README.md

metadata

license: apache-2.0
library_name: diffusers
pipeline_tag: text-to-image
widget:
  - text: city neighborhood
    output:
      url: 7d3aebe3-a08d-4a31-b5ac-06408e0c835a.jpeg
  - text: resort in hawaii
    output:
      url: a983aca6-9deb-41f5-8e8a-b7932cc83ff4.jpeg
  - text: factory
    output:
      url: be2fb507-af99-4258-a90f-c0df2bbab3ce.jpeg
  - text: university campus
    output:
      url: 12bbc23c-850a-484b-99a6-478c19417993.jpeg

Model Card for Model ID

This is a StableDiffusion based model that synthesizes satellite images given text prompts. The base stable diffusion model used is stable-diffusion-2-1-base (v2-1_512-ema-pruned.ckpt).

• Use it with 🧨 diffusers
• Use it with stablediffusion repository

Model Sources [optional]

• Repository: stable-diffusion

Examples

from diffusers import StableDiffusionPipeline

pipe = StableDiffusionPipeline.from_pretrained("MVRL/GeoSynth")
pipe = pipe.to("cuda:0")

image = pipe(
    "Satellite image features a city neighborhood",
).images[0]

image.save("generated_city.jpg")