This is my first lora 👉🏻👈🏻
ill_V2:
Dataset: 20
ill_V1:
Resolution: 1024,1024
Checkpoint: Illustrious-XL v0.1
obsession_V1:
Dataset: 39
Checkpoint: Obsession (Illustrious-XL) v3.0
basePony_V2:
Dataset: 50
Resolution: 1600,1600
basePony_V1:
Old high-quality dataset (50 arts)
Checkpoint: Pony Diffusion V6 XL.
indigo_V5:
New dataset (50 arts).
indigo_V4:
New caption model.
indigo_V3:
Minor improvements.
indigo_V3_standart:
Trying to make lora with standard settings.
indigo_V2:
Resolution: 1024,1600
indigo_V1:
Resolution: 1024,1024
Checkpoint: Indigo Furry Mix XL 3.0
Trained on 100 arts.
Description
New dataset.
FAQ
Comments (6)
Idk why, but in civitai online generation this lora works bad
Assuming that the model was trained on Indigo and not base pony, if that is true then it'll basically only work with Indigo and things merged with Indigo, not anything base pony as is the better practice. I genned 3 images using base pony and they're absolutely unusable.
@Kaladae Hello, thank you for your comment! In one of the next versions I will try to use base pony!
@Moderator347 If you're really going for a specific look, you may consider making different versions, such as what people do with SD1.5/Pony/Ani/etc. Maybe have one for base pony, one for Autism/Indigo/whatever else. If you feel like the work anyways. Or rather, the time, lol.
@Kaladae thank you, after your comment yesterday, I decided to make several versions on different checkpoints for my lora! But there will still be no version for SD1.5.
@Moderator347 Awesome, won't bother me too much about SD1.5 as I really only ever use pony anymore. Glad to help lol. Edit I just noticed they were all for different versions of indigo. Not sure that actually makes that much of a difference... I meant more for different checkpoints entirely - but you may have also discovered something, no clue.