Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,96 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
import torch
|
| 3 |
+
from diffusers.utils import load_image
|
| 4 |
+
from controlnet_flux import FluxControlNetModel
|
| 5 |
+
from transformer_flux import FluxTransformer2DModel
|
| 6 |
+
from pipeline_flux_controlnet_inpaint import FluxControlNetInpaintingPipeline
|
| 7 |
+
from PIL import Image, ImageDraw
|
| 8 |
+
|
| 9 |
+
# Load models
|
| 10 |
+
controlnet = FluxControlNetModel.from_pretrained("alimama-creative/FLUX.1-dev-Controlnet-Inpainting-Alpha", torch_dtype=torch.bfloat16)
|
| 11 |
+
transformer = FluxTransformer2DModel.from_pretrained(
|
| 12 |
+
"black-forest-labs/FLUX.1-dev", subfolder='transformer', torch_dtype=torch.bfloat16
|
| 13 |
+
)
|
| 14 |
+
pipe = FluxControlNetInpaintingPipeline.from_pretrained(
|
| 15 |
+
"black-forest-labs/FLUX.1-dev",
|
| 16 |
+
controlnet=controlnet,
|
| 17 |
+
transformer=transformer,
|
| 18 |
+
torch_dtype=torch.bfloat16
|
| 19 |
+
).to("cuda")
|
| 20 |
+
pipe.transformer.to(torch.bfloat16)
|
| 21 |
+
pipe.controlnet.to(torch.bfloat16)
|
| 22 |
+
|
| 23 |
+
def prepare_image_and_mask(image, width, height, overlap_percentage):
|
| 24 |
+
# Resize the input image to fit within the target size
|
| 25 |
+
image.thumbnail((width, height), Image.LANCZOS)
|
| 26 |
+
|
| 27 |
+
# Create a new white background image of the target size
|
| 28 |
+
background = Image.new('RGB', (width, height), (255, 255, 255))
|
| 29 |
+
|
| 30 |
+
# Paste the resized image onto the background
|
| 31 |
+
offset = ((width - image.width) // 2, (height - image.height) // 2)
|
| 32 |
+
background.paste(image, offset)
|
| 33 |
+
|
| 34 |
+
# Create a mask
|
| 35 |
+
mask = Image.new('L', (width, height), 255)
|
| 36 |
+
draw = ImageDraw.Draw(mask)
|
| 37 |
+
|
| 38 |
+
# Calculate the overlap area
|
| 39 |
+
overlap_x = int(image.width * overlap_percentage / 100)
|
| 40 |
+
overlap_y = int(image.height * overlap_percentage / 100)
|
| 41 |
+
|
| 42 |
+
# Draw the mask (black area is where we want to inpaint)
|
| 43 |
+
draw.rectangle([
|
| 44 |
+
(offset[0] + overlap_x, offset[1] + overlap_y),
|
| 45 |
+
(offset[0] + image.width - overlap_x, offset[1] + image.height - overlap_y)
|
| 46 |
+
], fill=0)
|
| 47 |
+
|
| 48 |
+
return background, mask
|
| 49 |
+
|
| 50 |
+
def inpaint(image, prompt, width, height, overlap_percentage, num_inference_steps, guidance_scale):
|
| 51 |
+
# Prepare image and mask
|
| 52 |
+
image, mask = prepare_image_and_mask(image, width, height, overlap_percentage)
|
| 53 |
+
|
| 54 |
+
# Set up generator for reproducibility
|
| 55 |
+
generator = torch.Generator(device="cuda").manual_seed(42)
|
| 56 |
+
|
| 57 |
+
# Run inpainting
|
| 58 |
+
result = pipe(
|
| 59 |
+
prompt=prompt,
|
| 60 |
+
height=height,
|
| 61 |
+
width=width,
|
| 62 |
+
control_image=image,
|
| 63 |
+
control_mask=mask,
|
| 64 |
+
num_inference_steps=num_inference_steps,
|
| 65 |
+
generator=generator,
|
| 66 |
+
controlnet_conditioning_scale=0.9,
|
| 67 |
+
guidance_scale=guidance_scale,
|
| 68 |
+
negative_prompt="",
|
| 69 |
+
true_guidance_scale=guidance_scale
|
| 70 |
+
).images[0]
|
| 71 |
+
|
| 72 |
+
return result
|
| 73 |
+
|
| 74 |
+
# Gradio interface
|
| 75 |
+
with gr.Blocks() as demo:
|
| 76 |
+
gr.Markdown("# FLUX Outpainting Demo")
|
| 77 |
+
with gr.Row():
|
| 78 |
+
with gr.Column():
|
| 79 |
+
input_image = gr.Image(type="pil", label="Input Image")
|
| 80 |
+
prompt_input = gr.Textbox(label="Prompt")
|
| 81 |
+
width_slider = gr.Slider(label="Width", minimum=256, maximum=1024, step=64, value=768)
|
| 82 |
+
height_slider = gr.Slider(label="Height", minimum=256, maximum=1024, step=64, value=768)
|
| 83 |
+
overlap_slider = gr.Slider(label="Overlap Percentage", minimum=0, maximum=50, step=1, value=10)
|
| 84 |
+
steps_slider = gr.Slider(label="Inference Steps", minimum=1, maximum=100, step=1, value=28)
|
| 85 |
+
guidance_slider = gr.Slider(label="Guidance Scale", minimum=0.1, maximum=10.0, step=0.1, value=3.5)
|
| 86 |
+
run_button = gr.Button("Generate")
|
| 87 |
+
with gr.Column():
|
| 88 |
+
output_image = gr.Image(label="Output Image")
|
| 89 |
+
|
| 90 |
+
run_button.click(
|
| 91 |
+
fn=inpaint,
|
| 92 |
+
inputs=[input_image, prompt_input, width_slider, height_slider, overlap_slider, steps_slider, guidance_slider],
|
| 93 |
+
outputs=output_image
|
| 94 |
+
)
|
| 95 |
+
|
| 96 |
+
demo.launch()
|