add file
Browse files- app.py +82 -0
- requirements.txt +6 -0
app.py
ADDED
|
@@ -0,0 +1,82 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from PIL import Image
|
| 2 |
+
import numpy as np
|
| 3 |
+
from rembg import remove
|
| 4 |
+
import cv2
|
| 5 |
+
import os
|
| 6 |
+
from torchvision.transforms import GaussianBlur
|
| 7 |
+
import gradio as gr
|
| 8 |
+
import replicate
|
| 9 |
+
import requests
|
| 10 |
+
from io import BytesIO
|
| 11 |
+
|
| 12 |
+
def create_mask(input):
|
| 13 |
+
input_path = 'input.png'
|
| 14 |
+
bg_removed_path = 'bg_removed.png'
|
| 15 |
+
mask_name = 'blured_mask.png'
|
| 16 |
+
|
| 17 |
+
input.save(input_path)
|
| 18 |
+
bg_removed = remove(input)
|
| 19 |
+
bg_removed = bg_removed.resize((512, 512))
|
| 20 |
+
bg_removed.save(bg_removed_path)
|
| 21 |
+
|
| 22 |
+
img2_grayscale = bg_removed.convert('L')
|
| 23 |
+
img2_a = np.array(img2_grayscale)
|
| 24 |
+
|
| 25 |
+
mask = np.array(img2_grayscale)
|
| 26 |
+
threshhold = 0
|
| 27 |
+
mask[img2_a==threshhold] = 1
|
| 28 |
+
mask[img2_a>threshhold] = 0
|
| 29 |
+
|
| 30 |
+
strength = 1
|
| 31 |
+
d = int(255 * (1-strength))
|
| 32 |
+
mask *= 255-d
|
| 33 |
+
mask += d
|
| 34 |
+
|
| 35 |
+
mask = Image.fromarray(mask)
|
| 36 |
+
|
| 37 |
+
blur = GaussianBlur(11,20)
|
| 38 |
+
mask = blur(mask)
|
| 39 |
+
mask = mask.resize((512, 512))
|
| 40 |
+
|
| 41 |
+
mask.save(mask_name)
|
| 42 |
+
|
| 43 |
+
return Image.open(mask_name)
|
| 44 |
+
|
| 45 |
+
|
| 46 |
+
def generate_image(image, product_name, target_name):
|
| 47 |
+
mask = create_mask(image)
|
| 48 |
+
image = image.resize((512, 512))
|
| 49 |
+
mask = mask.resize((512,512))
|
| 50 |
+
guidance_scale=16
|
| 51 |
+
num_samples = 1
|
| 52 |
+
|
| 53 |
+
prompt = 'a photo of a ' + product_name + ' with ' + target_name + ' product photograpy'
|
| 54 |
+
|
| 55 |
+
model = replicate.models.get("cjwbw/stable-diffusion-v2-inpainting")
|
| 56 |
+
version = model.versions.get("f9bb0632bfdceb83196e85521b9b55895f8ff3d1d3b487fd1973210c0eb30bec")
|
| 57 |
+
output = version.predict(prompt=prompt, image=open("bg_removed.png", "rb"), mask=open("blured_mask.png", "rb"))
|
| 58 |
+
response = requests.get(output[0])
|
| 59 |
+
|
| 60 |
+
return Image.open(BytesIO(response.content))
|
| 61 |
+
|
| 62 |
+
with gr.Blocks() as demo:
|
| 63 |
+
gr.Markdown("# Advertise better with AI")
|
| 64 |
+
# with gr.Tab("Prompt Paint - Basic"):
|
| 65 |
+
with gr.Row():
|
| 66 |
+
|
| 67 |
+
with gr.Column():
|
| 68 |
+
input_image = gr.Image(label = "Upload your product's photo", type = 'pil')
|
| 69 |
+
|
| 70 |
+
product_name = gr.Textbox(label="Describe your product")
|
| 71 |
+
target_name = gr.Textbox(label="Where do you want to put your product?")
|
| 72 |
+
# result_prompt = product_name + ' in ' + target_name + 'product photograpy ultrarealist'
|
| 73 |
+
|
| 74 |
+
image_button = gr.Button("Generate")
|
| 75 |
+
|
| 76 |
+
with gr.Column():
|
| 77 |
+
image_output = gr.Image()
|
| 78 |
+
|
| 79 |
+
image_button.click(generate_image, inputs=[input_image, product_name, target_name ], outputs=image_output, api_name='test')
|
| 80 |
+
|
| 81 |
+
|
| 82 |
+
demo.launch()
|
requirements.txt
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
rembg==2.0.30
|
| 2 |
+
torchvision==0.13.1
|
| 3 |
+
numpy==1.23.5
|
| 4 |
+
Pillow==9.3.0
|
| 5 |
+
opencv-python==4.6.0.66
|
| 6 |
+
gradio==3.9.1
|