File size: 5,194 Bytes
781bdf4
b9f5366
 
 
b300a4b
b9f5366
781bdf4
c7ce508
d1185e1
781bdf4
b9f5366
 
 
 
6b3d712
 
 
 
 
 
 
d1185e1
6b3d712
c7ce508
44143c1
7dcf583
d1185e1
 
c7ce508
 
 
 
94cb0ca
6b3d712
94cb0ca
6b3d712
 
 
 
 
 
 
 
 
b9f5366
c7ce508
d1185e1
b9f5366
6b3d712
c7ce508
6b3d712
c7ce508
b9f5366
c7ce508
 
 
b9f5366
 
 
 
 
 
 
6b3d712
b9f5366
 
 
 
 
 
 
 
 
 
6b3d712
b9f5366
6b3d712
b9f5366
 
 
 
 
 
 
 
 
6b3d712
b9f5366
 
6b3d712
 
b9f5366
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6b3d712
b9f5366
 
 
 
d1185e1
b9f5366
d1185e1
b9f5366
 
 
 
 
94cb0ca
b9f5366
d1185e1
b9f5366
 
 
6b3d712
 
b9f5366
 
6b3d712
 
 
 
 
b9f5366
 
6b3d712
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
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
import gradio as gr
import numpy as np
import random
from diffusers import DiffusionPipeline
import torch
import spaces

# Ensuring the correct device (GPU or CPU)
device = "cuda" if torch.cuda.is_available() else "cpu"

MAX_SEED = np.iinfo(np.int32).max
MAX_IMAGE_SIZE = 1024

@spaces.GPU
def generate_image(prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps):
    try:
        # Randomize seed if selected
        if randomize_seed:
            seed = random.randint(0, MAX_SEED)

        # Initialize the generator with the seed
        generator = torch.Generator(device=device).manual_seed(seed)

        # Load the model and ensure it's running on the correct precision (float32 for safety)
        print(f"Loading model on device: {device}")
        pipe = DiffusionPipeline.from_pretrained("CompVis/stable-diffusion-v1-4", torch_dtype=torch.float32)
        pipe = pipe.to(device)  # Ensure the model is moved to the correct device
        pipe.enable_xformers_memory_efficient_attention()  # Enable memory-efficient attention if you want
        print("Model loaded successfully on device.")

        # Clear any unused GPU memory
        torch.cuda.empty_cache()

        # Generate the image
        print(f"Generating image with prompt: {prompt}")
        image = pipe(
            prompt=prompt,
            negative_prompt=negative_prompt,
            guidance_scale=guidance_scale,
            num_inference_steps=num_inference_steps,
            width=width,
            height=height,
            generator=generator
        ).images[0]
        
        print("Image generated successfully.")
        return image, None  # Return both the image and None for error message

    except Exception as e:
        # Catch any exceptions during image generation
        error_message = f"An error occurred while generating the image: {str(e)}"
        print(error_message)
        
        # Return black placeholder image as a fallback
        error_image = np.zeros((512, 512, 3), dtype=np.uint8)
        return error_image, error_message

examples = [
    "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k",
    "An astronaut riding a green horse",
    "A delicious ceviche cheesecake slice",
]

css = """
#col-container {
    margin: 0 auto;
    max-width: 520px;
}
"""

with gr.Blocks(css=css) as demo:
    with gr.Column(elem_id="col-container"):
        gr.Markdown(f"""
        # Text-to-Image Gradio Template
        Currently running on {device}.
        """)

        with gr.Row():
            prompt = gr.Text(
                label="Prompt",
                show_label=False,
                max_lines=1,
                placeholder="Enter your prompt",
                container=False,
            )
            
            run_button = gr.Button("Generate Image", scale=0)

        result = gr.Image(label="Result", show_label=False)
        error_message = gr.Textbox(label="Error Message", visible=False)  # For showing error messages
        
        with gr.Accordion("Advanced Settings", open=False):
            negative_prompt = gr.Text(
                label="Negative prompt",
                max_lines=1,
                placeholder="Enter a negative prompt",
                visible=False,
            )
            
            seed = gr.Slider(
                label="Seed",
                minimum=0,
                maximum=MAX_SEED,
                step=1,
                value=0,
            )
            
            randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
            
            with gr.Row():
                width = gr.Slider(
                    label="Width",
                    minimum=256,
                    maximum=MAX_IMAGE_SIZE,
                    step=32,
                    value=512,
                )
                
                height = gr.Slider(
                    label="Height",
                    minimum=256,
                    maximum=MAX_IMAGE_SIZE,
                    step=32,
                    value=512,
                )

            with gr.Row():
                guidance_scale = gr.Slider(
                    label="Guidance scale",
                    minimum=0.0,
                    maximum=20.0,  # Increased max value for better results
                    step=0.1,
                    value=7.5,  # Increased value for better image quality
                )
                
                num_inference_steps = gr.Slider(
                    label="Number of inference steps",
                    minimum=1,
                    maximum=50,
                    step=1,
                    value=25,  # Increased inference steps for better detail
                )
        
        gr.Examples(
            examples=examples,
            inputs=[prompt]
        )

    # Run the image generation on button click
    run_button.click(
        fn=generate_image,
        inputs=[prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps],
        outputs=[result, error_message]  # We now have two outputs: image and error message
    )

demo.queue().launch()