Text Encoder Stable Diffusion does not just generate an image, ... If each step processed in the latent space is output as an image by the decoder one by one, it looks like this.