How does Stable Diffusion generate images from text prompts?

15 Eyl 06:45

I’ve read that Stable Diffusion is a latent diffusion model that turns text prompts into images by iteratively denoising latent representations. Can someone break down the main components—like the encoder, UNet, and decoder—and explain how the diffusion process works in practice? Also, how does the model balance fidelity and creativity during generation? Curious about the underlying mechanics.