File size: 652 Bytes
b38a82c
 
d37c147
 
b38a82c
03b5a35
b38a82c
d37c147
 
03b5a35
d37c147
 
03b5a35
 
b38a82c
 
 
 
 
03b5a35
 
b38a82c
 
 
03b5a35
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
import gradio as gr
from PIL import Image
import torch
from diffusers import StableDiffusionUpscalePipeline

# UltraSharpV2
model_id = "Kim2091/UltraSharpV2"
device = "cuda" if torch.cuda.is_available() else "cpu"

pipe = StableDiffusionUpscalePipeline.from_pretrained(model_id)
pipe = pipe.to(device)

def enhance_image(image: Image.Image) -> Image.Image:
    return pipe(image).images[0]

demo = gr.Interface(
    fn=enhance_image,
    inputs=gr.Image(type="pil"),
    outputs=gr.Image(type="pil"),
    title="UltraSharpV2 Enhancer",
    description="Enhances and upscales images (faces, objects, text).",
    allow_flagging="never"
)

demo.launch()