Each character must replicate exactly the same facial expression from the input images Stable Diffusion prompts

no results