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]
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.
  • 175 unique users in the last 30 minutes (including lurkers)
    ■▼
    File: Black Widow Killed (02).png
    (1.13 MB, 1024x1024) ▶
    1179706
    No.4141 [Report] AS
    ▶
    [Reply]
    Black Widow asphyxiated.
    ■▲▼
    File: ClipboardImage.png
    (1.49 MB, 1024x1024) ▶
    1562364
    No.4058 [Report] AS
    ▶
    [Reply]
    She knelt confidently in front of the camera every day, her smirk an indication of her cocky assurances that these primitive assholes wouldn't *dare.* But she was delusional: Washington wouldn't pay up for some bitchy influencer. When the bullet hit her throat, her facial expression said it all.

    More where this comes from in my dedicated thread here.
    ■▲▼
    File: 00190-2518570171-masterpiece, best qual(…).png
    (1.19 MB, 1344x896) ▶
    1247056
    No.1209 [Report] AS
    ▶
    [Reply]
    Merry Christmas to everyone!

    Let's take a look at the Christmas gifts I've prepared.

    https://drive.google.com/file/d/1JPMp_fxGJgNM-fTdz_OHvHXwB9xv2AJE/view?usp=sharing

    In short, this is an experimental work based on "trying to challenge the barbecue LORA that failed in past training" and applying some training classification methods so far. Compared with the relatively simple fixed routine type LORA in the past, it will be more troublesome. It is conceivable that there will be many problems, and the probability of ghost pictures is probably more than half. However, using img2img can actually achieve a good final effect, at least it works for me.

    recommended weight 0.8~0.9.

    I distinguish the main concepts of nbbqV3 as:
    【butchered】a female body that has just been killed, bled, and has normal skin color but has not been roasted.


    Comment too long, view post No.1209 to see the full comment.
    35 posts omitted. Click Reply to view.
    >>
    No.3910 [Report]
    ▶
    File: 00131-1153188017.png
    (1.4 MB, 1344x896) ▶
    1465938
    >>3908

    Although I'm not sure whether it counts, I have a penishead LORA, which is barely close to the male homosexual content you want to see?

    In addition to being heterosexual myself, it is also because most of guro LORA’s training materials are dominated by women. If you want to train one, I suggest you find the materials, classify them, and label them yourself. If you have any questions, I believe the people here will help guide you.
    >>
    No.3911 [Report]
    ▶
    File: 04514-958188908-masterpiece, best quali(…).png
    (1.47 MB, 1344x896) ▶
    1544177
    >>
    No.3913 [Report]
    ▶
    >>3911
    >>3910

    Lol that's gold.
    >>
    No.3919 [Report]
    ▶
    >>3909
    Do you expect women to solve easily solvable problems instead of sitting around complaining until a man fixes it for them? I know this is a niche imageboard for weirdos, but come on, you've spoken to a woman before right?
    >>
    No.3924 [Report]
    ▶
    >>3919

    >Do you expect women to solve easily solvable problems instead of sitting around complaining until a man fixes it for them?

    touché...

    >but come on, you've spoken to a woman before right?

    I have, but most of them were not from California.
    ■▲▼
    File: 119261065_p0.png
    (434 KB, 496x944) ▶
    444565
    No.3384 [Report] AS
    ▶
    [Reply]
    Are there any Lora where the dead girl is laid in a metal autopsy table in a morgue? I tried prompting it and it never comes out right
    3 posts omitted. Click Reply to view.
    >>
    No.3424 [Report]
    ▶
    File: 115669106_p0_master1200.jpg
    (582 KB, 1200x822) ▶
    596957
    I need something like a LoRA that do something like this. Pardon the male model.
    >>
    No.3469 [Report]
    ▶
    >>3424

    Not sure if this was on civit or pixai or wherever the fuck, but I remember this was supposed to be a morgue Lora for Pony when I downloaded it. Tried it again recently but I'm not getting any results. Maybe you'll have more Luck with it: https://pixeldrain.com/u/3HfwgLNL pw: gurochan ...check the meta for triggers

    Other than that, I'd suggest you consider curating a dataset from the morgue thread in /g (there are hundreds of pics) and then ask here for someone to train on that set for you.
    >>
    No.3471 [Report]
    ▶
    >>3469
    If i have to train a whole new model, how do i create a dataset out of the images in that thread?
    >>
    No.3472 [Report]
    ▶
    >>3471

    You collect those pics of the best quality that are closest representing what you're looking for. At least 20 images, but much more is probably better in your case. The pics should have (at least) 1024x1024 pixels for a good Pony training. Take a look at this post here >>3099
    >>
    No.3907 [Report]
    ▶
    >>3424
    "Pardon the male model"
    Seriously..? One tiny dick and it gets censored out, and you're apologising for it..? May as well apologise for having a dick while you're at it.. God forbid males exist..

    I been getting eye-raped by huge disgusting tits ever since I came here. But Guro is so niche that it's not like I have any other options.

    Did I miss a memo somewhere about this being a female only board..?
    ■▲▼
    File: 2023-01-17_182649.jpg
    (210 KB, 1024x1024) ▶
    215333
    No.262 [Report] AS
    ▶
    [Reply] [Last 50 posts]
    Embed - https://files.catbox.moe/i82053.zip

    Example sheet 1 - https://files.catbox.moe/hceqed.png

    Example sheet 2 - https://files.catbox.moe/q6qkjr.png

    Used pretty old model, maybe it will be better with new ones.
    90 posts omitted. Click Reply to view.
    >>
    No.2892 [Report]
    ▶
    File: pixai-1766461054357518712-3.png
    (979 KB, 768x1280) ▶
    1002930
    >>
    No.2893 [Report]
    ▶
    File: pixai-1766479041714106847-2.png
    (987 KB, 768x1280) ▶
    1011142
    >>
    No.3735 [Report]
    ▶
    File: pixai-1794422920646419309.png
    (2.16 MB, 896x2048) ▶
    2265103
    >>
    No.3736 [Report]
    ▶
    File: pixai-1794446486701738364.png
    (2.5 MB, 1200x2048) ▶
    2620716
    >>
    No.3893 [Report]
    ▶
    File: pixai-1799084330600864952-3.png
    (1.64 MB, 832x1472) ▶
    1721251
    Does anyone like to see Black Tulip hanged?
    ■▲▼
    File: Screenshot_20241005_194356_Remix.jpg
    (451 KB, 1079x1068) ▶
    462771
    No.3845 [Report] AS
    ▶
    [Reply]
    I'm gonna put out an SOS on that actually can someone tell me what I gotta put in my prompt for the knife to go IN her heart? Not beside, not in front of, not near. Like chest stab. I've used all combos I can think of anything helps
    Remix/MonAi/SeaArt
    >>
    No.3847 [Report]
    ▶
    There isn't even a generic stabbing lora yet, at least no working one that I'm aware of. So achieving that with one of these non local generators might most likely not be possible. The only two things I can think of that may help are the prompt phrase "sticking out from". Not sure what basemodels you're using but in tests with some very old sd1.5 models I've had some miniscule success with that. So perhaps try something like "sword (or knife/dagger) sticking out from her chest" or "sword sticking out from her breasts" (or try "...from between her breasts/cleavage" etc.). Also there was a "boobasword" lora for sd1.5 (there still is one for pony on civit, but I can't find the older versions there anymore) at low weight and putting stuff like "electric" and "magical" in the negative I got some potentially useable results. But it was very gacha, maybe 1 decent gen out of 30-50 or so. ...If you can use that new "illustriousXL" -model (see thread a little down below) on any of these on site generators that'll be your best bet I think.
    >>
    No.3891 [Report]
    ▶
    >>3847
    Thank you so much I appreciate it. I'll have to try that
    ■▲▼
    File: 00029-3335348420.jpg
    (115 KB, 512x1024) ▶
    118727
    No.2981 [Report] AS
    ▶
    [Reply]
    Pony and Lora.
    41 posts omitted. Click Reply to view.
    >>
    No.3833 [Report]
    ▶
    File: 00080-771245110.jpg
    (117 KB, 1024x512) ▶
    120450
    Good results with the lora posted here. I try to combine them. Toxic Ramen, Disorder and Momotaaai with either Pony or Damn Pony XL Realistic model.

    Example prompt:
    score_9_up, score_8_up, from_rule34xxx, rating_explicit,
    <lora:LORA 1:VALUE1>,
    <lora:LORA 2:VALUE2>,

    1girl,
    TRIGGER WORDS,
    guro,blood stain, splatter,
    headshot,
    stunningly beautiful woman,, long eyelashes, freckles, long wavy hair, heavy makeup,
    ripped clothing,

    Comment too long, view post No.3833 to see the full comment.
    >>
    No.3834 [Report]
    ▶
    File: 00201-1832649355.jpg
    (239 KB, 768x768) ▶
    244887
    >>
    No.3835 [Report]
    ▶
    File: 00067-3528521136.jpg
    (117 KB, 512x512) ▶
    120483
    >>
    No.3836 [Report]
    ▶
    File: 00229-3795993555.jpg
    (169 KB, 512x1024) ▶
    173485
    >>
    No.3844 [Report]
    ▶
    >>3836
    Holy shit, I love this! Your work is just getting better and better.
    ■▲▼
    File: 00111-3503608387.jpg
    (1.5 MB, 1280x2048) ▶
    1573831
    No.3841 [Report] AS
    ▶
    [Reply]
    Someone recommended me to post here,download link in comments

    https://e-hentai.org/g/2928445/7da13650d4/ [Guro warning/AI generated] Hanging preparing/ Hanged execution/绞刑 [With Lora publish/Lora发布]

    https://e-hentai.org/g/2931567/c864cc6ef5/ [Guro warning/AI generated] Guillotine execution/ 断头台 [With Lora publish/Lora发布]

    https://e-hentai.org/g/2999606/1ebaadb6a5/ [Guro warning/AI generated] Beheading/Beheaded execution/斩首 [With Lora publish/Lora发布]

    https://e-hentai.org/g/2927477/7e5a033ff8/ [Guro warning/AI generated] Spear execution/Spit Roasting girls/少女穿刺&烧烤 [With Lora publish/Lora发布]
    ■▲▼
    File: 03.jpg
    (5.1 MB, 3181x4666) ▶
    5346636
    No.3806 [Report] AS
    ▶
    [Reply]
    Learning to use ai in my art versions. here was a mix of daz studio AI and photoshop. Comments are welcome.
    >>
    No.3809 [Report]
    ▶
    >>3806
    loved your stuff for a few years now.

    Not realy a fan of your high posterization approach. sure it hides the photoshop mistakes, but imho it is not ideal.

    Maybe having an img2img with a 5-10% denoise rate as final step instead would help
    ■▲▼
    File: {6+girls}, {{{{{{pile of corpses}}}}}}(…).png
    (3.21 MB, 1920x1088) ▶
    3362106
    No.1720 [Report] AS
    ▶
    [Reply] [Last 50 posts]
    My particular niche is combat guro involving loli gurls getting sliced and gutted, which turns out to be Very niche, so I had to make my own
    80 posts omitted. Click Reply to view.
    >>
    No.3753 [Report]
    ▶
    File: {{{{{{{1girl, solo, loli, aged down, p(…).png
    (2.54 MB, 1024x1536) ▶
    2666131
    I put 'full body' in positive prompt and 'feet out of frame' in negative prompt and used a dozen bracket each, but it is still mostly getting ignored
    >>
    No.3754 [Report]
    ▶
    File: {{{{{{{1girl, solo, loli, aged down, p(…).png
    (2.05 MB, 1024x1536) ▶
    2151432
    Seriously I'm getting sick of it. NovelAI could use lots of more upgrades now.
    >>
    No.3755 [Report]
    ▶
    File: {{{{{{{1girl, solo, loli, aged down, p(…).png
    (2.35 MB, 1024x1536) ▶
    2467603
    >>
    No.3756 [Report]
    ▶
    File: {{{{{{{1girl, solo, loli, aged down, p(…).png
    (2.34 MB, 1024x1536) ▶
    2456714
    And one magical girl cut in half
    >>
    No.3757 [Report]
    ▶
    >>3753

    Maybe try "full shot", but more importantly either try "barefoot" or prompt for some specific footwear (sandals, socks, balletshoes etc.). This at least works well for me in 1.5 and in pony. No idea if it works in novelai, but doesn't hurt to try, right? Also maybe try "midshot" in the neg prompt.
    ■▲▼
    File: 00012-1879697674.png
    (1.04 MB, 800x1200) ▶
    1090651
    No.3155 [Report] AS
    ▶
    [Reply]
    https://civitai.com/models/642937/kihyahya-style
    List of tags:
    guro, blood, no arms, no legs, no head, neck stump, behead
    Organs tags:
    internal out, behalf (behalf make internal out stronger)
    Head tags:
    head only, head stump, penis from mouth (have problems), head fuck
    Brain tags:
    brain, brainfuck (sometimes brain looks too big)
    Additional tags:
    hanged (unstable), impaled (unstable), ryona, severed head, grabbing another's hair,
    Image prompt:
    score_9, score_8_up, score_7_up, score_6_up,
    <lora:Kihyahya:1.1> kihyahya, 1girl, 1boy, cowgirl, reverse outfit, blood, trembling, puffy nipples, big breast, guro, ahegao, riding dick, brain, internal out, orgasm
    19 posts omitted. Click Reply to view.
    >>
    No.3581 [Report]
    ▶
    File: 00054-2798323556.png
    (985 KB, 896x1024) ▶
    1008761
    >>
    No.3582 [Report]
    ▶
    File: 00043-995256718.png
    (1.03 MB, 896x1024) ▶
    1080199
    >>
    No.3583 [Report]
    ▶
    File: 00023-643658870.png
    (1010 KB, 896x1024) ▶
    1034319
    >>
    No.3584 [Report]
    ▶
    File: 00048-3455276054.png
    (1.91 MB, 1488x1216) ▶
    2004113
    >>
    No.3717 [Report]
    ▶
    How do you get it to generate organs? I have had no luck without a brain hat Lora to get any. And it doesn't play nice with nude
    ■▲▼
    File: 00000-34489644.png
    (1.05 MB, 1344x896) ▶
    1103608
    No.2015 [Report] AS
    ▶
    [Reply] [Last 50 posts]
    https://mega.nz/folder/GZ8DnBjZ#oc70hM9RGslozRhZDQJkwA

    According to the number of training times, it is divided into five versions: 2, 4, 6, 8, and 10. The overfitting of 8 and 10 is strong but the composition is relatively stable. However, it needs to be matched with i2i high-resolution correction to produce pictures. For those who don’t care about the beheading effect, using versions 2, 4, and 6 should be enough.

    The main difference between the open version of LORAv2c in this training is the classification of four concepts.
    60 posts omitted. Click Reply to view.
    >>
    No.3556 [Report]
    ▶
    File: 00050-2612216810.png
    (1.03 MB, 1344x896) ▶
    1084312
    >>
    No.3708 [Report]
    ▶
    File: 00065-3816587955.png
    (1.17 MB, 1344x896) ▶
    1226976
    >>
    No.3709 [Report]
    ▶
    File: 00070-3101450789.png
    (1.18 MB, 1344x896) ▶
    1232153
    >>
    No.3710 [Report]
    ▶
    File: 00068-883823397.png
    (1.19 MB, 1344x896) ▶
    1248492
    >>
    No.3711 [Report]
    ▶
    File: 00034-3353927125.png
    (1.13 MB, 1280x832) ▶
    1184209
    ■▲▼
    File: o95uzqk4d2jd1.jpeg
    (110 KB, 666x578) ▶
    113006
    No.3592 [Report] AS
    ▶
    [Reply]
    Something good AI works
    ■▲▼
    File: KS.png
    (1.65 MB, 704x1440) ▶
    1734914
    No.3566 [Report] AS
    ▶
    [Reply]
    Katana stab
    >>
    No.3575 [Report]
    ▶
    NICE.
    ■▲
    File: 00034-4151020795.png
    (784 KB, 1152x896) ▶
    803688
    No.3544 [Report] AS
    ▶
    [Reply]
    https://files.catbox.moe/og732p.safetensors
    flux can do some gore stuff already. This lora just enhanced it.No specific trigger words needed.
    2 posts omitted. Click Reply to view.
    >>
    No.3547 [Report]
    ▶
    File: 00035-945242551.png
    (1.09 MB, 896x1152) ▶
    1144050
    This image is a Disney 3d render of a woman's head placed on an ornate plate wtih short brown hair. To the right of the head, is a human torso, devoid of limbs and head, suspended by ropes from a metal bar. The torso is anatomically detailed, showing a realistic representation of human skin and muscle structure. The torso has small, perky breasts and a defined waist. the body skin is smooth and shiny, with visible marks from the cooking process.<lora:flux_g0re_v1-000006:1>
    Steps: 8, Sampler: Euler, Schedule type: Beta, CFG scale: 1, Distilled CFG Scale: 3.5, Seed: 945242551, Size: 896x1152, Model hash: f184d014ac, Model: flux-hyp8-Q8_0
    >>
    No.3548 [Report]
    ▶
    File: 00062-602269190.png
    (666 KB, 1152x896) ▶
    682630
    This is a highly detailed digital illustration, character sheet of a woman bust . The torso hip stands on a simple, white, circular base with a thin, metallic pole.
    The face is detailed with soft, rounded features.
    The background is a plain, gradient gray that contrasts with the figure.
    multiple views of the character,front view, side view, back view.<lora:flux_g0re_v1-000006:1>
    Steps: 8, Sampler: Euler, Schedule type: Beta, CFG scale: 1, Distilled CFG Scale: 3.5, Seed: 602269190, Size: 1152x896, Model hash: f184d014ac, Model: flux-hyp8-Q8_0
    >>
    No.3549 [Report]
    ▶
    This si awesome. Gonna test. it

    i made a decap LORA (even if Flux can already do decaps)

    you can dowbload it here

    https://files.catbox.moe/exsi33.safetensors

    triggers severed head.
    >>
    No.3550 [Report]
    ▶
    >>3549
    >>3549
    Here some preview pics of the decap one.

    https://files.catbox.moe/dcctmd.rar

    password guro
    >>
    No.3554 [Report]
    ▶
    >>3548
    >>3547
    >>3546
    Wow. i tried them with realistic models and it works great.


    Any other prompts you would advise?

    I ususalllly like to run the prompt fo the dataset images, with minor variations


    What else was this trained to do?
    Delete post: []
    [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37]