hf-internal-testing/diffusers-images-numpy

1#!/usr/bin/env python32import torch3from diffusers import StableDiffusionPipeline4 5torch_device = "cuda"6 7model_id = "CompVis/stable-diffusion-v1-4"8pipe = StableDiffusionPipeline.from_pretrained(9    model_id,10    use_auth_token=True,11)12pipe.to(torch_device)13 14prompt = "A fantasy landscape, trending on artstation"15generator = torch.Generator(device=torch_device).manual_seed(0)16output = pipe(prompt=prompt, guidance_scale=7.5, generator=generator, output_type="np")17 18import ipdb; ipdb.set_trace()19