model accuracy now allow images to be generated with fewer steps.
Recommended Setting
Steps:20-30
CFG scale: 5-7
Sampler: Euler a/DPM++ 2M Karras
ADetailer face_yolov8n/s.pt use can fix eyes
For the example images
I used 1024x1360,
generated directly without AD or hires fix.
Positive Prompt
score_9, score_8_up, score_7_up,source_anime,Negative Prompt
worst quality,bad quality,jpeg artifacts, source_cartoon,
3d, (censor),monochrome,blurry, lowres,watermark,V13 released.
Increased body stability and accuracy.
overall balance adjustment.
Recommended Setting
Steps: 30
CFG scale: 7
Sampler: Euler a/DPM++ 2M Karras
ADetailer face_yolov8n/s.pt use can fix eyes
For the example images
I used 1024x1360,
generated directly without AD or hires fix.
Positive Prompt
score_9, score_8_up, score_7_up,source_anime,Negative Prompt
worst quality,bad quality,jpeg artifacts, source_cartoon,
3d, (censor),monochrome,blurry, lowres,watermark,V12 released.
Style adjustment, background adjustment, and increased stability.
Added more compositions.
Recommended Setting
Steps: 30
CFG scale: 7
Sampler: Euler a/DPM++ 2M Karras
ADetailer face_yolov8n/s.pt use can fix eyes
Positive Prompt
score_9, score_8_up, score_7_up,source_anime,Negative Prompt
score_6, score_5, score_4, source_cartoon,
3d, (censor),monochrome,blurry, lowres,watermark,v11
Weight adjustment, balancing the art style
All example images(Except the cover) are generated at 1024x1360, without using AD and hires fix.
Recommended Setting
Steps: 30
CFG scale: 7
Sampler: Euler a/DPM++ 2M Karras
ADetailer face_yolov8n/s.pt use can fix eyes
Positive Prompt
score_9, score_8_up, score_7_up,source_anime,Negative Prompt
worst quality,bad quality,jpeg artifacts, source_cartoon,
3d, (censor),monochrome,blurry, lowres,watermark, V10
Better Background
Clothes details up
Better mature female
Recommended Setting
Steps: 30
CFG scale: 7
Sampler: Euler a/DPM++ 2M Karras
ADetailer face_yolov8n/s.pt use can fix eyes
Positive Prompt
score_9, score_8_up, score_7_up,source_anime,Negative Prompt
worst quality,bad quality,jpeg artifacts, source_cartoon,
3d, (censor),monochrome,blurry, lowres,watermark, ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
V9 Hyper12step
Please refer to the instructions on the right side of V9 Hyper12step for usage.
V9 released.
v8->v9
adjusted the facial data to make it more versatile, allowing for easier generation of faces across various ages.
added more NSFW materials.
attempted to reduce the generation of some unnecessary details.
Recommended Setting
Steps: 30
CFG scale: 7
Sampler: Euler a/DPM++ 2M Karras
ADetailer face_yolov8n/s.pt use can fix eyes
Positive Prompt
score_9, score_8_up, score_7_up,source_anime,Negative Prompt
worst quality,bad quality,jpeg artifacts, source_cartoon,
3d, (censor),monochrome,blurry, lowres,watermark, __________________________________________________________________________________________________
hyper12step released.
about size
Various combinations from 768x768 to 1360x1360 can produce good images.
Recommended Setting
Steps: 12
CFG scale: 3.5
Sampler: Euler a
Not very necessary to use ADetailer correction
Positive Prompt
score_9, score_8_up, score_7_up,source_anime,Negative Prompt
score_6, score_5, score_4, source_cartoon,
3d, (censor),monochrome,blurry, lowres,watermark,PS.The example images use a size of 896x1192. no AD no Hires. fix
__________________________________________________________________________________________________
Version 8 released.
v7->v8
Added some training related to camera angles and backgrounds.
Added more NSFW materials for additional training.
Added some training for special effects.
Optimization of overall composition logic.
Fixed some minor issues.
better eyes and hands.
Recommended Setting
Same as v7
__________________________________________________________________________________________________
Version 7
v6->v7
Fixed the issue where images appeared entirely covered with snowflakes.
Revised the composition logic to be more suitable for general scenarios.
Added new assets for backgrounds, mature women, and clothing.
Ensured consistency with the previous model and further fixed the issue of blurry eyes in medium and long shots.
Fixed some minor issues.
V7 Recommended Setting
Steps: 30
CFG scale: 7
Sampler: Euler a/DPM++ 2M Karras
ADetailer face_yolov8n/s.pt use can fix eyes
Positive Prompt
score_9, score_8_up, score_7_up,source_anime,Negative Prompt
score_6, score_5, score_4, source_cartoon,
3d, (censor),monochrome,blurry, lowres,watermark,________________________________________________________________________________________________
Regarding the 8step model, it is a lightweight model that can generate images in as few as 8 steps. However, the quality and text comprehension will be relatively lower. If speed is not a major concern for you, it is recommended to continue using version 6.0.
8 STEP Setting
Steps: 8-12
CFG scale: 3
Sampler: Euler a
ADetailer face_yolov8n/s.pt use can fix eyes
Positive Prompt and Negative Prompt same 6.0
__________________________________________________________________________________________________
6.0
v5->v6 Change point
Better Background
better eyes
fix vae
Recommended Setting
Steps: 30
CFG scale: 7
Sampler: Euler a/DPM++ 2M Karras(better eyes)
ADetailer face_yolov8n/s.pt use can fix eyes
Positive Prompt
score_9, score_8_up, score_7_up,source_anime,BREAKNegative Prompt
score_6, score_5, score_4, source_cartoon,
3d, (censor),monochrome,blurry, lowres,watermark,You can view the Supported art styles from https://civarchive.com/articles/5715
Thank you, DeepDark_Fantasy514, for providing the test.
v4->v5 Change point
Better Background
Clothes details up
Add some mature female
Little better hands and eyes
Recommended Setting
Steps: 30
CFG scale: 7
Sampler: Euler a
ADetailer face_yolov8n/s.pt use can fix eyes
Positive Prompt
score_9, score_8_up, score_7_up,Negative Prompt (THX for snipsnapsnipsnap recommend )
Base:score_6, score_5, score_4, source_cartoon
Optional: source_furry, source_pony can bleaches skin
else: 3d, (censor),monochrome,blurry, lowres,watermark,
I often use it now
[score_6, score_5, score_4, source_cartoon,
3d, (censor),monochrome,blurry, lowres,watermark, ]
Description
The example images use a size of 896x1192. no Hires. fix
Recommended Setting
Steps: 12
CFG scale:3.5-5
Sampler: Euler a/DPM++ 2M Karras
ADetailer face_yolov8n/s.pt use can fix eyes
Positive Prompt
score_9, score_8_up, score_7_up,source_anime,Negative Prompt
worst quality,bad quality,jpeg artifacts, source_cartoon,
3d, (censor),monochrome,blurry, lowres,watermark, FAQ
Comments (63)
I know that the existing pony model does not apply worst quality, bad quality, and jpeg artifacts to Negative Prompt. Does it work for that model?
Instead, could there be a problem if score_4, score_5, score_6, source_pony, and source_furry were used?
This model is not simply a combination of several models, but has undergone several small-scale training sessions. Therefore, it uses a mix of negative prompts from SDXL and pony.
Actually, you can achieve good results without using any negative prompts. My recommended settings are just what I find to be the most convenient for generating high-quality images.
This model gives a pretty good results. Some times it generates extra limbs,but this is not critical. You deserved a thumbs up
Anyone had found what´s new about NSFW stuff? I´m always trying random danbooru tags. Any positions list or something ?
Danbooru itself has a tag list you can use:
https://danbooru.donmai.us/wiki_pages/tag_groups
If you look for "Sex" through the tag list, you'll find "Sex acts" and "Sexual positions"
@markmuligan Thanks, but I already knew about the page, I meant it would be nice to have something like an “update sheet” about NSFW content from the previous version. It's not a big deal anyway.
When I look at some of the posts here, I seriously wonder why you would use a checkpoint that is explicitly designed for anime style only to then use half a dozen Loras that contain different styles together with that checkpoint..., weird people here!
Because that's what LoRAs are for, to mix styles and ideas? You mix style A with style B and out comes C. Simple?
@Tozi_White Yes, you can do that, but demonstrating what the checkpoint can do together with all possible styles here on this checkpoint page is not the right place and this page is not intended for that either, it has the sole purpose of showing renders that were created ONLY with this checkpoint, everything else you can publish in your own gallery! I really hope you understood it this time!
@ArtistryUnleashed You must be very tired of browsing Civita, if such a thing annoys you that you had to point it out in a comment, because under every checkpoint it is exactly the same. Nowhere is it written that the render made on a particular model is to be made only on it and with nothing else. No offensive, of course, but it's more the ones you seem weird in this case. People are playing with AI the way they want to, and it's awesome.
even without Loras it does a good job of mixing other styles, I've done 3d, semi realism, and hyper realistic with great results, overall, a great checkpoint
So?
好色好色
love this so much! would try this later
Very good
I hope you don't get rid of censorship generation. I know it's annoying to some people, but I use 'black bar censorship' in my prompts, and it saves me a lot of time.
v9 is more realistic
Fantastic! Best anime Checkpoint.
Hello! can we add realistic, realistic style, semi realistic on the negative to get a better anime look ? For me the model make me big nose anime girl that's why i ask to cut it and put these words to make it anime type!
I personally prefer love v8, v9 makes me feels more like a Korean style anime character, more realistic
This comes down to personal preference. Starting from version 6, all the models are very good to use, so just pick the one you like.
✨🦄👍🏻👍🏻
v9比以前的版本对比度低了
V9 Hyper12step makes really good background which I look for on a model.
Is this a NoobAI merge?
this is a pony，not noob merge
This model produces very good results with draft mode, I prefer draft mode over the normal mode
How bruh, ain't no way
nah bro I just tried it and op kinda valid, results looked nice but were just super messy like a hyper low cfg (bad prompt adherence plus messing up everything).
like try a draft then use a hi-res fix with the max cfg scale on the best one, it is insane
@FarEmms FR?
Imma try it rn, but what about the steps and the samplers tho, are those still put on the lowest or something?
@Annongg Euler A and balanced steps (maybe 28?) is a proper safe bet.
draft mode is super creative like having a low cfg and its cheap so just select the best one and use a hi-res fix and in the hi-res max out the cfg so it adds all the detail that was missing in the drafts.
much better than generic results you get from balanced everything.
v9.0 is fantastic for semi-real anime style, great hands and bodies, but seems to struggle with dark skin.
This model is not designed for that type of images
The model is one of the best, in my opinion, but it has a problem with images, when you try to force that for example the character vampire bites a woman in the picture, this model does not understand the meaning of "bites in the neck, bite in the neck", etc. because of what is almost unrealistic (I have never succeeded) to depict such a picture without Lora, which also does not do it correctly. Maybe the author will read this message and in the next version will fix this point.
Also the vampire teeth bite marks on the neck this model can't make.
bite(s) in the neck, in proper Pony vocabulary is : neck biting
"this model" is a Pony model. To use any Pony model properly, you have to write things it understands, aka tags it's been trained on.
Use BooruDataSetTagManager. This'll help you know how to write the tags for your needs, each tag separated by a coma.
Don't describe a scene with sentences like you'd do with SDXL.
good!
Sadly the v9 gives totally different results when compared to v8 and v8 couldn't used on site anymore... 🥺
Civitai does not allow multiple online.If you want to use V8, you can go to the link https://tensor.art/models/773705566569798829
@WAI0731 can't do the same prompt there, it has a bunch of limitations
@WAI0731 Tried again tensor, everything need to be paid, limit everything, unusable site...
@CathyPowers The free users can only use up to 25 steps. If you modify that, it should be able to generate.
@WAI0731 steps, resolution, heaviness more than 1.5, upscale, etc
@CathyPowers He only provides the basics, and you'll need a membership for the other features. I still recommend that running it locally is better if possible.
@WAI0731 i live in Brazil, any computer that could make AI would require me to not eat for a few months
@WAI0731but yeah, generate locally is 100x better
Very versatile model !!!
Nice job
Nice
Can you train a lora with this checkpoint? Or it's better to stick with regular pony/autismmix? I really like this checkpoint anime style..
Good
Not so good, all the character traits are going wrong, even I used LORA.
Why i cant us it anymore on openart?
how flexible it with pony style loras?
its pretty flexible, but I honestly think v8 was more flexible
Fun model
Works well and has quite the flexability
Pretty epic
I never found a model that can compete with AutismMix, but gotdamn your model is VERY good ! Freaking like it V9/V9 Hyperstep. Very good job !
Very Flexible and can do male characters, Modeler always go with hardcore female bias and ruin base pony.