Please support the artist if possible, they are an OG since almost a decadeTrigger Word: puripurijet Trigger Words based on yearMid 2024 - Present (special request): jetpuri_newest 2023 - 2024: jetpuri_recent 2022-2023: jetpuri_modern 2021-2022: jetpuri_old 2021 &amp; before (i didn't create a huge dataset for this): jetpuri_oldest Trained on: Illustrious V2Tested on: Illustrious V2 Recommended Resolutions: 1344x1728 (best), 1216x1792, 896x1152, 832x1216, 1728x1344, 1152x896 You can use Hi-Res Fix 4x UltraSharp with 0.3-0.5 denoising (Enhances an already great image) Note: The model was trained on 1.5M resolution(1536px), and will perform better on 1.5M. Recommended Strength: 0.95 All "Parody" Characters are appropriately tagged. Please check examples All custom outfits are appropriately tagged. Please check examples Generation Recommendations: Sampler: Res Multistep Ancestral CFG++ Scheduler: SGM Uniform CFG Scale: 2.5-4 (I used 2.8) Resolution: 1344x1728 add "uncensored" tag if needed add "bar censor" to negative if needed add "greyscale" to negative to make the image extra sharp Another epoch(96) is available on hugging face if you want to try. Its extremely sharp but a bit overtrained: <a target="_blank" rel="ugc" href="https://huggingface.co/starplatinump/jetpuri-0.1-illxl/tree/main?not-for-all-audiences=true">https://huggingface.co/starplatinump/jetpuri-0.1-illxl/tree/main?not-for-all-audiences=true</a> Few rants: All these mixes and finetunes like NTR MIX, Hassaku, etc train on older illustrious versions. These guys don't know the power of 1536px resolution. Hence why I will keep using base ill 2.0. My brain was filled with nicotine while making this LoRa.

[プリプリＪＥＴ] PuriPuri JET (jetpuri) - Style

PuripuriJETJetpuri_01.safetensors

Please support the artist if possible, they are an OG since almost a decade

Trigger Word: puripurijet
Trigger Words based on year

Mid 2024 - Present (special request): jetpuri_newest
2023 - 2024: jetpuri_recent
2022-2023: jetpuri_modern
2021-2022: jetpuri_old
2021 & before (i didn't create a huge dataset for this): jetpuri_oldest

Trained on: Illustrious V2

Tested on: Illustrious V2

Recommended Resolutions: 1344x1728 (best), 1216x1792, 896x1152, 832x1216, 1728x1344, 1152x896


You can use Hi-Res Fix 4x UltraSharp with 0.3-0.5 denoising (Enhances an already great image)

Note: The model was trained on 1.5M resolution(1536px), and will perform better on 1.5M.

Recommended Strength: 0.95

All "Parody" Characters are appropriately tagged. Please check examples
All custom outfits are appropriately tagged. Please check examples

Generation Recommendations:
Sampler: Res Multistep Ancestral CFG++
Scheduler: SGM Uniform
CFG Scale: 2.5-4 (I used 2.8)
Resolution: 1344x1728

add "uncensored" tag if needed
add "bar censor" to negative if needed
add "greyscale" to negative to make the image extra sharp

Another epoch(96) is available on hugging face if you want to try. Its extremely sharp but a bit overtrained: https://huggingface.co/starplatinump/jetpuri-0.1-illxl/tree/main?not-for-all-audiences=true

Few rants:
All these mixes and finetunes like NTR MIX, Hassaku, etc train on older illustrious versions. These guys don't know the power of 1536px resolution. Hence why I will keep using base ill 2.0.

My brain was filled with nicotine while making this LoRa.

Files with SHA256: e961803cbc617988320beb283a30e2890450102ae1593af8c940c1f1a1b5d172