Download
1 variant available
370 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9
00 1 2 3 4 5 6 7 8 9
(7)
Sep 18, 2026

80 1 2 3 4 5 6 7 8 9
290 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9
1000 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9
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
Google's Gemma 4 model turned into a tiny little anime character made to do various things that shall not be described. Originated from a certain image board on the Internet. Trained on Anima Base 1.0, training params in Lora metadata.
General Character Description:
gemmachan, google hair ornament, serafuku, white sailor collar, suspender skirt, blue beret, hair intakes, blue hair, long hair, blue eyes, blue bowtie, pleated skirt, randoseruNotes:
"google hair ornament" is what gives her the Google logo on her hair. It (usually) doesn't show up if unprompted, but it can sometimes leak onto other stuff. It can be moved around or put on other objects though if prompted, for example "She is holding the Google hair ornament"
If you use randoseru, sometimes it'll randomly put weird designs on the keychain. Don't ask me why.
You can use "aqua hair" instead if her hair sometimes comes out dark blue. Same goes for her outfit overall if the colors come out wrong, just put in "navy blue".
Many of the images were in the same style and I believe all of them originally came from ChatGPT gens. I'd say the Lora still came out fine (I excluded quite a few bad gens from the dataset) but it's about what you would expect from training on AI datasets.

