dmillar commited on
Commit
289e7d0
1 Parent(s): 8e861af

added python snippet

Browse files
Files changed (1) hide show
  1. README.md +12 -0
README.md CHANGED
@@ -10,3 +10,15 @@ inference: true
10
  Readers of the Wall Street Journal (WSJ) are familiar with the distinctive style used to create portaits of their writers and subjects. This is a fine-tuned stable
11
  diffusion model that can be used to create hedcut-styled images using the prompt **_wsj hedcut of \<subject\>_**. This model can also be used in a DreamBooth environment
12
  to train a face or other subject for custom and unique hedcuts.
 
 
 
 
 
 
 
 
 
 
 
 
 
10
  Readers of the Wall Street Journal (WSJ) are familiar with the distinctive style used to create portaits of their writers and subjects. This is a fine-tuned stable
11
  diffusion model that can be used to create hedcut-styled images using the prompt **_wsj hedcut of \<subject\>_**. This model can also be used in a DreamBooth environment
12
  to train a face or other subject for custom and unique hedcuts.
13
+
14
+ ```python
15
+ #!pip install diffusers transformers scipy torch
16
+ from diffusers import StableDiffusionPipeline
17
+ import torch
18
+ model_id = "dmillar/wsj-hedcut-v1"
19
+ pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
20
+ pipe = pipe.to("cuda")
21
+ prompt = "wsj hedcut of a woman"
22
+ image = pipe(prompt).images[0]
23
+ image.save("./woman_hedcut.png")
24
+ ```