Type | |
Stats | 2,436 |
Reviews | (463) |
Published | Apr 30, 2023 |
Base Model | |
Training | Steps: 3,040 Epochs: 16 |
Hash | AutoV2 C9D2AA8C76 |
Created with a dataset of around 380 images. I only made this model for fun. Please support the original artist. https://twitter.com/Anmi_
Recommended Lora Block Weight:
<lyco:anmi-ver3:1:1:lbw=0.5,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.1,0.1,0.2,0.3,0.2,0,0.7,0.7,0.7,0.6,0.5,0.5,0.7,0.7,0.7,0.7,0.5,0.8>
I got great results with a weight of 0.9
, 1.0
is also possible and more closely resembles the art-style but starts to change things up with the image.
This model does not have any trigger words but you can try some of these keywords to get better results:
long hair, looking at viewer, blush, 1girl, solo, medium breasts, bare shoulders, dress, bow, cleavage, skirt, full body, thighhighs, ribbon
I mainly used abyssorangemix3 for generating the images which seemed to give the best results, but I also tested Hassaku, Counterfeit V2.5 and Cetus Mix which all seem to give good to meh results.