Training Classifiers.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 5,
  "metadata": {
    "colab": {
      "provenance": [],
      "collapsed_sections": [
        "137f4887-f9ce-4d88-8e8b-20dbad3c8486",
        "c3cca553-2ceb-49bd-9db1-7db2137dad84",
        "g7ab3k5roPq4",
        "Ynn1qpdnorCK",
        "lWUp82-Hp0QC",
        "io4DmhqIqp6x",
        "G9jIpTLYrvlJ",
        "bL_lTBfiuSK6",
        "92333c38-f758-4532-9e8a-833fe6ef475e",
        "fc3a4f36-1f39-4ab4-a54a-b4278802fc86",
        "2768f076-bb76-4601-a98a-18a615674ef2",
        "PKKqLb3_wFKL",
        "RDeQqL17zYxa"
      ]
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.8.0"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "c232caf944b04614954d0d521ee9d705": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_4a972ad5422c461198731885d5548170",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_23e951ad5fca46ea93fb25b8a507dbcd",
              "IPY_MODEL_a35cb8522fbd4a7095b2e5653e86aea0",
              "IPY_MODEL_9912afa54cae4adf9f54abb2948d281e"
            ]
          }
        },
        "4a972ad5422c461198731885d5548170": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "23e951ad5fca46ea93fb25b8a507dbcd": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_ab097c93812149629fab836be27b0093",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": "100%",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_50e7e489df504a3ba0a0de60213a36ae"
          }
        },
        "a35cb8522fbd4a7095b2e5653e86aea0": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_7699105472be4476b6aec631d41a0358",
            "_dom_classes": [],
            "description": "",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 46830571,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 46830571,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_f240398c019542fd87a473779645877f"
          }
        },
        "9912afa54cae4adf9f54abb2948d281e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_edc6839d31674265a789cafff49cfbd6",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 44.7M/44.7M [00:01&lt;00:00, 36.6MB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_07cfb9d9ec2d485cac1b1281793d8b72"
          }
        },
        "ab097c93812149629fab836be27b0093": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "50e7e489df504a3ba0a0de60213a36ae": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "7699105472be4476b6aec631d41a0358": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "f240398c019542fd87a473779645877f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "edc6839d31674265a789cafff49cfbd6": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "07cfb9d9ec2d485cac1b1281793d8b72": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        }
      }
    },
    "gpuClass": "standard"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# Loading data and splitting it"
      ],
      "metadata": {
        "id": "kPEDWLcmPmU2"
      },
      "id": "kPEDWLcmPmU2"
    },
    {
      "cell_type": "markdown",
      "source": [
        "The cell runs for about 4-5 minutes"
      ],
      "metadata": {
        "id": "T9p7AcuNYdfi"
      },
      "id": "T9p7AcuNYdfi"
    },
    {
      "cell_type": "code",
      "source": [
        "import gdown\n",
        "\n",
        "# dataset part 1\n",
        "url = \"https://drive.google.com/drive/folders/1AxqLHFCjtoDiJjD0lweiOzVuTDZgla3h\"\n",
        "gdown.download_folder(url, quiet=True, use_cookies=False)\n",
        "\n",
        "# dataset part 2\n",
        "url = \"https://drive.google.com/drive/folders/15z3pimTCFH9ywO9oNgpGC6Ff_04Z4O_z\"\n",
        "gdown.download_folder(url, quiet=True, use_cookies=False)\n",
        "\n",
        "!mv Dataset_Part_1/ dataset/\n",
        "!mv Dataset_Part_2/ dataset1/"
      ],
      "metadata": {
        "id": "s5eiWne_UGzB"
      },
      "id": "s5eiWne_UGzB",
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "The next two cells can run for about 8 - 15 minutes"
      ],
      "metadata": {
        "id": "aNnv1tDlZBjt"
      },
      "id": "aNnv1tDlZBjt"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5O_axGqSoD1w"
      },
      "source": [
        "!unzip dataset/ACDC_Test_Drops.zip -d /content/dataset/\n",
        "!rm dataset/ACDC_Test_Drops.zip\n",
        "!unzip dataset/ACDC_Train_Drops.zip -d /content/dataset/\n",
        "!rm dataset/ACDC_Train_Drops.zip\n",
        "!unzip dataset/ACDC_Val_Drops.zip -d /content/dataset/\n",
        "!rm dataset/ACDC_Val_Drops.zip\n",
        "!unzip dataset/clean_acdc.zip -d /content/dataset/clean_acdc\n",
        "!rm dataset/clean_acdc.zip\n",
        "!unzip dataset/clean_test_acdc.zip -d /content/dataset/clean_test_acdc\n",
        "!rm dataset/clean_test_acdc.zip\n",
        "!unzip dataset/clean_test_woodscape.zip -d /content/dataset/clean_test_woodscape\n",
        "!rm dataset/clean_test_woodscape.zip\n",
        "!unzip dataset/defocus_test_woodscape.zip -d /content/dataset/defocus_test_woodscape\n",
        "!rm dataset/defocus_test_woodscape.zip\n",
        "!unzip dataset/drops_test_woodscape.zip -d /content/dataset/drops_test_woodscape\n",
        "!rm dataset/drops_test_woodscape.zip\n",
        "!unzip dataset/full_loss_test_woodscape.zip -d /content/dataset/full_loss_test_woodscape\n",
        "!rm dataset/full_loss_test_woodscape.zip\n",
        "!unzip dataset/motion_test_acdc.zip -d /content/dataset/motion_test_acdc\n",
        "!rm dataset/motion_test_acdc.zip\n",
        "!unzip dataset/motion_test_woodscape.zip -d /content/dataset/motion_test_woodscape\n",
        "!rm dataset/motion_test_woodscape.zip\n",
        "!unzip dataset/partial_test_woodscape.zip -d /content/dataset/partial_test_woodscape\n",
        "!rm dataset/partial_test_woodscape.zip\n",
        "!unzip dataset/SoilingDatasetNotFishEye_Test.zip -d /content/dataset/\n",
        "!rm dataset/SoilingDatasetNotFishEye_Test.zip\n",
        "!unzip dataset/SoilingDatasetNotFishEye_Train.zip -d /content/dataset/\n",
        "!rm dataset/SoilingDatasetNotFishEye_Train.zip\n",
        "!unzip dataset/spatter_acdc.zip -d /content/dataset/spatter_acdc\n",
        "!rm dataset/spatter_acdc.zip\n",
        "!unzip dataset/spatter_test_acdc.zip -d /content/dataset/spatter_test_acdc\n",
        "!rm dataset/spatter_test_acdc.zip\n",
        "!unzip dataset/Tapmer_Test_FullLoss.zip -d /content/dataset/\n",
        "!rm dataset/Tapmer_Test_FullLoss.zip\n",
        "!unzip dataset/Tapmer_Train_FullLoss.zip -d /content/dataset/\n",
        "!rm dataset/Tapmer_Train_FullLoss.zip"
      ],
      "id": "5O_axGqSoD1w",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RUHkJZwBx9x2"
      },
      "source": [
        "!unzip dataset1/clean_driving.zip -d /content/dataset1/clean_driving\n",
        "!rm dataset1/clean_driving.zip\n",
        "!unzip dataset1/clean_woodscape.zip -d /content/dataset1/clean_woodscape\n",
        "!rm dataset1/clean_woodscape.zip\n",
        "!unzip dataset1/defocus_acdc.zip -d /content/dataset1/defocus_acdc\n",
        "!rm dataset1/defocus_acdc.zip\n",
        "!unzip dataset1/defocus_driving.zip -d /content/dataset1/defocus_driving\n",
        "!rm dataset1/defocus_driving.zip\n",
        "!unzip dataset1/defocus_woodscape.zip -d /content/dataset1/defocus_woodscape\n",
        "!rm dataset1/defocus_woodscape.zip\n",
        "!unzip dataset1/defocus_test_acdc.zip -d /content/dataset1/defocus_test_acdc\n",
        "!rm dataset1/defocus_test_acdc.zip\n",
        "!unzip dataset1/drops_woodscape.zip -d /content/dataset1/drops_woodscape\n",
        "!rm dataset1/drops_woodscape.zip\n",
        "!unzip dataset1/full_loss_acdc_1000.zip -d /content/dataset1/full_loss_acdc_1000\n",
        "!rm dataset1/full_loss_acdc_1000.zip\n",
        "!unzip dataset1/full_loss_woodscape.zip -d /content/dataset1/full_loss_woodscape\n",
        "!rm dataset1/full_loss_woodscape.zip\n",
        "!unzip dataset1/motion_acdc.zip -d /content/dataset1/motion_acdc\n",
        "!rm dataset1/motion_acdc.zip\n",
        "!unzip dataset1/motion_driving.zip -d /content/dataset1/motion_driving\n",
        "!rm dataset1/motion_driving.zip\n",
        "!unzip dataset1/motion_woodscape.zip -d /content/dataset1/motion_woodscape\n",
        "!rm dataset1/motion_woodscape.zip\n",
        "!unzip dataset1/partial_woodscape.zip -d /content/dataset1/partial_woodscape\n",
        "!rm dataset1/partial_woodscape.zip"
      ],
      "id": "RUHkJZwBx9x2",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1c3QPuKS284Z"
      },
      "source": [
        "!mv dataset/SoilingDatasetNotFishEye_Test /content/\n",
        "!mv dataset/SoilingDatasetNotFishEye_Train /content/\n",
        "\n",
        "!mv SoilingDatasetNotFishEye_Test test\n",
        "!mv SoilingDatasetNotFishEye_Train train\n",
        "\n",
        "!bash -c 'mv dataset/Tapmer_Train_FullLoss/*.* /content/train/full'\n",
        "!bash -c 'mv dataset/Tapmer_Test_FullLoss/*.* /content/test/full'"
      ],
      "id": "1c3QPuKS284Z",
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Question: Why we use dirs.sort()?\n",
        "\n",
        "[Answer](https://stackoverflow.com/questions/54773598/os-walk-different-folder-ordering-on-mac-and-linux)\n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "TLvKUJ6hM-8L"
      },
      "id": "TLvKUJ6hM-8L"
    },
    {
      "cell_type": "code",
      "source": [
        "import os\n",
        "\n",
        "def data_transfer(old_folder, new_folder, break_number):\n",
        "    root = \"\"\n",
        "    for root, dirs, files in os.walk(os.path.join(root, old_folder)):\n",
        "        dirs.sort()\n",
        "        i = 0\n",
        "        for file in files:\n",
        "            image_name = os.path.join(root, file)\n",
        "            try: os.replace(image_name, new_folder + image_name.split('/')[3])\n",
        "            except: os.replace(image_name, new_folder + image_name.split('/')[2])\n",
        "            i = i + 1\n",
        "            if i == break_number: break\n",
        "        print(i)\n",
        "        break"
      ],
      "metadata": {
        "id": "_zUkt19ocEpc"
      },
      "id": "_zUkt19ocEpc",
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Test Full"
      ],
      "metadata": {
        "id": "eyRpEeS2fRVw"
      },
      "id": "eyRpEeS2fRVw"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PXTFLVbx8it9"
      },
      "source": [
        "data_transfer(\"dataset1/full_loss_acdc_1000/full_loss\", \"test/full/\", 93)\n",
        "data_transfer(\"dataset1/full_loss_woodscape/full_loss\", \"test/full/\", 92)\n",
        "data_transfer(\"dataset/full_loss_test_woodscape/full_loss_test\", \"test/full/\", 92)"
      ],
      "id": "PXTFLVbx8it9",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Train Full"
      ],
      "metadata": {
        "id": "0Rgg10WAfa5K"
      },
      "id": "0Rgg10WAfa5K"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wfTeH_fH_9Oa"
      },
      "source": [
        "data_transfer(\"dataset1/full_loss_acdc_1000/full_loss\", \"train/full/\", -1)\n",
        "data_transfer(\"dataset1/full_loss_woodscape/full_loss\", \"train/full/\", -1)\n",
        "data_transfer(\"dataset/full_loss_test_woodscape/full_loss_test\", \"train/full/\", -1)"
      ],
      "id": "wfTeH_fH_9Oa",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c1lNzA79BkQE"
      },
      "source": [
        "Test Defocus"
      ],
      "id": "c1lNzA79BkQE"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kKyaAMR1Bd6r"
      },
      "source": [
        "!mkdir train/defocus\n",
        "!mkdir test/defocus\n",
        "\n",
        "data_transfer(\"dataset1/defocus_acdc/defocus\", \"test/defocus/\", 205)\n",
        "data_transfer(\"dataset/defocus_test_woodscape/defocus_test\", \"test/defocus/\", 205)\n",
        "data_transfer(\"dataset1/defocus_driving/defocus_driving\", \"test/defocus/\", 87)\n",
        "data_transfer(\"dataset1/defocus_test_acdc/defocus\", \"test/defocus/\", 102)"
      ],
      "id": "kKyaAMR1Bd6r",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Train Defocus"
      ],
      "metadata": {
        "id": "NXYAHHUbg5Lo"
      },
      "id": "NXYAHHUbg5Lo"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZBvR3Hz7Eama"
      },
      "source": [
        "data_transfer(\"dataset1/defocus_acdc/defocus\", \"train/defocus/\", -1)\n",
        "data_transfer(\"dataset/defocus_test_woodscape/defocus_test\", \"train/defocus/\", -1)\n",
        "data_transfer(\"dataset1/defocus_driving/defocus_driving\", \"train/defocus/\", -1)\n",
        "data_transfer(\"dataset1/defocus_test_acdc/defocus\", \"train/defocus/\", -1)\n",
        "data_transfer(\"dataset1/defocus_woodscape/defocus\", \"train/defocus/\", -1)"
      ],
      "id": "ZBvR3Hz7Eama",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gnZ_sdYbGHOO"
      },
      "source": [
        "Test Motion"
      ],
      "id": "gnZ_sdYbGHOO"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8A2xhphYGBMq"
      },
      "source": [
        "!mkdir train/motion\n",
        "!mkdir test/motion\n",
        "\n",
        "data_transfer(\"dataset1/motion_acdc/motion\", \"test/motion/\", 205)\n",
        "data_transfer(\"dataset/motion_test_woodscape/motion_test\", \"test/motion/\", 205)\n",
        "data_transfer(\"dataset1/motion_driving/motion_driving\", \"test/motion/\", 87)\n",
        "data_transfer(\"dataset/motion_test_acdc/motion\", \"test/motion/\", 102)"
      ],
      "id": "8A2xhphYGBMq",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bK9mNJyXiJtn"
      },
      "source": [
        "Train Motion"
      ],
      "id": "bK9mNJyXiJtn"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Iud-2u8jG7V2"
      },
      "source": [
        "data_transfer(\"dataset1/motion_acdc/motion\", \"train/motion/\", -1)\n",
        "data_transfer(\"dataset/motion_test_woodscape/motion_test\", \"train/motion/\", -1)\n",
        "data_transfer(\"dataset1/motion_driving/motion_driving\", \"train/motion/\", -1)\n",
        "data_transfer(\"dataset/motion_test_acdc/motion\", \"train/motion/\", -1)\n",
        "data_transfer(\"dataset1/motion_woodscape/motion\", \"train/motion/\", -1)"
      ],
      "id": "Iud-2u8jG7V2",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7XZc9VZH09rc"
      },
      "source": [
        "Test Partial"
      ],
      "id": "7XZc9VZH09rc"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UhGmJj970so3"
      },
      "source": [
        "data_transfer(\"dataset/spatter_test_acdc/spatter\", \"test/partial/\", 225)\n",
        "data_transfer(\"dataset/partial_test_woodscape/partial_test\", \"test/partial/\", 210)"
      ],
      "id": "UhGmJj970so3",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Train Partial"
      ],
      "metadata": {
        "id": "hWQusQzpjNhY"
      },
      "id": "hWQusQzpjNhY"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "61znsqpu3Hv0"
      },
      "source": [
        "data_transfer(\"dataset/spatter_acdc/spatter\", \"train/partial/\", -1)\n",
        "data_transfer(\"dataset/spatter_test_acdc/spatter\", \"train/partial/\", -1)\n",
        "data_transfer(\"dataset/partial_test_woodscape/partial_test\", \"train/partial/\", -1)\n",
        "data_transfer(\"dataset1/partial_woodscape/partial\", \"train/partial/\", -1)"
      ],
      "id": "61znsqpu3Hv0",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BB1gmK8d4Dta"
      },
      "source": [
        "Test Drops"
      ],
      "id": "BB1gmK8d4Dta"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gnSJ9VTE4Jbc"
      },
      "source": [
        "!mv dataset/ACDC_Val_Drops/*.* test/drops\n",
        "\n",
        "data_transfer(\"dataset/ACDC_Train_Drops\", \"test/drops/\", 181)\n",
        "data_transfer(\"dataset/drops_test_woodscape/drops_test\", \"test/drops/\", 181)"
      ],
      "id": "gnSJ9VTE4Jbc",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "O0o-SlVFkHyf"
      },
      "source": [
        "Train Drops"
      ],
      "id": "O0o-SlVFkHyf"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HtvujeWU555h"
      },
      "source": [
        "!mv dataset/ACDC_Test_Drops/*.* train/drops\n",
        "\n",
        "data_transfer(\"dataset/ACDC_Train_Drops\", \"train/drops/\", -1)\n",
        "data_transfer(\"dataset/drops_test_woodscape/drops_test\", \"train/drops/\", -1)\n",
        "data_transfer(\"dataset1/drops_woodscape/drops\", \"train/drops/\", -1)"
      ],
      "id": "HtvujeWU555h",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qQAkXH58-wc9"
      },
      "source": [
        "Test Clean"
      ],
      "id": "qQAkXH58-wc9"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MPhQ4fd7-yaw"
      },
      "source": [
        "data_transfer(\"dataset/clean_test_acdc/clean\", \"test/clean/\", 202)\n",
        "data_transfer(\"dataset/clean_test_woodscape/clean_test\", \"test/clean/\", 202)\n",
        "data_transfer(\"dataset1/clean_driving/clean_driving\", \"test/clean/\", 202)"
      ],
      "id": "MPhQ4fd7-yaw",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Train Clean"
      ],
      "metadata": {
        "id": "w0ms86jIldbO"
      },
      "id": "w0ms86jIldbO"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TD3kRpGs_uiF"
      },
      "source": [
        "data_transfer(\"dataset/clean_test_acdc/clean\", \"train/clean/\", -1)\n",
        "data_transfer(\"dataset/clean_acdc/clean\", \"train/clean/\", -1)\n",
        "data_transfer(\"dataset/clean_test_woodscape/clean_test\", \"train/clean/\", -1)\n",
        "data_transfer(\"dataset1/clean_driving/clean_driving\", \"train/clean/\", -1)\n",
        "data_transfer(\"dataset1/clean_woodscape/clean\", \"train/clean/\", -1)"
      ],
      "id": "TD3kRpGs_uiF",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "137f4887-f9ce-4d88-8e8b-20dbad3c8486"
      },
      "source": [
        "# Download the libraries and import them"
      ],
      "id": "137f4887-f9ce-4d88-8e8b-20dbad3c8486"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lOycWb6v7Udl"
      },
      "source": [
        "!pip install imgaug wandb\n",
        "!pip install ml-collections"
      ],
      "id": "lOycWb6v7Udl",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import torch\n",
        "import os\n",
        "import cv2\n",
        "import imgaug as ia\n",
        "import numpy as np\n",
        "import sys\n",
        "import json\n",
        "import wandb\n",
        "\n",
        "from imgaug import augmenters as iaa\n",
        "\n",
        "from collections import defaultdict\n",
        "\n",
        "from torch.utils.data import Dataset, DataLoader\n",
        "from torch import nn\n",
        "from torch.nn import functional as F\n",
        "\n",
        "from torchvision.models import resnet18 as resnet18\n",
        "from torchvision import transforms\n",
        "\n",
        "import matplotlib.pyplot as plt"
      ],
      "metadata": {
        "id": "FuOq2f1IaWef"
      },
      "id": "FuOq2f1IaWef",
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c3cca553-2ceb-49bd-9db1-7db2137dad84"
      },
      "source": [
        "# Data preparation"
      ],
      "id": "c3cca553-2ceb-49bd-9db1-7db2137dad84"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "e654ff26-4e73-4242-9564-8f9d46cf98ac"
      },
      "source": [
        "TRAIN_DATA_PATH = 'train/'\n",
        "TEST_DATA_PATH = 'test/'"
      ],
      "id": "e654ff26-4e73-4242-9564-8f9d46cf98ac",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-6sdAk5D7zOg"
      },
      "source": [
        "Config"
      ],
      "id": "-6sdAk5D7zOg"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0Aji7NR871Ks"
      },
      "source": [
        "BATCH_SIZE = 64\n",
        "TRAIN_LEN = 0\n",
        "TEST_LEN = 0\n",
        "IMAGE_WIDTH = 224\n",
        "IMAGE_HEIGHT = 224\n",
        "NUM_WORKERS = 4\n",
        "EPOCHS = 5\n",
        "DEVICE = 'cuda'\n",
        "SEED = 2020\n",
        "SPLIT_DATA = 0.2\n",
        "LR = 0.0001\n",
        "MODEL_DESC = ''"
      ],
      "id": "0Aji7NR871Ks",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "452f04ea-3fd2-41e8-b6f4-8159d8fa810d"
      },
      "source": [
        "class VehiclesDatasetDisk(Dataset):\n",
        "\n",
        "    def __init__(self, path, transform=None):\n",
        "        super().__init__()\n",
        "\n",
        "        self.path = path\n",
        "        self.transform = transform\n",
        "\n",
        "        self.data = []\n",
        "\n",
        "        self.class_name_to_idx = {}\n",
        "        self.idx_to_class_name = []\n",
        "\n",
        "        for class_idx, class_name in enumerate(sorted(os.listdir(path))):\n",
        "            self.class_name_to_idx[class_name] = class_idx\n",
        "            self.idx_to_class_name.append(class_name)\n",
        "\n",
        "            class_images_path = os.path.join(path, class_name)\n",
        "            for image_name in os.listdir(class_images_path):\n",
        "                image_path = os.path.join(class_images_path, image_name)\n",
        "                \n",
        "                # remember the path to the image, we will read it later\n",
        "                self.data.append((image_path, class_name))\n",
        "\n",
        "    def __getitem__(self, idx):\n",
        "        image_path, class_name = self.data[idx]\n",
        "        # disk reading and decoding jpeg\n",
        "        image = cv2.imread(image_path, cv2.IMREAD_COLOR)\n",
        "        if self.transform:\n",
        "            image = self.transform(image)\n",
        "        return image, self.class_name_to_idx[class_name]\n",
        "\n",
        "    def __len__(self):\n",
        "        return len(self.data)"
      ],
      "id": "452f04ea-3fd2-41e8-b6f4-8159d8fa810d",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "50636f17-efbe-4995-b5eb-fe23bf998b27"
      },
      "source": [
        "train_transform = transforms.Compose([\n",
        "    iaa.Sequential([\n",
        "        #iaa.Fliplr(0.5),\n",
        "        iaa.Resize({\"height\": IMAGE_HEIGHT, \"width\": IMAGE_WIDTH})\n",
        "    ]).augment_image,\n",
        "    transforms.ToTensor(),\n",
        "    transforms.Normalize((0.485, 0.456, 0.406), (0.229, 0.224, 0.225))\n",
        "])\n",
        "\n",
        "test_transform = transforms.Compose([\n",
        "    iaa.Sequential([\n",
        "        iaa.Resize({\"height\": IMAGE_HEIGHT, \"width\": IMAGE_WIDTH})\n",
        "    ]).augment_image,\n",
        "    transforms.ToTensor(),\n",
        "    transforms.Normalize((0.485, 0.456, 0.406), (0.229, 0.224, 0.225))\n",
        "])"
      ],
      "id": "50636f17-efbe-4995-b5eb-fe23bf998b27",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "231bcd51-bf68-4a8b-9c32-5bde992233b9",
        "outputId": "2e4f1a9c-731f-4983-ee79-44806e2de133"
      },
      "source": [
        "train_dataset = VehiclesDatasetDisk(TRAIN_DATA_PATH, train_transform)\n",
        "test_dataset = VehiclesDatasetDisk(TEST_DATA_PATH, test_transform)\n",
        "\n",
        "TRAIN_LEN = len(train_dataset)\n",
        "TEST_LEN = len(test_dataset)\n",
        "\n",
        "print('Train dataset size:', len(train_dataset))\n",
        "print('Test dataset size:', len(test_dataset))"
      ],
      "id": "231bcd51-bf68-4a8b-9c32-5bde992233b9",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Train dataset size: 18584\n",
            "Test dataset size: 4397\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 342
        },
        "id": "f76df791-bcf9-4b51-bdb6-d12824894f68",
        "outputId": "46498924-f2b2-40b5-8021-f9b2721e4e5e"
      },
      "source": [
        "image, class_idx = train_dataset[6]\n",
        "\n",
        "print('Image shape:', image.shape)\n",
        "print('Class idx:', class_idx)\n",
        "print('Class name:', train_dataset.idx_to_class_name[class_idx])\n",
        "\n",
        "plt.imshow(image[1], cmap='Greys_r')"
      ],
      "id": "f76df791-bcf9-4b51-bdb6-d12824894f68",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Image shape: torch.Size([3, 224, 224])\n",
            "Class idx: 0\n",
            "Class name: clean\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f5524311e50>"
            ]
          },
          "metadata": {},
          "execution_count": 61
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "57796277-b326-4472-a870-d9aad02c940a"
      },
      "source": [
        "train_dataloader = DataLoader(\n",
        "    train_dataset,\n",
        "    batch_size=BATCH_SIZE,\n",
        "    shuffle=True,\n",
        "    num_workers=NUM_WORKERS,\n",
        "    pin_memory=True,\n",
        "    drop_last=True\n",
        ")\n",
        "\n",
        "test_dataloader = DataLoader(\n",
        "    test_dataset,  \n",
        "    batch_size=BATCH_SIZE, \n",
        "    shuffle=False,\n",
        "    num_workers=NUM_WORKERS, \n",
        "    pin_memory=True, \n",
        "    drop_last=False\n",
        ")"
      ],
      "id": "57796277-b326-4472-a870-d9aad02c940a",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "g7ab3k5roPq4"
      },
      "source": [
        "# Prepare a model ResNet-18"
      ],
      "id": "g7ab3k5roPq4"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ae590c39-028a-45a8-8e94-75a9ae28b724",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 67,
          "referenced_widgets": [
            "c232caf944b04614954d0d521ee9d705",
            "4a972ad5422c461198731885d5548170",
            "23e951ad5fca46ea93fb25b8a507dbcd",
            "a35cb8522fbd4a7095b2e5653e86aea0",
            "9912afa54cae4adf9f54abb2948d281e",
            "ab097c93812149629fab836be27b0093",
            "50e7e489df504a3ba0a0de60213a36ae",
            "7699105472be4476b6aec631d41a0358",
            "f240398c019542fd87a473779645877f",
            "edc6839d31674265a789cafff49cfbd6",
            "07cfb9d9ec2d485cac1b1281793d8b72"
          ]
        },
        "outputId": "22a34ccb-b86d-43b2-ed01-65991360872d"
      },
      "source": [
        "from torchvision.models import resnet34 as resnet34\n",
        "\n",
        "model = resnet18(pretrained=True)"
      ],
      "id": "ae590c39-028a-45a8-8e94-75a9ae28b724",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Downloading: \"https://download.pytorch.org/models/resnet18-f37072fd.pth\" to /root/.cache/torch/hub/checkpoints/resnet18-f37072fd.pth\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "c232caf944b04614954d0d521ee9d705",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "  0%|          | 0.00/44.7M [00:00<?, ?B/s]"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ae3f8104-6ea2-4c52-8409-5fe8275099e1"
      },
      "source": [
        "num_ftrs = model.fc.in_features\n",
        "model.fc = nn.Linear(num_ftrs, 7)\n",
        "torch.nn.init.xavier_normal_(model.fc.weight)\n",
        "model"
      ],
      "id": "ae3f8104-6ea2-4c52-8409-5fe8275099e1",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "f145a4b1-b2a4-4c81-880e-98d4df0542e2"
      },
      "source": [
        "# Freeze all the weights of the model\n",
        "for param in model.parameters():\n",
        "    param.requires_grad = False\n",
        "\n",
        "# Unfreeze last linear layer\n",
        "for param in model.fc.parameters():\n",
        "    param.requires_grad = True"
      ],
      "id": "f145a4b1-b2a4-4c81-880e-98d4df0542e2",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "b65873f1-b228-4ef1-a704-c2464db1bcb9",
        "outputId": "50c00245-2639-41d7-d680-62069fc59fa1"
      },
      "source": [
        "print('Total number of parameters:', sum(param.numel() for param in model.parameters()))"
      ],
      "id": "b65873f1-b228-4ef1-a704-c2464db1bcb9",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Total number of parameters: 11180103\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "24636680-ca72-47dd-98aa-32e25e9661b7",
        "outputId": "c9c73ded-351f-43f4-b3d1-97cf6a64edb0"
      },
      "source": [
        "print('Number of trainable parameters:', sum(param.numel() for param in model.parameters() if param.requires_grad))"
      ],
      "id": "24636680-ca72-47dd-98aa-32e25e9661b7",
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Number of trainable parameters: 3591\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2b156465-8b4c-45c4-8343-75e944d02936"
      },
      "source": [
        "if torch.cuda.is_available():\n",
        "    model = model.cuda()"
      ],
      "id": "2b156465-8b4c-45c4-8343-75e944d02936",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ynn1qpdnorCK"
      },
      "source": [
        "# Prepare a model ResNet-34"
      ],
      "id": "Ynn1qpdnorCK"
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 67,
          "referenced_widgets": [
            "c232caf944b04614954d0d521ee9d705",
            "4a972ad5422c461198731885d5548170",
            "23e951ad5fca46ea93fb25b8a507dbcd",
            "a35cb8522fbd4a7095b2e5653e86aea0",
            "9912afa54cae4adf9f54abb2948d281e",
            "ab097c93812149629fab836be27b0093",
            "50e7e489df504a3ba0a0de60213a36ae",
            "7699105472be4476b6aec631d41a0358",
            "f240398c019542fd87a473779645877f",
            "edc6839d31674265a789cafff49cfbd6",
            "07cfb9d9ec2d485cac1b1281793d8b72"
          ]
        },
        "outputId": "22a34ccb-b86d-43b2-ed01-65991360872d",
        "id": "PwQBSkc5orCL"
      },
      "source": [
        "from torchvision.models import resnet34 as resnet34\n",
        "\n",
        "model = resnet34(pretrained=True)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Downloading: \"https://download.pytorch.org/models/resnet18-f37072fd.pth\" to /root/.cache/torch/hub/checkpoints/resnet18-f37072fd.pth\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "c232caf944b04614954d0d521ee9d705",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "  0%|          | 0.00/44.7M [00:00<?, ?B/s]"
            ]
          },
          "metadata": {}
        }
      ],
      "id": "PwQBSkc5orCL"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SZiZVe3OorCL"
      },
      "source": [
        "num_ftrs = model.fc.in_features\n",
        "model.fc = nn.Linear(num_ftrs, 7)\n",
        "torch.nn.init.xavier_normal_(model.fc.weight)\n",
        "model"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "SZiZVe3OorCL"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-Nd4ryG3orCL"
      },
      "source": [
        "# Freeze all the weights of the model\n",
        "for param in model.parameters():\n",
        "    param.requires_grad = False\n",
        "\n",
        "# Unfreeze last linear layer\n",
        "for param in model.fc.parameters():\n",
        "    param.requires_grad = True"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "-Nd4ryG3orCL"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ix2RfPBhorCM"
      },
      "source": [
        "print('Total number of parameters:', sum(param.numel() for param in model.parameters()))"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "ix2RfPBhorCM"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AiqExSf_orCM"
      },
      "source": [
        "print('Number of trainable parameters:', sum(param.numel() for param in model.parameters() if param.requires_grad))"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "AiqExSf_orCM"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-LqqzrFworCM"
      },
      "source": [
        "if torch.cuda.is_available():\n",
        "    model = model.cuda()"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "-LqqzrFworCM"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lWUp82-Hp0QC"
      },
      "source": [
        "# Prepare a model ResNeXt50"
      ],
      "id": "lWUp82-Hp0QC"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3xVWC2LJp0QC"
      },
      "source": [
        "import torchvision\n",
        "\n",
        "model = torchvision.models.resnext50_32x4d(pretrained=True)"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "3xVWC2LJp0QC"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ty5ur6dgp0QD"
      },
      "source": [
        "num_ftrs = model.fc.in_features\n",
        "model.fc = nn.Linear(num_ftrs, 7)\n",
        "torch.nn.init.xavier_normal_(model.fc.weight)\n",
        "model"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "Ty5ur6dgp0QD"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2CSLCgEvp0QD"
      },
      "source": [
        "# Freeze all the weights of the model\n",
        "for param in model.parameters():\n",
        "    param.requires_grad = False\n",
        "\n",
        "# Unfreeze last linear layer\n",
        "for param in model.fc.parameters():\n",
        "    param.requires_grad = True"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "2CSLCgEvp0QD"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ky82x7dmp0QD"
      },
      "source": [
        "print('Total number of parameters:', sum(param.numel() for param in model.parameters()))"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "ky82x7dmp0QD"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "omncY2cMp0QE"
      },
      "source": [
        "print('Number of trainable parameters:', sum(param.numel() for param in model.parameters() if param.requires_grad))"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "omncY2cMp0QE"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FBwYDBPJp0QE"
      },
      "source": [
        "if torch.cuda.is_available():\n",
        "    model = model.cuda()"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "FBwYDBPJp0QE"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "io4DmhqIqp6x"
      },
      "source": [
        "# Prepare a model EfficientNet_B0"
      ],
      "id": "io4DmhqIqp6x"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5D7k3C6nuvvC"
      },
      "source": [
        "!pip install timm"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "5D7k3C6nuvvC"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HbhhAppecZMz",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8911566e-18a6-4bad-a929-e0eaed2ec5b4"
      },
      "source": [
        "import timm\n",
        "\n",
        "model = timm.create_model('efficientnet_b0', num_classes=7, pretrained=True)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Downloading: \"https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_b0_ra-3dd342df.pth\" to /root/.cache/torch/hub/checkpoints/efficientnet_b0_ra-3dd342df.pth\n"
          ]
        }
      ],
      "id": "HbhhAppecZMz"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_FCXU-LJ6lne"
      },
      "source": [
        "# Freeze all the weights of the model\n",
        "for param in model.parameters():\n",
        "    param.requires_grad = False\n",
        "\n",
        "# Unfreeze last linear layer\n",
        "model.head = nn.Linear(1280,7)"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "_FCXU-LJ6lne"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fhgYerj5qp6y"
      },
      "source": [
        "print('Total number of parameters:', sum(param.numel() for param in model.parameters()))"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "fhgYerj5qp6y"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ojl1m-sjqp6y"
      },
      "source": [
        "print('Number of trainable parameters:', sum(param.numel() for param in model.parameters() if param.requires_grad))"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "ojl1m-sjqp6y"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tJuJbCGfqp6y"
      },
      "source": [
        "if torch.cuda.is_available():\n",
        "    model = model.cuda()"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "tJuJbCGfqp6y"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "G9jIpTLYrvlJ"
      },
      "source": [
        "# Prepare a model ViT-Tiny"
      ],
      "id": "G9jIpTLYrvlJ"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4wD5SfkkrvlJ"
      },
      "source": [
        "!pip install timm"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "4wD5SfkkrvlJ"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-lNyuR4BrvlK"
      },
      "source": [
        "import timm\n",
        "\n",
        "model = timm.create_model('vit_tiny_patch16_224', num_classes=7, pretrained=True)"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "-lNyuR4BrvlK"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0TYNfkI3rvlK"
      },
      "source": [
        "# Freeze all the weights of the model\n",
        "for param in model.parameters():\n",
        "    param.requires_grad = False\n",
        "\n",
        "# Unfreeze last linear layer\n",
        "model.head = nn.Linear(192,7)"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "0TYNfkI3rvlK"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3l5wYbxurvlK"
      },
      "source": [
        "print('Total number of parameters:', sum(param.numel() for param in model.parameters()))"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "3l5wYbxurvlK"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yyR6kpTdrvlK"
      },
      "source": [
        "print('Number of trainable parameters:', sum(param.numel() for param in model.parameters() if param.requires_grad))"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "yyR6kpTdrvlK"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c-U3MjAOrvlL"
      },
      "source": [
        "if torch.cuda.is_available():\n",
        "    model = model.cuda()"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "c-U3MjAOrvlL"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bL_lTBfiuSK6"
      },
      "source": [
        "# Prepare a model DeiT-Tiny"
      ],
      "id": "bL_lTBfiuSK6"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1FGkyQOQuSK7"
      },
      "source": [
        "!pip install timm"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "1FGkyQOQuSK7"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sdJqsBa9uSK7"
      },
      "source": [
        "import timm\n",
        "\n",
        "model = timm.create_model('deit_tiny_patch16_224', num_classes=7, pretrained=True)"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "sdJqsBa9uSK7"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lZgkDN20uSK7"
      },
      "source": [
        "# Freeze all the weights of the model\n",
        "for param in model.parameters():\n",
        "    param.requires_grad = False\n",
        "\n",
        "# Unfreeze last linear layer\n",
        "model.head = nn.Linear(192,7)"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "lZgkDN20uSK7"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TEAAmNX_uSK7"
      },
      "source": [
        "print('Total number of parameters:', sum(param.numel() for param in model.parameters()))"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "TEAAmNX_uSK7"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h60TvTSnuSK8"
      },
      "source": [
        "print('Number of trainable parameters:', sum(param.numel() for param in model.parameters() if param.requires_grad))"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "h60TvTSnuSK8"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xBtemqssuSK8"
      },
      "source": [
        "if torch.cuda.is_available():\n",
        "    model = model.cuda()"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "xBtemqssuSK8"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "92333c38-f758-4532-9e8a-833fe6ef475e"
      },
      "source": [
        "# Prepapre a training function and initialize optimizer"
      ],
      "id": "92333c38-f758-4532-9e8a-833fe6ef475e"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8071ad23-0a62-4864-8a6a-a8355149df70"
      },
      "source": [
        "optimizer = torch.optim.Adam(model.parameters(), lr=LR)"
      ],
      "id": "8071ad23-0a62-4864-8a6a-a8355149df70",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "01191fff-79bd-4740-9925-81a0902d811e"
      },
      "source": [
        "As an indicator of quality, we will use 'Accuracy'"
      ],
      "id": "01191fff-79bd-4740-9925-81a0902d811e"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "a2f01db9-7d79-4f55-8df0-fa181e9f1bc5"
      },
      "source": [
        "def accuracy_wb(dataloader, model, label):\n",
        "    model = model.eval()\n",
        "    tp = 0\n",
        "    with torch.no_grad():\n",
        "        tracker = tqdm(dataloader, total=len(dataloader))\n",
        "        for batch in tracker:\n",
        "            image_batch, class_idx_batch = batch\n",
        "            if torch.cuda.is_available():\n",
        "                image_batch = image_batch.cuda()\n",
        "                class_idx_batch = class_idx_batch.cuda()\n",
        "            \n",
        "            logits = model(image_batch)\n",
        "            classes = torch.argmax(logits, dim=1)\n",
        "            tp += (class_idx_batch == classes).sum().item()\n",
        "\n",
        "        acc = tp / len(dataloader.dataset)\n",
        "        wandb.log({label: acc})\n",
        "        print(label, acc)\n",
        "    \n",
        "    return acc"
      ],
      "id": "a2f01db9-7d79-4f55-8df0-fa181e9f1bc5",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0b81d5d5-d756-42bb-84b4-fec99f862673"
      },
      "source": [
        "Start the training process"
      ],
      "id": "0b81d5d5-d756-42bb-84b4-fec99f862673"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "361d7f2c-f7b3-46cb-bd80-ee24cf9fb310"
      },
      "source": [
        "from tqdm import tqdm_notebook as tqdm\n",
        "\n",
        "def train(model, num_epochs, prefix):\n",
        "    \n",
        "    wandb.init(project=\"eye-vision\", config={\n",
        "    \"TRAIN_LEN\": TRAIN_LEN,\n",
        "    \"TEST_LEN\": TEST_LEN,\n",
        "    \"BATCH_SIZE\": BATCH_SIZE,\n",
        "    \"EPOCHS\": EPOCHS,\n",
        "    \"IMAGE_WIDTH\": IMAGE_WIDTH,\n",
        "    \"IMAGE_HEIGHT\": IMAGE_HEIGHT,\n",
        "    \"NUM_WORKERS\": NUM_WORKERS,\n",
        "    \"SPLIT_DATA\": SPLIT_DATA,\n",
        "    \"LR\": LR,\n",
        "    \"MODEL_DESC\": MODEL_DESC\n",
        "    })\n",
        "\n",
        "    wandb.watch(model)\n",
        "\n",
        "    best_acc = 0.2\n",
        "\n",
        "    for epoch in range(num_epochs):\n",
        "        print('\\nEpoch', epoch)\n",
        "        model = model.train()\n",
        "\n",
        "        tracker = tqdm(train_dataloader, total=len(train_dataloader))\n",
        "        for batch in tracker:\n",
        "            optimizer.zero_grad()\n",
        "            \n",
        "            image_batch, class_idx_batch = batch\n",
        "            if torch.cuda.is_available():\n",
        "                image_batch = image_batch.cuda()\n",
        "                class_idx_batch = class_idx_batch.cuda()\n",
        "\n",
        "            logits = model(image_batch)\n",
        "            loss = F.cross_entropy(logits, class_idx_batch)\n",
        "            loss.backward()\n",
        "\n",
        "            wandb.log({'train_loss': loss})\n",
        "\n",
        "            optimizer.step()\n",
        "\n",
        "        train_acc = accuracy_wb(train_dataloader, model, 'train_accuracy')\n",
        "        test_acc = accuracy_wb(test_dataloader, model, 'test_accuracy')\n",
        "\n",
        "        if (test_acc >= best_acc):\n",
        "          best_acc = test_acc\n",
        "          torch.save(model.state_dict(),'drive2/Diplom/ResNet18-64/' + prefix + 'model.pth')\n",
        "          torch.save(optimizer.state_dict(),'drive2/Diplom/ResNet18-64/' + prefix + 'optimizer.pth')"
      ],
      "id": "361d7f2c-f7b3-46cb-bd80-ee24cf9fb310",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fc3a4f36-1f39-4ab4-a54a-b4278802fc86"
      },
      "source": [
        "# Stage 1 - Training linear layer"
      ],
      "id": "fc3a4f36-1f39-4ab4-a54a-b4278802fc86"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "d7e40c5c-478b-4549-bc75-2f6507e040a1"
      },
      "source": [
        "prefix = '_stage_1_'\n",
        "MODEL_DESC = 'ResNet18' + prefix\n",
        "train(model, EPOCHS, MODEL_DESC)"
      ],
      "id": "d7e40c5c-478b-4549-bc75-2f6507e040a1",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2768f076-bb76-4601-a98a-18a615674ef2"
      },
      "source": [
        "# Stage 2 - Training last layer in main architecture from ResNet"
      ],
      "id": "2768f076-bb76-4601-a98a-18a615674ef2"
    },
    {
      "cell_type": "code",
      "source": [
        "import torchvision\n",
        "from torchvision.models import resnet34 as resnet34\n",
        "from torchvision.models import resnet18 as resnet18\n",
        "\n",
        "#model = torchvision.models.resnext50_32x4d(pretrained=True)\n",
        "#model = resnet34(pretrained=True)\n",
        "model = resnet18(pretrained=True)"
      ],
      "metadata": {
        "id": "wjaRSp0Bz5ui"
      },
      "id": "wjaRSp0Bz5ui",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rqAKN5xunvma"
      },
      "source": [
        "num_ftrs = model.fc.in_features\n",
        "model.fc = nn.Linear(num_ftrs, 7)\n",
        "torch.nn.init.xavier_normal_(model.fc.weight)"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "rqAKN5xunvma"
    },
    {
      "cell_type": "code",
      "source": [
        "model.load_state_dict(torch.load('drive2/Paper/ResNet18-64/ResNet18_stage_1_model.pth'))\n",
        "#model.load_state_dict(torch.load('drive2/Paper/ResNet34-64/ResNet34_stage_1_model.pth'))\n",
        "#model.load_state_dict(torch.load('drive2/Paper/ResNetX50-64/ResNetX50_stage_1_model.pth'))"
      ],
      "metadata": {
        "id": "F6TBPEh80KTQ"
      },
      "id": "F6TBPEh80KTQ",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MNXgRlDRnhZ8"
      },
      "source": [
        "# Freeze all the weights of the model\n",
        "for param in model.parameters():\n",
        "    param.requires_grad = False\n",
        "\n",
        "# Unfreeze last layer in the main architecture\n",
        "for param in model.layer4.parameters():\n",
        "    param.requires_grad = True\n",
        "\n",
        "# Unfreeze last linear layer\n",
        "for param in model.fc.parameters():\n",
        "    param.requires_grad = True"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "MNXgRlDRnhZ8"
    },
    {
      "cell_type": "markdown",
      "source": [
        "If you need to change the learning rate"
      ],
      "metadata": {
        "id": "jcQ4GAKFaI25"
      },
      "id": "jcQ4GAKFaI25"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aVRHgnYtnhZ9"
      },
      "source": [
        "sd = optimizer.state_dict()\n",
        "sd['param_groups'][0]['lr'] = 0.0003\n",
        "optimizer.load_state_dict(sd)\n",
        "LR = 0.0003"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "aVRHgnYtnhZ9"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bsnXYVvTnhZ9"
      },
      "source": [
        "prefix = '_stage_2_'\n",
        "MODEL_DESC = 'ResNet18' + prefix\n",
        "train(model, EPOCHS, MODEL_DESC)"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "bsnXYVvTnhZ9"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PKKqLb3_wFKL"
      },
      "source": [
        "# Stage 2.1 - Training last layer in main architecture from other models"
      ],
      "id": "PKKqLb3_wFKL"
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install timm"
      ],
      "metadata": {
        "id": "ZxcPxMo4xwCZ"
      },
      "id": "ZxcPxMo4xwCZ",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "EfficientNet_B0"
      ],
      "metadata": {
        "id": "ksXi21yXwZRJ"
      },
      "id": "ksXi21yXwZRJ"
    },
    {
      "cell_type": "code",
      "source": [
        "model = timm.create_model('efficientnet_b0', num_classes=7, pretrained=True)\n",
        "model.head = nn.Linear(1280,7)\n",
        "model.load_state_dict(torch.load('drive2/Paper/EfficNetB0-64/efficientnet_b0_stage_1_model.pth'))\n",
        "\n",
        "# Freeze all the weights of the model\n",
        "for param in model.parameters():\n",
        "    param.requires_grad = False\n",
        "\n",
        "# Unfreeze last layer in the main architecture\n",
        "i = 0\n",
        "for param in model.blocks.parameters():\n",
        "  i = i + 1\n",
        "  if i > 192: param.requires_grad = True\n",
        "\n",
        "# Unfreeze last linear layer\n",
        "model.head = nn.Linear(1280,7)"
      ],
      "metadata": {
        "id": "ffPRVFkEx-96"
      },
      "execution_count": null,
      "outputs": [],
      "id": "ffPRVFkEx-96"
    },
    {
      "cell_type": "code",
      "source": [
        "prefix = '_stage_2_'\n",
        "MODEL_DESC = 'EfficientNet_B0' + prefix\n",
        "train(model, EPOCHS, MODEL_DESC)"
      ],
      "metadata": {
        "id": "C-reEMsxxDmT"
      },
      "id": "C-reEMsxxDmT",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "ViT-Tiny"
      ],
      "metadata": {
        "id": "wDcP-uXPxrCi"
      },
      "id": "wDcP-uXPxrCi"
    },
    {
      "cell_type": "code",
      "source": [
        "model = timm.create_model('vit_tiny_patch16_224', num_classes=7, pretrained=True)\n",
        "model.head = nn.Linear(192,7)\n",
        "model.load_state_dict(torch.load('drive2/Paper/VIT-64/VIT_stage_1_model.pth'))\n",
        "\n",
        "# Freeze all the weights of the model\n",
        "for param in model.parameters():\n",
        "    param.requires_grad = False\n",
        "\n",
        "# Unfreeze last layer in the main architecture\n",
        "i = 0\n",
        "for param in model.blocks.parameters():\n",
        "  i = i + 1\n",
        "  if i > 132: param.requires_grad = True\n",
        "\n",
        "# Unfreeze last linear layer\n",
        "model.head = nn.Linear(192,7)"
      ],
      "metadata": {
        "id": "D3Jc5RNjx9Ma"
      },
      "execution_count": null,
      "outputs": [],
      "id": "D3Jc5RNjx9Ma"
    },
    {
      "cell_type": "code",
      "source": [
        "prefix = '_stage_2_'\n",
        "MODEL_DESC = 'vit_tiny' + prefix\n",
        "train(model, EPOCHS, MODEL_DESC)"
      ],
      "metadata": {
        "id": "XiqRul3jxrCj"
      },
      "execution_count": null,
      "outputs": [],
      "id": "XiqRul3jxrCj"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RDeQqL17zYxa"
      },
      "source": [
        "# Stage 3 - Testing and inference"
      ],
      "id": "RDeQqL17zYxa"
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install timm"
      ],
      "metadata": {
        "id": "oarCkjsx1yM0"
      },
      "execution_count": null,
      "outputs": [],
      "id": "oarCkjsx1yM0"
    },
    {
      "cell_type": "markdown",
      "source": [
        "ResNet.."
      ],
      "metadata": {
        "id": "q5FpHoDf1gz_"
      },
      "id": "q5FpHoDf1gz_"
    },
    {
      "cell_type": "code",
      "source": [
        "import torchvision\n",
        "from torchvision.models import resnet34 as resnet34\n",
        "from torchvision.models import resnet18 as resnet18\n",
        "\n",
        "#model = torchvision.models.resnext50_32x4d(pretrained=True)\n",
        "#model = resnet34(pretrained=True)\n",
        "model = resnet18(pretrained=True)\n",
        "num_ftrs = model.fc.in_features\n",
        "model.fc = nn.Linear(num_ftrs, 7)\n",
        "torch.nn.init.xavier_normal_(model.fc.weight)\n",
        "\n",
        "model.load_state_dict(torch.load('drive2/Paper/ResNet18-64/ResNet18_stage_2_model.pth'))\n",
        "#model.load_state_dict(torch.load('drive2/Paper/ResNet34-64/ResNet34_stage_2_model.pth'))\n",
        "#model.load_state_dict(torch.load('drive2/Paper/ResNetX50-64/ResNetX50_stage_2_model.pth'))"
      ],
      "metadata": {
        "id": "e78w5D5U1isF"
      },
      "id": "e78w5D5U1isF",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "EfficientNet_B0"
      ],
      "metadata": {
        "id": "2S13TcKS1PyT"
      },
      "id": "2S13TcKS1PyT"
    },
    {
      "cell_type": "code",
      "source": [
        "model = timm.create_model('efficientnet_b0', num_classes=7, pretrained=True)\n",
        "model.head = nn.Linear(1280,7)\n",
        "model.load_state_dict(torch.load('drive2/Paper/EfficNetB0-64/efficientnet_b0_stage_2_model.pth'))"
      ],
      "metadata": {
        "id": "UznaQmxu1PyU"
      },
      "execution_count": null,
      "outputs": [],
      "id": "UznaQmxu1PyU"
    },
    {
      "cell_type": "markdown",
      "source": [
        "ViT-Tiny"
      ],
      "metadata": {
        "id": "p0W9fydd1PyU"
      },
      "id": "p0W9fydd1PyU"
    },
    {
      "cell_type": "code",
      "source": [
        "model = timm.create_model('vit_tiny_patch16_224', num_classes=7, pretrained=True)\n",
        "model.head = nn.Linear(192,7)\n",
        "model.load_state_dict(torch.load('drive2/Paper/VIT-64/VIT_stage_2_model.pth'))"
      ],
      "metadata": {
        "id": "f5-4Rh-l1PyU"
      },
      "execution_count": null,
      "outputs": [],
      "id": "f5-4Rh-l1PyU"
    },
    {
      "cell_type": "markdown",
      "source": [
        "DeiT-Tiny"
      ],
      "metadata": {
        "id": "Jc7kHU-12AWc"
      },
      "id": "Jc7kHU-12AWc"
    },
    {
      "cell_type": "code",
      "source": [
        "model = timm.create_model('deit_tiny_patch16_224', num_classes=7, pretrained=True)\n",
        "model.head = nn.Linear(192,7)\n",
        "model.load_state_dict(torch.load('drive2/Paper/DeIT-64/DeIT_stage_1_model.pth'))"
      ],
      "metadata": {
        "id": "m1nrPHpv2AWc"
      },
      "execution_count": null,
      "outputs": [],
      "id": "m1nrPHpv2AWc"
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Important**"
      ],
      "metadata": {
        "id": "MkFr2v7h10RQ"
      },
      "id": "MkFr2v7h10RQ"
    },
    {
      "cell_type": "code",
      "source": [
        "model.eval()"
      ],
      "metadata": {
        "id": "uvnUfGv3ztR5"
      },
      "id": "uvnUfGv3ztR5",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "apw2grW3y5XR"
      },
      "source": [
        "**Load image**"
      ],
      "id": "apw2grW3y5XR"
    },
    {
      "cell_type": "code",
      "source": [
        "my_clsidx_to_labels = ['clean','defocus','drops','full', 'motion', 'partial', 'strong']"
      ],
      "metadata": {
        "id": "qwmegL2JzCgM"
      },
      "id": "qwmegL2JzCgM",
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 241
        },
        "outputId": "5c5453d9-6aaa-4522-a6d3-5453c5a17b63",
        "id": "xW09B0qTy5XR"
      },
      "source": [
        "from google.colab.patches import cv2_imshow\n",
        "\n",
        "image = cv2.imread(\"2101_RV.png\")\n",
        "h, w = image.shape[:2]\n",
        "image = cv2.resize(image, (224, 224))\n",
        "cv2_imshow(image)\n",
        "image_tensor = torch.Tensor(image).float() / 255\n",
        "mean = torch.Tensor([0.485, 0.456, 0.406])\n",
        "std = torch.Tensor([0.229, 0.224, 0.225])\n",
        "\n",
        "image_tensor = (image_tensor - mean) / std\n",
        "image_tensor = image_tensor.permute(2, 0, 1)\n",
        "batch = image_tensor.unsqueeze(0)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<PIL.Image.Image image mode=RGB size=224x224 at 0x7F59EE949110>"
            ]
          },
          "metadata": {}
        }
      ],
      "id": "xW09B0qTy5XR"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nM01S_tay5XS"
      },
      "source": [
        "**Get Prediciton**"
      ],
      "id": "nM01S_tay5XS"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pd8TG9Bqy5XS"
      },
      "source": [
        "if torch.cuda.is_available():\n",
        "    batch = batch.cuda()\n",
        "\n",
        "if torch.cuda.is_available():\n",
        "    model = model.cuda()"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "pd8TG9Bqy5XS"
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jeMMm4pzy5XS"
      },
      "source": [
        "with torch.no_grad():\n",
        "    logits = model(batch)"
      ],
      "execution_count": null,
      "outputs": [],
      "id": "jeMMm4pzy5XS"
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4eeaba51-6f14-4eb5-ab5a-158ffb4756dd",
        "id": "7n8Sg5poy5XS"
      },
      "source": [
        "probabilities = torch.nn.Softmax(dim=-1)(logits)\n",
        "print(probabilities)\n",
        "sortedProba = torch.argsort(probabilities, dim=-1, descending=True)\n",
        "print(sortedProba)\n",
        "text = \"\"\n",
        "for (i, idx) in enumerate(sortedProba[0, :3]):\n",
        "  if (i == 2):\n",
        "    text = text + \"{}: {:.2f}%\".format(my_clsidx_to_labels[idx.item()].strip(),probabilities[0, idx.item()] * 100)\n",
        "  else:\n",
        "    text = text + \"{}: {:.2f}% \\n\".format(my_clsidx_to_labels[idx.item()].strip(),probabilities[0, idx.item()] * 100)\n",
        "  \n",
        "print(text)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "tensor([[1.1050e-08, 4.0211e-12, 1.0573e-11, 6.4044e-13, 2.2486e-12, 1.0000e+00,\n",
            "         1.3114e-10]], device='cuda:0')\n",
            "tensor([[5, 0, 6, 2, 1, 4, 3]], device='cuda:0')\n",
            "partial: 100.00% \n",
            "clean: 0.00% \n",
            "strong: 0.00%\n"
          ]
        }
      ],
      "id": "7n8Sg5poy5XS"
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0k7M8mXSy5XS"
      },
      "source": [
        "**Inference time**"
      ],
      "id": "0k7M8mXSy5XS"
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ffe4e63c-4487-40bb-9b09-9bdff2bdc44c",
        "id": "vR7i_-fpy5XT"
      },
      "source": [
        "%%timeit\n",
        "with torch.no_grad():\n",
        "    logits = model(batch)\n",
        "    if torch.cuda.is_available():\n",
        "        torch.cuda.synchronize()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "100 loops, best of 5: 12.5 ms per loop\n"
          ]
        }
      ],
      "id": "vR7i_-fpy5XT"
    }
  ]
}