PixAI Membership
PixAI
PixAI
Generate
Sign in
Home
Generation
Models
Studio
NEW
Contest
Create with Mio.2
Toolbox
NEW
Guide

User LoRA

XL

Shisui Caster (FFXIV)

Create

  • AI Image Generator
  • AI Animation Generator
  • Toolbox
  • Thematic generators
  • Train LoRA
  • Mio.2 Agent
  • Studio

Discover

  • Trending Tags
  • Ranking
  • Model Market
  • Contest
  • News

About

  • Guide
  • PixAI Blog
  • Tsubaki.2
  • Meet Mio, Our Mascot
  • Content Rules

Pricing & Help

  • Membership
  • Buy Credits
  • Contact

Mobile App

  • App Store
  • Google Play
© 2026 PixAI
  • Terms of Service
  • Privacy Policy
  • Copyright Policy
  • Specified Commercial Transactions Law
Shisui Caster (FFXIV) - AI Model cover image
Uploaded at
Dec 8, 2024, 12:53 PM
Uses
237
Reviews
No reviews
Trigger Words

shisuicaster, detached sleeves, criss-cross halter, boots, toeless footwear, bandeau, black thighhighs, black gloves, black panties, hair flower , navel, bare shoulders, midriff, ribbon, jewelry,

Permissions
  • Allow image generation & sharing
  • Allow users to download this model
  • Allow commercial use of generated images

Recommended LoRAs

artstyle lock - AI Model cover image

User LoRA

artstyle lock
4
Miyuki Kujou, (Shomin Sample) Dragon Kings Special - AI Model cover image

User LoRA

Miyuki Kujou, (Shomin Sample) Dragon Kings Special
5
Girl_EV1 - AI Model cover image

User LoRA

Girl_EV1
11
22/7 Idol Member - AI Model cover image

User LoRA

22/7 Idol Member
2

Description

https://civitai.com/models/423421 Shisui Caster set from Final Fantasy XIV.

New illustrious trained version. NEW TRAINING PARAMETERS!!!! This is the best trained version so far in terms of eliminating the baked in Screen Shot style. I am still working to fine the optimal settings, but this one is encouraging that I'm on the right track with the settings. Continue to add "((cgi, 3d model,3d):1.5)" to the negative prompt to help control the screenshot style as it is still over baked. I hope to have to remove the need for it soon


PONY Version. Retrained with new keywords and training parameters, while using the old dataset. I'm still waiting for my plugins to update to be able to work on a new dataset. This new version should still have both the reduction of the screenshot look and improved compatibility with other LORA. Images generated with HiRes Fix. I also used Pony PDXL Negative Embeddings to help with the overall cleanliness on more realistic checkpoints. Adding "((cgi, 3d model,3d):1.5)" to the negative prompt will help control the screenshot look when using other LORA reducing the LORA weight will also help

the "jewelry" keyword is tied to the tassel on the back of the neck. Arm placement is a bit strict so arm prompts might generate an extra limb. I will need to clean this up once I can get a better dataset. My apologies. Enjoy and Good Luck! -Daraco

Comments