It is trained on Realistic Vision 5 but should work with other models too.
Check images for prompts.
Designer_Bra_V1.safetensors