from diffusers import StableDiffusionPipeline import torch

model_id = "runwayml/stable-diffusion-v1-5" pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16, revision="fp16") pipe = pipe.to("cuda")

prompt = "a photo of an astronaut riding a horse on mars" image = pipe(prompt).images[0]

image.save("astronaut_rides_horse.png")