Text-to-Image
Diffusers
Safetensors
text-generation-inference
DwanZhang commited on
Commit
2708f7e
1 Parent(s): 78de8e7

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +38 -1
README.md CHANGED
@@ -12,4 +12,41 @@ tags:
12
 
13
  Pretrained SD-1.5 weight for SePPO
14
 
15
- See Github Repo:[SePPO](https://github.com/DwanZhang-AI/SePPO/tree/main)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
12
 
13
  Pretrained SD-1.5 weight for SePPO
14
 
15
+ See Github Repo:[SePPO](https://github.com/DwanZhang-AI/SePPO/tree/main)
16
+
17
+ Inference Code:
18
+
19
+ ```
20
+ import os
21
+ import argparse
22
+ import numpy as np
23
+ import torch
24
+ from diffusers import StableDiffusionPipeline, UNet2DConditionModel
25
+ from PIL import Image
26
+
27
+ torch.set_grad_enabled(False)
28
+
29
+ if __name__ == "__main__":
30
+ parser = argparse.ArgumentParser(description="Generate images and calculate scores.")
31
+ parser.add_argument('--unet_checkpoint', type=str, required=True, help="Path to the UNet model checkpoint")
32
+ parser.add_argument('--prompt', type=str, required=True, help="Prompt")
33
+
34
+ args = parser.parse_args()
35
+
36
+ unet = UNet2DConditionModel.from_pretrained(args.unet_checkpoint, torch_dtype=torch.float16).to('cuda')
37
+
38
+ pipe = StableDiffusionPipeline.from_pretrained("pt-sk/stable-diffusion-1.5", torch_dtype=torch.float16)
39
+
40
+ pipe = pipe.to('cuda')
41
+ pipe.safety_checker = None
42
+ pipe.unet = unet
43
+ generator = torch.Generator(device='cuda').manual_seed(0)
44
+ gs = 7.5
45
+
46
+ ims = pipe(prompt=args.prompt, generator=generator, guidance_scale=gs).images[0]
47
+ img_path = os.path.join('SePPO', "0.png")
48
+
49
+ if isinstance(ims, np.ndarray):
50
+ ims = Image.fromarray(ims)
51
+ ims.save(img_path, format='PNG')
52
+ ```