Real_ESRGAN_Pytorch_Inference_WEB_UI.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "collapsed_sections": [
        "6BPxh_VmVVIu"
      ],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Nick088Official/Real-ESRGAN_Pytorch/blob/main/Real_ESRGAN_Pytorch_Inference_WEB_UI.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IRDbDYYMQt_Y"
      },
      "source": [
        "# Real-ESRGAN Pytorch Inference WEB UI\n",
        "\n",
        "[![arXiv](https://img.shields.io/badge/arXiv-Paper-<COLOR>.svg)](https://arxiv.org/abs/2107.10833)\n",
        "[![GitHub Stars](https://img.shields.io/github/stars/Nick088Official/Real-ESRGAN_Pytorch?style=social)](https://github.com/Nick088Official/Real-ESRGAN_Pytorch)\n",
        "\n",
        "This is a **Practical Image Restoration Demo** of the Pytorch ai-forever custom pretrained Real-ESRGAN models.\n",
        "We extend the powerful ESRGAN to a practical restoration application (namely, Real-ESRGAN), which is trained with pure synthetic data. <br>\n",
        "The following figure shows some real-life examples.\n",
        "\n",
        "Low quality image:\n",
        "\n",
        "![](https://github.com/Nick088Official/Real-ESRGAN-Pytorch/blob/main/inputs/lr_lion.png?raw=1)\n",
        "\n",
        "Real-ESRGAN_Pytorch result:\n",
        "\n",
        "![](https://github.com/Nick088Official/Real-ESRGAN-Pytorch/blob/main/results/sr_lion.png?raw=1)\n",
        "\n",
        "**Note that Real-ESRGAN may still fail in some cases as the real-world degradations are really too complex.**<br>\n",
        "Moreover, it **may not** perform well on human faces, you could try [GFPGAN](https://colab.research.google.com/github/Nick088Official/GFPGAN-Fix/blob/master/GFPGAN_fix_inference.ipynb) for that\n",
        "<br>\n",
        "\n",
        "This is the Pytorch Implementation of https://github.com/ai-forever/Real-ESRGAN\n",
        "\n",
        "**Credits:** [Nick088](https://linktr.ee/Nick088), Xinntao, Tencent, Geeve George, ai-forever, daroche"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GnpnrLfMV2jU",
        "cellView": "form"
      },
      "source": [
        "#@title Installation\n",
        "#@markdown Before running, make sure that you choose\n",
        "#@markdown * Runtime Type = Python 3\n",
        "#@markdown * Hardware Accelerator = GPU (Faster, free daily limit of gpu around 12 hours) or CPU (very slow)\n",
        "\n",
        "#@markdown in the **Runtime** menu -> **Change runtime type**\n",
        "\n",
        "#@markdown By running this, we clone the repository, set up the envrironment.\n",
        "\n",
        "\n",
        "from IPython.display import clear_output\n",
        "import torch\n",
        "\n",
        "if torch.cuda.is_available():\n",
        "    device = \"cuda\"\n",
        "    print(\"Using GPU\")\n",
        "else:\n",
        "    device = \"cpu\"\n",
        "    print(\"Using CPU\")\n",
        "\n",
        "!git clone https://github.com/Nick088Official/Real-ESRGAN_Pytorch/\n",
        "%cd Real-ESRGAN_Pytorch/Scripts\n",
        "!pip install -r requirements_ui.txt\n",
        "clear_output()\n",
        "print(\"Installed!\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Run UI\n",
        "\n",
        "#@markdown The type of tunnel you wanna use for seeing the public link, so that if one of them is down, you can use the other one.\n",
        "Tunnel = \"Gradio\" #@param [\"Gradio\", \"Ngrok\", \"Cloudfare\", \"LocalTunnel\"]\n",
        "\n",
        "#@markdown Also when using Ngrok ,Cloudfare or LocalTunnel, you need to wait for the Local URL to appear, and only after that click on the Public URL which is above.\n",
        "\n",
        "#@markdown Use the following option only if you chose Ngrok as the Tunnel:\n",
        "\n",
        "#@markdown You can get the Ngrok Tunnel Authtoken here: https://dashboard.ngrok.com/tunnels/authtokens/new.\n",
        "\n",
        "ngrok_tunnel_authtoken = \"\" #@param {type:\"string\"}\n",
        "\n",
        "!sed -i 's/import spaces/ /g'  /content/Real-ESRGAN_Pytorch/Scripts/Run_Inference_UI.py\n",
        "!sed -i 's/@spaces.GPU()/ /g'  /content/Real-ESRGAN_Pytorch/Scripts/Run_Inference_UI.py\n",
        "\n",
        "if Tunnel == \"Gradio\":\n",
        "  !sed -i 's/share=False/share=True/g'  /content/Real-ESRGAN_Pytorch/Scripts/Run_Inference_UI.py\n",
        "if Tunnel == \"Ngrok\":\n",
        "  !sed -i 's/share=True/share=False/g'  /content/Real-ESRGAN_Pytorch/Scripts/Run_Inference_UI.py\n",
        "  from pyngrok import ngrok\n",
        "  ngrok.set_auth_token(ngrok_tunnel_authtoken)\n",
        "  http_tunnel = ngrok.connect(7860, bind_tls=True)\n",
        "  print(\"Ngrok Tunnel Public URL:\", http_tunnel.public_url)\n",
        "elif Tunnel == \"Cloudfare\":\n",
        "  !sed -i 's/share=True/share=False/g'  /content/Real-ESRGAN_Pytorch/Scripts/Run_Inference_UI.py\n",
        "  # download cloudfare\n",
        "  !curl -LO https://github.com/cloudflare/cloudflared/releases/latest/download/cloudflared-linux-amd64.deb\n",
        "  !dpkg -i cloudflared-linux-amd64.deb\n",
        "  !rm -rf nohup.out\n",
        "  import time\n",
        "  # Run cloudflare\n",
        "  !nohup cloudflared tunnel --url localhost:7860 &\n",
        "  clear_output()\n",
        "  time.sleep(5)\n",
        "  # Find and print the Cloudflare URL with a prefix\n",
        "  cloudflare_url = !grep -oE \"https://[a-zA-Z0-9.-]+\\.trycloudflare\\.com\" nohup.out\n",
        "  print(f\"Cloudfare Tunnel Public URL: {cloudflare_url[0]}\")\n",
        "elif Tunnel == \"LocalTunnel\":\n",
        "  !sed -i 's/share=True/share=False/g'  /content/Real-ESRGAN_Pytorch/Scripts/Run_Inference_UI.py\n",
        "  # install\n",
        "  !npm install -g localtunnel\n",
        "  import time\n",
        "  import urllib\n",
        "  # run localtunnel\n",
        "  with open('url.txt', 'w') as file:\n",
        "        file.write('')\n",
        "\n",
        "  get_ipython().system_raw('lt --port 7860 >> url.txt 2>&1 &')\n",
        "\n",
        "  time.sleep(4)\n",
        "\n",
        "  endpoint_ip = urllib.request.urlopen('https://ipv4.icanhazip.com').read().decode('utf8').strip(\"\\n\")\n",
        "\n",
        "  with open('url.txt', 'r') as file:\n",
        "      tunnel_url = file.read()\n",
        "      tunnel_url = tunnel_url.replace(\"your url is: \", \"\")\n",
        "\n",
        "  clear_output()\n",
        "  print(f\"LocalTunnel Tunnel Public URL: \\033[0m\\033[93m{tunnel_url}\\033[0m\", end=\"\\033[0m\")\n",
        "\n",
        "  print(f'LocalTunnel Password: {endpoint_ip}')\n",
        "\n",
        "\n",
        "!python Run_Inference_UI.py"
      ],
      "metadata": {
        "cellView": "form",
        "id": "JYB4eIk9f3P_"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}