Pretrained SD-1.5 weight for SePPO: Semi-Policy Preference Optimization for Diffusion Alignment
See Github Repo: SePPO
Paper Report: Daily Paper
Inference Code:
import os
import argparse
import numpy as np
import torch
from diffusers import StableDiffusionPipeline, UNet2DConditionModel
from PIL import Image
torch.set_grad_enabled(False)
if __name__ == "__main__":
parser = argparse.ArgumentParser(description="Generate images and calculate scores.")
parser.add_argument('--unet_checkpoint', type=str, required=True, help="Path to the UNet model checkpoint")
parser.add_argument('--prompt', type=str, required=True, help="Prompt")
args = parser.parse_args()
unet = UNet2DConditionModel.from_pretrained(args.unet_checkpoint, torch_dtype=torch.float16).to('cuda')
pipe = StableDiffusionPipeline.from_pretrained("pt-sk/stable-diffusion-1.5", torch_dtype=torch.float16)
pipe = pipe.to('cuda')
pipe.safety_checker = None
pipe.unet = unet
generator = torch.Generator(device='cuda').manual_seed(0)
gs = 7.5
ims = pipe(prompt=args.prompt, generator=generator, guidance_scale=gs).images[0]
img_path = os.path.join('SePPO', "0.png")
if isinstance(ims, np.ndarray):
ims = Image.fromarray(ims)
ims.save(img_path, format='PNG')
- Downloads last month
- 43
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.
Model tree for DwanZhang/SePPO
Base model
stable-diffusion-v1-5/stable-diffusion-v1-5