File size: 2,897 Bytes
2733832
5550052
0a0bb7f
2733832
 
 
 
 
 
 
 
 
0a0bb7f
 
 
 
 
 
 
5550052
2733832
 
5550052
2733832
 
 
 
 
 
 
 
 
0a0bb7f
 
2733832
5550052
2733832
0a0bb7f
 
5550052
2733832
0a0bb7f
2733832
 
 
 
 
 
 
d93bc7b
 
 
 
 
 
2733832
 
 
 
 
 
 
 
0a0bb7f
2733832
 
 
 
 
 
 
d93bc7b
0a0bb7f
2733832
 
 
 
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
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
import gradio as gr
from openai import OpenAI
import httpx

# List of popular styles
STYLES = [
    "Photorealistic", "Oil Painting", "Watercolor", "Anime", 
    "Studio Ghibli", "Black and White", "Polaroid", "Sketch", 
    "3D Render", "Pixel Art", "Cyberpunk", "Steampunk", 
    "Art Nouveau", "Pop Art", "Minimalist"
]

# Default negative prompt
DEFAULT_NEGATIVE_PROMPT = """
ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, 
extra limbs, disfigured, deformed, body out of frame, bad anatomy, watermark, signature, 
cut off, low contrast, underexposed, overexposed, bad art, beginner, amateur, distorted face
"""

def enhance_prompt(client, prompt, style):
    enhanced_prompt_request = f"Enhance the following prompt for DALL-E 3 image generation in the style of {style}. Make it more detailed and vivid, while keeping the original intent: '{prompt}'"
    
    response = client.chat.completions.create(
        model="gpt-4",
        messages=[
            {"role": "system", "content": "You are an expert at creating detailed, vivid prompts for image generation."},
            {"role": "user", "content": enhanced_prompt_request}
        ]
    )
    
    return response.choices[0].message.content.strip()

def generate_image(api_key, prompt, style, negative_prompt):
    client = OpenAI(api_key=api_key, http_client=httpx.Client())
    
    enhanced_prompt = enhance_prompt(client, prompt, style)
    
    full_prompt = f"{enhanced_prompt}\nNegative prompt: {negative_prompt}"
    
    response = client.images.generate(
        model="dall-e-3",
        prompt=full_prompt,
        size="1024x1024",
        quality="standard",
        n=1
    )
    
    return response.data[0].url, enhanced_prompt

def process_and_generate(api_key, prompt, style, negative_prompt):
    client = OpenAI(api_key=api_key, http_client=httpx.Client())
    enhanced_prompt = enhance_prompt(client, prompt, style)
    image_url, _ = generate_image(api_key, enhanced_prompt, style, negative_prompt)
    return image_url, enhanced_prompt

with gr.Blocks() as demo:
    gr.Markdown("# DALL-E 3 Image Generator")
    
    with gr.Row():
        with gr.Column(scale=1):
            api_key = gr.Textbox(label="OpenAI API Key", type="password")
            prompt = gr.Textbox(label="Prompt")
            style = gr.Dropdown(label="Style", choices=STYLES)
            negative_prompt = gr.Textbox(label="Negative Prompt", value=DEFAULT_NEGATIVE_PROMPT)
            generate_btn = gr.Button("Generate Image")
        
        with gr.Column(scale=1):
            image_output = gr.Image(label="Generated Image")
            enhanced_prompt_output = gr.Textbox(label="Enhanced Prompt")
    
    generate_btn.click(
        process_and_generate,
        inputs=[api_key, prompt, style, negative_prompt],
        outputs=[image_output, enhanced_prompt_output]
    )

demo.launch()