Original prompt as shared from the source.
Make craiyon.com screenshot of an image of “prompt here” generated by an early 2022 early diffusion model, like stylegan, dall e mini, where ai was older, less advanced and weirder, more blurry, less knowledge and less coherent. Bad at faces. Never makes readable text. 256x256 output(s). Muddy, melted, incoherent, oversmoothed, artifacted, blown-out, uncanny, low-fidelity, lower num_steps, smaller dataset.
