File size: 1,018 Bytes
b060e0a 85d3ae6 b060e0a a810cc1 b060e0a a810cc1 b060e0a 5894220 b060e0a 1adaec6 b060e0a 390aa75 b060e0a 0de02d5 b060e0a 85d3ae6 b060e0a |
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 |
import gradio as gr
import numpy as np
from tensorflow.keras.preprocessing.image import load_img, img_to_array
from tensorflow.keras.models import load_model
from PIL import Image
import matplotlib.pyplot as plt
inputs = gr.inputs.Image(shape=(256, 256))
o1 = gr.outputs.Image()
o2 = gr.outputs.Image()
gen_model = load_model('model256.h5')
def colorify(pixels):
pixels = (pixels - 127.5) / 127.5
pixels = np.expand_dims(pixels, 0)
gen_image = gen_model.predict(pixels)
gen_image = (gen_image + 1) / 2
return Image.fromarray((gen_image[0] * 255.0).astype(np.uint8))
title = "Colorify"
description = "Recolor your images using this lite version of PIX2PIX GAN , model is trained on 700 randomly collected images from the internet with 256*256 pixels"
examples=[['example1.png'],['example2.jpg']]
article = "<p style='text-align: center'>"
gr.Interface(fn=colorify, inputs=inputs, outputs=o1, title=title, description=description, article=article, examples=examples, enable_queue=True).launch()
|