Type | |
Stats | 184 2,924 |
Reviews | (43) |
Published | Feb 16, 2025 |
Base Model | |
Training | Steps: 702 Epochs: 18 |
Usage Tips | Clip Skip: 2 |
Trigger Words | COLIG_ST |
Hash | AutoV2 2814E62669 |
Hello welcome!
Here I leave a small introduction of the LoRA:
I've been playing with image-based styles (with no tags other than the trigger).
Many of them didn't work and I retrained them until they did, anyway, here are the best of those experiments.
PS: I will upload them as I generate some examples.
PS2: I only made them to generate on the site, if you like I can reduce their size since I still have the datasets. ~ XD
PMN_STL:
I was deleting some images and chose some to train, it has many images of Pomni (I never made a lora for her and Pony and Illustrious already have her integrated design).
INKTA_ST | IL V1:
Inkta is trained to generate a new version in Tooncity, but the style is good and works with other Loras so I hope it helps you, I like how it works with impasto styles.
COLIG_ST | IL V1:
Colig is a difficult style to use, it tends to turn boys/girls into Frieza and I will see if I improve it for later but it is Inhuman's sister Lora.
InHUMN_ST | IL V1:
This style is based on dragon ball villains, men without eyebrows and alien girls, yup... I was playing with this a little.
NEWTRVL_ST | IL V1:
NEWTRVL is a lora based on the AGGA_ST042 data set (This is a style I never published)...