Author a user prompt for impromptu generation

[pramitchoudhary]

Guided impromptu generation using Stable Diffuse

• https://github.com/CompVis/stable-diffusion

• https://github.com/huggingface/diffusers

• [ ] Basic code:

  
  import requests
  import torch
  from torch import autocast
  from PIL import Image
  from io import BytesIO
  from diffusers import StableDiffusionImg2ImgPipeline

Upload image

from PIL import Image image_input = Image.open("...").convert("RGB") init_image = image_input.resize((768, 512))

device = "cuda" model_id_or_path = "CompVis/stable-diffusion-v1-4" pipe = StableDiffusionImg2ImgPipeline.from_pretrained("../diffuser_fun/stable-diffusion-v1-4" pipe = pipe.to(device)

prompt = "oil painting" with autocast(device): images = pipe(prompt=prompt, init_image=init_image, strength=0.75, guidance_scale=7.5)["sample"]

images[0].save("oil_paiting.png")

from matplotlib.pyplot import imshow import numpy as np %matplotlib inline

pil_im = Image.open('oil_paiting.png', 'r') imshow(np.asarray(pil_im))

- [ ] With re-training to customize

[pramitchoudhary]