Description: Agentsimon, a site looking at content generated with Stable diffusion models
exploration (840) umbrella (273) stable diffusion (170) criticism (152) computer art (123) world map (65) generated art (4) fake photo (2) agentsimon (1) simon hyde (1)
Looking at what the machine model Stable Diffusion 2 does with prompts you can see some of the generated images of are rendered bear no relation to reality. Take a look at the Caribbean street scene images and you'll see the architecture is similar. In fact, take a look at Montserrat's capital city, Plymouth and instead of being a city hit by volcanic dust the generator made a picturesque something else.
Also take note of the artifacts created by the model, in particular umbrellas , hands and legs . It also does some pretty strange faces as well.
Stable Diffusion is a deep learning , text-to-image model released in 2022. It is primarily used to generate detailed images conditioned on text descriptions, though it can also be applied to other tasks such as inpainting , outpainting, and generating image-to-image translations guided by a text prompt . [3]