all / g / s / fur / art / ai / 3dcg / f / rp / req / p2p / dis / lit / kaki chat
[Settings] [Home] [Catalog] [Search] [Thread list] [Stats] [Reports] [Watcher] [PMs] [Admin]
[Return] [Bottom]

Posting mode: Reply
BBCode
(for deletion)
  • Allowed file types are: gif, jpg, jpeg, png, webp, webm, mp4
  • Maximum file size allowed is 25000 KB.
  • Images greater than 255 * 255 pixels will be thumbnailed.
  • 157 unique users in the last 30 minutes (including lurkers)
    File: 00360-4239618668 v3.jpg
    (197 KB, 960x1264) ▶
    202187
    No.13910 [OP tools off] AS
    ▶
    Finally a new version of the GuRO Checkpoint Model is here.
    And this time it also possible to download it as a seperate LoRA.


    I have successfully trained a guro checkpoint built upon the WAI 14.0 base.
    I added 1000 synthetic images to better teach it some concepts. Bleeding of concepts decreased compared to the previous version.
    This model was trained on a total dataset of 8,000 images.

    A comprehensive list of the tags used during training can be found here:
    https://rentry.co/r-18G_illustrious_tags_V2

    It is not perfect but much better than the previous iterations. Including, but not limited to:
    Hanging.
    decapitation and amputee
    Head used as a sex toy from both directions,
    eye penetration
    nipple penetration
    ear penetration
    brain penetrarion
    Cannibalism cooked body. with face and not
    impalement
    impaled and roasted over fire
    harakiri and evisceration
    debreasting
    drowning
    and more.

    For the complete list of tags consult here:

    Fine tune checkpoint.
    Civitai: https://civitai.com/models/1794928?modelVersionId=2499924
    Link to the LoRA. Reminder that the LoRA does not work as well as the fine tune, reccomended weight is 0.8. Depending on the tags it can vary between 0.6 and 1.0.
    Civitai: https://civitai.com/models/2221022


    I have been banned on pixiv for the time being.

    My old account got banned on pixiv.
    You can contact me through this one
    https://www.pixiv.net/en/users/101800918
    Viewing last 50 posts
    >>
    No.19357
    ▶
    >>19331
    mostly manually.

    I used one of the old waify captioners and the Joy cpation.
    https://huggingface.co/spaces/fancyfeast/joy-caption-beta-one

    Results were meh, so i manually remveod or added necessary captions.
    Mos tof the images were taken form years of browsing guro chan, and from specialized guro chan threads, like decap threads on /g/ etc.
    >>
    No.19428
    ▶
    >>19055
    you're super good at this. can you do shego next?
    >>
    No.19683
    ▶
    File: _Illustrious_20260514-161443_00001_.jpg
    (270 KB, 1552x2592) ▶
    276717
    >>19428
    >>
    No.19684
    ▶
    File: _Illustrious_20260514-163354_00001_.jpg
    (275 KB, 1552x2592) ▶
    281901
    >>19428
    >>
    No.19717
    ▶
    >>19683
    >>19684
    you're literally the best right now.
    >>
    No.19724
    ▶
    Can someone upload the checkpoint please
    >>
    No.20817
    ▶
    Please come back.
    >>
    No.20896
    ▶
    >>19724

    Here you go fren
    https://mega.nz/folder/fiozmSAA#NFA18E1PEYNJ4Zvt1kXjaw
    >>
    No.21087
    ▶
    File: ComfyUI_00421_.mp4
    (732 KB, 640x640) ▶
    749756
    Nice Lora
    >>
    No.21173
    ▶
    File: _Illustrious_20260610-160228_00001_2.jpg
    (959 KB, 4464x2712) ▶
    982185
    >>
    No.21180
    ▶
    >>14069
    These look real clean- great job. Would love to see different characters in different debreasting / torture scenarios.
    >>
    No.21260
    ▶
    >>21173
    that is a cool one
    >>
    No.21275
    ▶
    File: 1780144545760.png
    [Animated PNG](326 KB, 882x1240) ▶
    334768
    How difficult do you like it? Can I request May getting a hose inserted into her butthole and her belly inflated big and round until it explodes into gore. If you could make it look like an outtake of the show that'd be AMAZING.
    >>
    No.21359
    ▶
    Thanks for the Model.
    But can someone reupload the Lora? Cant find it anywhere anymore :(
    >>
    No.21428
    ▶
    >>21359

    Same.
    >>
    No.21431
    ▶
    >>21359
    Here is lora:
    https://gofile.io/d/e1iyVg
    >>
    No.21432
    ▶
    >>21431
    Looks like it is another lora...
    You need Big_r-18G.safetensors
    >>
    No.21449
    ▶
    >>21432
    https://gofile.io/d/LVF8tD I've got this, is it the one you want?
    >>
    No.21475
    ▶
    Yeah! That's it thanks a lot!
    >>
    No.24241
    ▶
    Are there LoRA and checkpoints for Big_R18G?
    Where can I get the full-size version?
    Or has everyone switched to CompilativeG?
    >>
    No.24255
    ▶
    Can you make an anima of lora?
    >>
    No.24281
    ▶
    I got ahead of myself when requesting >>21275
    sorry.
    >>
    No.25363
    ▶
    File: _Illustrious_20260821-192837_00001_.jpg
    (240 KB, 1824x2208) ▶
    245865
    >>
    No.28991
    ▶
    File: _Illustrious_20260928-125702_00001_.jpg
    (239 KB, 1552x2592) ▶
    245298
    >>
    No.29001
    ▶
    >>28991
    that's hot
    >>
    No.29084
    ▶
    >>13910
    >>
    No.29214
    ▶
    File: _Illustrious_20260930-133448_00001_.jpg
    (208 KB, 1552x2592) ▶
    213464
    >>
    No.29277
    ▶
    It's all 404 even on civred....
    >>
    No.29617
    ▶
    >>21449
    Can u share the model one more time?That sharing link is invalid
    >>
    No.30050
    ▶
    https://tungsten.run/model/FiKReNdeJM?model_version=EjvBxKFRfz
    This link is still available.
    >>
    No.30213
    ▶
    >>14044
    >>
    No.31377
    ▶
    Can anyone share a link to the LoRa file that's almost 2 GB and was called something like Big R-18?
    >>
    No.31496
    ▶
    Please come back.
    >>
    No.31811
    ▶
    File: ComfyUI_145968_.png
    [Animated PNG](1.34 MB, 1024x1024) ▶
    1400972
    >>31496
    Thanks for callign me back.


    What did iI miss, what happened in the last 6 months in AI field?
    >>
    No.31812
    ▶
    File: ComfyUI_145875_.png
    [Animated PNG](1.29 MB, 1024x1024) ▶
    1348453
    >>
    No.31813
    ▶
    File: ComfyUI_145541_.png
    [Animated PNG](1.25 MB, 832x1216) ▶
    1311434
    >>31812
    >>
    No.31814
    ▶
    File: ComfyUI_145526_.png
    [Animated PNG](1.44 MB, 832x1216) ▶
    1505625
    >>31813
    >>
    No.31815
    ▶
    >>31814
    I am back.

    What is happening now?

    Should I red the model for illustrious? Should I fientune anima or MOSS?

    What are people using nowdays? I have been out of loop for too long
    >>
    No.31844
    ▶
    >>31815Krea 2 Main Model
    >>
    No.31871
    ▶
    >>31844
    I would not recommend Krea2 despite the fact that it innately supports gore. The dataset cut-out is earlier than roughly 8/20/2025 (i.e. Eri from Blue Archive is not supported), and the base model requires a NSFW LoRA to properly support sex in general.
    >>
    No.31906
    ▶
    >>31815
    I'd go for MOSS. Seems to handle guro out of the box and hasn't thrown up any censor roadblocks.
    >>
    No.31910
    ▶
    File: MOSS_SFT_00127_.png
    (3.99 MB, 1408x1664) ▶
    4181636
    >>31906
    yeha playing with MOSS now.

    It is very ice. but describing complex mechanisms takes too many iterations.


    Sharing prompts is fundamental here.
    >>
    No.31911
    ▶
    File: MOSS_SFT_00128_.png
    (2.35 MB, 960x2304) ▶
    2463440
    >>
    No.31918
    ▶
    File: MOSS_SFT_00162_.png
    (2.56 MB, 1152x1846) ▶
    2685339
    >>31911
    >>
    No.31972
    ▶
    File: MOSS_SFT_00112_.png
    (2.32 MB, 960x2304) ▶
    2428924
    And I hit a wall.

    MOSS either works very well, or you need to iterate forever by changing th epormpt just a bit to make it do what you want.

    I think I will be back in 1 week.

    I have the old data, I might either train a ayaswna LoRA, or a guro fien tune.

    If there is interest
    >>
    No.32022
    ▶
    File: AnimaTest_2615.png
    [Animated PNG](821 KB, 1024x1024) ▶
    840834
    >>31972
    I use anima. Very modular, loras train good, you'll have better luck than I do with your better data. Captioning is similar, use tags and remove what isn't on the image and what isn't wanted. Training simple too. Would recommend you try, can't do guro out of the box but its very easy to add with a few images.
    >>
    No.32068
    ▶
    >>32022
    There is a specific anima based guro checkpoint posted on this board with good results. You should check it out, v20 is the most powerful version. >>18332
    >>
    No.32086
    ▶
    Hi. I'm trying to find a LoRa for “Illustrious” called “Big R-18”; it was about 1.3 GB. Does anyone have this LoRa, and could you share it with me?
    I also can't find any LoRAs for “Anima” on the same theme that actually work. I'm most into the severed limbs and heads fetish.
    >>
    No.32164
    ▶
    File: AnimaTest_2629.png
    [Animated PNG](1.18 MB, 896x1152) ▶
    1232352
    >>32068
    Yeah, its a fine finetune but the maker didn't provide a list of tags, just gave a file and said "ask grok to tell you" which is kinda eh. Would much rather have what Mrik did. Not to discredit the finetune maker, just some things could be improved. As well, I like what Mrik focuses on, I'd love to have their data set so I can the debreasting images and make a lora myself.
    >>
    No.32222
    ▶
    Mrik here.

    I have the giant illustrious LoRA and i have the big dataset i completed 1 year ago.

    I will be back to my PC this weekend, and will share them both this weekend.
    [Top]
    Delete post: []