Text-to-image model has been recently improved to generate the semantically rich high-quality images by strengthening natural language processing via transformer in a stable diffusion process. However ...