kenken999's picture
te
5528d9e
raw
history blame
369 Bytes
import torch
from diffusers import StableDiffusion3Pipeline
pipe = StableDiffusion3Pipeline.from_pretrained("stabilityai/stable-diffusion-3-medium-diffusers", torch_dtype=torch.float16)
pipe = pipe.to("cuda")
image = pipe(
"A cat holding a sign that says hello world",
negative_prompt="",
num_inference_steps=28,
guidance_scale=7.0,
).images[0]
image