Verified: 20 days ago
SafeTensor
Type | |
Stats | 85 222 |
Reviews | (23) |
Published | Mar 19, 2025 |
Base Model | |
Training | Steps: 2,925 Epochs: 12 |
Usage Tips | Clip Skip: 2 |
Trigger Words | tatsuma_daisuke_style |
Hash | AutoV2 499043B77D |
LoRA that was trained on an images of Tatsuma Daisuke. This is my first self-made LoRA for Pony so it has some very strong biases in v1. Probably going to make another version soon
19\03\2025
UPDATE:
I finally did another version! This time it's for Illustrious based models.
V1 - version for style blends - less steps, still retains style features, less biases
V2 - version for maximum style adherence - strong biases and changes style significantlly even at lower weights
V3 - middle ground between V1 and V2 for those, who want more style featured in a blend