Spaces:
Runtime error
Runtime error
File size: 2,879 Bytes
b260763 37f377e b260763 3e3cc54 b260763 4ae241e cf3577e b7c177d 4ae241e 3e3cc54 4ae241e 11d6ee6 4ae241e 11d6ee6 4ae241e 5c7822d b260763 7bf4347 4ae241e cf3577e 4ae241e |
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 |
import gradio as gr
VALID_CHOICES = [
"Bald",
"Young",
"Mustache",
"Eyeglasses",
"Hat",
"Smiling",
"Gray_Hair",
]
ENABLE_GPU = False
MODEL_NAME = "stylegan_ffhq"
OUTPUT_LIST = [
gr.outputs.Image(type="pil", label="Generated Images"),
gr.outputs.Image(type="pil", label="Modified Images"),
]
# description = """
# This is an interactive demo of an extension of the <a href="https://genforce.github.io/interfacegan/">CVPR2020 InterfaceGAN paper</a>, by adding other attributes such as Hat, Bald, etc. to the generated images. Here is a step-by-step guide to use this interface:
# + 🌾 Select the Random seed you want to use to generate the images
# + 🗂 Select the list of attributes you want to modify (of course, you can mix several attributes)
# + 🛠 Select the coefficient of modification (higher value means stronger modification)
# + 🔥 Submit!
# Check the <a href="https://github.com/genforce/interfacegan">original repo</a> as well as the <a href="https://github.com/younesbelkada/interfacegan">extended version of the work</a>.
# ⭕ This method is biased on the data it has been trained for attribute recognition. E.g. if you decide to modify the "Bald" attribute on Female faces, the method will turn it into Male faces. Future work may focus more on this direction to try to have unbiased results of the modifications.
#Check the <a href="https://github.com/genforce/interfacegan">original repo</a> as well as the <a href="https://github.com/younesbelkada/interfacegan">extended version of the work</a>.
#
# </center>
# </p>
# """
description = """
<p>
This is an interactive demo of an extension of the <a href="https://genforce.github.io/interfacegan/">CVPR2020 InterfaceGAN paper</a>, which adds other attributes such as Hat, Bald, etc. to the generated images. Here is a step-by-step guide to use this space:
<ul>
<li>🌾 Select the Random seed you want to use to generate the images</li>
<li>🗂 Select the list of attributes you want to modify (of course, you can mix several attributes)</li>
<li>🛠 Select the modification scale (higher value means stronger modification)</li>
<li>🔥 Submit!</li>
</ul>
⭕ This method is biased on the data it has been trained for attribute recognition. E.g. if you decide to modify the "Bald" attribute on Female faces, the method will turn it into Male faces. Future work may focus more on this direction to try to have unbiased results of the modifications.
</p>
"""
css = """
ul {
list-style-type: none;
}
ul.no-bullets li
{
list-style-type: none;
}
"""
title = "InterfaceGAN++ Demo"
article = "<div style='text-align: center;'><a href='https://github.com/genforce/interfacegan' target='_blank'>Original Paper</a> | <a href='https://github.com/younesbelkada/interfacegan' target='_blank'>Extended Version</a></div>"
|