Hassaku aims to be a model with a bright, clear anime style. Model focus are nsfw images, but also with a high emphasis for good looking sfw images as well. My <a target="_blank" rel="ugc" href="https://discord.gg/zSR5FcYWWE">Discord</a> for everything related to anime models and art. My <a target="_blank" rel="ugc" href="https://discord.gg/zSR5FcYWWE">Discord</a> for everything related to anime models and art. If you'd like to support my work, you can do so at: <a target="_new" rel="ugc" href="https://subscribestar.adult/citrus-models">SubscribeStar</a>, every bit helps and is truly appreciated!____________________________________________________________ Supporters:Thanks to my supporters tamashiicolle, pttcot, SETI and Kodokuna_____________________________________________________Using the model:Use mostly <a target="_blank" rel="ugc" href="https://danbooru.donmai.us/">danbooru</a> tags. No extra vae needed. For better promting on it, use this <a target="_blank" rel="ugc" href="https://aituts.com/novelai-anime-prompt-techniques/">LINK</a> or <a target="_blank" rel="ugc" href="https://lunarmimi.net/freebies/novelai-anime-girl-prompt-guide/#1basic">LINK</a>. But instead of {}, use (), <a target="_blank" rel="ugc" href="https://github.com/AUTOMATIC1111/stable-diffusion-webui">stable-diffusion-webui</a> use (). Use "masterpiece" and "best quality" in positive, "worst quality" and "low quality" in negative.My negative ones are: (low quality, worst quality:1.4) with extra monochrome, signature, text or logo when needed.Use a clip skip 1 or 2. Clip 2 is better for private parts, img2img and prompt following. Clip 1 is visually better, because i assume, the model has more time and freedom there. I use clip2.Don't use face restore and underscores _, type red eyes and not red_eyes.Don't go to really high resolutions. Every model, like hassaku, get lost in the vastness of big images and has a much higher chance to greate, as example, a second anus._____________________________________________________<a target="_blank" rel="ugc" href="https://rentry.org/hdgpromptassist#loras">Loras</a>:Every LoRA that is build to function on anyV3 or orangeMixes, works on hassaku too. Some can be found <a target="_blank" rel="ugc" href="https://rentry.org/hdglorarepo">here</a>, <a target="_blank" rel="ugc" href="https://gitgud.io/gayshit/makesomefuckingporn#lora-list">here</a> or on civit by <a target="_blank" rel="ugc" href="https://civarchive.com/user/lottalewds">lottalewds</a><a target="_blank" rel="ugc" href="https://civarchive.com/user/fuggy">, </a><a target="_blank" rel="ugc" href="https://civarchive.com/user/Trauter">Trauter, </a><a target="_blank" rel="ugc" href="https://civarchive.com/user/Your_computer">Your_computer</a>, <a target="_blank" rel="ugc" href="https://civarchive.com/user/ekune">ekune</a> or<a target="_blank" rel="ugc" href="https://civarchive.com/user/fuggy"> </a><a target="_blank" rel="ugc" href="https://civarchive.com/user/lykon">lykon</a><a target="_blank" rel="ugc" href="https://civarchive.com/user/fuggy">.</a>_____________________________________________________Black result fix (vae bug in web ui): Use --no-half-vae in your command line argumentsI use a Eta noise seed delta of 31337 or 0, with a clip skip of 2 for the example images. Model quality mostly proved with sampler DDIM and DPM++ SDE Karras. I love DDIM the most (because it is the fastest).

Hassaku (SD1.5)

Biggest change is a more drawn style with less extreme bright skinWhat is better:- Slightly more open pussy and anus (with highress fix)- Anuses are less forgotten- More responsive to nsfw tags- Less realistic faces, without face defining prompts (ugly big lips faces, not good semirealistic)- Slightly less chances of double limbsWhat is better/worse for some:- A little bit less "shiny", more drawn and less extreme bright skin (especially on highresfix)- Clothes are more revealing- Bigger possibility of motion lines, motion blur and floating symbolsWhat is worse:- More chance of double anuses- Less precise to create menHard to determine: - Hands seems better then in 1.1, but hands are hands- Highressfix seems more controllable (less changes on highress) Trained with 30k NSFW images with later on fine-tune iterations. But needed heavy merging to make hands better again and reduce motion lines, motion blur and floating texts and hearts which are prominent in NSFW images. Problem with those, on highress, motion lines blend with the body and do spikes, motion blur extens to blurryness and floating texts/hearts morph into objects. The end result is much less changed as anticipated, sometimes identical, i decided to put it up as 1.2 version.

Files with SHA256: d3cd6ac55ae99c4848af5bde90b47306b17ee0a3a7b1aff4a7187284d3cdbf0e