AnyHentai is a merge of multiple models
for sampling methods, use Euler a (best), DDIM (second best), or DPM++ DM Karras
DOWNLOAD/INSTALLATION
Step 1: download SAFETENSORS and VAE files.
Step 2: put the SAFETENSORS file under "stable-diffusion-webui\models\Stable-diffusion"
Step 3: put the VAE file under "stable-diffusion-webui\models\VAE"
Step 4: Done! enjoy the model
USAGE
-use a minimal negative prompt for best results
-use Euler A and 20-25 steps for best results
-use danbooru tags
-I used a clip skip of 2 (optional)
-I also used the upscaler Latent (nearest-exact) with a highres step of 20 and denoise of 0.5 to improve image quality and detail (optional)
DO NOT USE A DENOISE STRENGTH BELOW 4.5 FOR BEST RESULTS
EXAMPLE PROMPT
((Masterpiece)), (best quality), (1girl), red hair, beautiful red eyes, medium breasts, classroom, black glasses, school uniform
VAE
The VAE is not required. you can use any VAE you like. I have found that VAE makes more vibrant and crisp images, but this is just my testing
for the VAE I used the same one as grapefruit
I will try to improve and update this model by adding other images, although I'm not too familiar with SD and training models, so I will most likely stick to only merging models.
_______
EXTRA INFO
-Loras have not been tested yet, but they should most likely work
-use the upscaler Latent (nearest-exact) for best results
- I will try to update the model at least once per week
-Image generation on the A1111 WebUI normally took around ~10 seconds using a 3080 TI with 12 gigabytes of RAM
CREDIT TO THE PEOPLE WHO MADE THE MODELS ORANGEMIX, HENTAI DIFFUSION, ANYTHING V3, ANYTHING V5, DREAMSHAPER, KENSHI, COCOA, CORNEO'S 7TH HEAVEN MIX, PROTOGEN V2.2 (ANIME), KOTOSMIX, COUNTERFEIT, MIXPROV3 AND GRAPEFRUIT
Description
Much better overall
Trained with mew model (AnythingV5)
Model now comes with both CKPT and SAFETENSORS files
FAQ
Comments (24)
OMG NEW VERSION LETS GO
The grapefruit model no longer lists a VAE. Could you upload the one you had?
EDIT: After some research, i've found out that both VAEs (Anything's and AOM's) are just a renamed NovelAI VAE. This is the one that grapefruit used originally. So it might just be the one you are looking for: https://huggingface.co/WarriorMama777/OrangeMixs/blob/main/VAEs/orangemix.vae.pt
No VAE is needed for version 2.0
@asdpro123 ahh so it is baked in ? cheers
when i tried inpainting with this today, i kept getting a "NansException: A tensor with all NaNs was produced in VAE." error, and when i disabled nand check, instead of generating something in the inpainted area, it would just make the inpaint black color more blurry and make it part of the pic. no clue what's goin on, had to use a completely different model to actually do anything.
has this model ever seen a male's face in its life? lol
prolly not
-={ How to Prune the model }=-
-={ Without breaking it }=-
My Model Converter Settings are like this:
- Precision: fp16 ( Already halves model without pruning if the model was fp32 )
- Pruning Methods: no-ema ( When using this you wont be able to merge with other models anymore or will notice problems, ema-only breaks Model )
- Checkpoint Format: Safetensors ( ckpt will also work safetensors loads faster into GPU )
- Extra Options: ( I didn't mess around with these. might become useful after OP adds baked vae or something )
NOTE: This will cause the results of your model to be different even if you have the same seed, CFG, prompt, steps, clip skip or sampler.
Model Converter A1111 Extension: Github
Instruction Video: Link
can you provide a download link for the pruned version?
where's the VAE for 2.0?
it's always the same VAE, NovelAI/AnyHentai/Anything/AnimeFull .. they are all just the same renamed, is the original of NovelAI in any case, same hash/CRC for all of them, just keep one and use it :).
Why the example prompt is using () between some words ?
That is to emphasize and give more importance to those prompts. You can have a maximum of 3 on each side "(((example )))". After that, you may encounter corrupted results.
(((red))) or (red:1.2) or (red:0.5) .you can try it. more effect.
Where is the VAE?
IIRC, this version has a VAE baked in.
@JAS2023 Getting an occasional error that suggests something is wrong with the VAE: NansException: A tensor with all NaNs was produced in VAE. This could be because there's not enough precision to represent the picture. Try adding --no-half-vae commandline argument to fix this. Use --disable-nan-check commandline argument to disable this check.
Too bad it won't fit on my google colab. Perhaps you should try to make prune version of it
Can i make a suggestion. I don't think its good to separate your pruned version and put them on another username. Keep all version of this together. Just giving you feedback. Makes it hard to track your updates if the same products is getting split up.
thats so great
can you do 1.9 pruned please thanks
so big！well done，thank your model
this was one of my fav SD 1.5 checkpoints before Pony and Illustrious... but none of them could nail this aesthetic so far :(
Is there a kind soul who could point me in the right direction pls?
Getting frustrated at how Pony's too western looking, but Illustrious has samey anime faces unless using style loras
Details
Files
anyhentai_20.safetensors
Mirrors
anyhentai_20.safetensors
5706_anyhentai_20.safetensors
AnyHentai2.0.safetensors
anyhentai_20.safetensors
anyhentai_20.safetensors
anyhentai_20.safetensors
anyhentai_20.safetensors
anyhentai_20.safetensors
anyhentai_20.safetensors
anyhentai_20.safetensors
anyhentai_20.safetensors
anyhentai_20.safetensors
anyhentai_20.safetensors
anyhentai_20.safetensors
nyhnt.safetensors
AnyHentai v2.safetensors
AnyHentai.safetensors
anyhentai_20.safetensors
anyhentai_20.safetensors
anyhentai.safetensors