This is a model used to quickly generate images on civitai ,
you can generate 4 images with 5 buzz.(12 step)
If the quality is not enough, you can increase it to 18 steps. (1 image with 2 buzz )
The advantage of this model is that it can generate more stable images at lower resolutions, and the diversity in SFW situations is also very good. If you'd like to generate more NSFW images, please use this
https://civarchive.com/models/827184/wai-nsfw-illustrious-sdxl
For the example images
I used 832x1216, 12step,
generated directly without ADetailer and hires fix.
Recommended settings:
Steps: 12-18
CFG scale: 5
Sampler: Euler a
The VAE is already integrated, please do not ask such questions anymore.
There are four safety rating tags: general, sensitive, nsfw,explicit.
Users are expected to consciously add "nsfw" to negative prompts to filter inappropriate content.
Positive Prompt
masterpiece,best quality,amazing quality,Negative Prompt
bad quality,worst quality,worst detail,sketch,censor,Description
FAQ
Comments (31)
FIRST ! I'm gon try this ASAP !!! WAI FTW !!
Edit: call me crazy but the NSFW v11 model generates images with just as good a quality at 12 steps, similar cost (for 1 or few imgs) but has superior NSFW understanding so not sure what this one has over it.
Sorry, I made a naming mistake when I just woke up, but it has been corrected. This version will better align with normal usage, with a more balanced weight, and better diversity at lower resolutions.
The V11 NSFW weight there is a bit too high, making it difficult to generate images suitable for normal with very short prompts
Loving this checkpoint so far, generates excellent images for being around 10 buzz to generate 2 images!
I'm very astonished! It creates images at a fast speed (every 3 minutes), even with my ancient GPU.
With no loras and at the recommended 12 steps, you can generate 2 images with 4 buzz. And the quality of the image gens look very decent too. This is extremely cost effective.
You can check if the tip option is turned off. According to my settings 12step, it should be 5 buzz for 4 images.
@WAI0731 that's right, I have creator tips enabled in my settings
@im_new_to_this Actually, you don't need to tip because BUZZ can't be converted into cash, and it has no use for me.
@WAI0731 ohh I didn't know that. I thought it would help creators like you
@im_new_to_this Thank you for your donation. Others may not know, but it doesn't really have any use for me.
ちょっと軽くテスト出力したいときに、時間も節約、品質も維持してくれる、最高です、ありがとう!
It's really easy to use.
This model proves that fast and cheap can also be good.
Can't wait to test on my 3060.
Local environment.
896x1152 with 4070ti can be generated in less than 6 seconds. If you use faster LORA + other models, you may use it regularly instead.
ローカル環境。
4070tiで896x1152なら6秒未満で生成可能。高速化LORA+他のモデルを使うなら、代わりに常用してもいいかもしれない。
works with my loras, phenomenal!
bro，15steps 5.129s？歪师傅你是这个👍
very efficient for the quality & still being able to use a (relatively) high cfg without cooking
and yeah probably the best to farm extra buzz from daily generator feedback for buzzlets since you'll spend a whole lot less on image generation than you gain from feedbac, could be useful
Can anyone confirm if it works correctly on 512x512 res (local run)?
this is a sdxl model
@derpmagician yeah I tried it recently on another site that allows custom resolutions and the results were not ideal which is to be expected from non SD1.5 based models. I was only curious cuz in the description it mentioned working well even with low res.
Hello! Civitai finally added "NoobAI" category. Please change on it.
This is an incredible model that's compatible with almost any lora. Why can't it be used in Civitai anymore?
civitai disabled generation for most models, creators cant do anything about it
When they rolled back the all checkpoints update only the 197 most used checkpoints plus illustrious 1, 1.1, and 2.0 were kept.
@sci Wasn't this checkpoint available for the on-site generator before full checkpoint coverage was implemented?
this used to be one of only 2 models i could use for illustrious just the other day, and now it's no longer selectable for use here. how the turn tables
Yeah, I've been missing it because this checkpoint is no longer available here
请问什么时候更新呀，很期待这个模型继续更新下去qwq
Apparently the major part of the users of this site are allergic to use models in the CivitAI's generator that are able to generate an image in less than 20 steps.
I say this because because is unbelievable the fact ChromaXL Mix and this model aren't winning in the auctions for the ability of being used in the Image generator of this site lately.
My 3060 12gb takes 20-22 seconds to process each picture in steps of 30, and 16-18 in just 10-12 seconds. I'd like to see more quality models that can produce the same result in fewer steps.... ٩(｡•́‿•̀｡)۶
WAI-CheapFast is so much better then ChromaXL. ChromaXL is hot dookie.