Download
1 variant available
License:
AnimaThe Anima Model is licensed by CircleStone Labs LLC. Copyright CircleStone Labs LLC. IN NO EVENT SHALL CIRCLESTONE LABS LLC BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH USE OF THIS MODEL.
Built on NVIDIA Cosmos
Hi, this is my first LoRA. If you got here and havent seen the movie? Go watch, its good. Then be sad theres 6 novels and only 1 anime adaptation. Anyway...
v1.0 - First attempt at a LoRA, not bad for not knowing much of anything.
Trigger word: suito
Strength: 1.7-1.85 (starts showing around 1.4, 2.0+ gets "cooked")
v2.0 - This one hits harder. Im pretty happy with how this turned out. Added outfit and hair triggers to help not force her uniform. Epoc 3 instead of 2 with smoother training so 3 actually improved rather than falling off.
Triggers: suito, officers uniform, brown pilot jacket, wet hair, slicked back hair
Strength: 1.25-1.50, 1.30-1.35 seems best
Note: glasses, can be used in either version to force her glasses. The training data has a good enough mix of her with and without that it lands pretty perfectly on hers with just the one tag.
I do this for fun and to learn. Like, comments and reactions are nice to see if my efforts are worth while and posting. I dont really care about the buzz or whatever but drop me a comment and like if you enjoy what Ive done. <3
Training data: Both versions use 43 screenshots from the movie. Various lighting, outfits, colour tones, and scenes. Should generally settle in the middle but has range if prompted correctly. Training done in Kohya on my own gpu. (No idea if that matters or not?)