You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
199 lines
8.3 KiB
199 lines
8.3 KiB
{
|
|
"nbformat": 4,
|
|
"nbformat_minor": 0,
|
|
"metadata": {
|
|
"colab": {
|
|
"name": "Face Depixelizer Eng",
|
|
"provenance": [],
|
|
"private_outputs": true,
|
|
"collapsed_sections": [],
|
|
"include_colab_link": true
|
|
},
|
|
"kernelspec": {
|
|
"name": "python3",
|
|
"display_name": "Python 3"
|
|
},
|
|
"accelerator": "GPU"
|
|
},
|
|
"cells": [
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {
|
|
"id": "view-in-github",
|
|
"colab_type": "text"
|
|
},
|
|
"source": [
|
|
"<a href=\"https://colab.research.google.com/github/tg-bomze/Face-Depixelizer/blob/master/Face_Depixelizer_Eng.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {
|
|
"id": "siqzcgRRyr_n",
|
|
"colab_type": "text"
|
|
},
|
|
"source": [
|
|
"<b><font color=\"black\" size=\"+4\">Face Depixelizer</font></b>\n",
|
|
"\n",
|
|
"Given a low-resolution input image, Face Depixelizer searches the outputs of a generative model (here, StyleGAN) for high-resolution images that are perceptually realistic and downscale correctly.\n",
|
|
"\n",
|
|
"<b><font color=\"black\" size=\"+2\">Based on:</font></b>\n",
|
|
"\n",
|
|
"**GitHub repository**: [PULSE](https://github.com/adamian98/pulse)\n",
|
|
"\n",
|
|
"Article: [PULSE: Self-Supervised Photo Upsampling via Latent Space Exploration of Generative Models](https://arxiv.org/abs/2003.03808)\n",
|
|
"\n",
|
|
"Creators: **[Alex Damian](https://github.com/adamian98), [Sachit Menon](mailto:sachit.menon@duke.edu).**\n",
|
|
"\n",
|
|
"<b><font color=\"black\" size=\"+2\">Colab created by:</font></b>\n",
|
|
"\n",
|
|
"GitHub: [@tg-bomze](https://github.com/tg-bomze),\n",
|
|
"Telegram: [@bomze](https://t.me/bomze),\n",
|
|
"Twitter: [@tg_bomze](https://twitter.com/tg_bomze).\n",
|
|
"\n",
|
|
"---\n",
|
|
"##### <font color='red'>Currently using Google Drive to store model weights and it has a daily cap on downloads, therefore, you may receive an error message saying \"Google Drive Quota Exceeded\" or \"No such file or directory: '/content/pulse/runs/face.png'\". If you are experiencing this error please try again later in the day or come back tomorrow.</font>\n",
|
|
"\n",
|
|
"```\n",
|
|
"To get started, click on the button (where the red arrow indicates). After clicking, wait until the execution is complete.\n",
|
|
"```\n",
|
|
"\n"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"metadata": {
|
|
"id": "fU0aGtD4Nl4W",
|
|
"colab_type": "code",
|
|
"cellView": "form",
|
|
"colab": {}
|
|
},
|
|
"source": [
|
|
"#@title <b><font color=\"red\" size=\"+3\">←</font><font color=\"black\" size=\"+3\"> Let's ROCK!</font></b>\n",
|
|
"#@markdown **After starting this block, you will need to scroll down page and upload a pixel square photo into which the whole human head fits. Neural network works best on images where people are directly facing the camera. Example:**\n",
|
|
"\n",
|
|
"#@markdown \n",
|
|
"\n",
|
|
"#@markdown *You can crop the photo [HERE](https://www.iloveimg.com/crop-image)*\n",
|
|
"\n",
|
|
"#@markdown ---\n",
|
|
"import torch\n",
|
|
"import torchvision\n",
|
|
"from pathlib import Path\n",
|
|
"if not Path(\"PULSE.py\").exists():\n",
|
|
" if Path(\"pulse\").exists():\n",
|
|
" %cd /content/pulse\n",
|
|
" else:\n",
|
|
" !git clone https://github.com/adamian98/pulse\n",
|
|
" %cd /content/pulse\n",
|
|
" !mkdir input/\n",
|
|
" toPIL = torchvision.transforms.ToPILImage()\n",
|
|
" toTensor = torchvision.transforms.ToTensor()\n",
|
|
" from bicubic import BicubicDownSample\n",
|
|
" D = BicubicDownSample(factor=1)\n",
|
|
"\n",
|
|
"import os\n",
|
|
"from io import BytesIO\n",
|
|
"import matplotlib.pyplot as plt\n",
|
|
"import matplotlib.image as mpimg\n",
|
|
"from PIL import Image\n",
|
|
"from PULSE import PULSE\n",
|
|
"from google.colab import files\n",
|
|
"from bicubic import BicubicDownSample\n",
|
|
"from IPython import display\n",
|
|
"from IPython.display import display\n",
|
|
"from IPython.display import clear_output\n",
|
|
"import numpy as np\n",
|
|
"from drive import open_url\n",
|
|
"from mpl_toolkits.axes_grid1 import ImageGrid\n",
|
|
"%matplotlib inline\n",
|
|
"\n",
|
|
"#@markdown ## Basic settings:\n",
|
|
"#@markdown ##### *If you have already uploaded a photo and just want to experiment with the settings, then uncheck the following checkbox*:\n",
|
|
"upload_new_photo = True #@param {type:\"boolean\"}\n",
|
|
"\n",
|
|
"\n",
|
|
"if upload_new_photo == True:\n",
|
|
" !rm -rf /content/pulse/input/face.png\n",
|
|
" clear_output()\n",
|
|
" uploaded = files.upload()\n",
|
|
" for fn in uploaded.keys():\n",
|
|
" print('User uploaded file \"{name}\" with length {length} bytes'.format(\n",
|
|
" name=fn, length=len(uploaded[fn])))\n",
|
|
" os.rename(fn, fn.replace(\" \", \"\"))\n",
|
|
" fn = fn.replace(\" \", \"\")\n",
|
|
"\n",
|
|
" if(len(uploaded.keys())!=1): raise Exception(\"You need to upload only one image.\")\n",
|
|
"\n",
|
|
" face = Image.open(fn)\n",
|
|
" face = face.resize((1024, 1024), Image.ANTIALIAS)\n",
|
|
" face = face.convert('RGB')\n",
|
|
" face_name = 'face.png'\n",
|
|
" face.save(face_name)\n",
|
|
" %cp $face_name /content/pulse/input/\n",
|
|
"\n",
|
|
" images = []\n",
|
|
" imagesHR = []\n",
|
|
" imagesHR.append(face)\n",
|
|
" face = toPIL(D(toTensor(face).unsqueeze(0).cuda()).cpu().detach().clamp(0,1)[0])\n",
|
|
" images.append(face)\n",
|
|
"\n",
|
|
"#@markdown ---\n",
|
|
"#@markdown ## Advanced settings:\n",
|
|
"#@markdown ##### *If you want to make a more accurate result, then modify the following* **DEFAULT** *variables*:\n",
|
|
"\n",
|
|
"input_dir = '/content/pulse/input/'\n",
|
|
"output_dir = '/content/pulse/runs/'\n",
|
|
"seed = 100 #@param {type:\"integer\"}\n",
|
|
"epsilon = 0.02 #@param {type:\"slider\", min:0.01, max:0.03, step:0.01}\n",
|
|
"noise_type = 'trainable' #@param ['zero', 'fixed', 'trainable']\n",
|
|
"optimizer = 'adam' #@param ['sgd', 'adam','sgdm', 'adamax']\n",
|
|
"learning_rate = 0.4 #@param {type:\"slider\", min:0, max:1, step:0.05}\n",
|
|
"learning_rate_schedule = 'linear1cycledrop' #@param ['fixed', 'linear1cycle', 'linear1cycledrop']\n",
|
|
"steps = 100 #@param {type:\"slider\", min:100, max:2000, step:50}\n",
|
|
"clear_output()\n",
|
|
"\n",
|
|
"seed = abs(seed)\n",
|
|
"print('Estimated Runtime: {}s.\\n'.format(round(0.23*steps)+6))\n",
|
|
"!python run.py \\\n",
|
|
" -input_dir $input_dir \\\n",
|
|
" -output_dir $output_dir \\\n",
|
|
" -seed $seed \\\n",
|
|
" -noise_type $noise_type \\\n",
|
|
" -opt_name $optimizer \\\n",
|
|
" -learning_rate $learning_rate \\\n",
|
|
" -steps $steps \\\n",
|
|
" -eps $epsilon \\\n",
|
|
" -lr_schedule $learning_rate_schedule\n",
|
|
"\n",
|
|
"#@markdown ---\n",
|
|
"#@markdown *If there is an error during execution or the \"**Browse**\" button is not active, try running this block again*\n",
|
|
"\n",
|
|
"fig, (ax1, ax2) = plt.subplots(1, 2)\n",
|
|
"ax1.imshow(mpimg.imread('/content/pulse/input/face.png'))\n",
|
|
"ax1.set_title('Original')\n",
|
|
"ax2.imshow(mpimg.imread('/content/pulse/runs/face.png'))\n",
|
|
"ax2.set_title('Result')\n",
|
|
"plt.show()"
|
|
],
|
|
"execution_count": null,
|
|
"outputs": []
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"metadata": {
|
|
"id": "DUfP6_7vTK3b",
|
|
"colab_type": "code",
|
|
"cellView": "form",
|
|
"colab": {}
|
|
},
|
|
"source": [
|
|
"#@title <b><font color=\"red\" size=\"+3\">←</font><font color=\"black\" size=\"+3\"> Download result</font></b>\n",
|
|
"try: files.download('/content/pulse/runs/face.png')\n",
|
|
"except: raise Exception(\"No result image\")"
|
|
],
|
|
"execution_count": null,
|
|
"outputs": []
|
|
}
|
|
]
|
|
}
|