Dedicated to the retired model あずまひかり (azuma hikari).
Recommended setting:
Lora weight: 0.75 - 0.80
Sampler: DPM++ SDE Karras or Euler a
CFG: 7-8
Steps: 20-25
Description
Fixing prompt in model and retrain with different settings
Details
Downloads
285
Platform
CivitAI
Platform Status
Available
Created
6/12/2023
Updated
11/12/2025
Deleted
-
Trigger Words:
hikari
Files
Hikari-v2.1.safetensors
Mirrors
CivitAI (1 mirrors)