Verified: 9 months ago
SafeTensor
Type | |
Stats | 126 0 |
Reviews | (9) |
Published | Aug 18, 2024 |
Base Model | |
Training | Steps: 1,103 Epochs: 21 |
Usage Tips | Clip Skip: 2 |
Trigger Words | lioness Fulllionessbody |
Hash | AutoV2 42009A14D3 |
Created on Civitai
Lora trained on 30 images of lionesses. I create this lora cuz i couldn't find lioness for SDXL. I am not a specialist but i think lora works fine in my opinion however, the shots from the back are unfortunately buggy. If you want angry lioness face "Fangs" works best.
I noticed that some sampling methods have artifacts. A good solution its to increase the number of sampling steps or change sampling method for a more detailed one.