CivArchive
    Artist Style - Hell and Heaven - LoRA - SD1.5 / SDXL / Pony / Illustrious - SD1.5 v1.0
    NSFW
    Preview 426399
    Preview 426411
    Preview 426403
    Preview 426408
    Preview 426401
    Preview 426404
    Preview 426410
    Preview 426405
    Preview 426409

    Artist Style - Hell and Heaven - LoRA

    My Patreon - For Commissions


    Update: 04.11.2025 - Illustrious v3.0

    • More Flexible

    • Better Style

    • New Train Script

    • Support High Resolution


    Illustrious v3.0

    • Trigger: sysdeep_hah

    • Trigger need: Yes

    • Base Model: Illustrious-XL

    • Generate Info: Dont use Base Model as Generation use Custom Models

    • Image Source: Screencaps

    • Train with NSFW Images: YES

    • Samples: 832x1216 + HRF

    • Sample Model: animemix 7.5

    For Stable Diffusion WebUi (v3)

    Under Downloads you find a Zip file (Training Data)- With Cover Image + LoRA Info (Trigger Word backed in)


    Illustrious v2.0

    Trigger: sysdeep_hah

    Train Image Size: 1024x1024 , 1216x832, 832x1216

    SD Version: SDXL

    Clip Skip: 1

    Trained on: Illustrious-XL

    Illustrious v1.0

    Trigger: sysdeep_hah

    Train Image Size: 1024x1024 , 1216x832, 832x1216

    SD Version: SDXL

    Clip Skip: 1

    Trained on: Illustrious-XL

    SDXL 1.0

    Trigger: sysdeep_hah

    Train Image Size: 1024x1024 , 1216x832, 832x1216

    SD Version: SDXL

    Clip Skip: 1

    Trained on: Animagine XL V3

    Pony Diffusion [SDXL]

    Trigger: sysdeep_hah

    Train Image Size: 1024x1024 , 1216x832, 832x1216

    SD Version: Pony [SDXL]

    Clip Skip: 2

    Trained on: Pony Diffusion V6 XL

    SD1.5 2.0:

    Trigger: hellandheaven

    Train on: 768x768

    Trainstype: Style and Concepts of the Artist

    Hard to use? Yes change weight and gen more Images then you get good results

    In the Gallery you find Grids of difference weights: 0.0 - 0.4 - 0.6 - 0.8 and 1.0

    SD1.5 Preview Images Info:

    Sampling: DPM++ SDE Karras

    Steps: 20

    LoRA strenght: 0.6

    Image Size: 480x720

    CFG: 7

    Hires. fix: yes

    Upscale by: 2

    Upscaler: R-ESRGAN 4x+ Anime6B

    Denoising: 0.6

    Description

    FAQ

    Comments (4)

    AIShitPosterOct 2, 2023
    CivitAI

    This one needs an update as it barely functions anymore

    SysDeep
    Author
    Oct 2, 2023

    Starting News version

    SysDeep
    Author
    Oct 4, 2023

    @AIShitPoster check v2 :)

    AIShitPosterOct 5, 2023

    @SysDeep  Nice, Hey got a question how do you train loras?
    Ive been using the tool on Pixai.art but it doesnt let me specify what each picture is or allow me to run 1000 steps on each pic, and it only lets me train a lora from 100 pictures, it works well for what it is as i have my own loras for varuious styles and characters but it doesnt let me specify each detail.

    I want to create a lora where its a pack of female pubic (pubic, anal, bellybutton trail and armpit) hair of different colors, as well as a lora for breast milk and large erect nipples,

    Seems to be a lack of effective loras for these

    LORA
    SD 1.5

    Details

    Downloads
    1,208
    Platform
    CivitAI
    Platform Status
    Available
    Created
    4/6/2023
    Updated
    5/16/2026
    Deleted
    -
    Trigger Words:
    hell_and_heaven

    Files

    hell_and_heaven_v1.safetensors

    Mirrors

    HuggingFace (1 mirrors)