what is improved in v1.3? :3

Will 1.3 have inpainting model too?

does it iclude inpainting?

Sorry, I'm wnoder which base model you use to traind?

I checked hassakuV1.3 by model toolkit.

https://github.com/arenasys/stable-diffusion-webui-model-toolkit.git

Then the following was displayed:

Model is 1.99 GB. Model type identified as SD-v1. Model components are: VAE-v1-SD, CLIP-v1-SD, UNET-v1-SD.

Contains no junk data. CLIP had incorrect positions, fixed: 41.

Uses the NAI VAE.

Model will be fixed (1 changes).

The CLIP seems incorrect, is this what you intended? Or is it a your mistake?1

Hello.

Thanks for all your work.

Is there a flat-2D version (Koji style) planned for your Hassaku model?

I don't know why but whenever I use this model, the images it generates are always very bright, I have tried everything and although the results are good, in the end the image is spoiled because there is too much brightness.

Can someone give me some advice?

Does anyone use this for vorish themes? I would like to learn how to make photos like that.

Tysm! Your checkpoint is really good, it fits almost every loras

Why would my generated image have a mosaic or censor on pussy?

I noticed that my default photography prompt for any model preview I prepare doesn't work great with the models Hassaku, Sudachi and Koji. Great models otherwise!

The model is very good, it is one of the best I have tried but I would like the pussy to have more effect, not that to have a more realistic effect you have to add other lora and lower the quality of the image.

"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"
I think you are underestimating the fetishes of ppl, while not my thing there are likely ppl who'd like that XD

Why won't it show up in the search? Are mods deboosting "bad" models?

when using the model i get like abstract art, for some reason, can any one explain. I just coping prompts and settings from images in the gallery.

Anyone use these anime models with realistic loras and get cool outputs?

Is this Checkpoint based on Anythingv3?

This is by far my favorite 2d model. I do everything in it, and things that are not possible ehehe.

Thanks for this, and keep up the good work

I find Hassaku to be just the worst to try and generate results with. I always, ALWAYS get mangled eyes, like it's trying to draw realistic eyes INSIDE anime eyes, even when I'm copying prompts. Even when I drag and drop images from this page, created with this model, with no other LoRAs (or LoRAs I have), keeping every single setting the same; I get the same image, but the eyes are mangled.

what are the poses that work without needing any loras?

Just found Hassaku after checking if Grapefruit had any updates. I guess I'll be upgrading to this one. Thanks

The model has trouble with males in almost any context. Would suggest improving that for future. Otherwise great model!

Would love a little update to this as most of my loras are still not on XL and this is my favourite

this is probably one of the best non XL model around

Do you have plan to make Dora model for SD 1.5. Or you can open training data for the model?

I don't know what happened, but this model seems to have undergone some changes. I tried my best to exclude the influence of various functions, but the painting style of this model has changed a lot, and it has even become worse. I downloaded and used this model online using cloab, I'm not clear what happened, did you make any changes to this model?

I just wanna say good...

good!I used it to produce MISAKA, it was quite stable and my face didn’t collapse! Thank you.

I went back to this for funnsies and i really still love it. Pony models suck at booru tags :x

Great model.

Please tell me if this downloaded model is placed in Checkpoint or in VAE

hello, can you tell the resolution you had your best results? I had amazing results in 1280x720, but now is starting to crwate another character, an extra limb (even though I put it on negative). thanks

any suggestions for the negative prompt to not create another character? I have a character but it's starting to render two or three of them on the same image. My negative prompt bellow:

score_5, score_4, score_3, score_2, score_1, lowres, (bad quality, worst quality:1.2), (bad anatomy:1.0), (extra fingers: 1.5), sketch, jpeg artifacts, ugly, poorly drawn, signature, watermark, blurry, low quality, distorted anatomy, poorly drawn face, (extra limbs:0.7), overly exaggerated proportions, messy background, dull colors, super saturated colors, malformed limbs, multiple people, crowd, group, two people, three people, extra hands, extra legs, extra heads, twins, reflections, duplicates, shadows of people

what the heck happened, why isn't it useable anymore

Despite its rather advanced age, this model is really still quite good among other anime models. Good anatomy and responsiveness, a pleasant anime look, while there is no excessive female focus. It is not surprising that it is still very popular, because good things do not become obsolete. Great model! Thanks!

Too bad i cant use it to generate from model source :(