YOLOv4_Training.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "YOLOv4_Training.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyPOCj/xtqvhgatM/XM7r7HL",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/aliabdelmageed/Weapons-Detection-YOLOv4/blob/master/YOLOv4_Training.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Lfq-84YAGtSv",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 122
        },
        "outputId": "52d1dd79-135d-4c45-c8ef-42c08e08f679"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/gdrive')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/gdrive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6PC_POgOGzZp",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "b76128b3-1eaf-421c-e9a3-d54e30e3d85e"
      },
      "source": [
        "%cd /content/gdrive/My Drive/weapon detector/"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/content/gdrive/My Drive/weapon detector\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wwUto0d-HHvh",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 176
        },
        "outputId": "66e29c1a-d12e-4924-ef45-64411662741f"
      },
      "source": [
        "!pip install wget"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting wget\n",
            "  Downloading https://files.pythonhosted.org/packages/47/6a/62e288da7bcda82b935ff0c6cfe542970f04e29c756b0e147251b2fb251f/wget-3.2.zip\n",
            "Building wheels for collected packages: wget\n",
            "  Building wheel for wget (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for wget: filename=wget-3.2-cp36-none-any.whl size=9682 sha256=6da19d65f83ec1187b7362cd512b851c8823e129e6f6f782e00f25d53ad5e66e\n",
            "  Stored in directory: /root/.cache/pip/wheels/40/15/30/7d8f7cea2902b4db79e3fea550d7d7b85ecb27ef992b618f3f\n",
            "Successfully built wget\n",
            "Installing collected packages: wget\n",
            "Successfully installed wget-3.2\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XfMT-capHXB_",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 123
        },
        "outputId": "f39c0098-42fc-46bb-c9db-e4f640a3cf04"
      },
      "source": [
        "# clone darknet repo\n",
        "!git clone https://github.com/AlexeyAB/darknet"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Cloning into 'darknet'...\n",
            "remote: Enumerating objects: 14073, done.\u001b[K\n",
            "remote: Total 14073 (delta 0), reused 0 (delta 0), pack-reused 14073\u001b[K\n",
            "Receiving objects: 100% (14073/14073), 12.72 MiB | 6.62 MiB/s, done.\n",
            "Resolving deltas: 100% (9574/9574), done.\n",
            "Checking out files: 100% (2013/2013), done.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yOUCnamzHjyi",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "66a7f468-cb22-414c-8593-848a64aba29f"
      },
      "source": [
        "# clone OIDv4_ToolKit to download open images dataset\n",
        "!git clone https://github.com/theAIGuysCode/OIDv4_ToolKit.git"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "fatal: destination path 'OIDv4_ToolKit' already exists and is not an empty directory.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1zFXIUBkHnnU",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "ac2ba398-c508-426d-86ae-12faceee3f23"
      },
      "source": [
        "%cd /content/gdrive/My Drive/weapon detector/OIDv4_ToolKit"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/content/gdrive/My Drive/weapon detector/OIDv4_ToolKit\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ju3iWEhbHtc4",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 675
        },
        "outputId": "0e7b4605-e648-4cc4-b165-daa6d0682cc6"
      },
      "source": [
        "!pip3 install -r requirements.txt"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: pandas in /usr/local/lib/python3.6/dist-packages (from -r requirements.txt (line 1)) (1.0.5)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.6/dist-packages (from -r requirements.txt (line 2)) (1.18.5)\n",
            "Collecting awscli\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/df/c3/7c49e3729c291b008f57fdd86267d6b1cfd593e8b2942677ca8af7dfc175/awscli-1.18.105-py2.py3-none-any.whl (3.3MB)\n",
            "\u001b[K     |████████████████████████████████| 3.3MB 2.7MB/s \n",
            "\u001b[?25hRequirement already satisfied: urllib3 in /usr/local/lib/python3.6/dist-packages (from -r requirements.txt (line 4)) (1.24.3)\n",
            "Requirement already satisfied: tqdm in /usr/local/lib/python3.6/dist-packages (from -r requirements.txt (line 5)) (4.41.1)\n",
            "Requirement already satisfied: opencv-python in /usr/local/lib/python3.6/dist-packages (from -r requirements.txt (line 6)) (4.1.2.30)\n",
            "Requirement already satisfied: python-dateutil>=2.6.1 in /usr/local/lib/python3.6/dist-packages (from pandas->-r requirements.txt (line 1)) (2.8.1)\n",
            "Requirement already satisfied: pytz>=2017.2 in /usr/local/lib/python3.6/dist-packages (from pandas->-r requirements.txt (line 1)) (2018.9)\n",
            "Requirement already satisfied: PyYAML<5.4,>=3.10; python_version != \"3.4\" in /usr/local/lib/python3.6/dist-packages (from awscli->-r requirements.txt (line 3)) (3.13)\n",
            "Collecting botocore==1.17.28\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/23/6d/0e17d14779c049161b3bbaa6b622d6bf68209082e579376a6cdb8214ec4c/botocore-1.17.28-py2.py3-none-any.whl (6.4MB)\n",
            "\u001b[K     |████████████████████████████████| 6.4MB 28.9MB/s \n",
            "\u001b[?25hCollecting colorama<0.4.4,>=0.2.5; python_version != \"3.4\"\n",
            "  Downloading https://files.pythonhosted.org/packages/c9/dc/45cdef1b4d119eb96316b3117e6d5708a08029992b2fee2c143c7a0a5cc5/colorama-0.4.3-py2.py3-none-any.whl\n",
            "Collecting rsa<=4.5.0,>=3.1.2; python_version != \"3.4\"\n",
            "  Downloading https://files.pythonhosted.org/packages/26/f8/8127fdda0294f044121d20aac7785feb810e159098447967a6103dedfb96/rsa-4.5-py2.py3-none-any.whl\n",
            "Requirement already satisfied: docutils<0.16,>=0.10 in /usr/local/lib/python3.6/dist-packages (from awscli->-r requirements.txt (line 3)) (0.15.2)\n",
            "Requirement already satisfied: s3transfer<0.4.0,>=0.3.0 in /usr/local/lib/python3.6/dist-packages (from awscli->-r requirements.txt (line 3)) (0.3.3)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.6/dist-packages (from python-dateutil>=2.6.1->pandas->-r requirements.txt (line 1)) (1.15.0)\n",
            "Requirement already satisfied: jmespath<1.0.0,>=0.7.1 in /usr/local/lib/python3.6/dist-packages (from botocore==1.17.28->awscli->-r requirements.txt (line 3)) (0.10.0)\n",
            "Requirement already satisfied: pyasn1>=0.1.3 in /usr/local/lib/python3.6/dist-packages (from rsa<=4.5.0,>=3.1.2; python_version != \"3.4\"->awscli->-r requirements.txt (line 3)) (0.4.8)\n",
            "Installing collected packages: botocore, colorama, rsa, awscli\n",
            "  Found existing installation: botocore 1.17.24\n",
            "    Uninstalling botocore-1.17.24:\n",
            "      Successfully uninstalled botocore-1.17.24\n",
            "  Found existing installation: rsa 4.6\n",
            "    Uninstalling rsa-4.6:\n",
            "      Successfully uninstalled rsa-4.6\n",
            "Successfully installed awscli-1.18.105 botocore-1.17.28 colorama-0.4.3 rsa-4.5\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.colab-display-data+json": {
              "pip_warning": {
                "packages": [
                  "rsa"
                ]
              }
            }
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "plwzeI5M5fah",
        "colab_type": "text"
      },
      "source": [
        "# Step 1: Gathering and Labeling a Custom Dataset\n",
        "I ran the following commands within the toolkit to create my custom training dataset. I chose to use 2000 images within my dataset."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KbkvTkh4ItAL",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 514
        },
        "outputId": "926a28a6-f04c-4569-9ebe-0ef3b975cc8e"
      },
      "source": [
        "# Download training data for Shotgun class\n",
        "!python main.py downloader --classes 'Shotgun' --type_csv train --limit 2000"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[92m\n",
            "\t\t   ___   _____  ______            _    _    \n",
            "\t\t .'   `.|_   _||_   _ `.         | |  | |   \n",
            "\t\t/  .-.  \\ | |    | | `. \\ _   __ | |__| |_  \n",
            "\t\t| |   | | | |    | |  | |[ \\ [  ]|____   _| \n",
            "\t\t\\  `-'  /_| |_  _| |_.' / \\ \\/ /     _| |_  \n",
            "\t\t `.___.'|_____||______.'   \\__/     |_____|\n",
            "\t\u001b[0m\n",
            "\u001b[92m\n",
            "             _____                    _                 _             \n",
            "            (____ \\                  | |               | |            \n",
            "             _   \\ \\ ___  _ _ _ ____ | | ___   ____  _ | | ____  ____ \n",
            "            | |   | / _ \\| | | |  _ \\| |/ _ \\ / _  |/ || |/ _  )/ ___)\n",
            "            | |__/ / |_| | | | | | | | | |_| ( ( | ( (_| ( (/ /| |    \n",
            "            |_____/ \\___/ \\____|_| |_|_|\\___/ \\_||_|\\____|\\____)_|    \n",
            "                                                          \n",
            "        \u001b[0m\n",
            "    [INFO] | Downloading Shotgun.\u001b[0m\n",
            "\n",
            "\u001b[95mShotgun\u001b[0m\n",
            "    [INFO] | Downloading train images.\u001b[0m\n",
            "    [INFO] | [INFO] Found 429 online images for train.\u001b[0m\n",
            "    [INFO] | Limiting to 2000 images.\u001b[0m\n",
            "    [INFO] | Download of 429 images in train.\u001b[0m\n",
            "100% 429/429 [04:17<00:00,  1.66it/s]\n",
            "    [INFO] | Done!\u001b[0m\n",
            "    [INFO] | Creating labels for Shotgun of train.\u001b[0m\n",
            "    [INFO] | Labels creation completed.\u001b[0m\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fjG4RmaYI2oO",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 514
        },
        "outputId": "fa506d5b-061d-4050-c8c7-2472dd144831"
      },
      "source": [
        "# Download testing data for Shotgun class\n",
        "!python main.py downloader --classes 'Shotgun' --type_csv validation --limit 1000"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[92m\n",
            "\t\t   ___   _____  ______            _    _    \n",
            "\t\t .'   `.|_   _||_   _ `.         | |  | |   \n",
            "\t\t/  .-.  \\ | |    | | `. \\ _   __ | |__| |_  \n",
            "\t\t| |   | | | |    | |  | |[ \\ [  ]|____   _| \n",
            "\t\t\\  `-'  /_| |_  _| |_.' / \\ \\/ /     _| |_  \n",
            "\t\t `.___.'|_____||______.'   \\__/     |_____|\n",
            "\t\u001b[0m\n",
            "\u001b[92m\n",
            "             _____                    _                 _             \n",
            "            (____ \\                  | |               | |            \n",
            "             _   \\ \\ ___  _ _ _ ____ | | ___   ____  _ | | ____  ____ \n",
            "            | |   | / _ \\| | | |  _ \\| |/ _ \\ / _  |/ || |/ _  )/ ___)\n",
            "            | |__/ / |_| | | | | | | | | |_| ( ( | ( (_| ( (/ /| |    \n",
            "            |_____/ \\___/ \\____|_| |_|_|\\___/ \\_||_|\\____|\\____)_|    \n",
            "                                                          \n",
            "        \u001b[0m\n",
            "    [INFO] | Downloading Shotgun.\u001b[0m\n",
            "\n",
            "\u001b[95mShotgun\u001b[0m\n",
            "    [INFO] | Downloading validation images.\u001b[0m\n",
            "    [INFO] | [INFO] Found 19 online images for validation.\u001b[0m\n",
            "    [INFO] | Limiting to 1000 images.\u001b[0m\n",
            "    [INFO] | Download of 19 images in validation.\u001b[0m\n",
            "100% 19/19 [00:14<00:00,  1.30it/s]\n",
            "    [INFO] | Done!\u001b[0m\n",
            "    [INFO] | Creating labels for Shotgun of validation.\u001b[0m\n",
            "    [INFO] | Labels creation completed.\u001b[0m\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Y-PEs0bBR4St",
        "colab_type": "text"
      },
      "source": [
        "# Converting Labels to YOLOv4 Format\n",
        "The labels that we get from the toolkit are not in the proper YOLOv4 format. Luckily, with a few simple commands we can convert them to the proper format.\n",
        "\n",
        "Within the root OIDv4_ToolKit folder open the file classes.txt and edit it to have the classes you just downloaded, one per line.\n",
        "\n",
        "#### Open classes.txt file and add Shotgun "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oExKvMy06wxM",
        "colab_type": "text"
      },
      "source": [
        "Now run this command to convert the labels generated by OIDv4_Toolkit into YOLOv4 labels (this should convert both train and validation dataset labels):"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "e_jPONmpJNzE",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 123
        },
        "outputId": "94fda434-2c9e-4513-bb9d-245fb0307391"
      },
      "source": [
        "!python convert_annotations.py"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Currently in subdirectory: train\n",
            "Converting annotations for class:  Shotgun\n",
            "100% 429/429 [00:13<00:00, 30.82it/s]\n",
            "Currently in subdirectory: validation\n",
            "Converting annotations for class:  Shotgun\n",
            "100% 19/19 [00:00<00:00, 35.14it/s]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kecCAjo567EN",
        "colab_type": "text"
      },
      "source": [
        "Remove the old 'Label' folder in the train and validation folders which contains the non YOLOv4 formatted labels by running the following:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bp9__Z14JeBk",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!rm -r OID/Dataset/train/'Shotgun'/Label/"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7ofJYs3hJkAh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!rm -r OID/Dataset/validation/'Shotgun'/Label/"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1gU_AMOJSPOh",
        "colab_type": "text"
      },
      "source": [
        "#Step 2 Move train & valid into darknet/data, and rename it to obj and test"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aTORZ6tvOidU",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "04b919ba-3a6a-472e-e4d8-ccd8e7d123ab"
      },
      "source": [
        "%cd /content/gdrive/My Drive/weapon detector/darknet"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/content/gdrive/My Drive/weapon detector/darknet\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "y_m5t3y98EKi",
        "colab_type": "text"
      },
      "source": [
        "#Step 3: Configuring Files for Training\n",
        "## i) Cfg File\n",
        "Now you need to edit the .cfg to fit your needs based on your object detector. Open it up in a code or text editor to do so.\n",
        "\n",
        "I recommend having **batch = 64** and **subdivisions = 16** for ultimate results. If you run into any issues then up subdivisions to 32.\n",
        "\n",
        "Make the rest of the changes to the cfg based on how many classes you are training your detector on.\n",
        "\n",
        "**Note:** \n",
        "I set my **max_batches = 6000**, **steps = 4800, 5400**, I changed the **classes = 1** in the three YOLO layers and **filters = 18** in the three convolutional layers before the YOLO layers.\n",
        "\n",
        "How to Configure Your Variables:\n",
        "\n",
        "width = 416\n",
        "\n",
        "height = 416\n",
        "**(these can be any multiple of 32, 416 is standard, you can sometimes improve results by making value larger like 608 but will slow down training)**\n",
        "\n",
        "max_batches = (# of classes) * 2000\n",
        "**(but no less than 6000 so if you are training for 1, 2, or 3 classes it will be 6000, however detector for 5 classes would have max_batches=10000)**\n",
        "\n",
        "steps = (80% of max_batches), (90% of max_batches)\n",
        "**(so if your max_batches = 10000, then steps = 8000, 9000)**\n",
        "\n",
        "filters = (# of classes + 5) * 3\n",
        "**(so if you are training for one class then your filters = 18, but if you are training for 4 classes then your filters = 27)**\n",
        "\n",
        "\n",
        "**Optional:** If you run into memory issues or find the training taking a super long time. In each of the three yolo layers in the cfg, change one line from random = 1 to **random = 0** to speed up training but slightly reduce accuracy of model. Will also help save memory if you run into any memory issues."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "17QRiMg38_CB",
        "colab_type": "text"
      },
      "source": [
        "## ii) obj.names and obj.data \n",
        "Create a new file within a code or text editor called **obj.names** where you will have one class name per line in the same order as your classes.txt from the dataset generation step.\n",
        "\n",
        "You will also create a **obj.data** file and fill it in like this (change your number of classes accordingly, as well as your backup location)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "K2vayyls7oIK",
        "colab_type": "text"
      },
      "source": [
        "## iii) Generating train.txt and test.txt\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XEYzFhfdJ8pG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!python generate_train.py\n",
        "!python generate_test.py"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ckl2Edw19rXj",
        "colab_type": "text"
      },
      "source": [
        "# Step 4: Download pre-trained weights for the convolutional layers.\n",
        "This step downloads the weights for the convolutional layers of the YOLOv4 network. By using these weights it helps your custom object detector to be way more accurate and not have to train as long."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UAABdbvQHZ6k",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 321
        },
        "outputId": "d8821e91-7212-4844-d5f2-415704693c81"
      },
      "source": [
        "#weights\n",
        "!wget https://github.com/AlexeyAB/darknet/releases/download/darknet_yolo_v3_optimal/yolov4.conv.137"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2020-07-27 02:33:34--  https://github.com/AlexeyAB/darknet/releases/download/darknet_yolo_v3_optimal/yolov4.weights\n",
            "Resolving github.com (github.com)... 140.82.112.4\n",
            "Connecting to github.com (github.com)|140.82.112.4|:443... connected.\n",
            "HTTP request sent, awaiting response... 302 Found\n",
            "Location: https://github-production-release-asset-2e65be.s3.amazonaws.com/75388965/ba4b6380-889c-11ea-9751-f994f5961796?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=AKIAIWNJYAX4CSVEH53A%2F20200727%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Date=20200727T023334Z&X-Amz-Expires=300&X-Amz-Signature=1b609d50e84cdaea4a9d2a1168ded7172e43972e37fe0eace28879cc6d9ddb36&X-Amz-SignedHeaders=host&actor_id=0&repo_id=75388965&response-content-disposition=attachment%3B%20filename%3Dyolov4.weights&response-content-type=application%2Foctet-stream [following]\n",
            "--2020-07-27 02:33:34--  https://github-production-release-asset-2e65be.s3.amazonaws.com/75388965/ba4b6380-889c-11ea-9751-f994f5961796?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=AKIAIWNJYAX4CSVEH53A%2F20200727%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Date=20200727T023334Z&X-Amz-Expires=300&X-Amz-Signature=1b609d50e84cdaea4a9d2a1168ded7172e43972e37fe0eace28879cc6d9ddb36&X-Amz-SignedHeaders=host&actor_id=0&repo_id=75388965&response-content-disposition=attachment%3B%20filename%3Dyolov4.weights&response-content-type=application%2Foctet-stream\n",
            "Resolving github-production-release-asset-2e65be.s3.amazonaws.com (github-production-release-asset-2e65be.s3.amazonaws.com)... 52.216.26.228\n",
            "Connecting to github-production-release-asset-2e65be.s3.amazonaws.com (github-production-release-asset-2e65be.s3.amazonaws.com)|52.216.26.228|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 257717640 (246M) [application/octet-stream]\n",
            "Saving to: ‘yolov4.weights’\n",
            "\n",
            "yolov4.weights      100%[===================>] 245.78M  58.5MB/s    in 4.2s    \n",
            "\n",
            "2020-07-27 02:33:38 (58.6 MB/s) - ‘yolov4.weights’ saved [257717640/257717640]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RzNMUmyq-xdx",
        "colab_type": "text"
      },
      "source": [
        "# Step 5: Train Your Custom Object Detector!"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "w4G19PhS-79A",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!sed -i 's/OPENCV=0/OPENCV=1/' Makefile\n",
        "!sed -i 's/GPU=0/GPU=1/' Makefile\n",
        "!sed -i 's/CUDNN=0/CUDNN=1/' Makefile\n",
        "!sed -i 's/CUDNN_HALF=0/CUDNN_HALF=1/' Makefile"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pGF1QLb6-8vC",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!make"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6P8j0zBy-x7M",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!./darknet detector train data/obj.data cfg/yolov4-custom.cfg yolov4.conv.137 -dont_show -map"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ANk-qz4h_M_k",
        "colab_type": "text"
      },
      "source": [
        "**TRICK**: If for some reason you get an error or your Colab goes idle during training, you have not lost your partially trained model and weights!\n",
        "We can kick off training from our last saved weights file so that we don't have to restart! Just run the following command but with your backup location."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "f1y5c_d1_L9n",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!./darknet detector train data/obj.data cfg/yolov4-custom.cfg ./backup/yolov4-last.weights -dont_show -map"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ay4VlhVl_tGF",
        "colab_type": "text"
      },
      "source": [
        "# Step 6: Checking the Mean Average Precision (mAP) of the Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jJ2EnHjk_yr-",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#evaluate\n",
        "!./darknet detector map data/obj.data cfg/yolov4-custom.cfg ./backup/yolov4-custom_last.weights -thresh 0.4"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9hvmY7a5_9ab",
        "colab_type": "text"
      },
      "source": [
        "# Step 7: Run Custom Object Detector!!!"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZFjd60St_9qR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!./darknet detector demo data/obj.data cfg/yolov4-custom.cfg ./backup/yolov4-custom_last.weights -dont_show gun5.mp4 -i 0 -out_filename gun5-detection-0.4.mp4 -thresh 0.4"
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}