Type | |
Stats | 136 1,797 |
Reviews | (31) |
Published | Nov 4, 2024 |
Base Model | |
Training | Steps: 900 Epochs: 150 |
Usage Tips | Clip Skip: 2 Strength: 0.8 |
Hash | AutoV2 25EFCBDD1F |
This model is designed for use with its embeddings, as it was trained using OneTrainer with pivotal tuning.
Notes
Character and outfit embeddings can be used together for the total look, or one at a time to combine them with different themes!
If you use ComfyUI, check out my node suite as it features a node that simplifies the way LoRAs trained with pivotal tuning are loaded.
The ComfyUI workflow for sample generation is embedded in the images, so feel free to check the settings/nodes if needed.
It works best at a strength of 0.8. A setting of 1.0 can be too strong for some models, though others may handle it just fine.
There's something off with the hands, probably caused by swords gripped in training data,,, Not sure.
The Wolf's head medallion is definetely off. :(
Use solo, old, 1boy/1girl to guide the model (positive or negative depending on what you need).