Trained on 147 images of Dhole (no repeats) from Kemono Friends 3 on Anything v3.0 Pruned FP32 (however it works on every model) This time I added the tag "dolr" to the training so that it associates "dolr" with many features of her, and fixes the association with "dhole" and "kemono (in parenthesis!)" messing up the generation Of course, dhole (kemono friends) still works Stats: 147 512x images (no repeats) for 20 epochs Network dimension: 32 Base model: Anything-V3.0-pruned-fp32.ckpt Batch size of 3 on RTX 3060 Generation tags: dolr, 1girl, animal ears, tail, dog ears, dog tail, white gloves, brown hair, short hair, dog girl, brown pleated skirt, brown sleeveless shirt, white sleeveless shirt, two-tone shirt, bare shoulders, bangs, brown eyes, yellow eyes, bare arms, white boots, white fur collar, white fur bow tie, brown thighhighs, sidelocks, hair between eyes, breast pocket, fur trim Sample images are 512x512 latent upscaled to 1024x1024 and generated on Anything v3.0 Pruned FP32, addnet strength 1 |