Type | |
Stats | 401 1,103 |
Reviews | (61) |
Published | Apr 16, 2024 |
Base Model | |
Training | Epochs: 10 |
Usage Tips | Clip Skip: 2 Strength: 1 |
Trigger Words | wood wooden wooden clothes wooden dress, wooden clothes wooden armor, wooden clothes wooden skin, colored skin (joints:0.7) |
Hash | AutoV2 34503D5835 |
"Wood or wood not?"
This model is part of my [Materials Series]! Materials Series is a collection of LoRAs for Pony XL that lets you represent a certain type of material or surface in your images.
Explore woodworking with this LoRA. Trained high quality images of wood and wooden objects, this LoRA has been created so that your Pony XL model is able to reproduce the texture of worked and varnished wood. It's still in an experimental phase, so don't hesitate to leave a comment!
Recommended parameters
Checkpoint: Pony XL v6, AutismMix, js2Prony, etc.
LoRA strength:
0.4 - 1.0 for base Pony.
0.5 - 1.2 for derivatives (AutismMix, js2Prony...).
Trigger keywords:
This model should work well without any trigger word, but you can optionally add some guidance using those words:
wood, wooden
wooden clothes
wooden skin, colored skin
wooden dress
wooden armor
joints
leaves
tree
brown ...
...
Tips
Make sure you adjust the LoRA strength according to your expectations. A high value can sometimes lead to weird results.
You may need to modify the strength when using multiple LoRAs, especially for characters. LoRAs for characters are generally very strong, so it's best to reduce the strength of the LoRAs and increase the strength of the Wood Work Style LoRA, depending on the other. Please refer to the image of Zelda or Shrek on the post above for the example.