Try this prompt for the first.
Tested on PonyV6TurboDPO. 800x1200, Euler_a, CFG=2, Steps=8.
score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, rating_explicit, corpsexl_injury_art, neck_cut, 1girl, solo, a young woman wearing white shirt and black shorts is lying down on the ground, spreading arms and legs, blood, bleeding, wounds, splatter, bruises, full body view, upright view, <lora:corpsexl_injury_art_v1Pony_LoCon:1>
The grid images shown are by this prompt and without inpainting.

## Detailed Prompting Guide ##
score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, rating_explicit: Wisdom spell for Pony
1girl, 1boy, solo, couple, hetero, etc.: Adding these improve the image. Start from "1girl, solo".
corpsexl_injury_art: Trigger word
headshot: Increasing chance to spill blood around the head
destroyed_head: Increasing chance to spill brain matters. Good with "no_head".
no_head: Increasing chance for the head not shown. It only works with high weights.
beheaded: Increasing chance for the head getting seperated from body (not very high chance)
neck_cut: Increasing chance for the neck getting cut. Adding "solo focus, knife" is very good.
damaged_torso: Increaasing chance for the body getting blooded.
slashed_body: Increasing chance for the body getting slashed (not works well)
gib: Increasing chance to splatter internal organs. It works well with "damaged_torso".
amputee_arms: Increasing chance for the arms getting separated from the body.
amputee_legs: Increasing chance for the legs getting separated from the body.
torn cloths: Increasing chance for the cloths getting damaged or stripped.
blood, bleeing, wounds, splatter, bruises: Increasing the amount of blood. Highly recommended.
peeing self, wet herself, yellow pee, pee puddle: Add this if you like incontinence.
large breasts: Add this if you like large tits.
vaginal, creampie, penis: Add this for sex. Use together with "solo focus".
full body view: Zoom out
portrait: Zoom in
upright view OR view from left OR view from right: Camera position. Start from "upright view".
lying left OR lying right OR upside down: The angle of the body lying. Use these with the landscape resolution.
white background, gray background: Add this to negative. Otherwise the background is relatively simple.
grayscale: Add this to positive or negative. It only changes the art style.

Some examples.
Here is the simple prompt for corpse fucking. 800x1200.
score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, rating_explicit, corpsexl_injury_art, neck_cut, vaginal, creampie, penis, 1girl, 1boy, solo focus, a young female athlete wearing white sports bra and white panties is lying down on a dirt, and getting fucked a man, panties aside, blonde hair, ponytail hair, blood, bleeding, wounds, splatter, bruises, full body view, upright view <lora:corpsexl_injury_art_v1Pony_LoCon:1.5>

Another example.
Here is the simple prompt for stabbing neck cut. 800x1200.
Positive: score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, rating_explicit, corpsexl_injury_art, neck_cut, knife, 1girl, 1boy, solo focus, a schoolgirl wearing white blouse with blue ribbon, indigo pleated skirt, white socks, and black loafers, is kneeling down on a dirt, and getting beheaded from back by a man, black hair, long hair, blood, bleeding, wounds, splatter, bruises, full body view, view from left, <lora:corpsexl_injury_art_v1Pony_LoCon:1.3>
Negative: no_head, damaged_torso, gib

Extra example.
Dissembered body. 1200x800 (landscape ratio).
Positive: score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, rating_explicit, corpsexl_injury_art, damaged_body, gib, amputee_arms, amputee_legs, torn cloths, 1girl, solo, a young woman wearing white dress is lying down on a dirt, spreading arms and legs wide, brown hair, short hair, blood, bleeding, wounds, splatter, bruises, full body view, lying left, <lora:corpsexl_injury_art_v1Pony_LoCon:1.6>
Negative: no_head

I have not explored the full potential of the LoRA yet.
Posting your image here is welcome. Enjoy!

Amazing work. I'll be testing and using this for sure. Thank you for your hard work

In case anyone interested on the "realistic" guro, you may find these LoRAs useful.
I don't post the result here because it is prohibited.
They work with "realistic" style or variant of Pony Diffusion.
They are trained by actual gore photo in the shithole of the Internet...

corpsexl_injury_v4Pony-16
https://mega.nz/file/AJRGmLDK#1tXs2SF_2_Ha_78SucLwGCNDiztZ5bawYGh2-C_Ncrs
Realistic head injury LoRA. Recommended weight: 1.0-2.0.
Suggested to add blood, bleeding, wounds, splatter, to the prompt.

corpsexl_rotten_v4Pony-12
https://mega.nz/file/ocgiUYKZ#HG5CPG430Z276StSttfu0uycV51tDooGnhV2DpOXwd8
Realistic rotten corpse LoRA. Recommended weight: 1.0-2.0.
Suggested to add dirty, nasty, putrid to the prompt.

These LoRAs are also very powerful. Strictly use only for your personal joy. Do not use it to harm anyone. Do not profit from these. Thanks.

Yo this is amazing, the new loras are indeed quite strong but the results are beautiful. One question tho, when they're combined with style or character loras the face features often become very blurry (without them the aesthetic is quite boring), any suggestion?

>>2529
Which realistic pony model do you use?


Or just the default pony model?

>>2544
I tried Pony base model, Pony Everclear and Pony Realism, tho I'd say base model has the best results.

>>2542
Unfortunately, it basically does not mix well with other LoRAs. But there is some way to improve the result.
By my LoRA themselves, the images shift to realistic already.
Adding them to a fully realistic PONY checkpoints sometimes causes an over-saturation.
I get the best result with base Pony checkpoint + My LoRA + Realistic style LoRA with weight around 0.5.
I usually use 2.5DRealisticV2.safetensors for this purpose.It is on CivitAI.

>>2550
Thanks, well the problem for me is not the results are not realistic, they are, well done on that aspect. The problem is the lora kinda turns every style into same aesthetic, and every character's face a bit blurry. It will be great if these two great concepts can be more flexible. But all in all, good loras.

Nice work. Have you planned to do other loras? Maybe we can have a proper beheading/severed head lora with clear neck stump?

>>2550
Thanks. I have plan to upgrade the lora in the future.

>>2552
I am collecting the beheading images for the training too. But right now there are only few images related to ISIS and Mexican mafia. If anyone can help collecting enough amount of images, it is much appreciated. I am also feeling that I should include some CG beheading images.

>>2258
Hm, nice lora but I was unable to make it show brain whatever way I played with tags. destroyed_head, no_head, adding description or removing it, trying simple and complicated prompts. Needed something like quarter of scull missing and brain visible through the hole. I guess I can achieve the effect with inpaint but was hopink to get something similar to first picture.

Just a better picture than one where I was talking about the brain, more loras used.

>>2841
Apparently my LoRA is not working on your image. Check your environment.

Thank you for sharing, the effect of lora is really great！

Still experimental but Flux version of realistic head injury LoRA is here.
Keyword: corpse_injury
Weight: 1.1
Add "large wounds at the head" may help a bit.
The effect of the LoRA is relatively weak. Success rate is 10-20%.
https://mega.nz/file/8VwUSCRK#_bdWKJkCyBLv8DZdlWShJc7rN7ZcskSgjnFKipC4kj0

>>3337
I love that we are making FLux LoRAs now.


Can you link a folder with some example images.

I'm working on a baheading model. will sahre when the results are bettetr.

Progress so far
https://files.catbox.moe/dpunsv.png
https://files.catbox.moe/cty2r0.png
https://files.catbox.moe/96s4rw.png

WIP.
pixiv.net/en/artworks/122130332

>>3422
hot if true.

how is progress?

>>3450
Not bad.
Improving anatomy is the issue right now.

>>3605
Dangit, what prompts, or loras if not listed here? I can't get stabbin to work for shit.

Snuffwan LoRA for Wan 2.1 T2V 14b
https://mega.nz/file/FQwk0LZI#VBPcNbYbZC_iG_fu8pmE-snPzOynAFkX3c94sWmfxLY

Prompt examples
https://www.pixiv.net/en/artworks/129334005

Sadly, LoRA for I2V is a mess.
It is not coming soon.

will you make an illustrious version of this?

>>8021
Don't expect anything perfect, but you can try this. It should work for any illustrious variant.
https://mega.nz/file/VR5niRyA#886u_gzhUil-DgP6sM5Kmhf0brplwrTOYwprDX3WmWg

>>8060
have you managed to generate any good pics?