谢焕宜_Ch_5.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "谢焕宜_Ch_5",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "jupytext": {
      "split_at_heading": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "def8024bae5345eebbe5ef6a55a46b6c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "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_0de0b5ad7f78457cbd6ea204ed61a131",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_924f9d97696b40a29d01951b478f4b7d",
              "IPY_MODEL_f3aa44456aee4fb5a090acb54214993f"
            ]
          }
        },
        "0de0b5ad7f78457cbd6ea204ed61a131": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "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
          }
        },
        "924f9d97696b40a29d01951b478f4b7d": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_5b2e4dfc755d4b419114a658ee1d93b8",
            "_dom_classes": [],
            "description": "100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 87306240,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 87306240,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_2b90641b790c4cb8b756dd27a474294e"
          }
        },
        "f3aa44456aee4fb5a090acb54214993f": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_6512233b89a84ed38000ea388d6f93b8",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 83.3M/83.3M [00:01&lt;00:00, 76.1MB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_0c63b82843744875a82af5039653a659"
          }
        },
        "5b2e4dfc755d4b419114a658ee1d93b8": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_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"
          }
        },
        "2b90641b790c4cb8b756dd27a474294e": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "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
          }
        },
        "6512233b89a84ed38000ea388d6f93b8": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "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"
          }
        },
        "0c63b82843744875a82af5039653a659": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "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
          }
        },
        "f6e65be4a85a4f69961315456fa64674": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "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_99da4055787f401ba6977ce8724f736f",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_dbf279950da247719598d6ffced0213b",
              "IPY_MODEL_4c3436bced7a411787abb81c191e44dd"
            ]
          }
        },
        "99da4055787f401ba6977ce8724f736f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "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
          }
        },
        "dbf279950da247719598d6ffced0213b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_306bd262954f47a1a318c7ddf665738a",
            "_dom_classes": [],
            "description": "100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 102502400,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 102502400,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_be8e6a1518334f529ea5e6a9713d174e"
          }
        },
        "4c3436bced7a411787abb81c191e44dd": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_d0ec09c4c37c4610a9e5603270c478f4",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 97.8M/97.8M [10:44&lt;00:00, 159kB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_f3b13a3ba52e4dc3922cdfbf0cbd0b8d"
          }
        },
        "306bd262954f47a1a318c7ddf665738a": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_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"
          }
        },
        "be8e6a1518334f529ea5e6a9713d174e": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "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
          }
        },
        "d0ec09c4c37c4610a9e5603270c478f4": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "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"
          }
        },
        "f3b13a3ba52e4dc3922cdfbf0cbd0b8d": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "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
          }
        },
        "4963a4429f4d418ba8ef420cfc8ffc52": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "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_922b36ce7b674e02ba6ba9ec843c45ec",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_ae780f942fe84f95a956b24bd4594d7e",
              "IPY_MODEL_cacb1dd327414bdf94fbaddb16be3c01"
            ]
          }
        },
        "922b36ce7b674e02ba6ba9ec843c45ec": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "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
          }
        },
        "ae780f942fe84f95a956b24bd4594d7e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_40f30d58d8104e94b1cd03c3c76e8ac3",
            "_dom_classes": [],
            "description": "100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 46827520,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 46827520,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_492c2e96587f4cc5a07697aeaa7b25d4"
          }
        },
        "cacb1dd327414bdf94fbaddb16be3c01": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_bd303a2ae9524a3f950e01a6b1f1d33b",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 44.7M/44.7M [00:00&lt;00:00, 54.4MB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_1eef947766084bf9a94d30dd07a66cdd"
          }
        },
        "40f30d58d8104e94b1cd03c3c76e8ac3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_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"
          }
        },
        "492c2e96587f4cc5a07697aeaa7b25d4": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "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
          }
        },
        "bd303a2ae9524a3f950e01a6b1f1d33b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "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"
          }
        },
        "1eef947766084bf9a94d30dd07a66cdd": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "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
          }
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/StuPidMRE/assignment/blob/main/%E8%B0%A2%E7%84%95%E5%AE%9C_Ch_5.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cEVFdjmwBQhF"
      },
      "source": [
        "!pip install -Uqq fastbook\n",
        "import fastbook\n",
        "fastbook.setup_book()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gfDVABiEBQhK"
      },
      "source": [
        "from fastbook import *"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "D9ExBEs5BQhL"
      },
      "source": [
        "# Image Classification"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HuflJ6JZBQhL"
      },
      "source": [
        "Now that you understand what deep learning is, what it's for, and how to create and deploy a model, it's time for us to go deeper! In an ideal world deep learning practitioners wouldn't have to know every detail of how things work under the hood… But as yet, we don't live in an ideal world. The truth is, to make your model really work, and work reliably, there are a lot of details you have to get right, and a lot of details that you have to check. This process requires being able to look inside your neural network as it trains, and as it makes predictions, find possible problems, and know how to fix them.\n",
        "\n",
        "So, from here on in the book we are going to do a deep dive into the mechanics of deep learning. What is the architecture of a computer vision model, an NLP model, a tabular model, and so on? How do you create an architecture that matches the needs of your particular domain? How do you get the best possible results from the training process? How do you make things faster? What do you have to change as your datasets change?\n",
        "\n",
        "We will start by repeating the same basic applications that we looked at in the first chapter, but we are going to do two things:\n",
        "\n",
        "- Make them better.\n",
        "- Apply them to a wider variety of types of data.\n",
        "\n",
        "In order to do these two things, we will have to learn all of the pieces of the deep learning puzzle. This includes different types of layers, regularization methods, optimizers, how to put layers together into architectures, labeling techniques, and much more. We are not just going to dump all of these things on you, though; we will introduce them progressively as needed, to solve actual problems related to the projects we are working on."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7g27b08OBQhL"
      },
      "source": [
        "## From Dogs and Cats to Pet Breeds"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9P3IWgTKBQhM"
      },
      "source": [
        "In our very first model we learned how to classify dogs versus cats. Just a few years ago this was considered a very challenging task—but today, it's far too easy! We will not be able to show you the nuances of training models with this problem, because we get a nearly perfect result without worrying about any of the details. But it turns out that the same dataset also allows us to work on a much more challenging problem: figuring out what breed of pet is shown in each image.\n",
        "\n",
        "In `chapter_intro` we presented the applications as already-solved problems. But this is not how things work in real life. We start with some dataset that we know nothing about. We then have to figure out how it is put together, how to extract the data we need from it, and what that data looks like. For the rest of this book we will be showing you how to solve these problems in practice, including all of the intermediate steps necessary to understand the data that you are working with and test your modeling as you go.\n",
        "\n",
        "We already downloaded the Pet dataset, and we can get a path to this dataset using the same code as in `chapter_intro`:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4KlVtPS3BQhM"
      },
      "source": [
        "from fastai.vision.all import *\n",
        "path = untar_data(URLs.PETS)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LLLQvpeZBQhM"
      },
      "source": [
        "Individual files representing items of data, such as text documents or images, possibly organized into folders or with filenames representing information about those items\n",
        "A table of data, such as in CSV format, where each row is an item which may include filenames providing a connection between the data in the table and data in other formats, such as text documents and images\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UFC1_cqDBQhN"
      },
      "source": [
        "Path.BASE_PATH = path"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZnrF_bOsBQhN",
        "outputId": "b51ae070-4677-40ff-b440-732e8b0de7f9"
      },
      "source": [
        "path.ls()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(#2) [Path('annotations'),Path('images')]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5CNCwYOoBQhO"
      },
      "source": [
        "We can see that this dataset provides us with *images* and *annotations* directories. The [website](https://www.robots.ox.ac.uk/~vgg/data/pets/) for the dataset tells us that the *annotations* directory contains information about where the pets are rather than what they are. In this chapter, we will be doing classification, not localization, which is to say that we care about what the pets are, not where they are. Therefore, we will ignore the *annotations* directory for now. So, let's have a look inside the *images* directory:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FozaDQYnBQhO",
        "outputId": "fc09c75e-4aff-4d8a-cdb8-def7dc1fd0dc"
      },
      "source": [
        "(path/\"images\").ls()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(#7393) [Path('images/saint_bernard_94.jpg'),Path('images/newfoundland_48.jpg'),Path('images/shiba_inu_206.jpg'),Path('images/Russian_Blue_83.jpg'),Path('images/scottish_terrier_78.jpg'),Path('images/samoyed_21.jpg'),Path('images/english_setter_56.jpg'),Path('images/miniature_pinscher_52.jpg'),Path('images/wheaten_terrier_76.jpg'),Path('images/boxer_49.jpg')...]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eB_RQRDrBQhO"
      },
      "source": [
        "Most functions and methods in fastai that return a collection use a class called `L`. `L` can be thought of as an enhanced version of the ordinary Python `list` type, with added conveniences for common operations. For instance, when we display an object of this class in a notebook it appears in the format shown there. The first thing that is shown is the number of items in the collection, prefixed with a `#`. You'll also see in the preceding output that the list is suffixed with an ellipsis. This means that only the first few items are displayed—which is a good thing, because we would not want more than 7,000 filenames on our screen!\n",
        "\n",
        "By examining these filenames, we can see how they appear to be structured. Each filename contains the pet breed, and then an underscore (`_`), a number, and finally the file extension. We need to create a piece of code that extracts the breed from a single `Path`. Jupyter notebooks make this easy, because we can gradually build up something that works, and then use it for the entire dataset. We do have to be careful to not make too many assumptions at this point. For instance, if you look carefully you may notice that some of the pet breeds contain multiple words, so we cannot simply break at the first `_` character that we find. To allow us to test our code, let's pick out one of these filenames:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eqQmNyWZBQhO",
        "outputId": "695eab16-2dde-4e57-9c6b-7d938bd784ce"
      },
      "source": [
        "fname = (path/\"images\").ls()[0]\n",
        "fname"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Path('images/saint_bernard_94.jpg')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 27
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bl8NQDnUBQhP"
      },
      "source": [
        "The most powerful and flexible way to extract information from strings like this is to use a *regular expression*, also known as a *regex*. A regular expression is a special string, written in the regular expression language, which specifies a general rule for deciding if another string passes a test (i.e., \"matches\" the regular expression), and also possibly for plucking a particular part or parts out of that other string. \n",
        "\n",
        "In this case, we need a regular expression that extracts the pet breed from the filename.\n",
        "\n",
        "We do not have the space to give you a complete regular expression tutorial here, but there are many excellent ones online and we know that many of you will already be familiar with this wonderful tool. If you're not, that is totally fine—this is a great opportunity for you to rectify that! We find that regular expressions are one of the most useful tools in our programming toolkit, and many of our students tell us that this is one of the things they are most excited to learn about. So head over to Google and search for \"regular expressions tutorial\" now, and then come back here after you've had a good look around. The [book's website](https://book.fast.ai/) also provides a list of our favorites.\n",
        "\n",
        "> a: Not only are regular expressions dead handy, but they also have interesting roots. They are \"regular\" because they were originally examples of a \"regular\" language, the lowest rung within the Chomsky hierarchy, a grammar classification developed by linguist Noam Chomsky, who also wrote _Syntactic Structures_, the pioneering work searching for the formal grammar underlying human language. This is one of the charms of computing: it may be that the hammer you reach for every day in fact came from a spaceship.\n",
        "\n",
        "> b: Online tool to test regex rules on text [here](https://regex101.com/)\n",
        "\n",
        "When you are writing a regular expression, the best way to start is just to try it against one example at first. Let's use the `findall` method to try a regular expression against the filename of the `fname` object:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RsaqgblaBQhP",
        "outputId": "02a726ff-907a-4201-df10-5cfe71ca8248"
      },
      "source": [
        "re.findall(r'(.+)_\\d+.jpg$', fname.name)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['saint_bernard']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XzYgD8X3BQhP"
      },
      "source": [
        "This regular expression plucks out all the characters leading up to the last underscore character, as long as the subsequence characters are numerical digits and then the JPEG file extension.\n",
        "\n",
        "Now that we confirmed the regular expression works for the example, let's use it to label the whole dataset. fastai comes with many classes to help with labeling. For labeling with regular expressions, we can use the `RegexLabeller` class. In this example we use the data block API we saw in `chapter_production` (in fact, we nearly always use the data block API—it's so much more flexible than the simple factory methods we saw in `chapter_intro`):"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yumiaR-pBQhP"
      },
      "source": [
        "pets = DataBlock(blocks = (ImageBlock, CategoryBlock),\n",
        "                 get_items=get_image_files, \n",
        "                 splitter=RandomSplitter(seed=42),\n",
        "                 get_y=using_attr(RegexLabeller(r'(.+)_\\d+.jpg$'), 'name'),\n",
        "                 item_tfms=Resize(460),\n",
        "                 batch_tfms=aug_transforms(size=224, min_scale=0.75))\n",
        "dls = pets.dataloaders(path/\"images\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UrFkoKclBQhQ"
      },
      "source": [
        "One important piece of this `DataBlock` call that we haven't seen before is in these two lines:\n",
        "\n",
        "```python\n",
        "item_tfms=Resize(460),\n",
        "batch_tfms=aug_transforms(size=224, min_scale=0.75)\n",
        "```\n",
        "\n",
        "These lines implement a fastai data augmentation strategy which we call *presizing*. Presizing is a particular way to do image augmentation that is designed to minimize data destruction while maintaining good performance."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xlS3sQpCBQhQ"
      },
      "source": [
        "## Presizing"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4G-K_3EBBQhQ"
      },
      "source": [
        "We need our images to have the same dimensions, so that they can collate into tensors to be passed to the GPU. We also want to minimize the number of distinct augmentation computations we perform. The performance requirement suggests that we should, where possible, compose our augmentation transforms into fewer transforms (to reduce the number of computations and the number of lossy operations) and transform the images into uniform sizes (for more efficient processing on the GPU).\n",
        "\n",
        "The challenge is that, if performed after resizing down to the augmented size, various common data augmentation transforms might introduce spurious empty zones, degrade data, or both. For instance, rotating an image by 45 degrees fills corner regions of the new bounds with emptiness, which will not teach the model anything. Many rotation and zooming operations will require interpolating to create pixels. These interpolated pixels are derived from the original image data but are still of lower quality.\n",
        "\n",
        "To work around these challenges, presizing adopts two strategies that are shown in `Presizing on the training set`:\n",
        "\n",
        "1. Resize images to relatively \"large\" dimensions—that is, dimensions significantly larger than the target training dimensions. \n",
        "1. Compose all of the common augmentation operations (including a resize to the final target size) into one, and perform the combined operation on the GPU only once at the end of processing, rather than performing the operations individually and interpolating multiple times.\n",
        "\n",
        "The first step, the resize, creates images large enough that they have spare margin to allow further augmentation transforms on their inner regions without creating empty zones. This transformation works by resizing to a square, using a large crop size. On the training set, the crop area is chosen randomly, and the size of the crop is selected to cover the entire width or height of the image, whichever is smaller.\n",
        "\n",
        "In the second step, the GPU is used for all data augmentation, and all of the potentially destructive operations are done together, with a single interpolation at the end."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_GLsUbVQBQhR"
      },
      "source": [
        "*Presizing on the training set*\n",
        "\n",
        "<img alt=\"Presizing on the training set\" width=\"600\" caption=\"Presizing on the training set\" id=\"presizing\" src=\"https://raw.githubusercontent.com/vtecftwy/fastbook/master/images/att_00060.png\">"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6bCZ2TndBQhR"
      },
      "source": [
        "This picture shows the two steps:\n",
        "\n",
        "1. *Crop full width or height*: This is in `item_tfms`, so it's applied to each individual image before it is copied to the GPU. It's used to ensure all images are the same size. On the training set, the crop area is chosen randomly. On the validation set, the center square of the image is always chosen.\n",
        "2. *Random crop and augment*: This is in `batch_tfms`, so it's applied to a batch all at once on the GPU, which means it's fast. On the validation set, only the resize to the final size needed for the model is done here. On the training set, the random crop and any other augmentations are done first.\n",
        "\n",
        "To implement this process in fastai you use `Resize` as an item transform with a large size, and `RandomResizedCrop` as a batch transform with a smaller size. `RandomResizedCrop` will be added for you if you include the `min_scale` parameter in your `aug_transforms` function, as was done in the `DataBlock` call in the previous section. Alternatively, you can use `pad` or `squish` instead of `crop` (the default) for the initial `Resize`.\n",
        "\n",
        "`Interpolations` shows the difference between an image that has been zoomed, interpolated, rotated, and then interpolated again (which is the approach used by all other deep learning libraries), shown here on the right, and an image that has been zoomed and rotated as one operation and then interpolated just once on the left (the fastai approach), shown here on the left."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 210
        },
        "id": "I8GFLiKDBQhS",
        "outputId": "e8fb4246-2bc1-4da8-ca47-3dcd0a033646"
      },
      "source": [
        "# interpolations\n",
        "# A comparison of fastai's data augmentation strategy (left) and the traditional approach (right).\n",
        "\n",
        "dblock1 = DataBlock(blocks=(ImageBlock(), CategoryBlock()),\n",
        "                    get_y=parent_label,\n",
        "                    item_tfms=Resize(460))\n",
        "\n",
        "# Place an image in the 'images/grizzly.jpg' subfolder where this notebook is located before running this\n",
        "p2dir = Path.cwd() / 'images'\n",
        "os.makedirs(p2dir, exist_ok=True)\n",
        "download_url('https://raw.githubusercontent.com/vtecftwy/fastbook/master/images/grizzly.jpg', p2dir / 'grizzly.jpg')\n",
        "\n",
        "\n",
        "# Create dls\n",
        "dls1 = dblock1.dataloaders([(Path.cwd()/'images'/'grizzly.jpg')]*100, bs=8)\n",
        "dls1.train.get_idxs = lambda: Inf.ones\n",
        "x,y = dls1.valid.one_batch()\n",
        "fig, axs = subplots(1, 2)\n",
        "fig.suptitle('Interpolation:\\nA comparison of fastai\\'s data augmentation strategy (left) and the traditional approach (right).')\n",
        "\n",
        "x1 = TensorImage(x.clone())\n",
        "x1 = x1.affine_coord(sz=224)\n",
        "x1 = x1.rotate(draw=30, p=1.)\n",
        "x1 = x1.zoom(draw=1.2, p=1.)\n",
        "x1 = x1.warp(draw_x=-0.2, draw_y=0.2, p=1.)\n",
        "\n",
        "tfms = setup_aug_tfms([Rotate(draw=30, p=1, size=224), Zoom(draw=1.2, p=1., size=224),\n",
        "                       Warp(draw_x=-0.2, draw_y=0.2, p=1., size=224)])\n",
        "x = Pipeline(tfms)(x)\n",
        "#x.affine_coord(coord_tfm=coord_tfm, sz=size, mode=mode, pad_mode=pad_mode)\n",
        "\n",
        "TensorImage(x[0]).show(ctx=axs[0])\n",
        "TensorImage(x1[0]).show(ctx=axs[1]);"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x216 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9GLuCrMRBQhS"
      },
      "source": [
        "You can see that the image on the right is less well defined and has reflection padding artifacts in the bottom-left corner; also, the grass iat the top left has disappeared entirely. We find that in practice using presizing significantly improves the accuracy of models, and often results in speedups too.\n",
        "\n",
        "The fastai library also provides simple ways to check your data looks right before training a model, which is an extremely important step. We'll look at those next."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4CozDsJDBQhS"
      },
      "source": [
        "### Checking and Debugging a DataBlock"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8SwCTm5CBQhS"
      },
      "source": [
        "We can never just assume that our code is working perfectly. Writing a `DataBlock` is just like writing a blueprint. You will get an error message if you have a syntax error somewhere in your code, but you have no guarantee that your template is going to work on your data source as you intend. So, before training a model you should always check your data. You can do this using the `show_batch` method:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 195
        },
        "id": "1aKL67AGBQhT",
        "outputId": "e1473853-c62e-40c4-91c0-612be923f49f"
      },
      "source": [
        "dls.show_batch(nrows=1, ncols=3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 648x216 with 3 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "G_fwi7y0BQhT"
      },
      "source": [
        "Take a look at each image, and check that each one seems to have the correct label for that breed of pet. Often, data scientists work with data with which they are not as familiar as domain experts may be: for instance, I actually don't know what a lot of these pet breeds are. Since I am not an expert on pet breeds, I would use Google images at this point to search for a few of these breeds, and make sure the images look similar to what I see in this output.\n",
        "\n",
        "If you made a mistake while building your `DataBlock`, it is very likely you won't see it before this step. To debug this, we encourage you to use the `summary` method. It will attempt to create a batch from the source you give it, with a lot of details. Also, if it fails, you will see exactly at which point the error happens, and the library will try to give you some help. For instance, one common mistake is to forget to use a `Resize` transform, so you end up with pictures of different sizes and are not able to batch them. Here is what the summary would look like in that case (note that the exact text may have changed since the time of writing, but it will give you an idea):"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "0pfl3_bLBQhT",
        "outputId": "b949cf3c-48cb-423c-c928-3cfdd68da3f2"
      },
      "source": [
        "pets1 = DataBlock(blocks = (ImageBlock, CategoryBlock),\n",
        "                 get_items=get_image_files, \n",
        "                 splitter=RandomSplitter(seed=42),\n",
        "                 get_y=using_attr(RegexLabeller(r'(.+)_\\d+.jpg$'), 'name'))\n",
        "\n",
        "pets1.summary(path/\"images\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Setting-up type transforms pipelines\n",
            "Collecting items from /root/.fastai/data/oxford-iiit-pet/images\n",
            "Found 7390 items\n",
            "2 datasets of sizes 5912,1478\n",
            "Setting up Pipeline: PILBase.create\n",
            "Setting up Pipeline: partial -> Categorize -- {'vocab': None, 'sort': True, 'add_na': False}\n",
            "\n",
            "Building one sample\n",
            "  Pipeline: PILBase.create\n",
            "    starting from\n",
            "      /root/.fastai/data/oxford-iiit-pet/images/Maine_Coon_3.jpg\n",
            "    applying PILBase.create gives\n",
            "      PILImage mode=RGB size=241x247\n",
            "  Pipeline: partial -> Categorize -- {'vocab': None, 'sort': True, 'add_na': False}\n",
            "    starting from\n",
            "      /root/.fastai/data/oxford-iiit-pet/images/Maine_Coon_3.jpg\n",
            "    applying partial gives\n",
            "      Maine_Coon\n",
            "    applying Categorize -- {'vocab': None, 'sort': True, 'add_na': False} gives\n",
            "      TensorCategory(6)\n",
            "\n",
            "Final sample: (PILImage mode=RGB size=241x247, TensorCategory(6))\n",
            "\n",
            "\n",
            "Collecting items from /root/.fastai/data/oxford-iiit-pet/images\n",
            "Found 7390 items\n",
            "2 datasets of sizes 5912,1478\n",
            "Setting up Pipeline: PILBase.create\n",
            "Setting up Pipeline: partial -> Categorize -- {'vocab': None, 'sort': True, 'add_na': False}\n",
            "Setting up after_item: Pipeline: ToTensor\n",
            "Setting up before_batch: Pipeline: \n",
            "Setting up after_batch: Pipeline: IntToFloatTensor -- {'div': 255.0, 'div_mask': 1}\n",
            "\n",
            "Building one batch\n",
            "Applying item_tfms to the first sample:\n",
            "  Pipeline: ToTensor\n",
            "    starting from\n",
            "      (PILImage mode=RGB size=241x247, TensorCategory(6))\n",
            "    applying ToTensor gives\n",
            "      (TensorImage of size 3x247x241, TensorCategory(6))\n",
            "\n",
            "Adding the next 3 samples\n",
            "\n",
            "No before_batch transform to apply\n",
            "\n",
            "Collating items in a batch\n",
            "Error! It's not possible to collate your items in a batch\n",
            "Could not collate the 0-th members of your tuples because got the following shapes\n",
            "torch.Size([3, 247, 241]),torch.Size([3, 450, 338]),torch.Size([3, 500, 375]),torch.Size([3, 236, 300])\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "error",
          "ename": "RuntimeError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mRuntimeError\u001b[0m                              Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-32-ac0f22fa2260>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      4\u001b[0m                  get_y=using_attr(RegexLabeller(r'(.+)_\\d+.jpg$'), 'name'))\n\u001b[1;32m      5\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 6\u001b[0;31m \u001b[0mpets1\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msummary\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpath\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m\"images\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/fastai/data/block.py\u001b[0m in \u001b[0;36msummary\u001b[0;34m(self, source, bs, show_batch, **kwargs)\u001b[0m\n\u001b[1;32m    188\u001b[0m         \u001b[0mwhy\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_find_fail_collate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    189\u001b[0m         \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Make sure all parts of your samples are tensors of the same size\"\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mwhy\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mwhy\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 190\u001b[0;31m         \u001b[0;32mraise\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    191\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    192\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mf\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mf\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdls\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mafter_batch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfs\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mname\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0;34m'noop'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m!=\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/fastai/data/block.py\u001b[0m in \u001b[0;36msummary\u001b[0;34m(self, source, bs, show_batch, **kwargs)\u001b[0m\n\u001b[1;32m    182\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"\\nCollating items in a batch\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    183\u001b[0m     \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 184\u001b[0;31m         \u001b[0mb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdls\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcreate_batch\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    185\u001b[0m         \u001b[0mb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mretain_types\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0ms\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mis_listy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0ms\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    186\u001b[0m     \u001b[0;32mexcept\u001b[0m \u001b[0mException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/fastai/data/load.py\u001b[0m in \u001b[0;36mcreate_batch\u001b[0;34m(self, b)\u001b[0m\n\u001b[1;32m    141\u001b[0m         \u001b[0;32melif\u001b[0m \u001b[0ms\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m  \u001b[0;32mreturn\u001b[0m \u001b[0mnext\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mit\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    142\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;32mraise\u001b[0m \u001b[0mIndexError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Cannot index an iterable dataset numerically - must use `None`.\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 143\u001b[0;31m     \u001b[0;32mdef\u001b[0m \u001b[0mcreate_batch\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mb\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mfa_collate\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mfa_convert\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprebatched\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mb\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    144\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mdo_batch\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mb\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mretain\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcreate_batch\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbefore_batch\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mb\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mb\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    145\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mto\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdevice\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdevice\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdevice\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/fastai/data/load.py\u001b[0m in \u001b[0;36mfa_collate\u001b[0;34m(t)\u001b[0m\n\u001b[1;32m     48\u001b[0m     \u001b[0mb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     49\u001b[0m     return (default_collate(t) if isinstance(b, _collate_types)\n\u001b[0;32m---> 50\u001b[0;31m             \u001b[0;32melse\u001b[0m \u001b[0mtype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mt\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mfa_collate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ms\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0mt\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mSequence\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     51\u001b[0m             else default_collate(t))\n\u001b[1;32m     52\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/fastai/data/load.py\u001b[0m in \u001b[0;36m<listcomp>\u001b[0;34m(.0)\u001b[0m\n\u001b[1;32m     48\u001b[0m     \u001b[0mb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     49\u001b[0m     return (default_collate(t) if isinstance(b, _collate_types)\n\u001b[0;32m---> 50\u001b[0;31m             \u001b[0;32melse\u001b[0m \u001b[0mtype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mt\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mfa_collate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ms\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0mt\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mSequence\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     51\u001b[0m             else default_collate(t))\n\u001b[1;32m     52\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/fastai/data/load.py\u001b[0m in \u001b[0;36mfa_collate\u001b[0;34m(t)\u001b[0m\n\u001b[1;32m     47\u001b[0m     \u001b[0;34m\"A replacement for PyTorch `default_collate` which maintains types and handles `Sequence`s\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     48\u001b[0m     \u001b[0mb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 49\u001b[0;31m     return (default_collate(t) if isinstance(b, _collate_types)\n\u001b[0m\u001b[1;32m     50\u001b[0m             \u001b[0;32melse\u001b[0m \u001b[0mtype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mt\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mfa_collate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ms\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0mt\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mSequence\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     51\u001b[0m             else default_collate(t))\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/torch/utils/data/_utils/collate.py\u001b[0m in \u001b[0;36mdefault_collate\u001b[0;34m(batch)\u001b[0m\n\u001b[1;32m     53\u001b[0m             \u001b[0mstorage\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0melem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstorage\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_new_shared\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnumel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     54\u001b[0m             \u001b[0mout\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0melem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnew\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstorage\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 55\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstack\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbatch\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mout\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mout\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     56\u001b[0m     \u001b[0;32melif\u001b[0m \u001b[0melem_type\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__module__\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'numpy'\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0melem_type\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__name__\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0;34m'str_'\u001b[0m\u001b[0;31m \u001b[0m\u001b[0;31m\\\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     57\u001b[0m             \u001b[0;32mand\u001b[0m \u001b[0melem_type\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__name__\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0;34m'string_'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/fastai/torch_core.py\u001b[0m in \u001b[0;36m__torch_function__\u001b[0;34m(self, func, types, args, kwargs)\u001b[0m\n\u001b[1;32m    330\u001b[0m         \u001b[0mconvert\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    331\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0m_torch_handled\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_opt\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfunc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mconvert\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mtypes\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTensor\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 332\u001b[0;31m         \u001b[0mres\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msuper\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__torch_function__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfunc\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtypes\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0margs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkwargs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    333\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mconvert\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mres\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mconvert\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mres\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    334\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mres\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mTensorBase\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mres\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset_meta\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mas_copy\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/torch/tensor.py\u001b[0m in \u001b[0;36m__torch_function__\u001b[0;34m(cls, func, types, args, kwargs)\u001b[0m\n\u001b[1;32m    960\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    961\u001b[0m         \u001b[0;32mwith\u001b[0m \u001b[0m_C\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mDisableTorchFunction\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 962\u001b[0;31m             \u001b[0mret\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfunc\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    963\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0m_convert\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mret\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcls\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    964\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mRuntimeError\u001b[0m: stack expects each tensor to be equal size, but got [3, 247, 241] at entry 0 and [3, 450, 338] at entry 1"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wkW7mwu-BQhT"
      },
      "source": [
        "```\n",
        "Setting-up type transforms pipelines\n",
        "Collecting items from /home/sgugger/.fastai/data/oxford-iiit-pet/images\n",
        "Found 7390 items\n",
        "2 datasets of sizes 5912,1478\n",
        "Setting up Pipeline: PILBase.create\n",
        "Setting up Pipeline: partial -> Categorize\n",
        "\n",
        "Building one sample\n",
        "  Pipeline: PILBase.create\n",
        "    starting from\n",
        "      /home/sgugger/.fastai/data/oxford-iiit-pet/images/american_bulldog_83.jpg\n",
        "    applying PILBase.create gives\n",
        "      PILImage mode=RGB size=375x500\n",
        "  Pipeline: partial -> Categorize\n",
        "    starting from\n",
        "      /home/sgugger/.fastai/data/oxford-iiit-pet/images/american_bulldog_83.jpg\n",
        "    applying partial gives\n",
        "      american_bulldog\n",
        "    applying Categorize gives\n",
        "      TensorCategory(12)\n",
        "\n",
        "Final sample: (PILImage mode=RGB size=375x500, TensorCategory(12))\n",
        "\n",
        "Setting up after_item: Pipeline: ToTensor\n",
        "Setting up before_batch: Pipeline: \n",
        "Setting up after_batch: Pipeline: IntToFloatTensor\n",
        "\n",
        "Building one batch\n",
        "Applying item_tfms to the first sample:\n",
        "  Pipeline: ToTensor\n",
        "    starting from\n",
        "      (PILImage mode=RGB size=375x500, TensorCategory(12))\n",
        "    applying ToTensor gives\n",
        "      (TensorImage of size 3x500x375, TensorCategory(12))\n",
        "\n",
        "Adding the next 3 samples\n",
        "\n",
        "No before_batch transform to apply\n",
        "\n",
        "Collating items in a batch\n",
        "Error! It's not possible to collate your items in a batch\n",
        "Could not collate the 0-th members of your tuples because got the following \n",
        "shapes:\n",
        "torch.Size([3, 500, 375]),torch.Size([3, 375, 500]),torch.Size([3, 333, 500]),\n",
        "torch.Size([3, 375, 500])\n",
        "```"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4pAl84bUBQhT"
      },
      "source": [
        "You can see exactly how we gathered the data and split it, how we went from a filename to a *sample* (the tuple (image, category)), then what item transforms were applied and how it failed to collate those samples in a batch (because of the different shapes). \n",
        "\n",
        "Once you think your data looks right, we generally recommend the next step should be using it to train a simple model. We often see people put off the training of an actual model for far too long. As a result, they don't actually find out what their baseline results look like. Perhaps your problem doesn't need lots of fancy domain-specific engineering. Or perhaps the data doesn't seem to train the model at all. These are things that you want to know as soon as possible. For this initial test, we'll use the same simple model that we used in `chapter_intro`:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 260,
          "referenced_widgets": [
            "def8024bae5345eebbe5ef6a55a46b6c",
            "0de0b5ad7f78457cbd6ea204ed61a131",
            "924f9d97696b40a29d01951b478f4b7d",
            "f3aa44456aee4fb5a090acb54214993f",
            "5b2e4dfc755d4b419114a658ee1d93b8",
            "2b90641b790c4cb8b756dd27a474294e",
            "6512233b89a84ed38000ea388d6f93b8",
            "0c63b82843744875a82af5039653a659"
          ]
        },
        "id": "ERGn7_DBBQhT",
        "outputId": "e5a739bf-0d22-490e-f973-6c592675bd1a"
      },
      "source": [
        "learn = cnn_learner(dls, resnet34, metrics=error_rate)\n",
        "learn.fine_tune(2)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading: \"https://download.pytorch.org/models/resnet34-333f7ec4.pth\" to /root/.cache/torch/hub/checkpoints/resnet34-333f7ec4.pth\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "def8024bae5345eebbe5ef6a55a46b6c",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, max=87306240.0), HTML(value='')))"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>1.505133</td>\n",
              "      <td>0.356941</td>\n",
              "      <td>0.119756</td>\n",
              "      <td>01:56</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>0.469629</td>\n",
              "      <td>0.268987</td>\n",
              "      <td>0.081191</td>\n",
              "      <td>02:29</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1</td>\n",
              "      <td>0.309376</td>\n",
              "      <td>0.232454</td>\n",
              "      <td>0.071719</td>\n",
              "      <td>02:29</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Q4us1hgVBQhU"
      },
      "source": [
        "As we've briefly discussed before, the table shown when we fit a model shows us the results after each epoch of training. Remember, an epoch is one complete pass through all of the images in the data. The columns shown are the average loss over the items of the training set, the loss on the validation set, and any metrics that we requested—in this case, the error rate.\n",
        "\n",
        "Remember that *loss* is whatever function we've decided to use to optimize the parameters of our model. But we haven't actually told fastai what loss function we want to use. So what is it doing? fastai will generally try to select an appropriate loss function based on what kind of data and model you are using. In this case we have image data and a categorical outcome, so fastai will default to using *cross-entropy loss*."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "e5_G05qWBQhU"
      },
      "source": [
        "## Cross-Entropy Loss"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "R-TE30ITBQhU"
      },
      "source": [
        "*Cross-entropy loss* is a loss function that is similar to the one we used in the previous chapter, but (as we'll see) has two benefits:\n",
        "\n",
        "- It works even when our dependent variable has more than two categories.\n",
        "- It results in faster and more reliable training.\n",
        "\n",
        "In order to understand how cross-entropy loss works for dependent variables with more than two categories, we first have to understand what the actual data and activations that are seen by the loss function look like."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QI0JHNXYBQhU"
      },
      "source": [
        "### Viewing Activations and Labels"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wyjshTe8BQhU"
      },
      "source": [
        "Let's take a look at the activations of our model. To actually get a batch of real data from our `DataLoaders`, we can use the `one_batch` method:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qIB72dFsBQhU"
      },
      "source": [
        "x,y = dls.one_batch()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0gAmpri8BQhV"
      },
      "source": [
        "As you see, this returns the dependent and independent variables, as a mini-batch. Let's see what is actually contained in our dependent variable:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ljPXvQt2BQhV",
        "outputId": "824ed24a-aef2-427c-8e95-09340e0dea08"
      },
      "source": [
        "y"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "TensorCategory([29, 24, 33, 14, 15, 33, 33, 15, 10, 29,  4,  0, 32, 35, 32, 15, 22,  4,  0,  1, 32, 28, 25, 17, 28, 23, 11, 28,  8, 23, 12, 18,  8, 20,  2,  3, 35, 19, 18, 23, 24, 18, 35, 19, 28, 33,  1, 11,\n",
              "         6,  3, 22, 10, 18, 31, 25,  8, 13,  8, 18, 12,  4, 36, 17, 10], device='cuda:0')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 35
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xmmUCI2EBQhV"
      },
      "source": [
        "Our batch size is 64, so we have 64 rows in this tensor. Each row is a single integer between 0 and 36, representing our 37 possible pet breeds. We can view the predictions (that is, the activations of the final layer of our neural network) using `Learner.get_preds`. This function either takes a dataset index (0 for train and 1 for valid) or an iterator of batches. Thus, we can pass it a simple list with our batch to get our predictions. It returns predictions and targets by default, but since we already have the targets, we can effectively ignore them by assigning to the special variable `_`:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 88
        },
        "id": "q2R6EI7jBQhV",
        "outputId": "fe79e907-2b4c-4e9f-92aa-e9b5e9e316e1"
      },
      "source": [
        "preds,_ = learn.get_preds(dl=[(x,y)])\n",
        "preds[0]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              ""
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([6.9139e-07, 1.4340e-06, 4.1220e-06, 5.1548e-06, 1.0022e-05, 3.0446e-05, 1.6502e-06, 1.6776e-04, 1.6061e-05, 1.4616e-05, 2.3343e-04, 1.2112e-06, 1.9846e-05, 2.1602e-07, 4.9744e-06, 6.6713e-06,\n",
              "        6.6926e-06, 1.1636e-05, 3.8119e-06, 4.4585e-07, 1.1631e-06, 2.2413e-06, 7.2751e-05, 4.3172e-04, 8.7021e-06, 1.4319e-06, 2.8707e-06, 1.3726e-06, 9.9276e-06, 9.9886e-01, 6.6512e-06, 9.8673e-06,\n",
              "        1.5690e-06, 1.8722e-05, 9.6846e-06, 5.3797e-06, 1.1561e-05])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "en6QWG6BBQhV"
      },
      "source": [
        "The actual predictions are 37 probabilities between 0 and 1, which add up to 1 in total:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gAWPyfgSBQhV",
        "outputId": "2f2d8dd9-193b-4419-f261-28e8aa1641b4"
      },
      "source": [
        "len(preds[0]),preds[0].sum()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(37, tensor(1.0000))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 38
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mMuD0YdMBQhW"
      },
      "source": [
        "To transform the activations of our model into predictions like this, we used something called the *softmax* activation function."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gxC8jmbXBQhW"
      },
      "source": [
        "### Softmax"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UNhFNaV6BQhW"
      },
      "source": [
        "In our classification model, we use the softmax activation function in the final layer to ensure that the activations are all between 0 and 1, and that they sum to 1.\n",
        "\n",
        "Softmax is similar to the sigmoid function, which we saw earlier. As a reminder sigmoid looks like this:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 319
        },
        "id": "2xz3U4pvBQhW",
        "outputId": "822822f1-4fc7-491e-cd31-4b95f72d5054"
      },
      "source": [
        "plot_function(torch.sigmoid, min=-4,max=4)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/fastbook/__init__.py:73: UserWarning: Not providing a value for linspace's steps is deprecated and will throw a runtime error in a future release. This warning will appear only once per process. (Triggered internally at  /pytorch/aten/src/ATen/native/RangeFactories.cpp:23.)\n",
            "  x = torch.linspace(min,max)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "k91Sg-ChBQhW"
      },
      "source": [
        "We can apply this function to a single column of activations from a neural network, and get back a column of numbers between 0 and 1, so it's a very useful activation function for our final layer.\n",
        "\n",
        "Now think about what happens if we want to have more categories in our target (such as our 37 pet breeds). That means we'll need more activations than just a single column: we need an activation *per category*. We can create, for instance, a neural net that predicts 3s and 7s that returns two activations, one for each class—this will be a good first step toward creating the more general approach. Let's just use some random numbers with a standard deviation of 2 (so we multiply `randn` by 2) for this example, assuming we have 6 images and 2 possible categories (where the first column represents 3s and the second is 7s):"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MkoHY3C8BQhX"
      },
      "source": [
        "torch.random.manual_seed(42);"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tA5Xtl1XBQhX",
        "outputId": "af7f8c01-c009-4179-a3d9-06d8fbe41e96"
      },
      "source": [
        "acts = torch.randn((6,2))*2\n",
        "acts"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[ 0.6734,  0.2576],\n",
              "        [ 0.4689,  0.4607],\n",
              "        [-2.2457, -0.3727],\n",
              "        [ 4.4164, -1.2760],\n",
              "        [ 0.9233,  0.5347],\n",
              "        [ 1.0698,  1.6187]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 40
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Wo4aaSwjBQhX"
      },
      "source": [
        "We can't just take the sigmoid of this directly, since we don't get rows that add to 1 (i.e., we want the probability of being a 3 plus the probability of being a 7 to add up to 1):"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eZP_GQ3FBQhX",
        "outputId": "48ce34c5-f3f3-48c0-95a1-322aa6d96c3f"
      },
      "source": [
        "acts.sigmoid()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[0.6623, 0.5641],\n",
              "        [0.6151, 0.6132],\n",
              "        [0.0957, 0.4079],\n",
              "        [0.9881, 0.2182],\n",
              "        [0.7157, 0.6306],\n",
              "        [0.7446, 0.8346]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GxdJK9PkBQhX"
      },
      "source": [
        "In `chapter_mnist_basics`, our neural net created a single activation per image, which we passed through the `sigmoid` function. That single activation represented the model's confidence that the input was a 3. Binary problems are a special case of classification problems, because the target can be treated as a single boolean value, as we did in `mnist_loss`. But binary problems can also be thought of in the context of the more general group of classifiers with any number of categories: in this case, we happen to have two categories. As we saw in the bear classifier, our neural net will return one activation per category.\n",
        "\n",
        "So in the binary case, what do those activations really indicate? A single pair of activations simply indicates the *relative* confidence of the input being a 3 versus being a 7. The overall values, whether they are both high, or both low, don't matter—all that matters is which is higher, and by how much.\n",
        "\n",
        "We would expect that since this is just another way of representing the same problem, that we would be able to use `sigmoid` directly on the two-activation version of our neural net. And indeed we can! \n",
        "\n",
        "We can just take the *difference* between the neural net activations, because that reflects how much more sure we are of the input being a 3 than a 7, and then take the sigmoid of that:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ou1KkVfKBQhY",
        "outputId": "c2a191aa-b181-4c81-b05f-31417d2ecdc4"
      },
      "source": [
        "(acts[:,0]-acts[:,1]).sigmoid()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([0.6025, 0.5021, 0.1332, 0.9966, 0.5959, 0.3661])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 42
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QUSarM7aBQhY"
      },
      "source": [
        "The second column (the probability of it being a 7) will then just be that value subtracted from 1. Now, we need a way to do all this that also works for more than two columns. It turns out that this function, called `softmax`, is exactly that:\n",
        "\n",
        "``` python\n",
        "def softmax(x): return exp(x) / exp(x).sum(dim=1, keepdim=True)\n",
        "```"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oCIMytcVBQhY"
      },
      "source": [
        "> jargon: Exponential function (exp): Literally defined as `e**x`, where `e` is a special number approximately equal to 2.718. It is the inverse of the natural logarithm function. Note that `exp` is always positive, and it increases _very_ rapidly!"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "A0KoryuDBQhY"
      },
      "source": [
        "Let's check that `softmax` returns the same values as `sigmoid` for the first column, and those values subtracted from 1 for the second column:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9VtwTcJjBQhY",
        "outputId": "70c8e7b1-9726-47db-a84f-8d3219cef882"
      },
      "source": [
        "sm_acts = torch.softmax(acts, dim=1)\n",
        "sm_acts"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[0.6025, 0.3975],\n",
              "        [0.5021, 0.4979],\n",
              "        [0.1332, 0.8668],\n",
              "        [0.9966, 0.0034],\n",
              "        [0.5959, 0.4041],\n",
              "        [0.3661, 0.6339]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 43
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "55cGpiJFBQhY"
      },
      "source": [
        "`softmax` is the multi-category equivalent of `sigmoid`—we have to use it any time we have more than two categories and the probabilities of the categories must add to 1, and we often use it even when there are just two categories, just to make things a bit more consistent. We could create other functions that have the properties that all activations are between 0 and 1, and sum to 1; however, no other function has the same relationship to the sigmoid function, which we've seen is smooth and symmetric. Also, we'll see shortly that the softmax function works well hand-in-hand with the loss function we will look at in the next section.\n",
        "\n",
        "If we have three output activations, such as in our bear classifier, calculating softmax for a single bear image would then look like something like `Example of softmax on the bear classifier`."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oCwqWxJzBQhZ"
      },
      "source": [
        "*Example of softmax on the bear classifier*\n",
        "\n",
        "<img alt=\"Bear softmax example\" width=\"280\" id=\"bear_softmax\" caption=\"Example of softmax on the bear classifier\" src=\"https://raw.githubusercontent.com/vtecftwy/fastbook/master/images/att_00062.png\">"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Fb_ivNmQBQhZ"
      },
      "source": [
        "What does this function do in practice? Taking the exponential ensures all our numbers are positive, and then dividing by the sum ensures we are going to have a bunch of numbers that add up to 1. The exponential also has a nice property: if one of the numbers in our activations `x` is slightly bigger than the others, the exponential will amplify this (since it grows, well... exponentially), which means that in the softmax, that number will be closer to 1. \n",
        "\n",
        "Intuitively, the softmax function *really* wants to pick one class among the others, so it's ideal for training a classifier when we know each picture has a definite label. (Note that it may be less ideal during inference, as you might want your model to sometimes tell you it doesn't recognize any of the classes that it has seen during training, and not pick a class because it has a slightly bigger activation score. In this case, it might be better to train a model using multiple binary output columns, each using a sigmoid activation.)\n",
        "\n",
        "Softmax is the first part of the cross-entropy loss—the second part is log likelihood. "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UK8qwkZuBQhZ"
      },
      "source": [
        "### Log Likelihood"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PgQMlgYrBQhZ"
      },
      "source": [
        "When we calculated the loss for our MNIST example in the last chapter we used:\n",
        "\n",
        "```python\n",
        "def mnist_loss(inputs, targets):\n",
        "    inputs = inputs.sigmoid()\n",
        "    return torch.where(targets==1, 1-inputs, inputs).mean()\n",
        "```\n",
        "\n",
        "Just as we moved from sigmoid to softmax, we need to extend the loss function to work with more than just binary classification—it needs to be able to classify any number of categories (in this case, we have 37 categories). Our activations, after softmax, are between 0 and 1, and sum to 1 for each row in the batch of predictions. Our targets are integers between 0 and 36.\n",
        "\n",
        "In the binary case, we used `torch.where` to select between `inputs` and `1-inputs`. When we treat a binary classification as a general classification problem with two categories, it actually becomes even easier, because (as we saw in the previous section) we now have two columns, containing the equivalent of `inputs` and `1-inputs`. So, all we need to do is select from the appropriate column. Let's try to implement this in PyTorch. For our synthetic 3s and 7s example, let's say these are our labels:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oTK2Gv6dBQhZ"
      },
      "source": [
        "targ = tensor([0,1,0,1,1,0])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8RIj8UuLBQhZ"
      },
      "source": [
        "and these are the softmax activations:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XuRrVMbuBQhZ",
        "outputId": "01969747-ae1a-4d4c-bd4d-9212b38c1429"
      },
      "source": [
        "sm_acts"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[0.6025, 0.3975],\n",
              "        [0.5021, 0.4979],\n",
              "        [0.1332, 0.8668],\n",
              "        [0.9966, 0.0034],\n",
              "        [0.5959, 0.4041],\n",
              "        [0.3661, 0.6339]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 45
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-dmUVtJBBQhZ"
      },
      "source": [
        "Then for each item of `targ` we can use that to select the appropriate column of `sm_acts` using tensor indexing, like so:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9bcUj6IvBQha",
        "outputId": "22c60e03-aff4-4cb0-8d5d-3c5acafc32b9"
      },
      "source": [
        "idx = range(6)\n",
        "sm_acts[idx, targ]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([0.6025, 0.4979, 0.1332, 0.0034, 0.4041, 0.3661])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 46
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FtoQ1NUkBQha"
      },
      "source": [
        "To see exactly what's happening here, let's put all the columns together in a table. Here, the first two columns are our activations, then we have the targets, the row index, and finally the result shown immediately above:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 166
        },
        "id": "gQv7ypUsBQha",
        "outputId": "26330f60-e4bf-4a5b-f488-ac53bc0eb3b7"
      },
      "source": [
        "from IPython.display import HTML\n",
        "df = pd.DataFrame(sm_acts, columns=[\"3\",\"7\"])\n",
        "df['targ'] = targ\n",
        "df['idx'] = idx\n",
        "df['loss'] = sm_acts[range(6), targ]\n",
        "t = df.style.hide_index()\n",
        "#To have html code compatible with our script\n",
        "html = t._repr_html_().split('</style>')[1]\n",
        "html = re.sub(r'<table id=\"([^\"]+)\"\\s*>', r'<table >', html)\n",
        "display(HTML(html))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<table ><thead>    <tr>        <th class=\"col_heading level0 col0\" >3</th>        <th class=\"col_heading level0 col1\" >7</th>        <th class=\"col_heading level0 col2\" >targ</th>        <th class=\"col_heading level0 col3\" >idx</th>        <th class=\"col_heading level0 col4\" >loss</th>    </tr></thead><tbody>\n",
              "                <tr>\n",
              "                                <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row0_col0\" class=\"data row0 col0\" >0.602469</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row0_col1\" class=\"data row0 col1\" >0.397531</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row0_col2\" class=\"data row0 col2\" >0</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row0_col3\" class=\"data row0 col3\" >0</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row0_col4\" class=\"data row0 col4\" >0.602469</td>\n",
              "            </tr>\n",
              "            <tr>\n",
              "                                <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row1_col0\" class=\"data row1 col0\" >0.502065</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row1_col1\" class=\"data row1 col1\" >0.497935</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row1_col2\" class=\"data row1 col2\" >1</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row1_col3\" class=\"data row1 col3\" >1</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row1_col4\" class=\"data row1 col4\" >0.497935</td>\n",
              "            </tr>\n",
              "            <tr>\n",
              "                                <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row2_col0\" class=\"data row2 col0\" >0.133188</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row2_col1\" class=\"data row2 col1\" >0.866811</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row2_col2\" class=\"data row2 col2\" >0</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row2_col3\" class=\"data row2 col3\" >2</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row2_col4\" class=\"data row2 col4\" >0.133188</td>\n",
              "            </tr>\n",
              "            <tr>\n",
              "                                <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row3_col0\" class=\"data row3 col0\" >0.996640</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row3_col1\" class=\"data row3 col1\" >0.003360</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row3_col2\" class=\"data row3 col2\" >1</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row3_col3\" class=\"data row3 col3\" >3</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row3_col4\" class=\"data row3 col4\" >0.003360</td>\n",
              "            </tr>\n",
              "            <tr>\n",
              "                                <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row4_col0\" class=\"data row4 col0\" >0.595949</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row4_col1\" class=\"data row4 col1\" >0.404051</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row4_col2\" class=\"data row4 col2\" >1</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row4_col3\" class=\"data row4 col3\" >4</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row4_col4\" class=\"data row4 col4\" >0.404051</td>\n",
              "            </tr>\n",
              "            <tr>\n",
              "                                <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row5_col0\" class=\"data row5 col0\" >0.366118</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row5_col1\" class=\"data row5 col1\" >0.633882</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row5_col2\" class=\"data row5 col2\" >0</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row5_col3\" class=\"data row5 col3\" >5</td>\n",
              "                        <td id=\"T_5196729c_c1eb_11eb_a073_0242ac1c0002row5_col4\" class=\"data row5 col4\" >0.366118</td>\n",
              "            </tr>\n",
              "    </tbody></table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "l4tPxRRABQha"
      },
      "source": [
        "Looking at this table, you can see that the final column can be calculated by taking the `targ` and `idx` columns as indices into the two-column matrix containing the `3` and `7` columns. That's what `sm_acts[idx, targ]` is actually doing.\n",
        "\n",
        "The really interesting thing here is that this actually works just as well with more than two columns. To see this, consider what would happen if we added an activation column for every digit (0 through 9), and then `targ` contained a number from 0 to 9. As long as the activation columns sum to 1 (as they will, if we use softmax), then we'll have a loss function that shows how well we're predicting each digit.\n",
        "\n",
        "We're only picking the loss from the column containing the correct label. We don't need to consider the other columns, because by the definition of softmax, they add up to 1 minus the activation corresponding to the correct label. Therefore, making the activation for the correct label as high as possible must mean we're also decreasing the activations of the remaining columns.\n",
        "\n",
        "PyTorch provides a function that does exactly the same thing as `sm_acts[range(n), targ]` (except it takes the negative, because when applying the log afterward, we will have negative numbers), called `nll_loss` (*NLL* stands for *negative log likelihood*):"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JM3obc9_BQha",
        "outputId": "dadcc143-c416-41a0-fcfa-480cce47a925"
      },
      "source": [
        "-sm_acts[idx, targ]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([-0.6025, -0.4979, -0.1332, -0.0034, -0.4041, -0.3661])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 48
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sT1cR63zBQha",
        "outputId": "21c5d941-b19e-4db5-bffc-2ca371e0767b"
      },
      "source": [
        "F.nll_loss(sm_acts, targ, reduction='none')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([-0.6025, -0.4979, -0.1332, -0.0034, -0.4041, -0.3661])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 49
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rW_Ha-HlBQhb"
      },
      "source": [
        "Despite its name, this PyTorch function does not take the log. We'll see why in the next section, but first, let's see why taking the logarithm can be useful."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3Fh4itXLBQhb"
      },
      "source": [
        "### Taking the Log"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-Ux9oVkDBQhb"
      },
      "source": [
        "The function we saw in the previous section works quite well as a loss function, but we can make it a bit better. The problem is that we are using probabilities, and probabilities cannot be smaller than 0 or greater than 1. That means that our model will not care whether it predicts 0.99 or 0.999. Indeed, those numbers are so close together—but in another sense, 0.999 is 10 times more confident than 0.99. So, we want to transform our numbers between 0 and 1 to instead be between negative infinity and infinity. There is a mathematical function that does exactly this: the *logarithm* (available as `torch.log`). It is not defined for numbers less than 0, and looks like this:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "3A4tj4edBQhb",
        "outputId": "74cdda59-6b0a-412f-c37f-b963b3f8cb69"
      },
      "source": [
        "plot_function(torch.log, min=0,max=4)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "969lAuUVBQhb"
      },
      "source": [
        "Does \"logarithm\" ring a bell? The logarithm function has this identity:\n",
        "\n",
        "```\n",
        "y = b**a\n",
        "a = log(y,b)\n",
        "```\n",
        "\n",
        "In this case, we're assuming that `log(y,b)` returns *log y base b*. However, PyTorch actually doesn't define `log` this way: `log` in Python uses the special number `e` (2.718...) as the base.\n",
        "\n",
        "Perhaps a logarithm is something that you have not thought about for the last 20 years or so. But it's a mathematical idea that is going to be really critical for many things in deep learning, so now would be a great time to refresh your memory. The key thing to know about logarithms is this relationship:\n",
        "\n",
        "    log(a*b) = log(a)+log(b)\n",
        "\n",
        "When we see it in that format, it looks a bit boring; but think about what this really means. It means that logarithms increase linearly when the underlying signal increases exponentially or multiplicatively. This is used, for instance, in the Richter scale of earthquake severity, and the dB scale of noise levels. It's also often used on financial charts, where we want to show compound growth rates more clearly. Computer scientists love using logarithms, because it means that multiplication, which can create really really large and really really small numbers, can be replaced by addition, which is much less likely to result in scales that are difficult for our computers to handle."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4qTbfhCpBQhb"
      },
      "source": [
        "> s: It's not just computer scientists that love logs! Until computers came along, engineers and scientists used a special ruler called a \"slide rule\" that did multiplication by adding logarithms. Logarithms are widely used in physics, for multiplying very big or very small numbers, and many other fields."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LgMjcglFBQhb"
      },
      "source": [
        "Taking the mean of the positive or negative log of our probabilities (depending on whether it's the correct or incorrect class) gives us the *negative log likelihood* loss. **In PyTorch, `nll_loss` assumes that you already took the log of the softmax, so it doesn't actually do the logarithm for you**."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OI-rU1AZBQhc"
      },
      "source": [
        "> warning: Confusing Name, Beware: The nll in **`nll_loss`** stands for \"negative log likelihood,\" but it doesn't actually take the log at all! It assumes you have _already_ taken the log. PyTorch has a function called **`log_softmax`** that combines `log` and `softmax` in a fast and accurate way. `nll_loss` is designed to be used after `log_softmax`."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Am8cYS11BQhc"
      },
      "source": [
        "When we first take the softmax, and then the log likelihood of that, that combination is called *cross-entropy loss*. In PyTorch, this is available as `nn.CrossEntropyLoss` (which, in practice, actually does `log_softmax` and then `nll_loss`):"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "60_9gaxGBQhc"
      },
      "source": [
        "loss_func = nn.CrossEntropyLoss()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Y_CmKRxxBQhc"
      },
      "source": [
        "As you see, this is a class. Instantiating it gives you an object which behaves like a function:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vNdizRyxBQhc",
        "outputId": "0f735088-4a51-4c7b-d5eb-89aa79d3f2eb"
      },
      "source": [
        "loss_func(acts, targ)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor(1.8045)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 52
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "62LKD7avBQhc"
      },
      "source": [
        "All PyTorch loss functions are provided in two forms, the class just shown above, and also a plain functional form, available in the `F` namespace:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6fLfjGv5BQhc",
        "outputId": "3155ba9f-531f-4e9f-ff94-6c59fb35a925"
      },
      "source": [
        "F.cross_entropy(acts, targ)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor(1.8045)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 53
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "V0ESaU1_BQhd"
      },
      "source": [
        "Either one works fine and can be used in any situation. We've noticed that most people tend to use the class version, and that's more often used in PyTorch's official docs and examples, so we'll tend to use that too.\n",
        "\n",
        "By default PyTorch loss functions take the mean of the loss of all items. You can use `reduction='none'` to disable that:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CuO74R9eBQhd",
        "outputId": "8953cf86-96be-47e8-a02a-ca3cb1403b2e"
      },
      "source": [
        "nn.CrossEntropyLoss(reduction='none')(acts, targ)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([0.5067, 0.6973, 2.0160, 5.6958, 0.9062, 1.0048])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 54
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gIHddn6GBQhd"
      },
      "source": [
        "> s: An interesting feature about cross-entropy loss appears when we consider its gradient. The gradient of `cross_entropy(a,b)` is just `softmax(a)-b`. Since `softmax(a)` is just the final activation of the model, that means that the gradient is proportional to the difference between the prediction and the target. This is the same as mean squared error in regression (assuming there's no final activation function such as that added by `y_range`), since the gradient of `(a-b)**2` is `2*(a-b)`. Because the gradient is linear, that means we won't see sudden jumps or exponential increases in gradients, which should lead to smoother training of models."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7ayoUSCiBQhd"
      },
      "source": [
        "We have now seen all the pieces hidden behind our loss function. But while this puts a number on how well (or badly) our model is doing, it does nothing to help us know if it's actually any good. Let's now see some ways to interpret our model's predictions."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Qusd6uqVBQhd"
      },
      "source": [
        "## Model Interpretation"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LNnenVLEBQhd"
      },
      "source": [
        "It's very hard to interpret loss functions directly, because they are designed to be things computers can differentiate and optimize, not things that people can understand. That's why we have metrics. These are not used in the optimization process, but just to help us poor humans understand what's going on. In this case, our accuracy is looking pretty good already! So where are we making mistakes?\n",
        "\n",
        "We saw in `chapter_intro` that we can use a confusion matrix to see where our model is doing well, and where it's doing badly:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 743
        },
        "id": "Kinb3rT9BQhd",
        "outputId": "8ff2f194-6844-423c-8512-8d54bfbee0f1"
      },
      "source": [
        "interp = ClassificationInterpretation.from_learner(learn)\n",
        "interp.plot_confusion_matrix(figsize=(12,12), dpi=60)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              ""
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cfMpKW1tBQhd"
      },
      "source": [
        "Oh dear—in this case, a confusion matrix is very hard to read. We have 37 different breeds of pet, which means we have 37×37 entries in this giant matrix! Instead, we can use the `most_confused` method, which just shows us the cells of the confusion matrix with the most incorrect predictions (here, with at least 5 or more):"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EVvc1D7lBQhe",
        "outputId": "24d269b3-b1b0-4b17-c8ee-56deee2f2d6b"
      },
      "source": [
        "interp.most_confused(min_val=5)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[('Ragdoll', 'Birman', 7),\n",
              " ('american_pit_bull_terrier', 'staffordshire_bull_terrier', 7),\n",
              " ('miniature_pinscher', 'chihuahua', 6),\n",
              " ('staffordshire_bull_terrier', 'american_pit_bull_terrier', 5)]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 56
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iSaahGrJBQhe"
      },
      "source": [
        "Since we are not pet breed experts, it is hard for us to know whether these category errors reflect actual difficulties in recognizing breeds. So again, we turn to Google. A little bit of Googling tells us that the most common category errors shown here are actually breed differences that even expert breeders sometimes disagree about. So this gives us some comfort that we are on the right track.\n",
        "\n",
        "We seem to have a good baseline. What can we do now to make it even better?"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "54BMRXwgBQhe"
      },
      "source": [
        "## Improving Our Model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DpQx_vAdBQhe"
      },
      "source": [
        "We will now look at a range of techniques to improve the training of our model and make it better. While doing so, we will explain a little bit more about transfer learning and how to fine-tune our pretrained model as best as possible, without breaking the pretrained weights.\n",
        "\n",
        "The first thing we need to set when training a model is the learning rate. We saw in the previous chapter that it needs to be just right to train as efficiently as possible, so how do we pick a good one? fastai provides a tool for this."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Qii7H-ssBQhe"
      },
      "source": [
        "### The Learning Rate Finder"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "U1dmZt9EBQhe"
      },
      "source": [
        "One of the most important things we can do when training a model is to make sure that we have the right learning rate. If our learning rate is too low, it can take many, many epochs to train our model. Not only does this waste time, but it also means that we may have problems with overfitting, because every time we do a complete pass through the data, we give our model a chance to memorize it.\n",
        "\n",
        "So let's just make our learning rate really high, right? Sure, let's try that and see what happens:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 143
        },
        "id": "t_eziClgBQhe",
        "outputId": "e3632bee-4796-4d01-ee73-6721a6eadc16"
      },
      "source": [
        "learn = cnn_learner(dls, resnet34, metrics=error_rate)\n",
        "learn.fine_tune(1, base_lr=0.1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>2.678281</td>\n",
              "      <td>2.865460</td>\n",
              "      <td>0.355886</td>\n",
              "      <td>01:56</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>4.507296</td>\n",
              "      <td>3.446720</td>\n",
              "      <td>0.906631</td>\n",
              "      <td>02:30</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DK3_xEktBQhf"
      },
      "source": [
        "That doesn't look good. Here's what happened. The optimizer stepped in the correct direction, but it stepped so far that it totally overshot the minimum loss. Repeating that multiple times makes it get further and further away, not closer and closer!\n",
        "\n",
        "What do we do to find the perfect learning rate—not too high, and not too low? In 2015 the researcher Leslie Smith came up with a brilliant idea, called the *learning rate finder*. His idea was to start with a very, very small learning rate, something so small that we would never expect it to be too big to handle. We use that for one mini-batch, find what the losses are afterwards, and then increase the learning rate by some percentage (e.g., doubling it each time). Then we do another mini-batch, track the loss, and double the learning rate again. We keep doing this until the loss gets worse, instead of better. This is the point where we know we have gone too far. We then select a learning rate a bit lower than this point. Our advice is to pick either:\n",
        "\n",
        "- One order of magnitude less than where the minimum loss was achieved (i.e., the minimum divided by 10)\n",
        "- The last point where the loss was clearly decreasing \n",
        "\n",
        "The learning rate finder computes those points on the curve to help you. Both these rules usually give around the same value. In the first chapter, we didn't specify a learning rate, using the default value from the fastai library (which is 1e-3):"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        },
        "id": "-tHV1mWbBQhf",
        "outputId": "d3decb9a-5c28-41b9-b085-c0648f55aafd"
      },
      "source": [
        "learn = cnn_learner(dls, resnet34, metrics=error_rate)\n",
        "lr_min,lr_steep = learn.lr_find()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              ""
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "n77ywAqGBQhf",
        "outputId": "56d1f489-e24f-4e1c-d0cc-6b2405326b10"
      },
      "source": [
        "print(f\"Minimum/10: {lr_min:.2e}, steepest point: {lr_steep:.2e}\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Minimum/10: 1.00e-02, steepest point: 6.31e-03\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WZ4CHI4oBQhf"
      },
      "source": [
        "We can see on this plot that in the range 1e-6 to 1e-3, nothing really happens and the model doesn't train. Then the loss starts to decrease until it reaches a minimum, and then increases again. We don't want a learning rate greater than 1e-1 as it will give a training that diverges like the one before (you can try for yourself), but 1e-1 is already too high: at this stage we've left the period where the loss was decreasing steadily.\n",
        "\n",
        "In this learning rate plot it appears that a learning rate around 3e-3 would be appropriate, so let's choose that:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 174
        },
        "id": "Cf2ZYQkUBQhf",
        "outputId": "c7ef0a3a-b24c-42ea-aa7f-cc6f738c7557"
      },
      "source": [
        "learn = cnn_learner(dls, resnet34, metrics=error_rate)\n",
        "learn.fine_tune(2, base_lr=3e-3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>1.361727</td>\n",
              "      <td>0.342064</td>\n",
              "      <td>0.114344</td>\n",
              "      <td>01:56</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>0.534960</td>\n",
              "      <td>0.384935</td>\n",
              "      <td>0.104871</td>\n",
              "      <td>02:30</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1</td>\n",
              "      <td>0.325205</td>\n",
              "      <td>0.217436</td>\n",
              "      <td>0.066306</td>\n",
              "      <td>02:30</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Mb8zzTpnBQhf"
      },
      "source": [
        "> Note: Logarithmic Scale: The learning rate finder plot has a logarithmic scale, which is why the middle point between 1e-3 and 1e-2 is between 3e-3 and 4e-3. This is because we care mostly about the order of magnitude of the learning rate."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0cxbexUSBQhg"
      },
      "source": [
        "It's interesting that the learning rate finder was only discovered in 2015, while neural networks have been under development since the 1950s. Throughout that time finding a good learning rate has been, perhaps, the most important and challenging issue for practitioners. The soltuon does not require any advanced maths, giant computing resources, huge datasets, or anything else that would make it inaccessible to any curious researcher. Furthermore, Leslie Smith, was not part of some exclusive Silicon Valley lab, but was working as a naval researcher. All of this is to say: breakthrough work in deep learning absolutely does not require access to vast resources, elite teams, or advanced mathematical ideas. There is lots of work still to be done that requires just a bit of common sense, creativity, and tenacity."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5HB0jhZ3BQhg"
      },
      "source": [
        "Now that we have a good learning rate to train our model, let's look at how we can fine-tune the weights of a pretrained model."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bsNoMrFlBQhg"
      },
      "source": [
        "### Unfreezing and Transfer Learning"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pOqGa-91BQhg"
      },
      "source": [
        "We discussed briefly in `chapter_intro` how transfer learning works. We saw that the basic idea is that a pretrained model, trained potentially on millions of data points (such as ImageNet), is fine-tuned for some other task. But what does this really mean?\n",
        "\n",
        "We now know that a convolutional neural network consists of many linear layers with a nonlinear activation function between each pair, followed by one or more final linear layers with an activation function such as softmax at the very end. The final linear layer uses a matrix with enough columns such that the output size is the same as the number of classes in our model (assuming that we are doing classification).\n",
        "\n",
        "This final linear layer is unlikely to be of any use for us when we are fine-tuning in a transfer learning setting, because it is specifically designed to classify the categories in the original pretraining dataset. So when we do transfer learning we remove it, throw it away, and replace it with a new linear layer with the correct number of outputs for our desired task (in this case, there would be 37 activations).\n",
        "\n",
        "This newly added linear layer will have entirely random weights. Therefore, our model prior to fine-tuning has entirely random outputs. But that does not mean that it is an entirely random model! All of the layers prior to the last one have been carefully trained to be good at image classification tasks in general. As we saw in the images from the [Zeiler and Fergus paper](https://arxiv.org/pdf/1311.2901.pdf) in `chapter_intro` (see `img_layer1` through `img_layer4`), the first few layers encode very general concepts, such as finding gradients and edges, and later layers encode concepts that are still very useful for us, such as finding eyeballs and fur.\n",
        "\n",
        "We want to train a model in such a way that we allow it to remember all of these generally useful ideas from the pretrained model, use them to solve our particular task (classify pet breeds), and only adjust them as required for the specifics of our particular task.\n",
        "\n",
        "Our challenge when fine-tuning is to replace the random weights in our added linear layers with weights that correctly achieve our desired task (classifying pet breeds) without breaking the carefully pretrained weights and the other layers. There is actually a very simple trick to allow this to happen: tell the optimizer to only update the weights in those randomly added final layers. Don't change the weights in the rest of the neural network at all. This is called *freezing* those pretrained layers."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8mHdp1FFBQhg"
      },
      "source": [
        "When we create a model from a pretrained network fastai automatically freezes all of the pretrained layers for us. When we call the `fine_tune` method fastai does two things:\n",
        "\n",
        "- Trains the randomly added layers for one epoch, with all other layers frozen\n",
        "- Unfreezes all of the layers, and trains them all for the number of epochs requested\n",
        "\n",
        "Although this is a reasonable default approach, it is likely that for your particular dataset you may get better results by doing things slightly differently. The `fine_tune` method has a number of parameters you can use to change its behavior, but it might be easiest for you to just call the underlying methods directly if you want to get some custom behavior. Remember that you can see the source code for the method by using the following syntax:\n",
        "\n",
        "    learn.fine_tune??\n",
        "\n",
        "So let's try doing this manually ourselves. First of all we will train the randomly added layers for three epochs, using `fit_one_cycle`. As mentioned in `chapter_intro`, `fit_one_cycle` is the suggested way to train models without using `fine_tune`. We'll see why later in the book; in short, what `fit_one_cycle` does is to start training at a low learning rate, gradually increase it for the first section of training, and then gradually decrease it again for the last section of training."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HDGTYqqJBQhg"
      },
      "source": [
        "learn.fine_tune??"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SMhUEcz8IZRD"
      },
      "source": [
        "``` python\n",
        "Signature: learn.fine_tune(epochs, base_lr=0.002, freeze_epochs=1, lr_mult=100, pct_start=0.3, div=5.0, \n",
        "                           lr_max=None, div_final=100000.0, wd=None, moms=None, cbs=None, reset_opt=False)\n",
        "Source:   \n",
        "@patch\n",
        "@delegates(Learner.fit_one_cycle)\n",
        "def fine_tune(self:Learner, epochs, base_lr=2e-3, freeze_epochs=1, lr_mult=100,\n",
        "              pct_start=0.3, div=5.0, **kwargs):\n",
        "    \"Fine tune with `freeze` for `freeze_epochs` then with `unfreeze` from `epochs` using discriminative LR\"\n",
        "    self.freeze()\n",
        "    self.fit_one_cycle(freeze_epochs, slice(base_lr), pct_start=0.99, **kwargs)\n",
        "    base_lr /= 2\n",
        "    self.unfreeze()\n",
        "    self.fit_one_cycle(epochs, slice(base_lr/lr_mult, base_lr), pct_start=pct_start, div=div, **kwargs)\n",
        "```\n",
        "source on [github](https://github.com/fastai/fastai/blob/master/fastai/callback/schedule.py#L104)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 142
        },
        "id": "0FRxfZqFBQhg",
        "outputId": "0cb64d36-4098-4183-b131-ccd69078119e"
      },
      "source": [
        "learn = cnn_learner(dls, resnet34, metrics=error_rate)\n",
        "learn.fit_one_cycle(3, 3e-3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>1.118552</td>\n",
              "      <td>0.313647</td>\n",
              "      <td>0.106901</td>\n",
              "      <td>01:55</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1</td>\n",
              "      <td>0.539335</td>\n",
              "      <td>0.239023</td>\n",
              "      <td>0.073748</td>\n",
              "      <td>01:56</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>2</td>\n",
              "      <td>0.327730</td>\n",
              "      <td>0.211937</td>\n",
              "      <td>0.066306</td>\n",
              "      <td>01:57</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oES5Ju5VBQhh"
      },
      "source": [
        "Then we'll unfreeze the model:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KbUg7udGBQhh"
      },
      "source": [
        "learn.unfreeze()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rbD6fj3-BQhh"
      },
      "source": [
        "and run `lr_find` again, because having more layers to train, and weights that have already been trained for three epochs, means our previously found learning rate isn't appropriate any more:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 300
        },
        "id": "auit8bmxBQhh",
        "outputId": "f476b447-33d8-48c0-9a5e-c6356f8e21c1"
      },
      "source": [
        "learn.lr_find()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              ""
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "SuggestedLRs(lr_min=9.999999747378752e-07, lr_steep=7.585775847473997e-07)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 64
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "farKlbo8BQhh"
      },
      "source": [
        "Note that the graph is a little different from when we had random weights: we don't have that sharp descent that indicates the model is training. That's because our model has been trained already. Here we have a somewhat flat area before a sharp increase, and we should take a point well before that sharp increase—for instance, 1e-5. The point with the maximum gradient isn't what we look for here and should be ignored.\n",
        "\n",
        "Let's train at a suitable learning rate:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 235
        },
        "id": "i1mClkIXBQhh",
        "outputId": "cd25af9a-8dc6-48a9-f0bd-dd9129e34207"
      },
      "source": [
        "learn.fit_one_cycle(6, lr_max=1e-5)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>0.249038</td>\n",
              "      <td>0.197518</td>\n",
              "      <td>0.061570</td>\n",
              "      <td>02:31</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1</td>\n",
              "      <td>0.246790</td>\n",
              "      <td>0.188312</td>\n",
              "      <td>0.059540</td>\n",
              "      <td>02:31</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>2</td>\n",
              "      <td>0.223633</td>\n",
              "      <td>0.186148</td>\n",
              "      <td>0.058187</td>\n",
              "      <td>02:31</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>3</td>\n",
              "      <td>0.222257</td>\n",
              "      <td>0.177872</td>\n",
              "      <td>0.056834</td>\n",
              "      <td>02:30</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>4</td>\n",
              "      <td>0.192824</td>\n",
              "      <td>0.173329</td>\n",
              "      <td>0.050744</td>\n",
              "      <td>02:30</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>5</td>\n",
              "      <td>0.188006</td>\n",
              "      <td>0.176515</td>\n",
              "      <td>0.054127</td>\n",
              "      <td>02:30</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3LlNcHhsBQhi"
      },
      "source": [
        "This has improved our model a bit, but there's more we can do. The deepest layers of our pretrained model might not need as high a learning rate as the last ones, so we should probably use different learning rates for those—this is known as using *discriminative learning rates*."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7QRAq4e3BQhi"
      },
      "source": [
        "### Discriminative Learning Rates"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dvWriTZxBQhi"
      },
      "source": [
        "Even after we unfreeze, we still care a lot about the quality of those pretrained weights. We would not expect that the best learning rate for those pretrained parameters would be as high as for the randomly added parameters, even after we have tuned those randomly added parameters for a few epochs. Remember, the pretrained weights have been trained for hundreds of epochs, on millions of images.\n",
        "\n",
        "In addition, do you remember the images we saw in `chapter_intro`, showing what each layer learns? The first layer learns very simple foundations, like edge and gradient detectors; these are likely to be just as useful for nearly any task. The later layers learn much more complex concepts, like \"eye\" and \"sunset,\" which might not be useful in your task at all (maybe you're classifying car models, for instance). So it makes sense to let the later layers fine-tune more quickly than earlier layers.\n",
        "\n",
        "Therefore, fastai's default approach is to use discriminative learning rates. This was originally developed in the ULMFiT approach to NLP transfer learning that we will introduce in `chapter_nlp`. Like many good ideas in deep learning, it is extremely simple: use a lower learning rate for the early layers of the neural network, and a higher learning rate for the later layers (and especially the randomly added layers). The idea is based on insights developed by [Jason Yosinski](https://arxiv.org/abs/1411.1792), who showed in 2014 that with transfer learning different layers of a neural network should train at different speeds, as seen in `Impact of different layers and training methods on transfer learning (Yosinski)`."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "drkVpQUjBQhi"
      },
      "source": [
        "*Impact of different layers and training methods on transfer learning (Yosinski)*\n",
        "\n",
        "<img alt=\"Impact of different layers and training methods on transfer learning (Yosinski)\" width=\"680\" caption=\"Impact of different layers and training methods on transfer learning (courtesy of Jason Yosinski et al.)\" id=\"yosinski\" src=\"https://raw.githubusercontent.com/vtecftwy/fastbook/master/images/att_00039.png\">"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "I1taU3FpBQhi"
      },
      "source": [
        "fastai lets you pass a Python `slice` object anywhere that a learning rate is expected. The first value passed will be the learning rate in the earliest layer of the neural network, and the second value will be the learning rate in the final layer. The layers in between will have learning rates that are multiplicatively equidistant throughout that range. Let's use this approach to replicate the previous training, but this time we'll only set the *lowest* layer of our net to a learning rate of 1e-6; the other layers will scale up to 1e-4. Let's train for a while and see what happens:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 546
        },
        "id": "HQzDX989BQhi",
        "outputId": "388186fb-2f31-4e74-f827-7ff6dac28033"
      },
      "source": [
        "learn = cnn_learner(dls, resnet34, metrics=error_rate)\n",
        "learn.fit_one_cycle(3, 3e-3)\n",
        "learn.unfreeze()\n",
        "learn.fit_one_cycle(12, lr_max=slice(1e-6,1e-4))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>1.126902</td>\n",
              "      <td>0.363168</td>\n",
              "      <td>0.117727</td>\n",
              "      <td>01:56</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1</td>\n",
              "      <td>0.515993</td>\n",
              "      <td>0.249321</td>\n",
              "      <td>0.082544</td>\n",
              "      <td>01:56</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>2</td>\n",
              "      <td>0.326551</td>\n",
              "      <td>0.218320</td>\n",
              "      <td>0.068336</td>\n",
              "      <td>01:55</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>0.266782</td>\n",
              "      <td>0.215367</td>\n",
              "      <td>0.066306</td>\n",
              "      <td>02:30</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1</td>\n",
              "      <td>0.248027</td>\n",
              "      <td>0.206848</td>\n",
              "      <td>0.058863</td>\n",
              "      <td>02:31</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>2</td>\n",
              "      <td>0.242612</td>\n",
              "      <td>0.199328</td>\n",
              "      <td>0.064953</td>\n",
              "      <td>02:31</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>3</td>\n",
              "      <td>0.210681</td>\n",
              "      <td>0.195977</td>\n",
              "      <td>0.063599</td>\n",
              "      <td>02:31</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>4</td>\n",
              "      <td>0.190988</td>\n",
              "      <td>0.199967</td>\n",
              "      <td>0.059540</td>\n",
              "      <td>02:31</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>5</td>\n",
              "      <td>0.176642</td>\n",
              "      <td>0.195279</td>\n",
              "      <td>0.062246</td>\n",
              "      <td>02:31</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>6</td>\n",
              "      <td>0.159173</td>\n",
              "      <td>0.191133</td>\n",
              "      <td>0.058863</td>\n",
              "      <td>02:31</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>7</td>\n",
              "      <td>0.145557</td>\n",
              "      <td>0.191110</td>\n",
              "      <td>0.058187</td>\n",
              "      <td>02:31</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>8</td>\n",
              "      <td>0.129775</td>\n",
              "      <td>0.195845</td>\n",
              "      <td>0.062923</td>\n",
              "      <td>02:31</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>9</td>\n",
              "      <td>0.128027</td>\n",
              "      <td>0.193581</td>\n",
              "      <td>0.058863</td>\n",
              "      <td>02:31</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>10</td>\n",
              "      <td>0.128468</td>\n",
              "      <td>0.193144</td>\n",
              "      <td>0.057510</td>\n",
              "      <td>02:31</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>11</td>\n",
              "      <td>0.132766</td>\n",
              "      <td>0.194398</td>\n",
              "      <td>0.059540</td>\n",
              "      <td>02:31</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mMBYzFibBQhi"
      },
      "source": [
        "Now the fine-tuning is working great!\n",
        "\n",
        "fastai can show us a graph of the training and validation loss:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "yOAG1U1qBQhj",
        "outputId": "d395af7e-9b54-415c-e119-8b43de4fbe68"
      },
      "source": [
        "learn.recorder.plot_loss()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2ThouSewBQhj"
      },
      "source": [
        "As you can see, the training loss keeps getting better and better. But notice that eventually the validation loss improvement slows, and sometimes even gets worse! This is the point at which the model is starting to over fit. In particular, the model is becoming overconfident of its predictions. But this does *not* mean that it is getting less accurate, necessarily. Take a look at the table of training results per epoch, and you will often see that the accuracy continues improving, even as the validation loss gets worse. In the end what matters is your accuracy, or more generally your chosen metrics, not the loss. The loss is just the function we've given the computer to help us to optimize."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SOOqGBbyBQhj"
      },
      "source": [
        "Another decision you have to make when training the model is for how long to train for. We'll consider that next."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RkV3jWvSBQhj"
      },
      "source": [
        "### Selecting the Number of Epochs"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DPUZV6KiBQhj"
      },
      "source": [
        "Often you will find that you are limited by time, rather than generalization and accuracy, when choosing how many epochs to train for. So your first approach to training should be to simply pick a number of epochs that will train in the amount of time that you are happy to wait for. Then look at the training and validation loss plots, as shown above, and in particular your metrics, and if you see that they are still getting better even in your final epochs, then you know that you have not trained for too long.\n",
        "\n",
        "On the other hand, you may well see that the metrics you have chosen are really getting worse at the end of training. Remember, it's not just that we're looking for the validation loss to get worse, but the actual metrics. Your validation loss will first get worse during training because the model gets overconfident, and only later will get worse because it is incorrectly memorizing the data. We only care in practice about the latter issue. Remember, our loss function is just something that we use to allow our optimizer to have something it can differentiate and optimize; it's not actually the thing we care about in practice.\n",
        "\n",
        "Before the days of 1cycle training it was very common to save the model at the end of each epoch, and then select whichever model had the best accuracy out of all of the models saved in each epoch. This is known as *early stopping*. However, this is very unlikely to give you the best answer, because those epochs in the middle occur before the  has had a chance to reach the small values, where it can really find the best result. Therefore, if you find that you have overfit, what you should actually do is retrain your model from scratch, and this time select a total number of epochs based on where your previous best results were found.\n",
        "\n",
        "If you have the time to train for more epochs, you may want to instead use that time to train more parameters—that is, use a deeper architecture."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rjiAi93oBQhj"
      },
      "source": [
        "### Deeper Architectures"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6jpizpAoBQhj"
      },
      "source": [
        "In general, a model with more parameters can model your data more accurately. (There are lots and lots of caveats to this generalization, and it depends on the specifics of the architectures you are using, but it is a reasonable rule of thumb for now.) For most of the architectures that we will be seeing in this book, you can create larger versions of them by simply adding more layers. However, since we want to use pretrained models, we need to make sure that we choose a number of layers that have already been pretrained for us.\n",
        "\n",
        "This is why, in practice, architectures tend to come in a small number of variants. For instance, the ResNet architecture that we are using in this chapter comes in variants with 18, 34, 50, 101, and 152 layer, pretrained on ImageNet. A larger (more layers and parameters; sometimes described as the \"capacity\" of a model) version of a ResNet will always be able to give us a better training loss, but it can suffer more from overfitting, because it has more parameters to overfit with.\n",
        "\n",
        "In general, a bigger model has the ability to better capture the real underlying relationships in your data, and also to capture and memorize the specific details of your individual images.\n",
        "\n",
        "However, using a deeper model is going to require more GPU RAM, so you may need to lower the size of your batches to avoid an *out-of-memory error*. This happens when you try to fit too much inside your GPU and looks like:\n",
        "\n",
        "```\n",
        "Cuda runtime error: out of memory\n",
        "```\n",
        "\n",
        "You may have to restart your notebook when this happens. The way to solve it is to use a smaller batch size, which means passing smaller groups of images at any given time through your model. You can pass the batch size you want to the call creating your `DataLoaders` with `bs=`.\n",
        "\n",
        "The other downside of deeper architectures is that they take quite a bit longer to train. One technique that can speed things up a lot is *mixed-precision training*. This refers to using less-precise numbers (*half-precision floating point*, also called *fp16*) where possible during training. As we are writing these words in early 2020, nearly all current NVIDIA GPUs support a special feature called *tensor cores* that can dramatically speed up neural network training, by 2-3x. They also require a lot less GPU memory. To enable this feature in fastai, just add `to_fp16()` after your `Learner` creation (you also need to import the module).\n",
        "\n",
        "You can't really know ahead of time what the best architecture for your particular problem is—you need to try training some. So let's try a ResNet-50 now with mixed precision:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3j-gCVapBQhk",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 446,
          "referenced_widgets": [
            "f6e65be4a85a4f69961315456fa64674",
            "99da4055787f401ba6977ce8724f736f",
            "dbf279950da247719598d6ffced0213b",
            "4c3436bced7a411787abb81c191e44dd",
            "306bd262954f47a1a318c7ddf665738a",
            "be8e6a1518334f529ea5e6a9713d174e",
            "d0ec09c4c37c4610a9e5603270c478f4",
            "f3b13a3ba52e4dc3922cdfbf0cbd0b8d"
          ]
        },
        "outputId": "e3dd10e5-8af4-44d3-dca9-ed6187ad8424"
      },
      "source": [
        "from fastai.callback.fp16 import *\n",
        "learn = cnn_learner(dls, resnet50, metrics=error_rate).to_fp16()\n",
        "learn.fine_tune(6, freeze_epochs=3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading: \"https://download.pytorch.org/models/resnet50-19c8e357.pth\" to /root/.cache/torch/hub/checkpoints/resnet50-19c8e357.pth\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "f6e65be4a85a4f69961315456fa64674",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, max=102502400.0), HTML(value='')))"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>nan</td>\n",
              "      <td>nan</td>\n",
              "      <td>0.968200</td>\n",
              "      <td>02:12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1</td>\n",
              "      <td>nan</td>\n",
              "      <td>nan</td>\n",
              "      <td>0.968200</td>\n",
              "      <td>02:12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>2</td>\n",
              "      <td>nan</td>\n",
              "      <td>nan</td>\n",
              "      <td>0.968200</td>\n",
              "      <td>02:12</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>nan</td>\n",
              "      <td>nan</td>\n",
              "      <td>0.968200</td>\n",
              "      <td>02:50</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1</td>\n",
              "      <td>nan</td>\n",
              "      <td>nan</td>\n",
              "      <td>0.968200</td>\n",
              "      <td>02:52</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>2</td>\n",
              "      <td>nan</td>\n",
              "      <td>nan</td>\n",
              "      <td>0.968200</td>\n",
              "      <td>02:52</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>3</td>\n",
              "      <td>nan</td>\n",
              "      <td>nan</td>\n",
              "      <td>0.968200</td>\n",
              "      <td>02:51</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>4</td>\n",
              "      <td>nan</td>\n",
              "      <td>nan</td>\n",
              "      <td>0.968200</td>\n",
              "      <td>02:51</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>5</td>\n",
              "      <td>nan</td>\n",
              "      <td>nan</td>\n",
              "      <td>0.968200</td>\n",
              "      <td>02:52</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pJEJ4J9xBQhk"
      },
      "source": [
        "You'll see here we've gone back to using `fine_tune`, since it's so handy! We can pass `freeze_epochs` to tell fastai how many epochs to train for while frozen. It will automatically change learning rates appropriately for most datasets.\n",
        "\n",
        "In this case, we're not seeing a clear win from the deeper model. This is useful to remember—bigger models aren't necessarily better models for your particular case! Make sure you try small models before you start scaling up."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "afBGzdARBQhk"
      },
      "source": [
        "## Conclusion"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vxi_laTJBQhk"
      },
      "source": [
        "In this chapter you learned some important practical tips, both for getting your image data ready for modeling (presizing, data block summary) and for fitting the model (, unfreezing, discriminative learning rates, setting the number of epochs, and using deeper architectures). Using these tools will help you to build more accurate image models, more quickly.\n",
        "\n",
        "We also discussed cross-entropy loss. This part of the book is worth spending plenty of time on. You aren't likely to need to actually implement cross-entropy loss from scratch yourself in practice, but it's really important you understand the inputs to and output from that function, because it (or a variant of it, as we'll see in the next chapter) is used in nearly every classification model. So when you want to debug a model, or put a model in production, or improve the accuracy of a model, you're going to need to be able to look at its activations and loss, and understand what's going on, and why. You can't do that properly if you don't understand your loss function.\n",
        "\n",
        "If cross-entropy loss hasn't \"clicked\" for you just yet, don't worry—you'll get there! First, go back to the last chapter and make sure you really understand `mnist_loss`. Then work gradually through the cells of the notebook for this chapter, where we step through each piece of cross-entropy loss. Make sure you understand what each calculation is doing, and why. Try creating some small tensors yourself and pass them into the functions, to see what they return.\n",
        "\n",
        "Remember: the choices made in the implementation of cross-entropy loss are not the only possible choices that could have been made. Just like when we looked at regression we could choose between mean squared error and mean absolute difference (L1). If you have other ideas for possible functions that you think might work, feel free to give them a try in this chapter's notebook! (Fair warning though: you'll probably find that the model will be slower to train, and less accurate. That's because the gradient of cross-entropy loss is proportional to the difference between the activation and the target, so SGD always gets a nicely scaled step for the weights.)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IbsvwiQ5PI4B"
      },
      "source": [
        "# unpackAI Assignments Section"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-e0_ErxnUAWm"
      },
      "source": [
        "| Week | Chapters                  | Assignment                                                            | Comment |\n",
        "|:----:|---------------------------|-----------------------------------------------------------------------|:-------:|\n",
        "| 1    | Warm-up Session           | No assignment                                                         | Done\n",
        "| 2    | Chapter 4, MNIST Basic    | Make a model using SGD for the full MNIST dataset. Built from scratch | Done\n",
        "|      | Chapter 5, Pet Breeds     | Read sections on Cross-Entropy Loss to use in model                   | Done\n",
        "| 3    | Chapter 5, Pet Breeds     | Read notebook, run it and perform assingments at the bottom           | NOW\n",
        "|      | Chapter 6, Multi-Category | Read notebook, run it and perform assingments at the bottom           | NOW\n",
        "|      | Chapter 7, Sizing and TTA | Read notebook, run it and perform assingments at the bottom           | NOW"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nTuZtDEPckhq"
      },
      "source": [
        "## Prepare environment"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KMl3iiBigLjq"
      },
      "source": [
        "To make test faster, we will use a smaller image set first. Once we see a trend, it is still possible to use the full set later on."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TRp54tNnZyJ6",
        "outputId": "c73fae52-e739-4d7b-9180-f3535363967c"
      },
      "source": [
        "number_images = 1000\n",
        "\n",
        "all_images = (path/'images').ls()\n",
        "imgs_small = path / 'img-small'\n",
        "if imgs_small.is_dir():\n",
        "    shutil.rmtree(imgs_small)\n",
        "os.makedirs(imgs_small, exist_ok=True)\n",
        "for f in all_images[:number_images]:\n",
        "    shutil.copy(src=f, dst=imgs_small / f.name)\n",
        "imgs_small.ls()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(#1000) [Path('img-small/saint_bernard_94.jpg'),Path('img-small/newfoundland_48.jpg'),Path('img-small/shiba_inu_206.jpg'),Path('img-small/Russian_Blue_83.jpg'),Path('img-small/scottish_terrier_78.jpg'),Path('img-small/samoyed_21.jpg'),Path('img-small/english_setter_56.jpg'),Path('img-small/miniature_pinscher_52.jpg'),Path('img-small/wheaten_terrier_76.jpg'),Path('img-small/boxer_49.jpg')...]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 69
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KBDqpLH8gWAy"
      },
      "source": [
        "We will run the model several times, but we want to compare results. To make the test reproducible, it is important to ensure all the random keys are seeded to the same value. This is how to do that (see reference [here](https://github.com/fastai/fastai/issues/2832#issuecomment-698759541)).\n",
        "- use `set_seed()` function as `set_seed(value, True)`\n",
        "- for each time we want to run the model:\n",
        "    - set the seed\n",
        "    - create the dls from the datablock\n",
        "    - create the learner\n",
        "    - train the learner\n",
        "\n",
        "for example:\n",
        "``` python\n",
        "set_seed(88, True)\n",
        "dls = pets.dataloaders(path/\"img-small\")\n",
        "learn = cnn_learner(dls, resnet18, metrics=error_rate)\n",
        "learn.fine_tune(2, base_lr=3e-3)\n",
        "```\n",
        "and then again\n",
        "``` python\n",
        "set_seed(88, True)\n",
        "dls = pets.dataloaders(path/\"img-small\")\n",
        "learn = cnn_learner(dls, resnet18, metrics=error_rate)\n",
        "learn.fine_tune(2, base_lr=3e-3)\n",
        "```\n",
        "\n",
        "This will give twice exactly the same result. If you change one parameter in one of the two models, you know any difference is due to that change only. If you remove the `set_seed()` function, it will not be the same result."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1PuAMFR_UGpV"
      },
      "source": [
        "##  Assignment 1: Rate Finder\n",
        "\n",
        "How effective is the learn rate finder (`lr_find()` [doc](https://docs.fast.ai/callback.schedule.html#Learner.lr_find))? Test this by running the same model three times with different rates, for instance: lr in the optimum area, lr too small and lr too big.\n",
        "\n",
        "*Reminder on how to use the learning rate finder*:\n",
        "- Run `learn.lr_finder()` which will print out suggestions and one plot:\n",
        "<img src=\"https://raw.githubusercontent.com/vtecftwy/fastbook/master/images/chapter5_lr_find_plot.png\">\n",
        "\n",
        "- Optimum LR: use a value in the recommended area (between where the loss function curve is the steepest and one thenth of where the loss function curve is the lowest). Both values are roughly the same. `lr` for the steepest point is smaller than `lr` for the lowest point.\n",
        "\n",
        "The suggested values `SuggestedLRs(lr_min=0.017378008365631102, lr_steep=0.0030199517495930195)` are the two options between which you can pick a learning rate. `lr_min` is one tenth of the `lr` at the lowest point in the curve and `lr_steep` is the `lr` where the curve has the steepest downward slope.\n",
        "\n",
        "You also can save `lr_min` and `lr_steep` by using \n",
        "``` python\n",
        "    lr_min, lr_steep = learn.lr_find()\n",
        "```"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZUGovIyFVqDk"
      },
      "source": [
        "# Prepare DataBlock that will be used for each step below\n",
        "pets = DataBlock(blocks = (ImageBlock, CategoryBlock),\n",
        "                 get_items=get_image_files, \n",
        "                 splitter=RandomSplitter(seed=42),\n",
        "                 get_y=using_attr(RegexLabeller(r'(.+)_\\d+.jpg$'), 'name'),\n",
        "                 item_tfms=Resize(460),\n",
        "                 batch_tfms=aug_transforms(size=224, min_scale=0.75))\n",
        "\n",
        "seed_value = 1988"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Fldw9Yakke4y"
      },
      "source": [
        "Use `lr_find` to pick the values, then train the model and plot the graph of train and validation losses."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 369,
          "referenced_widgets": [
            "4963a4429f4d418ba8ef420cfc8ffc52",
            "922b36ce7b674e02ba6ba9ec843c45ec",
            "ae780f942fe84f95a956b24bd4594d7e",
            "cacb1dd327414bdf94fbaddb16be3c01",
            "40f30d58d8104e94b1cd03c3c76e8ac3",
            "492c2e96587f4cc5a07697aeaa7b25d4",
            "bd303a2ae9524a3f950e01a6b1f1d33b",
            "1eef947766084bf9a94d30dd07a66cdd"
          ]
        },
        "id": "PxA1R5kKkife",
        "outputId": "c2d9c7d0-0512-422f-e4f4-69f72e5dd1d6"
      },
      "source": [
        "set_seed(seed_value, True)\n",
        "dls = pets.dataloaders(path/\"img-small\")\n",
        "learn = cnn_learner(dls, resnet18, metrics=error_rate)\n",
        "lr_min, lr_steep = learn.lr_find()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading: \"https://download.pytorch.org/models/resnet18-5c106cde.pth\" to /root/.cache/torch/hub/checkpoints/resnet18-5c106cde.pth\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "4963a4429f4d418ba8ef420cfc8ffc52",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, max=46827520.0), HTML(value='')))"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              ""
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9uHxcV2LwHpP",
        "outputId": "5abaf0f7-3361-4209-d328-dd433f08bbc5"
      },
      "source": [
        "print(f\"lr_min: {lr_min}; lr_steep: {lr_steep}\")\n",
        "\n",
        "# format float numbers to make them easier to read\n",
        "print(f\"lr_min: {lr_min:.1e}; lr_steep: {lr_steep:.1e}\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "lr_min: 0.012022644281387329; lr_steep: 0.005248074419796467\n",
            "lr_min: 1.2e-02; lr_steep: 5.2e-03\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DBNHIz7SUGgV"
      },
      "source": [
        "#### LR in Optimum Area"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qyx684frVK78",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 253
        },
        "outputId": "c7f53015-b7c4-4fca-a898-318c074c3bf5"
      },
      "source": [
        "lr = 1.2e-2# add your value\n",
        "print(f\"Using lr = {lr:.2e}\")\n",
        "\n",
        "set_seed(seed_value, True)\n",
        "dls = pets.dataloaders(path/\"img-small\")\n",
        "learn = cnn_learner(dls, resnet18, metrics=error_rate)\n",
        "\n",
        "learn.fine_tune(epochs=4, base_lr=lr)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using lr = 1.20e-02\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>2.955107</td>\n",
              "      <td>1.101228</td>\n",
              "      <td>0.280000</td>\n",
              "      <td>00:13</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>0.641772</td>\n",
              "      <td>0.760738</td>\n",
              "      <td>0.220000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1</td>\n",
              "      <td>0.669410</td>\n",
              "      <td>1.471529</td>\n",
              "      <td>0.355000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>2</td>\n",
              "      <td>0.644232</td>\n",
              "      <td>1.148305</td>\n",
              "      <td>0.290000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>3</td>\n",
              "      <td>0.550168</td>\n",
              "      <td>0.850822</td>\n",
              "      <td>0.225000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "biZlk_tRVTKH",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "outputId": "59bccb89-8546-47dc-c06f-66429ee9337b"
      },
      "source": [
        "learn.recorder.plot_loss()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Po-WzoK0UGV4"
      },
      "source": [
        "#### LR too small"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QuK1bHlkBQhl",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 253
        },
        "outputId": "c755eefe-344d-4e08-92b6-08c32aa6afa5"
      },
      "source": [
        "lr_too_small = 1e-7 # add your value\n",
        "print(f\"Using lr = {lr_too_small:.2e}\")\n",
        "\n",
        "set_seed(seed_value, True)\n",
        "dls = pets.dataloaders(path/\"img-small\")\n",
        "learn = cnn_learner(dls, resnet18, metrics=error_rate)\n",
        "\n",
        "learn.fine_tune(epochs=4, base_lr=lr_too_small)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using lr = 1.00e-07\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>5.333564</td>\n",
              "      <td>4.431967</td>\n",
              "      <td>0.980000</td>\n",
              "      <td>00:12</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>5.201937</td>\n",
              "      <td>4.300502</td>\n",
              "      <td>0.970000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1</td>\n",
              "      <td>5.278004</td>\n",
              "      <td>4.315574</td>\n",
              "      <td>0.970000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>2</td>\n",
              "      <td>5.278126</td>\n",
              "      <td>4.327958</td>\n",
              "      <td>0.970000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>3</td>\n",
              "      <td>5.336677</td>\n",
              "      <td>4.323915</td>\n",
              "      <td>0.970000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jFea8eOsVTfP",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "outputId": "cb64880f-fdb8-44b0-bc34-17eaa23263a4"
      },
      "source": [
        "learn.recorder.plot_loss()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Bf__-FDbVOg0"
      },
      "source": [
        "#### LR too large"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3h_tXdNbVRso",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 253
        },
        "outputId": "175c2765-69d5-442e-f84b-0a10fe51f388"
      },
      "source": [
        "lr_too_big = 1 # add your value\n",
        "print(f\"Using lr = {lr_too_big:.2e}\")\n",
        "\n",
        "set_seed(seed_value, True)\n",
        "dls = pets.dataloaders(path/\"img-small\")\n",
        "learn = cnn_learner(dls, resnet18, metrics=error_rate)\n",
        "\n",
        "learn.fine_tune(epochs=4, base_lr=lr_too_big)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using lr = 1.00e+00\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>28.818129</td>\n",
              "      <td>1136.027222</td>\n",
              "      <td>0.960000</td>\n",
              "      <td>00:12</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>79.201439</td>\n",
              "      <td>37250.199219</td>\n",
              "      <td>0.980000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1</td>\n",
              "      <td>66.022095</td>\n",
              "      <td>1154.632812</td>\n",
              "      <td>0.965000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>2</td>\n",
              "      <td>58.616680</td>\n",
              "      <td>674.372192</td>\n",
              "      <td>0.960000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>3</td>\n",
              "      <td>47.579521</td>\n",
              "      <td>60.055771</td>\n",
              "      <td>0.945000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qWSupFC3VT_k",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "outputId": "9c7d8ba1-faa8-4ade-96d5-66f6125aee85"
      },
      "source": [
        "learn.recorder.plot_loss()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ow5etdV_VrpP"
      },
      "source": [
        "#### Conclusions and Reflections on LR:\n",
        "- What do you see? \n",
        "- What is your intuition to explain what you see?\n",
        "- "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "i3y1fG3f_Oc6"
      },
      "source": [
        "the error rate goes extreamly high when learning rate is too large or too small\n",
        "\n",
        "Maybe it needs a proper learning rate to achieve a satisfying result, But if it's too large, it gets overfitted?"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kr-pDWiZ-1h-"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TNpn8UzPWFw9"
      },
      "source": [
        "## Assignment 2: Using pretrained models"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cHtEU389WLS3"
      },
      "source": [
        "We are using pretrained models. The training process to work with such model is:\n",
        "1. Train the last few layers of the model with the new dataset (only the parameters fo the last few layers are modified) \n",
        "2. Unfreeze the model and finetune the model by retraining using the new dataset (parameters of all the layers are modified)\n",
        "\n",
        "We can do this in two ways:\n",
        "1. use `.finetune()` ([doc](https://docs.fast.ai/callback.schedule.html#Learner.fine_tune))\n",
        "2. use `learn.fit_one_cycle` ([doc](https://docs.fast.ai/callback.schedule.html#Learner.fit_one_cycle)) then `learn.unfreeze` ([doc](https://docs.fast.ai/learner.html#Transfer-learning)) and again `learn.fit_one_cycle`\n",
        "\n",
        "In this assignement, you will build a model with the same dataset, using the two methods and reflect on the differences"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TfLwotYlXorp"
      },
      "source": [
        "# Prepare DataBlock that will be used for each steps below\n",
        "pets = DataBlock(blocks = (ImageBlock, CategoryBlock),\n",
        "                 get_items=get_image_files, \n",
        "                 splitter=RandomSplitter(seed=42),\n",
        "                 get_y=using_attr(RegexLabeller(r'(.+)_\\d+.jpg$'), 'name'),\n",
        "                 item_tfms=Resize(460),\n",
        "                 batch_tfms=aug_transforms(size=224, min_scale=0.75))\n",
        "\n",
        "seed_value = 1988"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xLDv_uuwXiGM"
      },
      "source": [
        "### `finetune`\n",
        "\n",
        "We want to train a model with enough epochs both for the last laywers and then all the laywers. Let's pick 6 epochs for each steps.\n",
        "\n",
        "*hint*: consider some of the arguments of  `fine_tune()`. There are more but these are the relevant ones:\n",
        "``` \n",
        "        Learner.fine_tune(epochs, base_lr=0.002, freeze_epochs=1)\n",
        "```"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 300
        },
        "id": "XQuR1zHoXhXT",
        "outputId": "79a02b0d-78e7-4fd5-b5f8-f5ed85e68f55"
      },
      "source": [
        "# Set model for reproducibility and find optimum lr\n",
        "set_seed(seed_value, True)\n",
        "dls = pets.dataloaders(path/\"img-small\")\n",
        "learn = cnn_learner(dls, resnet18, metrics=error_rate)\n",
        "lr_min, lr_steep = learn.lr_find()\n",
        "print(f\"lr_min: {lr_min:.1e}; lr_steep: {lr_steep:.1e}\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              ""
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "lr_min: 1.2e-02; lr_steep: 5.2e-03\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2i4seWJj9W0x",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b897dd51-7304-4d12-e4bd-02e6f19fb5b0"
      },
      "source": [
        "# Pick your optimum learning rate:\n",
        "lr = 1.2e-2# add your value\n",
        "print(f\"Using lr = {lr:.2e}\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using lr = 1.20e-02\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "da1TsnQ_hrAN",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 298
        },
        "outputId": "df752339-3b9d-4f8e-c366-28e5c64a4c78"
      },
      "source": [
        "# Define arguments to use optimum lr and 6 epochs both for training of last layers and for full model (total of 12)\n",
        "set_seed(seed_value, True)\n",
        "dls = pets.dataloaders(path/\"img-small\")\n",
        "learn = cnn_learner(dls, resnet18, metrics=error_rate)\n",
        "\n",
        "learn.fine_tune(epochs=6, base_lr=lr)\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>2.955107</td>\n",
              "      <td>1.101228</td>\n",
              "      <td>0.280000</td>\n",
              "      <td>00:13</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>0.607477</td>\n",
              "      <td>0.642282</td>\n",
              "      <td>0.180000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1</td>\n",
              "      <td>0.546239</td>\n",
              "      <td>1.031950</td>\n",
              "      <td>0.270000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>2</td>\n",
              "      <td>0.578255</td>\n",
              "      <td>1.093180</td>\n",
              "      <td>0.240000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>3</td>\n",
              "      <td>0.539418</td>\n",
              "      <td>1.156895</td>\n",
              "      <td>0.285000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>4</td>\n",
              "      <td>0.462502</td>\n",
              "      <td>0.912359</td>\n",
              "      <td>0.240000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>5</td>\n",
              "      <td>0.385166</td>\n",
              "      <td>0.817166</td>\n",
              "      <td>0.220000</td>\n",
              "      <td>00:14</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D5ZFW8N2iwvz",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "outputId": "723694ed-2dc9-4330-f04b-80d8652cff89"
      },
      "source": [
        "learn.recorder.plot_loss()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ow4sGZIXXlgT"
      },
      "source": [
        "### `fit_one_cycle`\n",
        "\n",
        "We want to do the same as above but with the additional optimisation of `fit_one_cycle`. The steps are:\n",
        "1. set the model and find optimum lr\n",
        "1. train for 6 epochs with the pretrained model using `fit_one_cycle` (remember that pretrained model come frozen)\n",
        "1. unfreeze the model\n",
        "1. perform a new `lr_find` step as now the model is very different (more layers). (*See the note below*).\n",
        "1. train again for 6 epochs with `fit_one_cycle`"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gOlhmPD1GBka"
      },
      "source": [
        "*Note on `lr_find` after unfreezing*: \n",
        "\n",
        "The shape of the loss curve will be very different when you use it after some training compared to the first time you use it.\n",
        "- The first time, you will usually see a \"deep valley\" after a flat period, at the moment the model starts really learning and loss drops dramatically.\n",
        "- After initial training, you will not see that valley at all. Instead you will see some noisy flat area, followed by a `lr` after which the training diverges and loss explodes. You will want to select a lr not too small but far enough from the \"explosion\" area. In the example below, a rate of 1e-4 or 2e-4 should work well.\n",
        "\n",
        "<img src=\"https://raw.githubusercontent.com/vtecftwy/fastbook/master/images/chapter5_lr_find_plot_random_model.png\">\n",
        "<img src=\"https://raw.githubusercontent.com/vtecftwy/fastbook/master/images/chapter5_lr_find_plot_after_unfreeze.png\">"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 300
        },
        "id": "lIj0q4SdXpf7",
        "outputId": "8e23b07e-864c-4623-cda1-e8a782068f00"
      },
      "source": [
        "# Set model for reproducibility and find optimum lr\n",
        "set_seed(seed_value, True)\n",
        "dls = pets.dataloaders(path/\"img-small\")\n",
        "learn = cnn_learner(dls, resnet18, metrics=error_rate)\n",
        "lr_min, lr_steep = learn.lr_find()\n",
        "print(f\"lr_min: {lr_min:.1e}; lr_steep: {lr_steep:.1e}\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              ""
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "lr_min: 1.2e-02; lr_steep: 5.2e-03\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "amtGRtieCv5X",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b6d19dce-afc7-4297-c696-ef99d0237ce2"
      },
      "source": [
        "# Pick your optimum learning rate. \n",
        "# In this case, because we have seeded the model, you should use the same lr as before for comparison.\n",
        "# But you can experiment with different one's\n",
        "lr = (lr_min + lr_steep) /2\n",
        "print(f\"Using lr = {lr:.2e}\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using lr = 8.64e-03\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zXvPeUyBijeW",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 235
        },
        "outputId": "07794d99-0a98-43be-b7ba-03119af91e88"
      },
      "source": [
        "# Define fit_one_cycle function and arguments to use optimum lr and 6 epochs\n",
        "\n",
        "learn.fine_tune??\n",
        "learn = cnn_learner(dls, resnet34, metrics=error_rate)\n",
        "learn.fit_one_cycle(6, lr)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>3.457027</td>\n",
              "      <td>0.966394</td>\n",
              "      <td>0.280000</td>\n",
              "      <td>00:16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1</td>\n",
              "      <td>1.920593</td>\n",
              "      <td>0.864308</td>\n",
              "      <td>0.205000</td>\n",
              "      <td>00:16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>2</td>\n",
              "      <td>1.317731</td>\n",
              "      <td>0.778791</td>\n",
              "      <td>0.205000</td>\n",
              "      <td>00:16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>3</td>\n",
              "      <td>0.971978</td>\n",
              "      <td>0.703349</td>\n",
              "      <td>0.200000</td>\n",
              "      <td>00:16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>4</td>\n",
              "      <td>0.739646</td>\n",
              "      <td>0.572107</td>\n",
              "      <td>0.160000</td>\n",
              "      <td>00:16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>5</td>\n",
              "      <td>0.580259</td>\n",
              "      <td>0.537628</td>\n",
              "      <td>0.160000</td>\n",
              "      <td>00:16</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xV16Ys4-j-Ja",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "outputId": "3500623d-02e0-4b57-ccc4-192e0bb5502a"
      },
      "source": [
        "learn.recorder.plot_loss()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWoAAAD4CAYAAADFAawfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deXiV5YH38e+dk33fyQYkYZEQ9k0sQl0ZXGsXxGp9u9ja6djFmW46navL+zozznSmrY5tZ7BqN9diHVvritUqFNkUMRCUJQkJIfu+b/f7xzmBAAkESPI855zf57rOdZ7znIVfIPxy5342Y61FRETcK8TpACIicnoqahERl1NRi4i4nIpaRMTlVNQiIi4XOh4fmpqaanNzc8fjo0VEAtLOnTvrrLVpwz03LkWdm5vLjh07xuOjRUQCkjGmbKTnNPUhIuJyKmoREZdTUYuIuNy4zFGLiJyN3t5eKioq6OrqcjrKuIuMjCQnJ4ewsLBRv0dFLSKOq6ioIC4ujtzcXIwxTscZN9Za6uvrqaioIC8vb9Tv09SHiDiuq6uLlJSUgC5pAGMMKSkpZ/2bg4paRFwh0Et60Ll8na4p6q7efta/cZBtJQ1ORxERcRXXFDXAw5tK+fcX96FzZIvIRGpqauJnP/vZWb/v6quvpqmpaRwSncg1RR0Z5uHLl01nR1kjb+yvczqOiASRkYq6r6/vtO97/vnnSUxMHK9Yx7imqAFuXDKZnKQo/vPl9zWqFpEJc9ddd3Hw4EEWLFjA0qVLWblyJddffz2zZ88G4IYbbmDx4sUUFhayfv36Y+/Lzc2lrq6O0tJSCgoK+MIXvkBhYSGrV6+ms7NzzPK5ave88NAQvnrZDL719G42Ftdw5exJTkcSkQn2gz/uYW9ly5h+5uyseL53XeGIz997770UFRWxa9cuXn/9da655hqKioqO7UL38MMPk5ycTGdnJ0uXLuXjH/84KSkpJ3zG/v37efzxx3nwwQe58cYbefrpp/nUpz41JvldNaIG+NiibHJTovnRKx8wMKBRtYhMvGXLlp2wn/P999/P/PnzWb58OeXl5ezfv/+U9+Tl5bFgwQIAFi9eTGlp6ZjlcdWIGiDUE8LXrpjB3z/5Li8UVXHNvEynI4nIBDrdyHeixMTEHFt+/fXX2bhxI1u2bCE6OppLLrlk2P2gIyIiji17PJ4xnfpw3Yga4Pr52UxPj+XHGz+gX6NqERlncXFxtLa2Dvtcc3MzSUlJREdHs2/fPt56660JTufSovaEGO68YgYHatr447uVTscRkQCXkpLCihUrmDNnDt/85jdPeG7NmjX09fVRUFDAXXfdxfLlyyc8nxmPvSuWLFliz/fCAQMDlqvvf5Ou3n42/sOHCfW48meKiIyB4uJiCgoKnI4xYYb7eo0xO621S4Z7vWvbLyTE8PdXzqS0voM/aFQtIkHMtUUNsHr2JGZlxPHAnw9orlpEgpari9oYw1cvn8Ghunae261RtYgEJ1cXNcCawgxmpMfywJ8PaL9qEQlKri/qkBDDVy6fwf6aNl7cU+V0HBGRCef6oga4Zm4m+Wkx3P/qfo2qRSTo+EVRe0IMX750OvuqWnmluNrpOCIixMbGAlBZWcknPvGJYV9zySWXcL67KoOfFDXA9fOzmJoSzf2v7teZ9UTENbKystiwYcO4/hl+U9ShnhDuuHQ6eypbeO39GqfjiEiAueuuu/jpT3967PH3v/997rnnHi6//HIWLVrE3LlzefbZZ095X2lpKXPmzAGgs7OTm266iYKCAj760Y+O2fk+XHdSptP56MJs7tu4n5+9dpDLZukUqCIB6YW7oOq9sf3MjLlw1b2nfcm6deu48847ueOOOwB46qmneOmll/jqV79KfHw8dXV1LF++nOuvv37E6x7+/Oc/Jzo6muLiYnbv3s2iRYvGJL7fjKgBwjwh3HZxHjvKGnnncKPTcUQkgCxcuJCamhoqKyt59913SUpKIiMjg3/8x39k3rx5XHHFFRw5coTq6pG3k73xxhvHzkE9b9485s2bNybZ/GpEDXDj0sn8eOMH/OLNEn56S5LTcURkrJ1h5Due1q5dy4YNG6iqqmLdunU8+uij1NbWsnPnTsLCwsjNzR32FKfjza9G1ACxEaHcvGwKLxQdpbyhw+k4IhJA1q1bxxNPPMGGDRtYu3Ytzc3NpKenExYWxmuvvUZZWdlp379q1Soee+wxAIqKiti9e/eY5PK7ogb4zIpcQozhkc2lTkcRkQBSWFhIa2sr2dnZZGZmcsstt7Bjxw7mzp3Lr3/9a2bNmnXa93/pS1+ira2NgoICvvvd77J48eIxyeV3Ux8AmQlRXDsvkye3H+ZrV8wgISrM6UgiEiDee+/4hszU1FS2bNky7Ova2toA7wVui4qKAIiKiuKJJ54Y80x+OaIG+PzKfNp7+nly+2Gno4iIjCu/Leo52Qksz0/mkc2l9PYPOB1HRGTc+G1RA3xhZT5Hm7t4/r2jTkcRkfMULEccn8vX6ddFfekF6eSnxfDgm4eC5h9ZJBBFRkZSX18f8P+PrbXU19cTGRl5Vu/zy42Jg0JCDLddnMd3niliR1kjS3OTnY4kIucgJyeHiooKamtrnY4y7iIjI8nJyTmr9/h1UYP3sPJ7n9/HY1sPq6hF/FRYWBh5eXlOx3Atv576AIgOD+WGhdn86b2jNLb3OB1HRGTM+X1RA3xy2RR6+gZ4+u0Kp6OIiIy5gCjq2VnxLJicyGPbDgf8xggRCT6jLmpjjMcY844x5rnxDHSubr5wCodq29la0uB0FBGRMXU2I+qvAcXjFeR8XTcvi7jIUB7bqiMVRSSwjKqojTE5wDXAL8Y3zrmLCvfwsYXZvFhURYM2KopIABntiPonwLeAEY/VNsbcbozZYYzZ4dS+kDdfOJWe/gE27Cx35M8XERkPZyxqY8y1QI21dufpXmetXW+tXWKtXZKWljZmAc/GBRlxLJ6axOPbyrVRUUQCxmhG1CuA640xpcATwGXGmN+Oa6rzcPOyKZTUtbPlYL3TUURExsQZi9pae7e1NsdamwvcBPzZWvupcU92jq6Zl0lCVBiPbtNGRREJDAGxH/VQkWEePrE4h5eKqqhpmfhrm4mIjLWzKmpr7evW2mvHK8xYueXCKfQNWJ7Yro2KIuL/Am5EDZCfFsvKGak8tvUwfbqogIj4uYAsaoBbl0+lqqWLjcXVTkcRETkvAVvUl81KJyshkt+8dfrLu4uIuF3AFnWoJ4SbL5zC5gP1HKhpczqOiMg5C9iiBli3dAphHsOjWzWqFhH/FdBFnRYXwVVzMtmws4KOnj6n44iInJOALmqAWy+aSmtXH8/uqnQ6iojIOQn4ol4yNYlZGXH8ZkuZzv8hIn4p4IvaGMOtF01l79EWdpY1Oh1HROSsBXxRg/dK5XGRoTzy11Kno4iInLWgKOro8FBuWjqZF4uqONrc6XQcEZGzEhRFDfB/LsrFWstvtmhXPRHxL0FT1JOTo7miYBKPbztMV2+/03FEREYtaIoa4LMr8mjs6OXZXUecjiIiMmpBVdTL85OZlRHHI5tLtaueiPiNoCpqYwyfXZHLvqpWthzSpbpExD8EVVEDfGRBNknRYfxyc6nTUURERiXoijoyzMMnl03hleJqyhs6nI4jInJGQVfU4D3/R4gx/EoHwIiIHwjKos5MiOKqORk8ub2ctm6dVU9E3C0oixrg8yvzae3u4yldAFdEXC5oi3rB5ESWTE3i4c0l9A9oVz0Rca+gLWqAz6/Mo6Kxk5f3VDkdRURkREFd1FfOzmBychS/2FTidBQRkREFdVF7QgyfW5HHzrJG3jmsc1WLiDsFdVEDrF0ymbjIUB7SqFpEXCroizo2IpSbl03hhaIqKhp1AIyIuE/QFzXApz+UC6ADYETElVTUQFZiFNfMzeSJbeW0dvU6HUdE5AQqap/Pr8zzHgCzo8LpKCIiJ1BR+8zLSWRpbhK//KsOgBERd1FRD/G5FXmUN3Tyyt5qp6OIiByjoh5idWEGOUlRPLxZu+qJiHuoqIfwhBg+86FctpU0UHSk2ek4IiKAivoUNy6dTEy4h4d1AIyIuISK+iTxkWGsXTKZP+6upKaly+k4IiIq6uF8dkUufQOW375V5nQUEZEzF7UxJtIYs80Y864xZo8x5gcTEcxJU1NiuKJgEr/depiu3n6n44hIkBvNiLobuMxaOx9YAKwxxiwf31jO+9yKPBrae3h21xGno4hIkDtjUVuvNt/DMN8t4I8IWZ6fTEFmPOvfOERP34DTcUQkiI1qjtoY4zHG7AJqgFestVvHN5bzjDF8Y/VMDta2s/6Ng07HEZEgNqqittb2W2sXADnAMmPMnJNfY4y53Rizwxizo7a2dqxzOuLygklcMy+T+189wMHatjO/QURkHJzVXh/W2ibgNWDNMM+tt9YusdYuSUtLG6t8jvv+dYVEhXu4+/fvMaBzgIiIA0az10eaMSbRtxwFXAnsG+9gbpEWF8F3ri5gW0kDT2wvdzqOiASh0YyoM4HXjDG7ge1456ifG99Y7rJ2SQ4fmpbCvz5fTLUOghGRCTaavT52W2sXWmvnWWvnWGv/70QEcxNjDP/y0bn09A/wvWf3OB1HRIKMjkwcpdzUGO68YiYv7qnixaIqp+OISBBRUZ+Fz6/MY1ZGHD/44x7au/ucjiMiQUJFfRbCPCH880fncLS5i/te3e90HBEJEirqs7R4ajI3LZ3MQ5tK2FfV4nQcEQkCKupz8O01s0iICuOfninSvtUiMu5U1OcgKSacu66axY6yRjbs1FXLRWR8qajP0ScW5bA0N4l/eaGYhvYep+OISABTUZ+jkBDDPTfMpa2rj3tfKHY6jogEMBX1ebggI47bVubx1I4KNu2vczqOiAQoFfV5+vsrZpKfFsO3n95Na1ev03FEJACpqM9TZJiH/1g7n6PNnfzL85oCEZGxp6IeA4umJHH7qmk8vq2cv3wQGOfiFhH3UFGPkTuvmMGM9Fi+vWE3zZ2aAhGRsaOiHiODUyC1bd3c89xep+OISABRUY+h+ZMT+dKHp/G7nRVs3FvtdBwRCRAq6jH2lcunU5AZzzc2vEtlU6fTcUQkAKiox1hEqIef3ryQ3r4BvvzY2/T2DzgdSUT8nIp6HOSnxXLvx+fx9uEmfvjS+07HERE/p6IeJ9fNz+JTy6ew/o1DvKL5ahE5DyrqcfRP18ymMCuerz+1i/KGDqfjiIifUlGPo8gwDz+7ZRHWwpcfe5vOnn6nI4mIH1JRj7OpKTH8cO18dh9p5ou/3Ul3n8paRM6OinoCrJmTwb0fm8sbH9Tylcfe0Z4gInJWVNQTZN3SKXzvutm8vLeab/zuXfp1CS8RGaVQpwMEk8+uyKOjp58fvvQ+UWEe/vVjczHGOB1LRFxORT3B7rh0Op09/Tzw2gHCQ0P4wfWFKmsROS0VtQO+vnomPf0DrH/jEL39ln++YQ4hISprERmeitoBxhjuvmoWYR7DT187SG//AP/28Xl4VNYiMgwVtUOMMXxj9QWEeUL4ycb99PYP8J9r5xPq0fZdETmRitpBxhjuvGImYZ4QfvjS+/T1W35y0wLCVNYiMoSK2gXuuHQ6EaEh3POnYiyW+25aqLIWkWNU1C7x+ZX5ANzzp2KM2cV96xZoGkREABW1q3x+ZT7Wwj8/X0yIMfz4Rs1Zi4iK2nW+sCqfAWv51xf2YYAfqaxFgp6K2oW++OFpDFj4txf3YfGWteasRYKXitqlvnTJNIyBe1/YR2dPHw/cvIjIMI/TsUTEARqmudjffnga/+8jhWwsruG2X22nvbvP6Ugi4gAVtcvdelEu/7l2PlsO1nPrQ1tp7ux1OpKITLAzFrUxZrIx5jVjzF5jzB5jzNcmIpgc9/HFOfzslkW8d6SZm9a/RVVzl9ORRGQCjWZE3Qd83Vo7G1gO3GGMmT2+seRka+Zk8tCnl1JW3851D2xiR2mD05FEZIKcsaittUettW/7lluBYiB7vIPJqVbNTOOZv1tBdLiHTz74Fo9uLXM6kohMgLOaozbG5AILga3DPHe7MWaHMWZHbW3t2KSTU1yQEccf7riYD01L5TvPFHH379/TdRhFAtyoi9oYEws8DdxprW05+Xlr7Xpr7RJr7ZK0tLSxzCgnSYgO4+HPLOVLl0zj8W2H+cgDm3ltXw3W6vJeIoFoVEVtjAnDW9KPWmt/P76RZDQ8IYZvr5nF+lsX09HTz2d/uZ0b/2cL2zV3LRJwRrPXhwEeAoqttT8a/0hyNlYXZvDq1z/MPTfMoay+g7X/vYXPPLKN8oYOp6OJyBgZzYh6BXArcJkxZpfvdvU455KzEOYJ4VPLp/KXb17K3VfNYmdZI9fc/yYv76lyOpqIjAEzHvOaS5YssTt27Bjzz5XRKW/o4I7H3mZ3RTNfWJnHt9bM0rlCRFzOGLPTWrtkuOf0vzcATU6O5nd/exGfvmgqD75Zwrr/2UJlU6fTsUTkHKmoA1REqIcffGQOD9y8kA+q27j2vzax+UCd07FE5ByoqAPctfOy+MOXV5AaG86tD23lv/9yULvxifgZFXUQyE+L5Zm/W8FVczO594V9/N2jb9OmM/GJ+A0VdZCIiQjlgU8u5DtXF/Dy3mo+8sAm9laectySiLiQijqIGGP4wqp8fnvbhTR39nH9A5v49xf30dWrQ9BF3ExFHYQumpbCxn9YxUcXZvOz1w9y1X1vsuVgvdOxRGQEKuoglRgdzg/Xzue3t11I/4Dlkw++xbc37Ka+rdvpaCJyEhV1kLt4Riov3bmKL67K5+m3K7jkP17nF28eoqdvwOloIuKjohaiwj3cfXUBL965ikVTkrjnT8Wsue8NXnu/xuloIoKKWoaYnh7Lrz63jEc+sxQsfPaR7dz60FaKjjQ7HU0kqKmo5RSXzkrnxTtX8U/XFPDekWau/a9NfOXxdyirb3c6mkhQ0kmZ5LRaunpZ/5dD/GLTIfr6LZ9cNoXbV+UzOTna6WgiAeV0J2VSUcuo1LR0cd+r+3lyezkWuHZeJl9cNY3ZWfFORxMJCCpqGTNHmzt56M0SHt92mPaeflbNTOMrl01naW6y09FE/JqKWsZcc0cvv91axiObS6hr6+HDM9P4xuoLmJuT4HQ0Eb+kopZx09nTz6+2lPLffzlIU0cvawoz+IfVM5k5Kc7paCJ+RUUt466lq5eH3izhF28eor2nn+X5yaxbOpmr5mQSGeZxOp6I66moZcI0tPfw2NYyntpRweGGDuIiQ7lhQTafWj6VCzI0yhYZiYpaJtzAgOWtknqe3F7OC0VV9PQNcNmsdL50yTRteBQZhopaHNXY3sNv3vJueGzs6GXx1CT+9sPTuGxWOp4Q43Q8EVdQUYsrdPT08dT2ch58s4QjTZ1kJ0axbulkblwymYyESKfjiThKRS2u0ts/wEt7qnhiWzmbDtQRYuCyWelcNz+LZXnJZCZEOR1RZMKdrqhDJzqMSJgnhGvnZXHtvCzK6tt5cns5v9tZwcZi79n6shOjWJqbxNK8ZK6dm0VCdJjDiUWcpRG1uEJf/wD7qlrZXtrguzVS29pNdLiHtYtz+OyKPHJTY5yOKTJuNPUhfsday57KFh7ZXMof3j1C34DlyoJJfGZFLsvzUgjRRkgJMCpq8Ws1LV38eksZj24to7Gjl5ykKD62MJuPL85haopG2RIYVNQSEDp7+nl5bxUbdlaw6UAd1sLS3CSunZfF6sJJ2ggpfk1FLQHnaHMnz7xzhGfePsL+mjYAFkxOZM2cDNYUZmg+W/yOiloC2oGaNl7aU8VLe6rYXeG9bNisjDhWF2bwN4WTmJ0ZjzGa0xZ3U1FL0Kho7ODlPdW8tKeK7aUNDFiYnBzF1XMzuW5eFoVZKm1xJxW1BKX6tm42FlfzQlEVm/bX0TdgyU+L4bp5WVw3P5Pp6TpJlLiHilqCXmN7Dy/uqeIPuyp5q6Qea71XXb9qTgZr5mRoekQcp6IWGaKmpYsXiqp4oego20q80yNTkqO5Zl4mNyzI1ulYxREqapER1Ld188reap4vqmLzgTr6BywFmfHcsCCL6+ZnkZWoXf5kYqioRUahtrWbP+2u5H93VbKrvAmAmZNiWTE9lYunp3JhfgqxETo9jowPFbXIWSqpa+elPd5R9raSBrr7BggNMSycksjF09O4eEYq83MSCPWEOB1VAoSKWuQ8dPX283ZZI28eqGPzgTreO9KMtRAXEcryaSlcmJfM0txkCrPiVdxyzs7rNKfGmIeBa4Eaa+2csQ4n4naRYR4+ND2VD01PBbx7kPz1YD2bDtTx14N1vLK3GoDocA+LpiSxYnoqV85OZ1parPYkkTFxxhG1MWYV0Ab8erRFrRG1BJPqli62lXhPz7qtpIF9Va0A5KXGcOXsSVw+K52FU5IID9VoW0Z23lMfxphc4DkVtciZHW3uZOPeal7eW81bh+rp7bdEhXlYNDWRC/O8UyXzJycSGeZxOqq4yIQUtTHmduB2gClTpiwuKys7p7AigaSlq5fN++vYWtLAW4fqj422Q0MMF2TEMTc7gbk5CczLTqQgM05z3EFMI+rRaCyFBy+DmLQTb7GDy+m++1SITYdwnZ1Nzl5TRw/bShp4p7yJoiPNvHekmaaOXgDiIkO5KD+FlTNSWTE9lbzUGM1xBxFdM3E0POEw+wZor/XeqnZDWy10Nw//+rDoIWWe7i3wk8t8sOCjkiBEIyWBxOhwVhdmsLowA/BeyaaisZN3ypvYcrCON/fX8bJv42R2YhSrZqayckYaH5qWQmJ0uJPRxUEaUZ9JX/fx8m6rPb58bF0NtNdBu+/e9p/6GcbjK/LTjdaHlHtoxMR/neIK1lrK6jt480Adm/bX8tcD9bR29xFiYG5OIhfle+e4F+cmER+pi/4GkvOa+jDGPA5cAqQC1cD3rLUPne49AVXUZ2NgADobfSVe47uv85X5MOXe2z7850QkHC/2kcp88HFkAujX44DV1z/AuxVNvPFBHZsO1LG7oonefkuIgYLMeJZMTWJOdgKFWQnMmBRLmOa4/ZYOeHGrnvZhRuo1w5d7RwMwzL+VJ/z4dMvg1MvJ5T74fFQyhEVO+JcpY6ezp593yhvZVuLdFXBXeRMdPd7f4sI9IczMiCU/NZacpCiyk6LISYr2LidGaS8Tl1NRB4L+PuioH91Ivb0G+nuG/5ywaG9hRyd5586jkiE6+cT7qKQT10UmQIj+k7tR/4CltL6dPZUt7DnSzN6jLZTVd1DZ1EnfwIn/t9PjIpic7C3u6WmxFGbHMzszgUnxEdpo6QIq6mBjLXS3nDqn3tnovXU0QGfDqct2YIQPNN6yHrHQk4Yv+fAYTcs4pH/AUt3SxZGmTsobOqho9N6XN3ZQ3tDJkabOY69NjglndmY8c3MSmJ+TwNycRLISIlXeE0xFLWc2MODdw6WzEToaveU9XKGfsK4RelpH/kxP+OgK/eR1Hm0kG2+tXb3sq2plb2ULeytbKKps5v2q1mOj8NTYcN8+3om+8k4gPU7TZuNJRS3jp6/HN1I/XaEP89xA78ifGR7nm5oZodAHS39wCicyUdMzY6Crt599Va3srmji3fJmio40s7+mlcEZlIz4SOZkxzM7K4E5WfEUZido5D2GtB+1jJ/QcIib5L2NlrXeDalDC32wzE8p+0ZoKvMudzUz7AbVQRHxx0s7yncfmXji8rHnTnpdmC4QEBnmYcHkRBZMToSLvOvau/vYe7SF3RXN7K5oYk9lC3/eV3OsvBOjw5idGU9hVjyzs+IpzEogPzVGR1iOMY2oxX8M9HvL+uRRe1czdDZ577uaTlwefG6kXSEHeSJGLvHTFXxkovcHRBAd0NTZ009x1fGNl3sqW9hX1UpPn3cbR3hoCLMy4ijI8JZ3QWY8MyfF6oCdM9DUh0hfj3cD67ESbxy+0Ecq+xE3tIJ3Y+toRvOJwz8XAAc49fYPcKi2nT2VzRQfbWHvUe/cd2PH8Smu9LgIZk6KY8akWGZOimPmpFimp8eREKVtEqCpDxHvFE1oqnd/8rNlLXS3jq7QBx/X7T++3Nd5+s8PjRq+4KNTID4b4rMgIce7HJfhyrn4ME8IF2TEnXBhYGstVS1d7KtqZX91K+9XtbG/ppXHtx2mq/f4D75J8RFMS4tlclI0k5OP7/udnxZLcoxG4aCiFjkzMzhijj+39/d1DynxppOWT37cDK1HoXafd5fK3o6TsnggLhMSsr3FPXh/bDnHe4CTC6ZijDFkJkSRmRDFpRekH1s/MOA9v8n+mlY+qPaW96Hadl7dV0NdW/cJn5EaG8FM3wg8Py2G9LhI0uMjSI+LIC0ugohQ9/3QGg+a+hBxK2u95d18BFqOQHOF7973eHC5/8RywxPuK/OcU0fkg2UenezKfdw7e/qP7ft9oKaND6pb+aCmjQPVrbT3nHoenUnxEeSnxpKfFkN+Wiz5qTHkpcaQkxTldxs0NUctEqis9R6xOlKJt1RAy9FTd4cMjRy5xAdH6S46j8zAgKWurZua1m5qWruoafEul9V3cKiujUO17TR3Hv8aQ0MMU5KjyUuNYXJyNNmJUWQleg+rz0qMJDUmgpAQd3xtgzRHLRKojPGdzyUVshYM/5qBAe9pBYYW+dBiL3nDO91y8gbT8FhvkQ9X4oPrIuKG/zPHWEiIIT0+kvT4SCDhlOettTS093Corp2SunZK69oprW/nUG07W0saaOvuO+H1oSGGtLgI0uMjmRQXQbZvTnxaqndk7rbD6jWiFhHvuWTaqoeZYqmAlkrvcls1p+zHHpFw0nz5YJlnHV92wT7qzZ29VDZ1cqSxk8rmTqpbuqhu6fbdd1He0Eln7/GplagwD5kJkUyKj2RSfAST4iPJSoxiako0uSkxZCdFjfmZCjWiFpHT84R6SzUhGyYvG/41fT3ekfdIUyyV70BH3anvi0o+tcTDY727JYZGeufUQyN9e+ZEevdpDx16G/qaiHPa6yUhKoyEqDAKMoffIDwwYKlu7eJQbTuHatsoqes4VuI7yhqpaemmp//4bxyhIYacpCjuvrqAv/FdBGI8qahFZHRCwyFpqvc2kt4uX4FXnjo6by6Hw1u8G0jPR7J3L0YAAAT0SURBVEjokDIfWvBDynzEHwLDvyYkNJLM0HAyQyNZkR4BmYOfHQehaVhPGPXdIZQ19VPa3Mehxl5KG7pImaDdB1XUIjJ2wiIhZZr3NpLeTujp8O6t0tflHan3dXlPzdvX5d2dsa97hHXdvvcNec0p67q9px7oO/nzh7znLBm8V05JBRYPrgwJg9KTfgjEZcDnXjz7v7czUFGLyMQKi3J23tpa6O8dxQ+Hk3+AnO6Hiu+HxTh9XSpqEQkuxvhGwf5z1KN/7REuIhKEVNQiIi6nohYRcTkVtYiIy6moRURcTkUtIuJyKmoREZdTUYuIuNy4nD3PGFMLlI35B5+/VGCYs8a4mjJPHH/M7Y+ZwT9zj3fmqdbatOGeGJeiditjzI6RTiPoVso8cfwxtz9mBv/M7WRmTX2IiLicilpExOWCrajXOx3gHCjzxPHH3P6YGfwzt2OZg2qOWkTEHwXbiFpExO+oqEVEXC4gi9oY87AxpsYYUzRkXbIx5hVjzH7ffZKTGU9mjJlsjHnNGLPXGLPHGPM133q35440xmwzxrzry/0D3/o8Y8xWY8wBY8yTxhjXnaXdGOMxxrxjjHnO99gfMpcaY94zxuwyxuzwrXP790iiMWaDMWafMabYGHORH2S+wPd3PHhrMcbc6VTugCxq4JfAmpPW3QW8aq2dAbzqe+wmfcDXrbWzgeXAHcaY2bg/dzdwmbV2PrAAWGOMWQ78G/Bja+10oBG4zcGMI/kaUDzksT9kBrjUWrtgyD69bv8euQ940Vo7C5iP9+/c1Zmtte/7/o4X4L1MYgfwDE7lttYG5A3IBYqGPH4fyPQtZwLvO53xDPmfBa70p9xANPA2cCHeI7hCfesvAl5yOt9JWXPw/ke7DHgO7/VLXZ3Zl6sUSD1pnWu/R4AEoATfjgv+kHmYr2E1sNnJ3IE6oh7OJGvtUd9yFTDJyTCnY4zJBRYCW/GD3L4phF1ADfAKcBBostb2+V5SAWQ7lW8EPwG+BQz4Hqfg/swAFnjZGLPTGHO7b52bv0fygFrgEd800y+MMTG4O/PJbgIe9y07kjuYivoY6/1x6Mr9Eo0xscDTwJ3W2pahz7k1t7W233p/RcwBlgGzHI50WsaYa4Eaa+1Op7Ocg4uttYuAq/BOj60a+qQLv0dCgUXAz621C4F2TpoucGHmY3zbKa4HfnfycxOZO5iKutoYkwngu69xOM8pjDFheEv6UWvt732rXZ97kLW2CXgN77RBojFm8Cr3OcARx4KdagVwvTGmFHgC7/THfbg7MwDW2iO++xq8c6bLcPf3SAVQYa3d6nu8AW9xuznzUFcBb1trq32PHckdTEX9B+DTvuVP450Ddg1jjAEeAoqttT8a8pTbc6cZYxJ9y1F459WL8Rb2J3wvc1Vua+3d1toca20u3l9r/2ytvQUXZwYwxsQYY+IGl/HOnRbh4u8Ra20VUG6MucC36nJgLy7OfJJPcnzaA5zK7fRE/ThN/j8OHAV68f5Evw3vHOSrwH5gI5DsdM6TMl+M99eo3cAu3+1qP8g9D3jHl7sI+K5vfT6wDTiA99fGCKezjpD/EuA5f8jsy/eu77YH+I5vvdu/RxYAO3zfI/8LJLk9sy93DFAPJAxZ50huHUIuIuJywTT1ISLil1TUIiIup6IWEXE5FbWIiMupqEVEXE5FLSLicipqERGX+/82h0w60uXonAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ogTew9w9J33a"
      },
      "source": [
        "# Unfreeze the mode\n",
        "\n",
        "learn.unfreeze()\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "N-_kXzmYkBrL",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        },
        "outputId": "c6995e15-419c-4b08-f8d2-4486be2c5bae"
      },
      "source": [
        "# Find new optimum lr\n",
        "lr_min, lr_steep = learn.lr_find()\n",
        "print(f\"lr_min: {lr_min:.1e}; lr_steep: {lr_steep:.1e}\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              ""
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "lr_min: 3.3e-05; lr_steep: 9.1e-07\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "t01xq7fmEiL6",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "477652a5-f6e5-4280-be71-f3574435de96"
      },
      "source": [
        "# Pick your optimum learning rate after unfreezing:\n",
        "\n",
        "lr = lr_min# add your value\n",
        "\n",
        "print(f\"Using lr = {lr:.2e}\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using lr = 3.31e-05\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "W1jihOXXkH76",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 235
        },
        "outputId": "142096b5-5560-467b-d02c-98bfcace66c0"
      },
      "source": [
        "# Define fit_one_cycle function and arguments to use optimum lr and 6 epochs\n",
        "\n",
        "#learn.fine_tune??\n",
        "#learn = cnn_learner(dls, resnet34, metrics=error_rate)\n",
        "learn.fit_one_cycle(6, lr)\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "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>error_rate</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>5.245749</td>\n",
              "      <td>4.142587</td>\n",
              "      <td>0.965000</td>\n",
              "      <td>00:16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1</td>\n",
              "      <td>5.141267</td>\n",
              "      <td>3.860997</td>\n",
              "      <td>0.935000</td>\n",
              "      <td>00:16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>2</td>\n",
              "      <td>5.018444</td>\n",
              "      <td>3.708949</td>\n",
              "      <td>0.915000</td>\n",
              "      <td>00:16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>3</td>\n",
              "      <td>4.878945</td>\n",
              "      <td>3.615644</td>\n",
              "      <td>0.905000</td>\n",
              "      <td>00:16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>4</td>\n",
              "      <td>4.805032</td>\n",
              "      <td>3.569284</td>\n",
              "      <td>0.905000</td>\n",
              "      <td>00:16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>5</td>\n",
              "      <td>4.709460</td>\n",
              "      <td>3.562125</td>\n",
              "      <td>0.910000</td>\n",
              "      <td>00:16</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8hIr7c9ik5ci",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "outputId": "399aada2-12ee-4a55-f454-8bdfdd9c6ac1"
      },
      "source": [
        "learn.recorder.plot_loss()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qgVibs7ZXxG7"
      },
      "source": [
        "### Conclusion and Reflections"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TlxmEoSk6RZS"
      },
      "source": [
        "What I've learnt:\n",
        "\n",
        "1.presizing, data block summary\n",
        "\n",
        "2.unfreezing, discriminative learning rates, setting the number of epochs, and using deeper architectures\n",
        "\n",
        "3.cross-entropy loss\n",
        "\n",
        "Reflections:\n",
        "\n",
        "It pretty busy last week. So I did not spend enough time learning fastai. Since I shall have more free time this week, I should pay more attendtion.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2WfGhamO6P-Z"
      },
      "source": [
        "## Assignment 3: Questionnaire\n",
        "\n",
        "Try to answer the questions below, with special attention to the questions in bold. We will address those in the session."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c7dlUSPU6P-Z"
      },
      "source": [
        "1. **Why do we first resize to a large size on the CPU, and then to a smaller size on the GPU?**\n",
        "1. If you are not familiar with regular expressions, find a regular expression tutorial, and some problem sets, and complete them. Have a look on the book's website for suggestions.\n",
        "1. **What are the two ways in which data is most commonly provided, for most deep learning datasets?**\n",
        "1. Look up the documentation for `L` and try using a few of the new methods that it adds.\n",
        "1. Look up the documentation for the Python `pathlib` module and try using a few methods of the `Path` class.\n",
        "1. **Give two examples of ways that image transformations can degrade the quality of the data.**\n",
        "1. **What method does fastai provide to view the data in a `DataLoaders`?**\n",
        "1. **What method does fastai provide to help you debug a `DataBlock`?**\n",
        "1. Should you hold off on training a model until you have thoroughly cleaned your data?\n",
        "1. **What are the two pieces that are combined into cross-entropy loss in PyTorch?**\n",
        "1. **What are the two properties of activations that softmax ensures? Why is this important?**\n",
        "1. When might you want your activations to not have these two properties?\n",
        "1. Calculate the `exp` and `softmax` columns of `bear_softmax` yourself (i.e., in a spreadsheet, with a calculator, or in a notebook).\n",
        "1. Why can't we use `torch.where` to create a loss function for datasets where our label can have more than two categories?\n",
        "1. What is the value of log(-2)? Why?\n",
        "1. **What are two good rules of thumb for picking a learning rate from the learning rate finder?**\n",
        "1. **What two steps does the `fine_tune` method do?**\n",
        "1. In Jupyter Notebook, how do you get the source code for a method or function?\n",
        "1. What are discriminative learning rates?\n",
        "1. How is a Python `slice` object interpreted when passed as a learning rate to fastai?\n",
        "1. Why is early stopping a poor choice when using 1cycle training?\n",
        "1. What is the difference between `resnet50` and `resnet101`?\n",
        "1. What does `to_fp16` do?"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "prCBx4fP6yej"
      },
      "source": [
        "1.Compose all of the common augmentation operations (including a resize to the final target size) into one, and perform the combined operation on the GPU only once at the end of processing, rather than performing the operations individually and interpolating multiple times.\n",
        "\n",
        "3.Individual files representing items of data, such as text documents or images, possibly organized into folders or with filenames representing information about those items\n",
        "\n",
        "A table of data, such as in CSV format, where each row is an item which may include filenames providing a connection between the data in the table and data in other formats, such as text documents and images\n",
        "\n",
        "6.rotating an image by a degree fills corner regions of the new bounds with emptiness, which will not teach the model anything. Many rotation and zooming operations will require interpolating to create pixels. These interpolated pixels are derived from the original image data but are still of lower quality.\n",
        "\n",
        "7.one_batch,\n",
        "\n",
        "8.summary()\n",
        "\n",
        "10.softmax and nll_loss\n",
        "\n",
        "11.the properties that all activations are between 0 and 1, and sum to 1\n",
        "\n",
        "16.a model with more parameters can model your data more accurately. you can create larger versions of them by simply adding more layers\n",
        "\n",
        "17.\n",
        "Trains the randomly added layers for one epoch, with all other layers frozen\n",
        "Unfreezes all of the layers, and trains them all for the number of epochs requested\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ysR8reaM6P-a"
      },
      "source": [
        "### Further Research"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XH4Ytqdk6P-a"
      },
      "source": [
        "1. Find the paper by Leslie Smith that introduced the learning rate finder, and read it.\n",
        "1. See if you can improve the accuracy of the classifier in this chapter. What's the best accuracy you can achieve? Look on the forums and the book's website to see what other students have achieved with this dataset, and how they did it."
      ]
    }
  ]
}