Judy Hopps (Qupo) New Anima version!

Trained with Anima-Standalone-Trainer
Mostly standard settings
max_train_epochs = 10
learning_rate = 0.0001

text_encoder_lr = 0.00005

optimizer_type = "AdamW8bit"
lr_scheduler = "cosine"
network_dim = 32

network_alpha = 16

network_train_unet_only = true


2590 steps
n_repeats: 7
37 image dataset
Resolution: 512x512
More data in metadata

Trigger words are not necessary, but influence style strength