File size: 1,919 Bytes
987344a
 
 
 
4f85ac9
fbba7f0
 
987344a
fbba7f0
 
 
 
 
 
 
 
 
 
 
98eb761
1766620
98eb761
987344a
 
4f85ac9
987344a
 
98eb761
987344a
 
 
 
 
4f85ac9
 
987344a
 
 
 
4f85ac9
 
0f6c29c
987344a
 
0f6c29c
 
 
 
 
4f85ac9
987344a
4f85ac9
987344a
 
 
4f85ac9
 
987344a
 
 
e5ef5c7
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
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
import gradio as gr
from diffusers import StableDiffusionXLPipeline
import torch

# Cache model between generations
pipe = None

def load_model():
    global pipe
    if pipe is None:
        model_id = "stabilityai/stable-diffusion-xl-base-1.0"
        pipe = StableDiffusionXLPipeline.from_pretrained(
            model_id,
            torch_dtype=torch.float16,
            variant="fp16",
            use_safetensors=True
        )
        if torch.cuda.is_available():
            pipe.to("cuda")
            # Enable memory optimizations
            pipe.enable_attention_slicing()
            pipe.enable_model_cpu_offload()
    return pipe

def generate_image(prompt, negative_prompt, steps=30, guidance_scale=7.5):
    pipe = load_model()
    
    # Generate image with reduced resolution for stability
    image = pipe(
        prompt=prompt,
        negative_prompt=negative_prompt,
        num_inference_steps=int(steps),
        guidance_scale=guidance_scale,
        width=1024,
        height=1024,
    ).images[0]
    
    return image

# Optimized Gradio interface
with gr.Blocks() as demo:
    gr.Markdown("# 🖼️ SDXL Image Generator (Stable)")
    with gr.Row():
        with gr.Column():
            prompt = gr.Textbox(label="Prompt", placeholder="Describe your image...", max_lines=2)
            negative_prompt = gr.Textbox(label="Negative Prompt", value="low quality, blurry", max_lines=2)
            with gr.Row():
                steps = gr.Slider(10, 50, value=30, label="Steps")
                guidance = gr.Slider(1.0, 15.0, value=7.5, label="Guidance Scale")
            submit = gr.Button("Generate", variant="primary")
        with gr.Column():
            output = gr.Image(label="Result", height=512)

    submit.click(
        generate_image,
        inputs=[prompt, negative_prompt, steps, guidance],
        outputs=output
    )

if __name__ == "__main__":
    demo.launch()