Trained with 20 images using base "anything-v4.0-pruned-fp32" model.

Works great with models mix "0.6(linneamixv74) + 0.4(0.6(realdosmix) + 0.4(ddosmix_V2))" and lora microwaist_03Lora