MOSS-Image guide

MOSS-Image prompts: the 17 official examples

The example prompts OpenMOSS ships with its demo, each with the image it produced, the seed and every setting. They are long: the model was trained on detailed captions, and these show what that buys you. Click one for the full prompt and, for JSON prompts, its boxes drawn over the image.

More official outputs

From the project's showcase. No prompts were published for these.

Images by the OpenMOSS team, from the MOSS-Image repository (Apache 2.0). Three examples from the set are left out here: one shows a real politician and two are suggestive.