MNIST_digit_classifier_part1.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "MNIST_digit_classifier_part1.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyPYg/VveU3kE6da6pjpsod0",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU",
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "85c74e8d056e4a7b89c487e6ec3703f4": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_4d68cfa0220a4759ad3e25d91d60ba32",
              "IPY_MODEL_aaa779e7437d49f5b97df3e7cab52bde",
              "IPY_MODEL_234c36fb6b0f409383e43cc4f0c05201"
            ],
            "layout": "IPY_MODEL_78bc6b0199a24fde832ff80e61103433"
          }
        },
        "4d68cfa0220a4759ad3e25d91d60ba32": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_df4fa31d0fc34b46acaaf31d3216df3c",
            "placeholder": "​",
            "style": "IPY_MODEL_e636bceeb0df46cbb00a7331f0b28d45",
            "value": "100%"
          }
        },
        "aaa779e7437d49f5b97df3e7cab52bde": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_4a6a4cb8711f4f0a94dd3c3b41a20931",
            "max": 87319819,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_1bdda5e2efdb40b69181aa958485ed41",
            "value": 87319819
          }
        },
        "234c36fb6b0f409383e43cc4f0c05201": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_6081239db21e421183d894d1967023db",
            "placeholder": "​",
            "style": "IPY_MODEL_29c857194b1d4ab19c988ee248dbf54a",
            "value": " 83.3M/83.3M [00:00&lt;00:00, 101MB/s]"
          }
        },
        "78bc6b0199a24fde832ff80e61103433": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "df4fa31d0fc34b46acaaf31d3216df3c": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "e636bceeb0df46cbb00a7331f0b28d45": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "4a6a4cb8711f4f0a94dd3c3b41a20931": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "1bdda5e2efdb40b69181aa958485ed41": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "6081239db21e421183d894d1967023db": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "29c857194b1d4ab19c988ee248dbf54a": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/ozvnnn/MNIST_fastai/blob/main/MNIST_digit_classifier_part1.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "In this notebook, I am going to make digit classifier with explanations. In this part, we'll create learner and export it. And, in part 2, we make predictions. First cell includes command for upgrade fastai on colab and import the necessary parts.\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "s1oDgppN36HP"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Yqgyk3O33nbi"
      },
      "outputs": [],
      "source": [
        "! [ -e /content ] && pip install -Uqq fastai  # upgrade fastai on colab\n",
        "! pip install fastbook\n",
        "from fastai.vision.all import * \n",
        "from fastbook import * \n",
        "import ipywidgets as widgets"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "First thing we should have done must be downloading necessary data and creating datablock. First cell downloads and extracts data to 'path'. In dataset, we have a different directory for testing and tranining sets and also for every digit.We can take a look what's inside of data with \"(path).ls()\"\n"
      ],
      "metadata": {
        "id": "P0YOj6uD5EF_"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "path = untar_data(URLs.MNIST)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 37
        },
        "id": "nWwGV_Fh68T0",
        "outputId": "ba92a139-a7ee-43e6-88a2-85e7b0332643"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "\n",
              "<style>\n",
              "    /* Turns off some styling */\n",
              "    progress {\n",
              "        /* gets rid of default border in Firefox and Opera. */\n",
              "        border: none;\n",
              "        /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
              "        background-size: auto;\n",
              "    }\n",
              "    .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
              "        background: #F44336;\n",
              "    }\n",
              "</style>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "\n",
              "    <div>\n",
              "      <progress value='15687680' class='' max='15683414' style='width:300px; height:20px; vertical-align: middle;'></progress>\n",
              "      100.03% [15687680/15683414 00:01<00:00]\n",
              "    </div>\n",
              "    "
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "To show training and testing:\n"
      ],
      "metadata": {
        "id": "GnTqCZVY9TWh"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "path.ls()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UqGXNDZ-9VF8",
        "outputId": "7c9e8b72-b5bd-47fc-8bfc-3a42c7b9b629"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(#2) [Path('/root/.fastai/data/mnist_png/training'),Path('/root/.fastai/data/mnist_png/testing')]"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "To show digit files:\n"
      ],
      "metadata": {
        "id": "oGffGY_h9W2G"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "(path/\"training\").ls()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UL04bnPH9YV4",
        "outputId": "9c82bd6a-12e0-4ca1-f565-7c8c609b48c4"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(#10) [Path('/root/.fastai/data/mnist_png/training/7'),Path('/root/.fastai/data/mnist_png/training/8'),Path('/root/.fastai/data/mnist_png/training/2'),Path('/root/.fastai/data/mnist_png/training/9'),Path('/root/.fastai/data/mnist_png/training/3'),Path('/root/.fastai/data/mnist_png/training/1'),Path('/root/.fastai/data/mnist_png/training/0'),Path('/root/.fastai/data/mnist_png/training/5'),Path('/root/.fastai/data/mnist_png/training/6'),Path('/root/.fastai/data/mnist_png/training/4')]"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now, we'll create our datablock.\n"
      ],
      "metadata": {
        "id": "wtE6yRNu9cbI"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "digit_block = DataBlock(\n",
        "        blocks=(ImageBlock, CategoryBlock),\n",
        "        get_items=get_image_files,\n",
        "        splitter=RandomSplitter(valid_pct=0.2, seed=2),\n",
        "        get_y=parent_label,\n",
        "        batch_tfms=aug_transforms(mult=2., do_flip=False)) ##do_flip = False is important here because flipping the digits may gives us unwanted results since digit are meanless when they are flipped."
      ],
      "metadata": {
        "id": "LG-DPJ3d9vgT"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Last cell created our datablock not dataset or dataloader since we didn't give it our data. Let's give our data and make a little observation."
      ],
      "metadata": {
        "id": "WAqe7Ar__AaY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "dls = digit_block.dataloaders(path/\"training\")\n",
        "dls.train.show_batch(max_n=4, nrows=1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 192
        },
        "id": "RucOLD8E_N91",
        "outputId": "74304c90-f896-4e7d-fcc6-650bcc302cc9"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x216 with 4 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAqwAAACvCAYAAAA4yYy3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAT0klEQVR4nO3deWxV5brH8eel0NICVRARyiwIZUhBbWWGgwODBhQhigMOGELAi8rwjwIxV0UxORHDOQIh6NUriKLViBdzCyKTA0gpocUqxaKAjDLPQ2HdP+45ifF5Nux2d3e/u/v7SU7OOb+s9a7Xsrp9XO5nPS4IAgEAAAB8VSPWGwAAAACuhIIVAAAAXqNgBQAAgNcoWAEAAOA1ClYAAAB4jYIVAAAAXqNgBQAAgNcoWCvAOXfqL/+55Jz7R6z3BYTLOTfSOfeTc+60c67UOdcn1nsCroTPXcQr59xq59y5P92722K9p3hUM9YbiEdBENT99/92ztUVkf0i8nHsdgSEzzl3l4i8LiIPisgPItIktjsCro7PXcS5/wiCYEGsNxHPKFgjN1xEDorIulhvBAjTf4rIS0EQrP/X/98Ty80AFcDnLpBg+EpA5B4Xkf8OmHGLOOCcSxKRbBG53jn3i3Pud+fcP51zqbHeG1AOfO4i3rzmnDvknPvWOfe3WG8mHjl+3yvOOddSRHaISNsgCH6N9X6Aq3HOZcj/P1HdJCJDROSiiHwuIquDIJgay70B4eBzF/HGOddNRIpF5IKIjBSRf4pI1yAISmO6sTjDE9bIjBKRb/jQRBw5+6///kcQBPuCIDgkIm+IyN0x3BNQHnzuIq4EQbAhCIKTQRCcD4LgPRH5VvjMLTcK1sg8JiLvxXoTQLiCIDgqIr+LyJ//1Qr/mgXxhM9dxLtARFysNxFvKFgryDnXU0SaCl2qiD//JSITnHONnHP1RWSiiPxPjPcEXBWfu4g3zrlrnXMDnXO1nXM1nXOPiEhfEfnfWO8t3vCWgIp7XEQ+DYLgZKw3ApTTyyLSUERKROSciCwRkRkx3REQHj53EW9qicgrIpIpIpdE5GcRuS8IgpKY7ioO0XQFAAAAr/GVAAAAAHiNghUAAABeo2AFAACA1yhYAQAA4LUrviXAOUdHFiIWBEGVv2+OexeVgXsXVSE/P19l2dnZEa1Z1fcu9y0qw5XuW56wAgAAwGsUrAAAAPAaBSsAAAC8RsEKAAAAr11x0hVfokZloHEl9jp37qyyRx55RGUFBQUq+/jjxB3bzr2LeEXTFeIRTVcAAACIWxSsAAAA8BoFKwAAALxGwQoAAACv0XSFqKNxJTrq1Klj5i+//LLKRo0apbJVq1ap7IEHHoh8Y9UI927s9e3bV2Vjx45VmdVEGM+GDRtm5p999llY59N0hXhE0xUAAADiFgUrAAAAvEbBCgAAAK9RsAIAAMBrFKwAAADwWs1YbwDA1VlvBFi7dq15rDWG9a233lLZpEmTIt8YEGXWfT59+nSVXb58WWWDBg0y11y+fHnkG4uynJwcMw/3LQFAdcMTVgAAAHiNghUAAABeo2AFAACA1yhYAQAA4DVGsyLqGG9ZPp06dVLZe++9p7IOHTqY58+dO1dlM2fOVNmhQ4cqsLvEwr0bPx588EGVLVy40Dz21VdfVdmLL75Y6XuKhNVEJiJSo0Z4z5kYzVr5nLN/pFaenJyssvr166vs1KlTYWUiIleq16oLRrMCAAAgblGwAgAAwGsUrAAAAPAaBSsAAAC8RtNVHGvRooXKTp48aR579OjRaG8nJBpXQmvatKnKiouLwzr3iSeeMHMm4VQe7t34lpWVZebW5DerIcaaGhdrVjOW1YhF01X4ateurbKMjAyV3XTTTeb51s//3nvvVVnfvn1V9sUXX6jMmuQmInLhwgUzr05ougIAAEDcomAFAACA1yhYAQAA4DUKVgAAAHitZqw3kChSUlLMvEmTJiq7+eabVTZ06FCVZWdnq2zixInmddatW6ey8+fPm8ei6vz+++8qs75YbzVY0VwFXFlhYaGZ9+nTR2WvvPKKyqwGpwkTJqjMauKKFqvBp6ysrMquHy+sRioRkYEDB6qsd+/eKhs2bJjKWrVqZa5pNa/v2LFDZYsWLVLZggULVJYIzVUVwRNWAAAAeI2CFQAAAF6jYAUAAIDXKFgBAADgNSZdRUFqaqrKQk3I6NGjh8oGDx6sspycHJVZk1mmTp1qXueDDz5Q2YEDB8xjK1uiTQuqVauWyoqKisxjmzdvrrJJkyap7O2331YZjRbRl2j3Lq4uLy9PZc7Zt8mAAQOivZ2QEmnSldWINmfOHPPYRx99VGVWU3RSUpLKQtVLJSUlKrMapX/55Zew10xUTLoCAABA3KJgBQAAgNcoWAEAAOA1ClYAAAB4jYIVAAAAXmM0awhW12ezZs1U1qVLF5V17NhRZaNHjzav07Jly7CuHW7H4qlTp8zrnD592sxR+VasWKEy689ZRGTatGkqe+edd1TGGwEAP0yZMkVlW7ZsMY/t1q2byjZs2FDpe0p08+bNU9k999xjHrt9+3aVtWvXTmXHjx9X2cGDB801ly1bprLS0lKV8UaAyPCEFQAAAF6jYAUAAIDXKFgBAADgNQpWAAAAeC2hmq6Sk5NVFqoZpm3btipr3769yoYMGaKyjIwMlTVt2tS8TrjNVJcvX1bZnj17VFazpv1HGqoZC5Gxxu/169dPZS+88IJ5/qxZs1Rm/VkD8EPjxo3DPvb2229XGU1XkZkxY4bK+vfvr7KxY8ea5588eVJl1oh0q3k2VHMdTbFVgyesAAAA8BoFKwAAALxGwQoAAACvUbACAADAa9W26cr6Ynznzp1VNnDgQPN8q3Gma9euKou0QebEiRMqq1u3rsrOnTunssLCQpUtXrw4ov2gfJ588kmV7dy5U2WzZ882z69uDVY1auh/BrYmt1lCTYGpbj8jxLfs7Oywj7322mujuJPqLy0tTWUjR45UWW5ursq+/PJLc03r82TNmjUV2B2qGk9YAQAA4DUKVgAAAHiNghUAAABeo2AFAACA17xourKajJo0aaKyNm3amOf36tVLZePGjQvrOqEaQqwGkEuXLqls9erVKlu1apXKNm7caF6nUaNGKnv//fdVlpKSojJrEgeix2q6sybZzJs3T2WnT5+Oyp4iYd371n1v3Xsi9s+jW7duKqtXr57KDh48qLLNmzeb19m2bZvKrGk1QFUoLi4O+9jdu3dHcSfVn/X3R8uBAwdUFqqJE/GLJ6wAAADwGgUrAAAAvEbBCgAAAK9RsAIAAMBrFKwAAADwWlTfEpCenq6yCRMmqGzo0KEqy8/PV1lGRoZ5nY4dO6osNTVVZWfPnlXZnj17zDV//fVXla1fv15l69atC+vc8+fPm9d55plnVHbx4kWVWW8Z2Lp1q8p87EavLp577jmVWff4zz//XOnXTkpKUlnNmvavr3Vs69atVWZ19JeUlIS1nkjo0Yd/ZY1CtO7xUPeu9UaARYsWqWzGjBlh7QeIxNGjR1V26NAh89hwxxLDtm/fPpUVFRWp7PHHH1fZrFmzzDV5e0D84gkrAAAAvEbBCgAAAK9RsAIAAMBrFKwAAADwWlSbrqwvnFuj1po3b66yrKyssNa7Uv5X1ujHJUuWmMcWFhaqzGqm2b9/v8pq1aqlsoceesi8jtUwZn0pfOXKlSrbuXOnyqxmFlSO4cOHq+zYsWMqs8YElkerVq1U1r17d5Xl5OSY51tjiWvXrq0ya+TplClTVFZQUGBeZ+LEiWHtKS0tTWXJyckqs0a4ioj06NFDZSNGjFCZ9XPPy8sz12RkJipq/PjxKrPGfouI/Pbbb1HeTfVmNStbP3+rWXTp0qXmmtZnXDQaZVH5eMIKAAAAr1GwAgAAwGsUrAAAAPAaBSsAAAC8FtWmq0uXLqnMmmDVq1cvlaWkpKjs+++/N6+zadMmlVlTqUpLS1UWqvni3LlzKgt3QoY1kStUg0zLli1V9vXXX6tszZo1Kgs1pQvRsXjxYpWNGjVKZdYEqs6dO5trdujQQWWvvfaayqxJVdbktlCsxilr79bviPV7LCLy0UcfqcxqDrR+l62sYcOG5nXuv/9+lWVnZ6ts8uTJKisrKzPXtBoyjh8/bh4L/JnVFBmqucqa0obI7N27V2X9+vVT2Ycffmiev3DhQpXNnj1bZZ9//rnK+IyILZ6wAgAAwGsUrAAAAPAaBSsAAAC8RsEKAAAAr0W16erUqVMq+/bbb1VmTb1p0KCBykJN3LGaj/bt2xfWfkI1lITLmiBkNdj07t3bPP/w4cMqs/46165dW4HdoTIVFxerrEYN/c98Tz/9tMpCNRR16dJFZUePHlWZ1UT4xhtvmGt++umnKrOmTZ05c0Zl4TYWitgNCJE0JaSmppq59XtrHfvwww+rzGrYEhG5fPmyypYtW3a1LaIasyYmtm/fXmXWRLU//vjDXHP16tUR7wtXZzVejxkzxjx2+vTpKps2bZrK7r77bpW9+eab5prW37MvXLhgHouK4wkrAAAAvEbBCgAAAK9RsAIAAMBrFKwAAADwWlSbrizWRBCr2cGa4uPjlIk6deqorFOnTirLzMw0z7earqxJV1ZjGl/qrlrz589X2bhx41Q2YMCAsNe0Gopyc3NVZjVYlZSUhH2d06dPh31sJJKSklRmNVBa93N6erq5pjUl7sYbb1SZ1UhlZSJ2c5nVdIPEYd2TVtOVle3YscNc02rKtaYoovKtW7fOzCdMmKCyqVOnqmz48OEq279/v7mm9Vl85MiRq20R5cQTVgAAAHiNghUAAABeo2AFAACA1yhYAQAA4DUKVgAAAHityt8SYHXthuq88401hjMjI0NlI0aMCHvNvLw8lVmj/y5evBj2mogOa5SpNapvzpw5Ya9Zt25dlWVlZansqaeeUtnChQvNNYuKisK6ttUV3bFjR5U1btzYPN96a4f1+229NeG+++5TWYsWLczrWGtabz2wRueuWLHCXNMa5Wi9mQSJw3rrSqNGjVTWtm1blR06dMhcs127diorLCyswO5QXqHGrlufE3PnzlWZNdbZGpuNqsMTVgAAAHiNghUAAABeo2AFAACA1yhYAQAA4LUqb7qKZ1aDTJ8+fVR23XXXqezYsWPmmvn5+Srbu3evyqxRkoi9jRs3qswarWo1aoiINGzYUGW33nqrym677TaVjRw50lxz69atKrNGB1r35C233KKy/v37m9dJS0tTmdXoYI2itJpUCgoKzOtYIxat61h/PcuXLzfXLC0tNXMkLusztkOHDiqzmrN2795trvnTTz9FvjFUqqZNm6ps7NixKrOarK3PexG7IReVjyesAAAA8BoFKwAAALxGwQoAAACvUbACAADAazRdlUPr1q1V1r17d5Vdf/31Kvvmm2/MNTdv3qwypmnED6vB6aWXXlJZs2bNzPOtSWmPPfaYynr27KmyUJOhrOZA61hrgtQNN9wQ1noi9j29Z88elaWnp6ts/fr1KgvV0PDVV1+prKyszDwWqKiUlBSVZWZmqsyaELdgwQJzTSYU+sdqIu3Xr5/Ktm/frrIffvjBXNNqLEXl4wkrAAAAvEbBCgAAAK9RsAIAAMBrFKwAAADwGk1X5WBNK7rrrrtUZk29KC4uNtcM1WiC+GB92b6wsDCsLJQlS5aozJp+ZU2lErEnUG3ZskVlVgPB2bNnVdamTRvzOklJSWFlVtOW1Wx4+PBh8zpAVbCmyTnnVGZNQGrevHlU9oSKS01NNfMhQ4aozGqOs5o9rQl9qDo8YQUAAIDXKFgBAADgNQpWAAAAeI2CFQAAAF5L+KYr6wv0IvZkoGeffVZlDRo0UNmPP/6osk8++STs61+6dMk8FonhxIkTKlu1alVYWXlY9541/aqgoCCi6wDxwGoutCa/WaymSBGRd999N5ItIUzWZ9nrr79uHmtNutq0aZPKcnNzI98YKhVPWAEAAOA1ClYAAAB4jYIVAAAAXqNgBQAAgNcoWAEAAOC1hH9LgNUVLWJ3h6akpKjs+PHjKlu5cqXKjhw5Yl7HGgkHVIVQ9z6QiKw3vmRmZoZ1bteuXSt7OyiHwYMHq2zMmDHmsbt371bZ3//+d5XxdhT/8IQVAAAAXqNgBQAAgNcoWAEAAOA1ClYAAAB4LeGbrkJp3769ytLT01VWWlqqMuvL2vv376+cjQEAKp31ub127VqV9enTR2VFRUVR2RO0tLQ0lT3//PMq27Vrl3n++PHjVbZ69WqVMSLdPzxhBQAAgNcoWAEAAOA1ClYAAAB4jYIVAAAAXkv4pqtrrrnGzFNTU1WWkZGhspKSEpVt27ZNZSdOnKjA7gAAVWHr1q0qmz9/vspycnJUlpeXZ65pTUc8f/58BXaHfztz5ozKevfurbI77rjDPH/Dhg0qo8EqPvCEFQAAAF6jYAUAAIDXKFgBAADgNQpWAAAAeC2hmq6ccypr2LCheWxWVpbKatWqpbL69eur7PDhwypLSkoyr8OXvQEg9k6ePKmy/Px8lU2ePFll1sRDEZEgCCLfGCpk5cqVsd4CKhlPWAEAAOA1ClYAAAB4jYIVAAAAXqNgBQAAgNcoWAEAAOC1hHpLgNWpH6qL0xr1Zr1lYM6cOSqzxrDyNgAA8Jc1MnXXrl0qy83NVVlycnJU9gStZ8+eKvvuu+9isBNUNZ6wAgAAwGsUrAAAAPAaBSsAAAC8RsEKAAAAryVU01VZWZnKQjVdbd68WWWZmZkqa9Kkicpq1kyoHysAxD2rqbZRo0YqGzRokMrq1atnrrl06VKVWY1cCN/o0aNVNnPmTJXdeeed5vkXLlyo9D2havCEFQAAAF6jYAUAAIDXKFgBAADgNQpWAAAAeM2FajoCAAAAfMATVgAAAHiNghUAAABeo2AFAACA1yhYAQAA4DUKVgAAAHiNghUAAABe+z9hdZ65vPXeIgAAAABJRU5ErkJggg==\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Picture shown above can be look weird since we apply augmentations on them. Next step is creating a learner. "
      ],
      "metadata": {
        "id": "JEztUzir8FFn"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "learn = cnn_learner(dls, resnet34, metrics=accuracy)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 67,
          "referenced_widgets": [
            "85c74e8d056e4a7b89c487e6ec3703f4",
            "4d68cfa0220a4759ad3e25d91d60ba32",
            "aaa779e7437d49f5b97df3e7cab52bde",
            "234c36fb6b0f409383e43cc4f0c05201",
            "78bc6b0199a24fde832ff80e61103433",
            "df4fa31d0fc34b46acaaf31d3216df3c",
            "e636bceeb0df46cbb00a7331f0b28d45",
            "4a6a4cb8711f4f0a94dd3c3b41a20931",
            "1bdda5e2efdb40b69181aa958485ed41",
            "6081239db21e421183d894d1967023db",
            "29c857194b1d4ab19c988ee248dbf54a"
          ]
        },
        "id": "Ro03h64m_7fF",
        "outputId": "325497c3-f194-48a1-cf21-b02952f82155"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Downloading: \"https://download.pytorch.org/models/resnet34-b627a593.pth\" to /root/.cache/torch/hub/checkpoints/resnet34-b627a593.pth\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "  0%|          | 0.00/83.3M [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "85c74e8d056e4a7b89c487e6ec3703f4"
            }
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "cnn_learner builds a convolutional neural network learner from dataloaders. In our case we use ResNet architecture with 34 layers. Then, we will make a check for optimal learning rate.\n"
      ],
      "metadata": {
        "id": "AkGSwvtrADMw"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "learn.lr_find()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 304
        },
        "id": "JX9EbyLvARsr",
        "outputId": "bfc241eb-4ca0-43c3-a0b8-09dcb5d920a2"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "\n",
              "<style>\n",
              "    /* Turns off some styling */\n",
              "    progress {\n",
              "        /* gets rid of default border in Firefox and Opera. */\n",
              "        border: none;\n",
              "        /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
              "        background-size: auto;\n",
              "    }\n",
              "    .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
              "        background: #F44336;\n",
              "    }\n",
              "</style>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              ""
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "SuggestedLRs(valley=0.0014454397605732083)"
            ]
          },
          "metadata": {},
          "execution_count": 8
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        ""
      ],
      "metadata": {
        "id": "RwHOlukFCSEn"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "learn.fine_tune(12, base_lr=0.0012022644514217973, cbs=[ShowGraphCallback()])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 986
        },
        "id": "VmfsmtiEBMK1",
        "outputId": "8ef49d9e-a628-44ae-99b8-6df7e2b82bd0"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "\n",
              "<style>\n",
              "    /* Turns off some styling */\n",
              "    progress {\n",
              "        /* gets rid of default border in Firefox and Opera. */\n",
              "        border: none;\n",
              "        /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
              "        background-size: auto;\n",
              "    }\n",
              "    .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
              "        background: #F44336;\n",
              "    }\n",
              "</style>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: left;\">\n",
              "      <th>epoch</th>\n",
              "      <th>train_loss</th>\n",
              "      <th>valid_loss</th>\n",
              "      <th>accuracy</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>1.233510</td>\n",
              "      <td>0.713048</td>\n",
              "      <td>0.769250</td>\n",
              "      <td>01:58</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "\n",
              "<style>\n",
              "    /* Turns off some styling */\n",
              "    progress {\n",
              "        /* gets rid of default border in Firefox and Opera. */\n",
              "        border: none;\n",
              "        /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
              "        background-size: auto;\n",
              "    }\n",
              "    .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
              "        background: #F44336;\n",
              "    }\n",
              "</style>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: left;\">\n",
              "      <th>epoch</th>\n",
              "      <th>train_loss</th>\n",
              "      <th>valid_loss</th>\n",
              "      <th>accuracy</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>0.471720</td>\n",
              "      <td>0.203093</td>\n",
              "      <td>0.933750</td>\n",
              "      <td>02:08</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1</td>\n",
              "      <td>0.271164</td>\n",
              "      <td>0.111953</td>\n",
              "      <td>0.963500</td>\n",
              "      <td>02:07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>2</td>\n",
              "      <td>0.175693</td>\n",
              "      <td>0.066175</td>\n",
              "      <td>0.979083</td>\n",
              "      <td>02:07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>3</td>\n",
              "      <td>0.154894</td>\n",
              "      <td>0.047589</td>\n",
              "      <td>0.985250</td>\n",
              "      <td>02:07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>4</td>\n",
              "      <td>0.133842</td>\n",
              "      <td>0.041241</td>\n",
              "      <td>0.987583</td>\n",
              "      <td>02:07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>5</td>\n",
              "      <td>0.092512</td>\n",
              "      <td>0.039037</td>\n",
              "      <td>0.987000</td>\n",
              "      <td>02:07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>6</td>\n",
              "      <td>0.086166</td>\n",
              "      <td>0.031547</td>\n",
              "      <td>0.989833</td>\n",
              "      <td>02:07</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>7</td>\n",
              "      <td>0.069900</td>\n",
              "      <td>0.029535</td>\n",
              "      <td>0.990833</td>\n",
              "      <td>02:08</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>8</td>\n",
              "      <td>0.057457</td>\n",
              "      <td>0.025965</td>\n",
              "      <td>0.991833</td>\n",
              "      <td>02:08</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>9</td>\n",
              "      <td>0.061980</td>\n",
              "      <td>0.022729</td>\n",
              "      <td>0.992833</td>\n",
              "      <td>02:08</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>10</td>\n",
              "      <td>0.054856</td>\n",
              "      <td>0.023145</td>\n",
              "      <td>0.992250</td>\n",
              "      <td>02:08</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>11</td>\n",
              "      <td>0.052097</td>\n",
              "      <td>0.022478</td>\n",
              "      <td>0.992417</td>\n",
              "      <td>02:08</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "interp = ClassificationInterpretation.from_learner(learn)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 17
        },
        "id": "3jaQpirVIlLn",
        "outputId": "3729c6e3-f024-4b2e-c630-d03130f52709"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "\n",
              "<style>\n",
              "    /* Turns off some styling */\n",
              "    progress {\n",
              "        /* gets rid of default border in Firefox and Opera. */\n",
              "        border: none;\n",
              "        /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
              "        background-size: auto;\n",
              "    }\n",
              "    .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
              "        background: #F44336;\n",
              "    }\n",
              "</style>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              ""
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "interp.plot_confusion_matrix()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 311
        },
        "id": "qH8s8y6RImob",
        "outputId": "9402b456-9cfc-4bce-ad10-518b2b9b730d"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "\n",
              "<style>\n",
              "    /* Turns off some styling */\n",
              "    progress {\n",
              "        /* gets rid of default border in Firefox and Opera. */\n",
              "        border: none;\n",
              "        /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
              "        background-size: auto;\n",
              "    }\n",
              "    .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
              "        background: #F44336;\n",
              "    }\n",
              "</style>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              ""
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "interp.plot_top_losses(10)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 585
        },
        "id": "Oyz3AtycIsRe",
        "outputId": "10cf9c82-dad1-416b-a3a3-645ab71aa2c2"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "\n",
              "<style>\n",
              "    /* Turns off some styling */\n",
              "    progress {\n",
              "        /* gets rid of default border in Firefox and Opera. */\n",
              "        border: none;\n",
              "        /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
              "        background-size: auto;\n",
              "    }\n",
              "    .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
              "        background: #F44336;\n",
              "    }\n",
              "</style>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              ""
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 648x648 with 9 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "learn.export(\"models/MNIST-99acc.pkl\")"
      ],
      "metadata": {
        "id": "4I_FIxAdI-Q5"
      },
      "execution_count": 13,
      "outputs": []
    }
  ]
}