Spaces:
Sleeping
Sleeping
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,120 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# app.py
|
2 |
+
|
3 |
+
import gradio as gr
|
4 |
+
import torch
|
5 |
+
from diffusers import AutoPipelineForInpainting
|
6 |
+
from PIL import Image
|
7 |
+
|
8 |
+
# --- Model Loading ---
|
9 |
+
# Load the model only once at the start of the application
|
10 |
+
# We use float16 for memory efficiency and speed on GPUs
|
11 |
+
# If no GPU is available, this will run on CPU (but it will be very slow)
|
12 |
+
try:
|
13 |
+
pipe = AutoPipelineForInpainting.from_pretrained(
|
14 |
+
"stabilityai/stable-diffusion-2-inpainting",
|
15 |
+
torch_dtype=torch.float16,
|
16 |
+
variant="fp16"
|
17 |
+
).to("cuda")
|
18 |
+
except Exception as e:
|
19 |
+
print(f"Could not load model on GPU: {e}. Falling back to CPU.")
|
20 |
+
pipe = AutoPipelineForInpainting.from_pretrained(
|
21 |
+
"stabilityai/stable-diffusion-2-inpainting"
|
22 |
+
)
|
23 |
+
|
24 |
+
# --- The Inpainting Function ---
|
25 |
+
# This is the core function that takes user inputs and generates the image
|
26 |
+
def inpaint_image(input_dict, prompt, negative_prompt, guidance_scale, num_steps):
|
27 |
+
"""
|
28 |
+
Performs inpainting on an image based on a mask and a prompt.
|
29 |
+
|
30 |
+
Args:
|
31 |
+
input_dict (dict): A dictionary from Gradio's Image component containing 'image' and 'mask'.
|
32 |
+
prompt (str): The text prompt describing what to generate in the masked area.
|
33 |
+
negative_prompt (str): The text prompt describing what to avoid.
|
34 |
+
guidance_scale (float): A value to control how much the generation follows the prompt.
|
35 |
+
num_steps (int): The number of inference steps.
|
36 |
+
|
37 |
+
Returns:
|
38 |
+
PIL.Image: The resulting image after inpainting.
|
39 |
+
"""
|
40 |
+
# Separate the image and the mask from the input dictionary
|
41 |
+
image = input_dict["image"].convert("RGB")
|
42 |
+
mask_image = input_dict["mask"].convert("RGB")
|
43 |
+
|
44 |
+
# The model works best with images of a specific size (e.g., 512x512)
|
45 |
+
# We can resize for consistency, but for user-friendliness, we'll let the pipeline handle it.
|
46 |
+
# However, it's good practice to inform the user that square images work best.
|
47 |
+
|
48 |
+
print(f"Starting inpainting with prompt: '{prompt}'")
|
49 |
+
|
50 |
+
# Run the inpainting pipeline
|
51 |
+
result_image = pipe(
|
52 |
+
prompt=prompt,
|
53 |
+
image=image,
|
54 |
+
mask_image=mask_image,
|
55 |
+
negative_prompt=negative_prompt,
|
56 |
+
guidance_scale=guidance_scale,
|
57 |
+
num_inference_steps=int(num_steps),
|
58 |
+
).images[0]
|
59 |
+
|
60 |
+
return result_image
|
61 |
+
|
62 |
+
# --- Gradio User Interface ---
|
63 |
+
with gr.Blocks(theme=gr.themes.Soft()) as demo:
|
64 |
+
gr.Markdown(
|
65 |
+
"""
|
66 |
+
# 🎨 AI Image Fixer (Inpainting)
|
67 |
+
|
68 |
+
Have an AI-generated image with weird hands, faces, or artifacts? Fix it here!
|
69 |
+
|
70 |
+
**How to use:**
|
71 |
+
1. Upload your image.
|
72 |
+
2. Use the brush tool to "paint" over the parts you want to replace. This is your mask.
|
73 |
+
3. Write a prompt describing what you want in the painted-over area.
|
74 |
+
4. Adjust the advanced settings if you want more control.
|
75 |
+
5. Click "Fix It!" and see the magic happen.
|
76 |
+
"""
|
77 |
+
)
|
78 |
+
|
79 |
+
with gr.Row():
|
80 |
+
# Input column
|
81 |
+
with gr.Column():
|
82 |
+
gr.Markdown("### 1. Upload & Mask Your Image")
|
83 |
+
# The Image component with a drawing tool for masking
|
84 |
+
input_image = gr.Image(
|
85 |
+
label="Upload Image & Draw Mask",
|
86 |
+
source="upload",
|
87 |
+
tool="brush",
|
88 |
+
type="pil" # We want to work with PIL images in our function
|
89 |
+
)
|
90 |
+
|
91 |
+
gr.Markdown("### 2. Describe Your Fix")
|
92 |
+
prompt = gr.Textbox(label="Prompt", placeholder="e.g., 'A beautiful, realistic human hand, detailed fingers'")
|
93 |
+
|
94 |
+
# Accordion for advanced settings to keep the UI clean
|
95 |
+
with gr.Accordion("Advanced Settings", open=False):
|
96 |
+
negative_prompt = gr.Textbox(label="Negative Prompt", placeholder="e.g., 'blurry, distorted, extra fingers, cartoon'")
|
97 |
+
guidance_scale = gr.Slider(minimum=0, maximum=20, value=8.0, label="Guidance Scale")
|
98 |
+
num_steps = gr.Slider(minimum=10, maximum=100, step=1, value=40, label="Inference Steps")
|
99 |
+
|
100 |
+
# Output column
|
101 |
+
with gr.Column():
|
102 |
+
gr.Markdown("### 3. Get Your Result")
|
103 |
+
output_image = gr.Image(
|
104 |
+
label="Resulting Image",
|
105 |
+
type="pil"
|
106 |
+
)
|
107 |
+
|
108 |
+
# The button to trigger the process
|
109 |
+
submit_button = gr.Button("Fix It!", variant="primary")
|
110 |
+
|
111 |
+
# Connect the button to the function
|
112 |
+
submit_button.click(
|
113 |
+
fn=inpaint_image,
|
114 |
+
inputs=[input_image, prompt, negative_prompt, guidance_scale, num_steps],
|
115 |
+
outputs=output_image
|
116 |
+
)
|
117 |
+
|
118 |
+
# Launch the Gradio app
|
119 |
+
if __name__ == "__main__":
|
120 |
+
demo.launch()
|