AIMC2_Handson_Fashion_MNIST_tutorial_with_pytorch.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "Fashion-MNIST_tutorial_with_pytorch.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.6"
    },
    "gpuClass": "standard",
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "8e607298272e4b6db8ce9c0e3159bc6d": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_0ab2982301e2483fb4b7f552821a8847",
              "IPY_MODEL_b68eaeb25dc143f9b8bacb8e0d7c0e82",
              "IPY_MODEL_f1b442c50512463ca08d04963c6f8019"
            ],
            "layout": "IPY_MODEL_3bc93915b5cb4314b719327a7230f584"
          }
        },
        "0ab2982301e2483fb4b7f552821a8847": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_38f4012b08c84219ad14b2fecc55eef2",
            "placeholder": "​",
            "style": "IPY_MODEL_f30f400428c143d98913f6530ad2838f",
            "value": "100%"
          }
        },
        "b68eaeb25dc143f9b8bacb8e0d7c0e82": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_00495ecf06944d8c95d0a3272f1b52e4",
            "max": 26421880,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_49a8038d370b40d7a01224ad6bfcdf87",
            "value": 26421880
          }
        },
        "f1b442c50512463ca08d04963c6f8019": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_bd1720c80da94882b01c894dd522f4c9",
            "placeholder": "​",
            "style": "IPY_MODEL_57ee79d0060445b7944bc25b2e3505ce",
            "value": " 26421880/26421880 [00:02&lt;00:00, 17538961.86it/s]"
          }
        },
        "3bc93915b5cb4314b719327a7230f584": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "38f4012b08c84219ad14b2fecc55eef2": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "f30f400428c143d98913f6530ad2838f": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "00495ecf06944d8c95d0a3272f1b52e4": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "49a8038d370b40d7a01224ad6bfcdf87": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "bd1720c80da94882b01c894dd522f4c9": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "57ee79d0060445b7944bc25b2e3505ce": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "fe87464d51214287a865313d3f6503b5": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_8d54e16c1f2d403a96793993c478248b",
              "IPY_MODEL_bc8017acebbb4e1daea1d16debcf112b",
              "IPY_MODEL_3612aed2dd66408e9f9559d6971d9918"
            ],
            "layout": "IPY_MODEL_e85c3b584dd24c4ab49ef44b1dda81f6"
          }
        },
        "8d54e16c1f2d403a96793993c478248b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_cb27556cad68447b9d1b27665b1ae60c",
            "placeholder": "​",
            "style": "IPY_MODEL_3331ecac16514cb391dd9673c25300f5",
            "value": "100%"
          }
        },
        "bc8017acebbb4e1daea1d16debcf112b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_df365b11a4d347ae83d18c16236bfe0e",
            "max": 29515,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_6aa75f0b6819427b8e7fdc86f7343bdb",
            "value": 29515
          }
        },
        "3612aed2dd66408e9f9559d6971d9918": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_5d3e8b785a014e88a9457be06d5549b6",
            "placeholder": "​",
            "style": "IPY_MODEL_dcfdc576bbce4377821e7f8b0165c193",
            "value": " 29515/29515 [00:00&lt;00:00, 209789.48it/s]"
          }
        },
        "e85c3b584dd24c4ab49ef44b1dda81f6": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "cb27556cad68447b9d1b27665b1ae60c": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "3331ecac16514cb391dd9673c25300f5": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "df365b11a4d347ae83d18c16236bfe0e": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "6aa75f0b6819427b8e7fdc86f7343bdb": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "5d3e8b785a014e88a9457be06d5549b6": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "dcfdc576bbce4377821e7f8b0165c193": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "9b7da931ad354203b7b4253f3107df8b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_7782931043c146aa8af2814ab6066c00",
              "IPY_MODEL_ed332bae787b4abfac54a78f92ef4e85",
              "IPY_MODEL_75cc26ea990146ce8e5007755acc0e64"
            ],
            "layout": "IPY_MODEL_c873ea044276451a8b5ecd236b8f9974"
          }
        },
        "7782931043c146aa8af2814ab6066c00": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_0d7bd0351b2f48a3881b5f8a9e2596c7",
            "placeholder": "​",
            "style": "IPY_MODEL_0b6801fa900c47d0bc7715b9354192fb",
            "value": "100%"
          }
        },
        "ed332bae787b4abfac54a78f92ef4e85": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_7e16cf580e8e411e85b54ca4ff17add3",
            "max": 4422102,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_91bad34abb334d7d9214590146074691",
            "value": 4422102
          }
        },
        "75cc26ea990146ce8e5007755acc0e64": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_bbd0eba42b4f4f8db9249248a0a06f12",
            "placeholder": "​",
            "style": "IPY_MODEL_ca661ae9bc39409bb53e137b995c8560",
            "value": " 4422102/4422102 [00:01&lt;00:00, 5321209.20it/s]"
          }
        },
        "c873ea044276451a8b5ecd236b8f9974": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "0d7bd0351b2f48a3881b5f8a9e2596c7": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "0b6801fa900c47d0bc7715b9354192fb": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "7e16cf580e8e411e85b54ca4ff17add3": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "91bad34abb334d7d9214590146074691": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "bbd0eba42b4f4f8db9249248a0a06f12": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "ca661ae9bc39409bb53e137b995c8560": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "e606284a47f545aba2fef6786f3f35ba": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_ef305e87bb404d5ea9e1f2480381d8a0",
              "IPY_MODEL_62b8398b036348ceb3b11309a8e6ddbe",
              "IPY_MODEL_330aec21439343c09386d050db0fd799"
            ],
            "layout": "IPY_MODEL_723ca6eed92e48a58903c5f1cc07f3bf"
          }
        },
        "ef305e87bb404d5ea9e1f2480381d8a0": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_42c6a1851b104a9ea8bd8723a133629d",
            "placeholder": "​",
            "style": "IPY_MODEL_058bdd86c26d4c5a8e8286e8a66c0950",
            "value": "100%"
          }
        },
        "62b8398b036348ceb3b11309a8e6ddbe": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_5dcde03f66f24f6eb011d02167023538",
            "max": 5148,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_77a1ecbf04444adbbafc212660c9ab03",
            "value": 5148
          }
        },
        "330aec21439343c09386d050db0fd799": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_04a7b752faa4477b9235db07d3b0e09b",
            "placeholder": "​",
            "style": "IPY_MODEL_0db68ce136564a1ca0bf334982efc13c",
            "value": " 5148/5148 [00:00&lt;00:00, 197088.95it/s]"
          }
        },
        "723ca6eed92e48a58903c5f1cc07f3bf": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "42c6a1851b104a9ea8bd8723a133629d": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "058bdd86c26d4c5a8e8286e8a66c0950": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "5dcde03f66f24f6eb011d02167023538": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "77a1ecbf04444adbbafc212660c9ab03": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "04a7b752faa4477b9235db07d3b0e09b": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "0db68ce136564a1ca0bf334982efc13c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "source": [
        "import torch\n",
        "from torch import nn\n",
        "print(torch.__version__)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jbsoLLNVoTp1",
        "outputId": "ef80a2bf-e1d4-4c15-b6dc-d03a0964999a"
      },
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1.12.1+cu113\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Get cpu or gpu device for training.\n",
        "device = \"cuda\" if torch.cuda.is_available() else \"cpu\"\n",
        "print(f\"Using {device} device\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YG5_42sn_DJa",
        "outputId": "26d5d942-2d1c-4c9c-9312-566b6bb5dc43"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Using cuda device\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WsAx4toWxKZE"
      },
      "source": [
        "![대체 텍스트](https://www.pyimagesearch.com/wp-content/uploads/2019/02/fashion_mnist_dataset_sample.png)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xhfmxR8o2sWd"
      },
      "source": [
        "\n",
        "# MNIST is too easy.\n",
        "Convolutional nets can achieve 99.7% on MNIST. Classic machine learning algorithms can also achieve 97% easily. \n",
        "# MNIST is overused. \n",
        "In this April 2017 Twitter thread, Google Brain research scientist and deep learning expert Ian Goodfellow calls for people to move away from MNIST.\n",
        "# MNIST can not represent modern CV tasks.\n",
        "\n",
        "# Fashion MNIST dataset\n",
        "Similar to the MNIST digit dataset, the Fashion MNIST dataset includes:\n",
        "\n",
        "60,000 training examples\n",
        "\n",
        "10,000 testing examples\n",
        "\n",
        "10 classes\n",
        "\n",
        "28×28 grayscale/single channel images"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sFD7RLgA3ouu"
      },
      "source": [
        "![대체 텍스트](https://www.pyimagesearch.com/wp-content/uploads/2019/02/fashion_mnist_obtaining.jpg)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "\n",
        "%matplotlib inline\n",
        "import matplotlib.pyplot as plt"
      ],
      "metadata": {
        "id": "mMq1b7jCozkn"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from torch.utils.data import DataLoader\n",
        "from torchvision import datasets\n",
        "from torchvision.transforms import ToTensor"
      ],
      "metadata": {
        "id": "S84_7q_-pKJ7"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "emDMGh2lWBHd"
      },
      "source": [
        "#STEP 1: Fashion MNIST 데이터 읽어들이기"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Download training data from open datasets.\n",
        "training_data = datasets.FashionMNIST(\n",
        "    root=\"data\",\n",
        "    train=True,\n",
        "    download=True,\n",
        "    transform=ToTensor(),\n",
        ")\n",
        "\n",
        "# Download test data from open datasets.\n",
        "test_data = datasets.FashionMNIST(\n",
        "    root=\"data\",\n",
        "    train=False,\n",
        "    download=True,\n",
        "    transform=ToTensor(),\n",
        ")\n",
        "\n",
        "# initialize the label names\n",
        "labelNames = [\"top\", \"trouser\", \"pullover\", \"dress\", \"coat\", \n",
        "             \"sandal\", \"shirt\", \"sneaker\", \"bag\", \"ankle boot\"]"
      ],
      "metadata": {
        "id": "KS5iII2Go8uV",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 443,
          "referenced_widgets": [
            "8e607298272e4b6db8ce9c0e3159bc6d",
            "0ab2982301e2483fb4b7f552821a8847",
            "b68eaeb25dc143f9b8bacb8e0d7c0e82",
            "f1b442c50512463ca08d04963c6f8019",
            "3bc93915b5cb4314b719327a7230f584",
            "38f4012b08c84219ad14b2fecc55eef2",
            "f30f400428c143d98913f6530ad2838f",
            "00495ecf06944d8c95d0a3272f1b52e4",
            "49a8038d370b40d7a01224ad6bfcdf87",
            "bd1720c80da94882b01c894dd522f4c9",
            "57ee79d0060445b7944bc25b2e3505ce",
            "fe87464d51214287a865313d3f6503b5",
            "8d54e16c1f2d403a96793993c478248b",
            "bc8017acebbb4e1daea1d16debcf112b",
            "3612aed2dd66408e9f9559d6971d9918",
            "e85c3b584dd24c4ab49ef44b1dda81f6",
            "cb27556cad68447b9d1b27665b1ae60c",
            "3331ecac16514cb391dd9673c25300f5",
            "df365b11a4d347ae83d18c16236bfe0e",
            "6aa75f0b6819427b8e7fdc86f7343bdb",
            "5d3e8b785a014e88a9457be06d5549b6",
            "dcfdc576bbce4377821e7f8b0165c193",
            "9b7da931ad354203b7b4253f3107df8b",
            "7782931043c146aa8af2814ab6066c00",
            "ed332bae787b4abfac54a78f92ef4e85",
            "75cc26ea990146ce8e5007755acc0e64",
            "c873ea044276451a8b5ecd236b8f9974",
            "0d7bd0351b2f48a3881b5f8a9e2596c7",
            "0b6801fa900c47d0bc7715b9354192fb",
            "7e16cf580e8e411e85b54ca4ff17add3",
            "91bad34abb334d7d9214590146074691",
            "bbd0eba42b4f4f8db9249248a0a06f12",
            "ca661ae9bc39409bb53e137b995c8560",
            "e606284a47f545aba2fef6786f3f35ba",
            "ef305e87bb404d5ea9e1f2480381d8a0",
            "62b8398b036348ceb3b11309a8e6ddbe",
            "330aec21439343c09386d050db0fd799",
            "723ca6eed92e48a58903c5f1cc07f3bf",
            "42c6a1851b104a9ea8bd8723a133629d",
            "058bdd86c26d4c5a8e8286e8a66c0950",
            "5dcde03f66f24f6eb011d02167023538",
            "77a1ecbf04444adbbafc212660c9ab03",
            "04a7b752faa4477b9235db07d3b0e09b",
            "0db68ce136564a1ca0bf334982efc13c"
          ]
        },
        "outputId": "948433e9-5357-4544-e6fd-20c73578d0ec"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/train-images-idx3-ubyte.gz\n",
            "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/train-images-idx3-ubyte.gz to data/FashionMNIST/raw/train-images-idx3-ubyte.gz\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "  0%|          | 0/26421880 [00:00<?, ?it/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "8e607298272e4b6db8ce9c0e3159bc6d"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting data/FashionMNIST/raw/train-images-idx3-ubyte.gz to data/FashionMNIST/raw\n",
            "\n",
            "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/train-labels-idx1-ubyte.gz\n",
            "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/train-labels-idx1-ubyte.gz to data/FashionMNIST/raw/train-labels-idx1-ubyte.gz\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "  0%|          | 0/29515 [00:00<?, ?it/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "fe87464d51214287a865313d3f6503b5"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting data/FashionMNIST/raw/train-labels-idx1-ubyte.gz to data/FashionMNIST/raw\n",
            "\n",
            "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/t10k-images-idx3-ubyte.gz\n",
            "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/t10k-images-idx3-ubyte.gz to data/FashionMNIST/raw/t10k-images-idx3-ubyte.gz\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "  0%|          | 0/4422102 [00:00<?, ?it/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "9b7da931ad354203b7b4253f3107df8b"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting data/FashionMNIST/raw/t10k-images-idx3-ubyte.gz to data/FashionMNIST/raw\n",
            "\n",
            "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/t10k-labels-idx1-ubyte.gz\n",
            "Downloading http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/t10k-labels-idx1-ubyte.gz to data/FashionMNIST/raw/t10k-labels-idx1-ubyte.gz\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "  0%|          | 0/5148 [00:00<?, ?it/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "e606284a47f545aba2fef6786f3f35ba"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting data/FashionMNIST/raw/t10k-labels-idx1-ubyte.gz to data/FashionMNIST/raw\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "training_data[0][0].shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MVzYrpJ4r4zl",
        "outputId": "5f430475-84d8-4554-c4dd-648095f5f94e"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([1, 28, 28])"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "training_data[0][1]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EeWhAmAhs4ZQ",
        "outputId": "8efd1e4a-41f7-4062-f82a-0e06abd81f88"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "9"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# flatten 28*28 images to a 784 vector for each image\n",
        "width = height = 28\n",
        "num_pixels = width * height\n",
        "\n",
        "num_classes = 10\n",
        "\n",
        "# 훈련셋과 검증셋 분리\n",
        "train_data, val_data = torch.utils.data.random_split(training_data, [50000, 10000])\n",
        "\n",
        "print ('train shape: \\t', len(train_data))\n",
        "print ('valid shape: \\t', len(val_data))\n",
        "print ('test shape: \\t', len(test_data))\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "545EKzYOoT-o",
        "outputId": "6ae6a384-f7a4-4eb0-cebf-a4868cb6e2b8"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "train shape: \t 50000\n",
            "valid shape: \t 10000\n",
            "test shape: \t 10000\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "batch_size = 100\n",
        "\n",
        "# Create data loaders.\n",
        "train_dataloader = DataLoader(train_data, batch_size=batch_size)\n",
        "val_dataloader = DataLoader(val_data, batch_size=batch_size)\n",
        "test_dataloader = DataLoader(test_data, batch_size=batch_size)\n",
        "\n",
        "for testX, testY in test_dataloader:\n",
        "    print(f\"Shape of X [N, C, H, W]: {testX.shape}\")\n",
        "    print(f\"Shape of y: {testY.shape} {testY.dtype}\")\n",
        "    break"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9h3ll-guprOm",
        "outputId": "6e3cf9bf-052c-47c4-a76a-4f9295fdb2d0"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Shape of X [N, C, H, W]: torch.Size([100, 1, 28, 28])\n",
            "Shape of y: torch.Size([100]) torch.int64\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test_dataloader.dataset.data.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "viY0n2RMqInj",
        "outputId": "ac611509-0cdc-46ab-92cb-71b64888a167"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([10000, 28, 28])"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "testX.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KoUKNs4zqx85",
        "outputId": "09e0b06d-819a-49b3-f4cc-c43865c5a468"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([100, 1, 28, 28])"
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8RwCuia7X_gD"
      },
      "source": [
        "#STEP 2: 데이터 살펴보기"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YFgKtGe4YFoT",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 730
        },
        "outputId": "2cfcaedc-0755-44c2-8d15-b3c692a05243"
      },
      "source": [
        "plt_row = 5\n",
        "plt_col = 5\n",
        "\n",
        "width = height = 28\n",
        "\n",
        "plt.rcParams[\"figure.figsize\"] = (15,15)\n",
        "\n",
        "f, axarr = plt.subplots(plt_row, plt_col)\n",
        "\n",
        "for i in range(plt_row*plt_col):\n",
        "\n",
        "    sub_plt = axarr[int(i/plt_row), i%plt_col]\n",
        "    sub_plt.axis('off')\n",
        "    sub_plt.imshow(testX[i].reshape(width, height), cmap='gray')\n",
        "    sub_plt_title = 'R: ' + labelNames[testY[i]]\n",
        "    sub_plt.set_title(sub_plt_title)\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 25 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aPIIgjLbc9Cu"
      },
      "source": [
        "#STEP 3: 딥러닝을 위한 데이터 전처리"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "testX[0]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "s6V70-JOnLHR",
        "outputId": "211b3b22-091d-47f3-c131-ccf455a946c0"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[[0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0118, 0.0039, 0.0000, 0.0000, 0.0275,\n",
              "          0.0000, 0.1451, 0.0000, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0039, 0.0078, 0.0000,\n",
              "          0.1059, 0.3294, 0.0431, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.4667, 0.0000, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0039, 0.0000, 0.0000,\n",
              "          0.3451, 0.5608, 0.4314, 0.0000, 0.0000, 0.0000, 0.0000, 0.0863,\n",
              "          0.3647, 0.4157, 0.0000, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0157, 0.0000, 0.2078,\n",
              "          0.5059, 0.4706, 0.5765, 0.6863, 0.6157, 0.6510, 0.5294, 0.6039,\n",
              "          0.6588, 0.5490, 0.0000, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0078, 0.0000, 0.0431, 0.5373,\n",
              "          0.5098, 0.5020, 0.6275, 0.6902, 0.6235, 0.6549, 0.6980, 0.5843,\n",
              "          0.5922, 0.5647, 0.0000, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0039, 0.0000,\n",
              "          0.0078, 0.0039, 0.0000, 0.0118, 0.0000, 0.0000, 0.4510, 0.4471,\n",
              "          0.4157, 0.5373, 0.6588, 0.6000, 0.6118, 0.6471, 0.6549, 0.5608,\n",
              "          0.6157, 0.6196, 0.0431, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0039, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0118, 0.0000, 0.0000, 0.3490, 0.5451, 0.3529,\n",
              "          0.3686, 0.6000, 0.5843, 0.5137, 0.5922, 0.6627, 0.6745, 0.5608,\n",
              "          0.6235, 0.6627, 0.1882, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0078, 0.0157,\n",
              "          0.0039, 0.0000, 0.0000, 0.0000, 0.3843, 0.5333, 0.4314, 0.4275,\n",
              "          0.4314, 0.6353, 0.5294, 0.5647, 0.5843, 0.6235, 0.6549, 0.5647,\n",
              "          0.6196, 0.6627, 0.4667, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0078, 0.0078, 0.0039, 0.0078, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.1020, 0.4235, 0.4588, 0.3882, 0.4353, 0.4588,\n",
              "          0.5333, 0.6118, 0.5255, 0.6039, 0.6039, 0.6118, 0.6275, 0.5529,\n",
              "          0.5765, 0.6118, 0.6980, 0.0000],\n",
              "         [0.0118, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0824,\n",
              "          0.2078, 0.3608, 0.4588, 0.4353, 0.4039, 0.4510, 0.5059, 0.5255,\n",
              "          0.5608, 0.6039, 0.6471, 0.6667, 0.6039, 0.5922, 0.6039, 0.5608,\n",
              "          0.5412, 0.5882, 0.6471, 0.1686],\n",
              "         [0.0000, 0.0000, 0.0902, 0.2118, 0.2549, 0.2980, 0.3333, 0.4627,\n",
              "          0.5020, 0.4824, 0.4353, 0.4431, 0.4627, 0.4980, 0.4902, 0.5451,\n",
              "          0.5216, 0.5333, 0.6275, 0.5490, 0.6078, 0.6314, 0.5647, 0.6078,\n",
              "          0.6745, 0.6314, 0.7412, 0.2431],\n",
              "         [0.0000, 0.2667, 0.3686, 0.3529, 0.4353, 0.4471, 0.4353, 0.4471,\n",
              "          0.4510, 0.4980, 0.5294, 0.5333, 0.5608, 0.4941, 0.4980, 0.5922,\n",
              "          0.6039, 0.5608, 0.5804, 0.4902, 0.6353, 0.6353, 0.5647, 0.5412,\n",
              "          0.6000, 0.6353, 0.7686, 0.2275],\n",
              "         [0.2745, 0.6627, 0.5059, 0.4078, 0.3843, 0.3922, 0.3686, 0.3804,\n",
              "          0.3843, 0.4000, 0.4235, 0.4157, 0.4667, 0.4706, 0.5059, 0.5843,\n",
              "          0.6118, 0.6549, 0.7451, 0.7451, 0.7686, 0.7765, 0.7765, 0.7333,\n",
              "          0.7725, 0.7412, 0.7216, 0.1412],\n",
              "         [0.0627, 0.4941, 0.6706, 0.7373, 0.7373, 0.7216, 0.6706, 0.6000,\n",
              "          0.5294, 0.4706, 0.4941, 0.4980, 0.5725, 0.7255, 0.7647, 0.8196,\n",
              "          0.8157, 1.0000, 0.8196, 0.6941, 0.9608, 0.9882, 0.9843, 0.9843,\n",
              "          0.9686, 0.8627, 0.8078, 0.1922],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0471, 0.2627, 0.4157, 0.6431, 0.7255,\n",
              "          0.7804, 0.8235, 0.8275, 0.8235, 0.8157, 0.7451, 0.5882, 0.3216,\n",
              "          0.0314, 0.0000, 0.0000, 0.0000, 0.6980, 0.8157, 0.7373, 0.6863,\n",
              "          0.6353, 0.6196, 0.5922, 0.0431],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000],\n",
              "         [0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000,\n",
              "          0.0000, 0.0000, 0.0000, 0.0000]]])"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6BhD2uuPfgZG"
      },
      "source": [
        "#STEP 4: 첫번째 인공지능 모델 (퍼셉트론)\n",
        "\n",
        "![대체 텍스트](https://www.simplilearn.com/ice9/free_resources_article_thumb/diagram-of-a-biological-neuron.jpg)\n",
        "\n",
        "![대체 텍스트](http://bit.ly/2ldH0Bg)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class LogisticRegression(nn.Module):\n",
        "  def __init__(self):\n",
        "    super(LogisticRegression, self).__init__()\n",
        "    self.flatten = nn.Flatten()\n",
        "    self.log_reg = nn.Sequential(\n",
        "        nn.Linear(num_pixels, num_classes),\n",
        "    )\n",
        "  def forward(self, x):\n",
        "    x = self.flatten(x)\n",
        "    logits = self.log_reg(x)\n",
        "    return logits"
      ],
      "metadata": {
        "id": "ReaVA8m37xSw"
      },
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model = LogisticRegression().to(device)"
      ],
      "metadata": {
        "id": "G_axoz_o-r81"
      },
      "execution_count": 46,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from torchsummary import summary\n",
        "summary(model, input_size=(1, 28, 28))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "G0WN4loAxJWS",
        "outputId": "0ff8f9f4-79b1-4767-a363-45690cef0284"
      },
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "----------------------------------------------------------------\n",
            "        Layer (type)               Output Shape         Param #\n",
            "================================================================\n",
            "           Flatten-1                  [-1, 784]               0\n",
            "            Linear-2                   [-1, 10]           7,850\n",
            "================================================================\n",
            "Total params: 7,850\n",
            "Trainable params: 7,850\n",
            "Non-trainable params: 0\n",
            "----------------------------------------------------------------\n",
            "Input size (MB): 0.00\n",
            "Forward/backward pass size (MB): 0.01\n",
            "Params size (MB): 0.03\n",
            "Estimated Total Size (MB): 0.04\n",
            "----------------------------------------------------------------\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DVwdrbt2gyWg"
      },
      "source": [
        "#STEP 5: 첫번째 인공지능 모델 학습!!!"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "loss_fn = nn.CrossEntropyLoss()\n",
        "optimizer = torch.optim.SGD(model.parameters(), lr=0.01, weight_decay=1e-6, momentum=0.9, nesterov=True)"
      ],
      "metadata": {
        "id": "mNWi2tnK_TLm"
      },
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def train(dataloader, model, loss_fn, optimizer):\n",
        "    size = len(dataloader.dataset)\n",
        "    num_batches = len(dataloader)\n",
        "    model.train()\n",
        "    train_loss, correct = 0, 0\n",
        "    for batch, (X, y) in enumerate(dataloader):\n",
        "        X, y = X.to(device), y.to(device)\n",
        "\n",
        "        # Compute prediction error\n",
        "        pred = model(X)\n",
        "        loss = loss_fn(pred, y)\n",
        "\n",
        "        train_loss += loss.item()\n",
        "        correct += (pred.argmax(1) == y).type(torch.float).sum().item()\n",
        "\n",
        "        # Backpropagation\n",
        "        optimizer.zero_grad()\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "\n",
        "        if batch % 100 == 0:\n",
        "            loss, current = loss.item(), batch * len(X)\n",
        "            print(f\"loss: {loss:>7f}  [{current:>5d}/{size:>5d}]\")\n",
        "\n",
        "    train_loss /= num_batches\n",
        "    correct /= size\n",
        "\n",
        "    print(f\"Train: \\n Accuracy: {(100*correct):>0.1f}%, Avg loss: {train_loss:>8f} \\n\")\n",
        "\n",
        "    return train_loss, 100*correct"
      ],
      "metadata": {
        "id": "FfeEcPafsfzC"
      },
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def test(dataloader, model, loss_fn):\n",
        "    size = len(dataloader.dataset)\n",
        "    num_batches = len(dataloader)\n",
        "    model.eval()\n",
        "    test_loss, correct = 0, 0\n",
        "    with torch.no_grad():\n",
        "        for X, y in dataloader:\n",
        "            X, y = X.to(device), y.to(device)\n",
        "            pred = model(X)\n",
        "            test_loss += loss_fn(pred, y).item()\n",
        "            correct += (pred.argmax(1) == y).type(torch.float).sum().item()\n",
        "    \n",
        "    test_loss /= num_batches\n",
        "    correct /= size\n",
        "    print(f\"Test: \\n Accuracy: {(100*correct):>0.1f}%, Avg loss: {test_loss:>8f} \\n\")\n",
        "\n",
        "    return test_loss, 100*correct"
      ],
      "metadata": {
        "id": "p4d3t_cdsiLJ"
      },
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# fix random seed for reproductibility\n",
        "seed = 7\n",
        "np.random.seed(seed)\n",
        "\n",
        "epochs = 20\n",
        "\n",
        "loss = []\n",
        "accuracy = []\n",
        "val_loss = []\n",
        "val_accuracy = []\n",
        "\n",
        "for t in range(epochs):\n",
        "    print(f\"Epoch {t+1}\\n-------------------------------\")\n",
        "    tmp_loss, tmp_accuracy = train(train_dataloader, model, loss_fn, optimizer)\n",
        "    loss.append(tmp_loss)\n",
        "    accuracy.append(tmp_accuracy)\n",
        "    tmp_loss, tmp_accuracy = test(val_dataloader, model, loss_fn)\n",
        "    val_loss.append(tmp_loss)\n",
        "    val_accuracy.append(tmp_accuracy)\n",
        "print(\"Done!\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "R2tlBrZ6_hSB",
        "outputId": "6f7dac32-8e11-4158-9799-da69667512c0"
      },
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1\n",
            "-------------------------------\n",
            "loss: 2.309973  [    0/50000]\n",
            "loss: 0.646663  [10000/50000]\n",
            "loss: 0.610393  [20000/50000]\n",
            "loss: 0.755498  [30000/50000]\n",
            "loss: 0.553119  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 78.0%, Avg loss: 0.678032 \n",
            "\n",
            "Test: \n",
            " Accuracy: 81.9%, Avg loss: 0.534517 \n",
            "\n",
            "Epoch 2\n",
            "-------------------------------\n",
            "loss: 0.534577  [    0/50000]\n",
            "loss: 0.460759  [10000/50000]\n",
            "loss: 0.524083  [20000/50000]\n",
            "loss: 0.686839  [30000/50000]\n",
            "loss: 0.486284  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 83.1%, Avg loss: 0.509938 \n",
            "\n",
            "Test: \n",
            " Accuracy: 83.3%, Avg loss: 0.488917 \n",
            "\n",
            "Epoch 3\n",
            "-------------------------------\n",
            "loss: 0.484022  [    0/50000]\n",
            "loss: 0.425178  [10000/50000]\n",
            "loss: 0.499515  [20000/50000]\n",
            "loss: 0.663066  [30000/50000]\n",
            "loss: 0.461344  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 84.0%, Avg loss: 0.477372 \n",
            "\n",
            "Test: \n",
            " Accuracy: 83.9%, Avg loss: 0.468246 \n",
            "\n",
            "Epoch 4\n",
            "-------------------------------\n",
            "loss: 0.458860  [    0/50000]\n",
            "loss: 0.408235  [10000/50000]\n",
            "loss: 0.484617  [20000/50000]\n",
            "loss: 0.651572  [30000/50000]\n",
            "loss: 0.447644  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 84.5%, Avg loss: 0.459725 \n",
            "\n",
            "Test: \n",
            " Accuracy: 84.2%, Avg loss: 0.455935 \n",
            "\n",
            "Epoch 5\n",
            "-------------------------------\n",
            "loss: 0.442636  [    0/50000]\n",
            "loss: 0.397764  [10000/50000]\n",
            "loss: 0.473714  [20000/50000]\n",
            "loss: 0.645106  [30000/50000]\n",
            "loss: 0.438837  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 84.8%, Avg loss: 0.448091 \n",
            "\n",
            "Test: \n",
            " Accuracy: 84.5%, Avg loss: 0.447649 \n",
            "\n",
            "Epoch 6\n",
            "-------------------------------\n",
            "loss: 0.430823  [    0/50000]\n",
            "loss: 0.390252  [10000/50000]\n",
            "loss: 0.465125  [20000/50000]\n",
            "loss: 0.640983  [30000/50000]\n",
            "loss: 0.432641  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 85.0%, Avg loss: 0.439623 \n",
            "\n",
            "Test: \n",
            " Accuracy: 84.8%, Avg loss: 0.441654 \n",
            "\n",
            "Epoch 7\n",
            "-------------------------------\n",
            "loss: 0.421615  [    0/50000]\n",
            "loss: 0.384342  [10000/50000]\n",
            "loss: 0.458107  [20000/50000]\n",
            "loss: 0.638039  [30000/50000]\n",
            "loss: 0.428007  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 85.2%, Avg loss: 0.433069 \n",
            "\n",
            "Test: \n",
            " Accuracy: 84.8%, Avg loss: 0.437100 \n",
            "\n",
            "Epoch 8\n",
            "-------------------------------\n",
            "loss: 0.414138  [    0/50000]\n",
            "loss: 0.379413  [10000/50000]\n",
            "loss: 0.452245  [20000/50000]\n",
            "loss: 0.635724  [30000/50000]\n",
            "loss: 0.424380  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 85.4%, Avg loss: 0.427780 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.0%, Avg loss: 0.433516 \n",
            "\n",
            "Epoch 9\n",
            "-------------------------------\n",
            "loss: 0.407900  [    0/50000]\n",
            "loss: 0.375150  [10000/50000]\n",
            "loss: 0.447273  [20000/50000]\n",
            "loss: 0.633760  [30000/50000]\n",
            "loss: 0.421439  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 85.5%, Avg loss: 0.423379 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.2%, Avg loss: 0.430618 \n",
            "\n",
            "Epoch 10\n",
            "-------------------------------\n",
            "loss: 0.402597  [    0/50000]\n",
            "loss: 0.371375  [10000/50000]\n",
            "loss: 0.443006  [20000/50000]\n",
            "loss: 0.632003  [30000/50000]\n",
            "loss: 0.418986  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 85.6%, Avg loss: 0.419631 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.2%, Avg loss: 0.428225 \n",
            "\n",
            "Epoch 11\n",
            "-------------------------------\n",
            "loss: 0.398024  [    0/50000]\n",
            "loss: 0.367981  [10000/50000]\n",
            "loss: 0.439309  [20000/50000]\n",
            "loss: 0.630372  [30000/50000]\n",
            "loss: 0.416895  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 85.8%, Avg loss: 0.416381 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.3%, Avg loss: 0.426215 \n",
            "\n",
            "Epoch 12\n",
            "-------------------------------\n",
            "loss: 0.394036  [    0/50000]\n",
            "loss: 0.364898  [10000/50000]\n",
            "loss: 0.436081  [20000/50000]\n",
            "loss: 0.628825  [30000/50000]\n",
            "loss: 0.415081  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 85.8%, Avg loss: 0.413521 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.4%, Avg loss: 0.424503 \n",
            "\n",
            "Epoch 13\n",
            "-------------------------------\n",
            "loss: 0.390524  [    0/50000]\n",
            "loss: 0.362076  [10000/50000]\n",
            "loss: 0.433242  [20000/50000]\n",
            "loss: 0.627336  [30000/50000]\n",
            "loss: 0.413484  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 85.9%, Avg loss: 0.410974 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.4%, Avg loss: 0.423027 \n",
            "\n",
            "Epoch 14\n",
            "-------------------------------\n",
            "loss: 0.387405  [    0/50000]\n",
            "loss: 0.359481  [10000/50000]\n",
            "loss: 0.430729  [20000/50000]\n",
            "loss: 0.625893  [30000/50000]\n",
            "loss: 0.412062  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.0%, Avg loss: 0.408684 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.4%, Avg loss: 0.421743 \n",
            "\n",
            "Epoch 15\n",
            "-------------------------------\n",
            "loss: 0.384612  [    0/50000]\n",
            "loss: 0.357082  [10000/50000]\n",
            "loss: 0.428492  [20000/50000]\n",
            "loss: 0.624487  [30000/50000]\n",
            "loss: 0.410785  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.0%, Avg loss: 0.406606 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.5%, Avg loss: 0.420617 \n",
            "\n",
            "Epoch 16\n",
            "-------------------------------\n",
            "loss: 0.382095  [    0/50000]\n",
            "loss: 0.354858  [10000/50000]\n",
            "loss: 0.426491  [20000/50000]\n",
            "loss: 0.623114  [30000/50000]\n",
            "loss: 0.409628  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.1%, Avg loss: 0.404708 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.5%, Avg loss: 0.419621 \n",
            "\n",
            "Epoch 17\n",
            "-------------------------------\n",
            "loss: 0.379811  [    0/50000]\n",
            "loss: 0.352790  [10000/50000]\n",
            "loss: 0.424692  [20000/50000]\n",
            "loss: 0.621773  [30000/50000]\n",
            "loss: 0.408574  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.2%, Avg loss: 0.402963 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.6%, Avg loss: 0.418735 \n",
            "\n",
            "Epoch 18\n",
            "-------------------------------\n",
            "loss: 0.377726  [    0/50000]\n",
            "loss: 0.350862  [10000/50000]\n",
            "loss: 0.423067  [20000/50000]\n",
            "loss: 0.620462  [30000/50000]\n",
            "loss: 0.407607  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.2%, Avg loss: 0.401350 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.6%, Avg loss: 0.417942 \n",
            "\n",
            "Epoch 19\n",
            "-------------------------------\n",
            "loss: 0.375813  [    0/50000]\n",
            "loss: 0.349059  [10000/50000]\n",
            "loss: 0.421593  [20000/50000]\n",
            "loss: 0.619180  [30000/50000]\n",
            "loss: 0.406717  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.3%, Avg loss: 0.399852 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.6%, Avg loss: 0.417231 \n",
            "\n",
            "Epoch 20\n",
            "-------------------------------\n",
            "loss: 0.374046  [    0/50000]\n",
            "loss: 0.347370  [10000/50000]\n",
            "loss: 0.420252  [20000/50000]\n",
            "loss: 0.617927  [30000/50000]\n",
            "loss: 0.405893  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.3%, Avg loss: 0.398455 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.6%, Avg loss: 0.416589 \n",
            "\n",
            "Done!\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 학습과정 살펴보기\n",
        "fig, loss_ax = plt.subplots()\n",
        "\n",
        "acc_ax = loss_ax.twinx()\n",
        "\n",
        "loss_ax.plot(loss, 'y', label='train loss')\n",
        "loss_ax.plot(val_loss, 'r', label='val loss')\n",
        "loss_ax.set_ylim([0.0, 1.5])\n",
        "\n",
        "acc_ax.plot(accuracy, 'b', label='train acc')\n",
        "acc_ax.plot(val_accuracy, 'g', label='val acc')\n",
        "acc_ax.set_ylim([50, 100])\n",
        "\n",
        "loss_ax.set_xlabel('epoch')\n",
        "loss_ax.set_ylabel('loss')\n",
        "acc_ax.set_ylabel('accuray')\n",
        "\n",
        "loss_ax.legend(loc='upper left')\n",
        "acc_ax.legend(loc='lower left')\n",
        "\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 683
        },
        "id": "fbdGyJjgx1Ch",
        "outputId": "2391f374-56f4-4e37-d5cd-7cb86d177434"
      },
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "J_i3PjlRjRSY"
      },
      "source": [
        "#STEP 6: 결과 확인 (테스트 데이터셋)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Final evaluation of the model\n",
        "test_loss, test_accuracy = test(test_dataloader, model, loss_fn)\n",
        "print(\"Baseline Error: %.2f%%\" % (100-test_accuracy))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3Rhb57UzyWK7",
        "outputId": "fea92a9d-188b-41eb-e2ec-7c7040b215e3"
      },
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test: \n",
            " Accuracy: 84.3%, Avg loss: 0.444806 \n",
            "\n",
            "Baseline Error: 15.69%\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "P9pKn3KijmbK"
      },
      "source": [
        "#STEP 7: 학습된 weight 살펴보기"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FmTwHrdQFobw",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 307
        },
        "outputId": "d0da73df-1ef5-46fa-f6ae-7721e939cfd8"
      },
      "source": [
        "# Visualize weights\n",
        "W = model.log_reg[0].weight.cpu().detach().numpy()\n",
        "print(\"W shape : \", W.shape)\n",
        "\n",
        "plt.figure(figsize=(15, 15), frameon=False)\n",
        "for ind, val in enumerate(W):\n",
        "    plt.subplot(5, 5, ind + 1)\n",
        "    im = val.reshape((28,28))\n",
        "    plt.axis(\"off\")\n",
        "    plt.imshow(im, cmap='gray',interpolation='nearest')"
      ],
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "W shape :  (10, 784)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 10 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hlxVSWn8kKXH"
      },
      "source": [
        "![대체 텍스트](https://www.pyimagesearch.com/wp-content/uploads/2019/02/fashion_mnist_obtaining.jpg)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "q-E36CT-koWF"
      },
      "source": [
        "#STEP 8: 두번째 인공지능 모델 (MLP)\n",
        "\n",
        "![대체 텍스트](https://www.researchgate.net/profile/Hadley_Brooks/publication/270274130/figure/fig3/AS:667886670594050@1536247999230/Architecture-of-a-multilayer-neural-network-with-one-hidden-layer-The-input-layer.png)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class MultiLayerPerceptron(nn.Module):\n",
        "  def __init__(self):\n",
        "    super(MultiLayerPerceptron, self).__init__()\n",
        "    self.flatten = nn.Flatten()\n",
        "    self.mlp = nn.Sequential(\n",
        "        nn.Linear(num_pixels, 256),\n",
        "        nn.Sigmoid(),\n",
        "        nn.Linear(256, num_classes),\n",
        "    )\n",
        "  def forward(self, x):\n",
        "    x = self.flatten(x)\n",
        "    logits = self.mlp(x)\n",
        "    return logits"
      ],
      "metadata": {
        "id": "fBWAmDrIzJE7"
      },
      "execution_count": 53,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ApdIJBsjFob1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "cf1f2127-b04d-444e-8b3e-029510ee4734"
      },
      "source": [
        "# build the model\n",
        "model = MultiLayerPerceptron().to(device)\n",
        "summary(model, input_size=(1, 28, 28))\n",
        "\n",
        "loss_fn = nn.CrossEntropyLoss()\n",
        "optimizer = torch.optim.SGD(model.parameters(), lr=0.01, weight_decay=1e-6, momentum=0.9, nesterov=True)\n",
        "\n",
        "# fix random seed for reproductibility\n",
        "seed = 7\n",
        "np.random.seed(seed)\n",
        "\n",
        "# Fit the model\n",
        "epochs = 20\n",
        "\n",
        "loss = []\n",
        "accuracy = []\n",
        "val_loss = []\n",
        "val_accuracy = []\n",
        "\n",
        "for t in range(epochs):\n",
        "    print(f\"Epoch {t+1}\\n-------------------------------\")\n",
        "    tmp_loss, tmp_accuracy = train(train_dataloader, model, loss_fn, optimizer)\n",
        "    loss.append(tmp_loss)\n",
        "    accuracy.append(tmp_accuracy)\n",
        "    tmp_loss, tmp_accuracy = test(val_dataloader, model, loss_fn)\n",
        "    val_loss.append(tmp_loss)\n",
        "    val_accuracy.append(tmp_accuracy)\n",
        "print(\"Done!\")"
      ],
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "----------------------------------------------------------------\n",
            "        Layer (type)               Output Shape         Param #\n",
            "================================================================\n",
            "           Flatten-1                  [-1, 784]               0\n",
            "            Linear-2                  [-1, 256]         200,960\n",
            "           Sigmoid-3                  [-1, 256]               0\n",
            "            Linear-4                   [-1, 10]           2,570\n",
            "================================================================\n",
            "Total params: 203,530\n",
            "Trainable params: 203,530\n",
            "Non-trainable params: 0\n",
            "----------------------------------------------------------------\n",
            "Input size (MB): 0.00\n",
            "Forward/backward pass size (MB): 0.01\n",
            "Params size (MB): 0.78\n",
            "Estimated Total Size (MB): 0.79\n",
            "----------------------------------------------------------------\n",
            "Epoch 1\n",
            "-------------------------------\n",
            "loss: 2.316170  [    0/50000]\n",
            "loss: 1.485126  [10000/50000]\n",
            "loss: 1.061400  [20000/50000]\n",
            "loss: 0.974858  [30000/50000]\n",
            "loss: 0.807665  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 64.6%, Avg loss: 1.135686 \n",
            "\n",
            "Test: \n",
            " Accuracy: 75.4%, Avg loss: 0.705990 \n",
            "\n",
            "Epoch 2\n",
            "-------------------------------\n",
            "loss: 0.710500  [    0/50000]\n",
            "loss: 0.641644  [10000/50000]\n",
            "loss: 0.660901  [20000/50000]\n",
            "loss: 0.798425  [30000/50000]\n",
            "loss: 0.602181  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 77.3%, Avg loss: 0.634777 \n",
            "\n",
            "Test: \n",
            " Accuracy: 79.5%, Avg loss: 0.580614 \n",
            "\n",
            "Epoch 3\n",
            "-------------------------------\n",
            "loss: 0.594276  [    0/50000]\n",
            "loss: 0.534785  [10000/50000]\n",
            "loss: 0.576280  [20000/50000]\n",
            "loss: 0.751221  [30000/50000]\n",
            "loss: 0.524660  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 80.5%, Avg loss: 0.550762 \n",
            "\n",
            "Test: \n",
            " Accuracy: 81.5%, Avg loss: 0.523847 \n",
            "\n",
            "Epoch 4\n",
            "-------------------------------\n",
            "loss: 0.534923  [    0/50000]\n",
            "loss: 0.480481  [10000/50000]\n",
            "loss: 0.537410  [20000/50000]\n",
            "loss: 0.722764  [30000/50000]\n",
            "loss: 0.486664  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 82.0%, Avg loss: 0.508289 \n",
            "\n",
            "Test: \n",
            " Accuracy: 82.6%, Avg loss: 0.492719 \n",
            "\n",
            "Epoch 5\n",
            "-------------------------------\n",
            "loss: 0.499838  [    0/50000]\n",
            "loss: 0.449705  [10000/50000]\n",
            "loss: 0.516090  [20000/50000]\n",
            "loss: 0.699266  [30000/50000]\n",
            "loss: 0.466167  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 82.9%, Avg loss: 0.482820 \n",
            "\n",
            "Test: \n",
            " Accuracy: 83.2%, Avg loss: 0.472561 \n",
            "\n",
            "Epoch 6\n",
            "-------------------------------\n",
            "loss: 0.478359  [    0/50000]\n",
            "loss: 0.429003  [10000/50000]\n",
            "loss: 0.501794  [20000/50000]\n",
            "loss: 0.678724  [30000/50000]\n",
            "loss: 0.453960  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 83.5%, Avg loss: 0.465009 \n",
            "\n",
            "Test: \n",
            " Accuracy: 83.7%, Avg loss: 0.457915 \n",
            "\n",
            "Epoch 7\n",
            "-------------------------------\n",
            "loss: 0.464215  [    0/50000]\n",
            "loss: 0.413495  [10000/50000]\n",
            "loss: 0.490659  [20000/50000]\n",
            "loss: 0.661423  [30000/50000]\n",
            "loss: 0.445693  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 84.1%, Avg loss: 0.451360 \n",
            "\n",
            "Test: \n",
            " Accuracy: 84.0%, Avg loss: 0.446547 \n",
            "\n",
            "Epoch 8\n",
            "-------------------------------\n",
            "loss: 0.453987  [    0/50000]\n",
            "loss: 0.401250  [10000/50000]\n",
            "loss: 0.481051  [20000/50000]\n",
            "loss: 0.647150  [30000/50000]\n",
            "loss: 0.439268  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 84.5%, Avg loss: 0.440308 \n",
            "\n",
            "Test: \n",
            " Accuracy: 84.3%, Avg loss: 0.437327 \n",
            "\n",
            "Epoch 9\n",
            "-------------------------------\n",
            "loss: 0.445852  [    0/50000]\n",
            "loss: 0.391209  [10000/50000]\n",
            "loss: 0.472296  [20000/50000]\n",
            "loss: 0.635374  [30000/50000]\n",
            "loss: 0.433700  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 84.8%, Avg loss: 0.431019 \n",
            "\n",
            "Test: \n",
            " Accuracy: 84.7%, Avg loss: 0.429605 \n",
            "\n",
            "Epoch 10\n",
            "-------------------------------\n",
            "loss: 0.438887  [    0/50000]\n",
            "loss: 0.382682  [10000/50000]\n",
            "loss: 0.464186  [20000/50000]\n",
            "loss: 0.625586  [30000/50000]\n",
            "loss: 0.428506  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 85.0%, Avg loss: 0.422995 \n",
            "\n",
            "Test: \n",
            " Accuracy: 84.9%, Avg loss: 0.422972 \n",
            "\n",
            "Epoch 11\n",
            "-------------------------------\n",
            "loss: 0.432629  [    0/50000]\n",
            "loss: 0.375188  [10000/50000]\n",
            "loss: 0.456670  [20000/50000]\n",
            "loss: 0.617367  [30000/50000]\n",
            "loss: 0.423455  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 85.3%, Avg loss: 0.415913 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.1%, Avg loss: 0.417157 \n",
            "\n",
            "Epoch 12\n",
            "-------------------------------\n",
            "loss: 0.426847  [    0/50000]\n",
            "loss: 0.368388  [10000/50000]\n",
            "loss: 0.449708  [20000/50000]\n",
            "loss: 0.610372  [30000/50000]\n",
            "loss: 0.418463  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 85.5%, Avg loss: 0.409555 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.3%, Avg loss: 0.411972 \n",
            "\n",
            "Epoch 13\n",
            "-------------------------------\n",
            "loss: 0.421418  [    0/50000]\n",
            "loss: 0.362052  [10000/50000]\n",
            "loss: 0.443248  [20000/50000]\n",
            "loss: 0.604327  [30000/50000]\n",
            "loss: 0.413528  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 85.7%, Avg loss: 0.403767 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.5%, Avg loss: 0.407292 \n",
            "\n",
            "Epoch 14\n",
            "-------------------------------\n",
            "loss: 0.416275  [    0/50000]\n",
            "loss: 0.356037  [10000/50000]\n",
            "loss: 0.437235  [20000/50000]\n",
            "loss: 0.599014  [30000/50000]\n",
            "loss: 0.408683  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 85.9%, Avg loss: 0.398438 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.5%, Avg loss: 0.403024 \n",
            "\n",
            "Epoch 15\n",
            "-------------------------------\n",
            "loss: 0.411377  [    0/50000]\n",
            "loss: 0.350261  [10000/50000]\n",
            "loss: 0.431613  [20000/50000]\n",
            "loss: 0.594262  [30000/50000]\n",
            "loss: 0.403965  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.1%, Avg loss: 0.393489 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.8%, Avg loss: 0.399101 \n",
            "\n",
            "Epoch 16\n",
            "-------------------------------\n",
            "loss: 0.406693  [    0/50000]\n",
            "loss: 0.344688  [10000/50000]\n",
            "loss: 0.426326  [20000/50000]\n",
            "loss: 0.589935  [30000/50000]\n",
            "loss: 0.399404  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.2%, Avg loss: 0.388857 \n",
            "\n",
            "Test: \n",
            " Accuracy: 86.0%, Avg loss: 0.395472 \n",
            "\n",
            "Epoch 17\n",
            "-------------------------------\n",
            "loss: 0.402202  [    0/50000]\n",
            "loss: 0.339309  [10000/50000]\n",
            "loss: 0.421327  [20000/50000]\n",
            "loss: 0.585921  [30000/50000]\n",
            "loss: 0.395016  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.4%, Avg loss: 0.384497 \n",
            "\n",
            "Test: \n",
            " Accuracy: 86.1%, Avg loss: 0.392095 \n",
            "\n",
            "Epoch 18\n",
            "-------------------------------\n",
            "loss: 0.397887  [    0/50000]\n",
            "loss: 0.334133  [10000/50000]\n",
            "loss: 0.416571  [20000/50000]\n",
            "loss: 0.582130  [30000/50000]\n",
            "loss: 0.390804  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.5%, Avg loss: 0.380374 \n",
            "\n",
            "Test: \n",
            " Accuracy: 86.2%, Avg loss: 0.388937 \n",
            "\n",
            "Epoch 19\n",
            "-------------------------------\n",
            "loss: 0.393739  [    0/50000]\n",
            "loss: 0.329174  [10000/50000]\n",
            "loss: 0.412020  [20000/50000]\n",
            "loss: 0.578493  [30000/50000]\n",
            "loss: 0.386768  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.7%, Avg loss: 0.376458 \n",
            "\n",
            "Test: \n",
            " Accuracy: 86.3%, Avg loss: 0.385973 \n",
            "\n",
            "Epoch 20\n",
            "-------------------------------\n",
            "loss: 0.389747  [    0/50000]\n",
            "loss: 0.324443  [10000/50000]\n",
            "loss: 0.407643  [20000/50000]\n",
            "loss: 0.574959  [30000/50000]\n",
            "loss: 0.382904  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.8%, Avg loss: 0.372728 \n",
            "\n",
            "Test: \n",
            " Accuracy: 86.3%, Avg loss: 0.383182 \n",
            "\n",
            "Done!\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 학습과정 살펴보기\n",
        "fig, loss_ax = plt.subplots()\n",
        "\n",
        "acc_ax = loss_ax.twinx()\n",
        "\n",
        "loss_ax.plot(loss, 'y', label='train loss')\n",
        "loss_ax.plot(val_loss, 'r', label='val loss')\n",
        "loss_ax.set_ylim([0.0, 1.5])\n",
        "\n",
        "acc_ax.plot(accuracy, 'b', label='train acc')\n",
        "acc_ax.plot(val_accuracy, 'g', label='val acc')\n",
        "acc_ax.set_ylim([50, 100])\n",
        "\n",
        "loss_ax.set_xlabel('epoch')\n",
        "loss_ax.set_ylabel('loss')\n",
        "acc_ax.set_ylabel('accuray')\n",
        "\n",
        "loss_ax.legend(loc='upper left')\n",
        "acc_ax.legend(loc='lower left')\n",
        "\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 683
        },
        "id": "dpiUdjTB-3dl",
        "outputId": "a67b953b-bb82-406e-b95f-0fcacc64673b"
      },
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dSEjknxsFob4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "adbea23f-7456-4063-e49e-9ccf6a21c613"
      },
      "source": [
        "# Final evaluation of the model\n",
        "test_loss, test_accuracy = test(test_dataloader, model, loss_fn)\n",
        "print(\"Baseline Error: %.2f%%\" % (100-test_accuracy))"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test: \n",
            " Accuracy: 85.8%, Avg loss: 0.420230 \n",
            "\n",
            "Baseline Error: 14.23%\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eLQtffyil8WK"
      },
      "source": [
        "#STEP 9: 세번째 인공지능 모델 (DEEP-MLP)\n",
        "\n",
        "![대체 텍스트](https://i.stack.imgur.com/OH3gI.png)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "26LDMwE1InL2"
      },
      "source": [
        "![대체 텍스트](http://www.saedsayad.com/images/ANN_Sigmoid.png)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class DeepLayerPerceptron(nn.Module):\n",
        "  def __init__(self):\n",
        "    super(DeepLayerPerceptron, self).__init__()\n",
        "    self.flatten = nn.Flatten()\n",
        "    self.mlp = nn.Sequential(\n",
        "        nn.Linear(num_pixels, 256),\n",
        "        nn.Sigmoid(),\n",
        "        nn.Linear(256, 256),\n",
        "        nn.Sigmoid(),        \n",
        "        nn.Linear(256, 256),\n",
        "        nn.Sigmoid(),        \n",
        "        nn.Linear(256, 256),\n",
        "        nn.Sigmoid(),        \n",
        "        nn.Linear(256, 256),\n",
        "        nn.Sigmoid(),        \n",
        "        nn.Linear(256, num_classes),\n",
        "    )\n",
        "  def forward(self, x):\n",
        "    x = self.flatten(x)\n",
        "    logits = self.mlp(x)\n",
        "    return logits"
      ],
      "metadata": {
        "id": "BMgdRHoNAFCN"
      },
      "execution_count": 56,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3SIj7p4dFob_",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a49bc8e7-e058-490a-cd35-0fd549699859"
      },
      "source": [
        "# build the model\n",
        "model = DeepLayerPerceptron().to(device)\n",
        "summary(model, input_size=(1, 28, 28))\n",
        "\n",
        "loss_fn = nn.CrossEntropyLoss()\n",
        "optimizer = torch.optim.SGD(model.parameters(), lr=0.01, weight_decay=1e-6, momentum=0.9, nesterov=True)\n",
        "\n",
        "# fix random seed for reproductibility\n",
        "seed = 7\n",
        "np.random.seed(seed)\n",
        "\n",
        "# Fit the model\n",
        "epochs = 20\n",
        "\n",
        "loss = []\n",
        "accuracy = []\n",
        "val_loss = []\n",
        "val_accuracy = []\n",
        "\n",
        "for t in range(epochs):\n",
        "    print(f\"Epoch {t+1}\\n-------------------------------\")\n",
        "    tmp_loss, tmp_accuracy = train(train_dataloader, model, loss_fn, optimizer)\n",
        "    loss.append(tmp_loss)\n",
        "    accuracy.append(tmp_accuracy)\n",
        "    tmp_loss, tmp_accuracy = test(val_dataloader, model, loss_fn)\n",
        "    val_loss.append(tmp_loss)\n",
        "    val_accuracy.append(tmp_accuracy)\n",
        "print(\"Done!\")"
      ],
      "execution_count": 57,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "----------------------------------------------------------------\n",
            "        Layer (type)               Output Shape         Param #\n",
            "================================================================\n",
            "           Flatten-1                  [-1, 784]               0\n",
            "            Linear-2                  [-1, 256]         200,960\n",
            "           Sigmoid-3                  [-1, 256]               0\n",
            "            Linear-4                  [-1, 256]          65,792\n",
            "           Sigmoid-5                  [-1, 256]               0\n",
            "            Linear-6                  [-1, 256]          65,792\n",
            "           Sigmoid-7                  [-1, 256]               0\n",
            "            Linear-8                  [-1, 256]          65,792\n",
            "           Sigmoid-9                  [-1, 256]               0\n",
            "           Linear-10                  [-1, 256]          65,792\n",
            "          Sigmoid-11                  [-1, 256]               0\n",
            "           Linear-12                   [-1, 10]           2,570\n",
            "================================================================\n",
            "Total params: 466,698\n",
            "Trainable params: 466,698\n",
            "Non-trainable params: 0\n",
            "----------------------------------------------------------------\n",
            "Input size (MB): 0.00\n",
            "Forward/backward pass size (MB): 0.03\n",
            "Params size (MB): 1.78\n",
            "Estimated Total Size (MB): 1.81\n",
            "----------------------------------------------------------------\n",
            "Epoch 1\n",
            "-------------------------------\n",
            "loss: 2.345807  [    0/50000]\n",
            "loss: 2.308258  [10000/50000]\n",
            "loss: 2.336035  [20000/50000]\n",
            "loss: 2.336331  [30000/50000]\n",
            "loss: 2.304713  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 10.1%, Avg loss: 2.312507 \n",
            "\n",
            "Test: \n",
            " Accuracy: 10.2%, Avg loss: 2.306398 \n",
            "\n",
            "Epoch 2\n",
            "-------------------------------\n",
            "loss: 2.315394  [    0/50000]\n",
            "loss: 2.310939  [10000/50000]\n",
            "loss: 2.338319  [20000/50000]\n",
            "loss: 2.333797  [30000/50000]\n",
            "loss: 2.301911  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 10.0%, Avg loss: 2.311425 \n",
            "\n",
            "Test: \n",
            " Accuracy: 10.2%, Avg loss: 2.305999 \n",
            "\n",
            "Epoch 3\n",
            "-------------------------------\n",
            "loss: 2.314630  [    0/50000]\n",
            "loss: 2.313044  [10000/50000]\n",
            "loss: 2.337975  [20000/50000]\n",
            "loss: 2.331681  [30000/50000]\n",
            "loss: 2.300233  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 10.1%, Avg loss: 2.310572 \n",
            "\n",
            "Test: \n",
            " Accuracy: 10.2%, Avg loss: 2.305929 \n",
            "\n",
            "Epoch 4\n",
            "-------------------------------\n",
            "loss: 2.314541  [    0/50000]\n",
            "loss: 2.314582  [10000/50000]\n",
            "loss: 2.334625  [20000/50000]\n",
            "loss: 2.329902  [30000/50000]\n",
            "loss: 2.299710  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 10.1%, Avg loss: 2.309802 \n",
            "\n",
            "Test: \n",
            " Accuracy: 10.2%, Avg loss: 2.306034 \n",
            "\n",
            "Epoch 5\n",
            "-------------------------------\n",
            "loss: 2.314678  [    0/50000]\n",
            "loss: 2.315493  [10000/50000]\n",
            "loss: 2.328755  [20000/50000]\n",
            "loss: 2.328303  [30000/50000]\n",
            "loss: 2.300205  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 10.0%, Avg loss: 2.309104 \n",
            "\n",
            "Test: \n",
            " Accuracy: 10.2%, Avg loss: 2.306253 \n",
            "\n",
            "Epoch 6\n",
            "-------------------------------\n",
            "loss: 2.314717  [    0/50000]\n",
            "loss: 2.315612  [10000/50000]\n",
            "loss: 2.321697  [20000/50000]\n",
            "loss: 2.326866  [30000/50000]\n",
            "loss: 2.301374  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 9.9%, Avg loss: 2.308463 \n",
            "\n",
            "Test: \n",
            " Accuracy: 10.2%, Avg loss: 2.306551 \n",
            "\n",
            "Epoch 7\n",
            "-------------------------------\n",
            "loss: 2.314527  [    0/50000]\n",
            "loss: 2.314856  [10000/50000]\n",
            "loss: 2.314963  [20000/50000]\n",
            "loss: 2.325645  [30000/50000]\n",
            "loss: 2.302825  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 9.9%, Avg loss: 2.307872 \n",
            "\n",
            "Test: \n",
            " Accuracy: 10.2%, Avg loss: 2.306847 \n",
            "\n",
            "Epoch 8\n",
            "-------------------------------\n",
            "loss: 2.314094  [    0/50000]\n",
            "loss: 2.313332  [10000/50000]\n",
            "loss: 2.309599  [20000/50000]\n",
            "loss: 2.324592  [30000/50000]\n",
            "loss: 2.304284  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 9.8%, Avg loss: 2.307329 \n",
            "\n",
            "Test: \n",
            " Accuracy: 10.2%, Avg loss: 2.307044 \n",
            "\n",
            "Epoch 9\n",
            "-------------------------------\n",
            "loss: 2.313427  [    0/50000]\n",
            "loss: 2.311302  [10000/50000]\n",
            "loss: 2.305982  [20000/50000]\n",
            "loss: 2.323574  [30000/50000]\n",
            "loss: 2.305616  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 9.9%, Avg loss: 2.306837 \n",
            "\n",
            "Test: \n",
            " Accuracy: 10.2%, Avg loss: 2.307104 \n",
            "\n",
            "Epoch 10\n",
            "-------------------------------\n",
            "loss: 2.312583  [    0/50000]\n",
            "loss: 2.309093  [10000/50000]\n",
            "loss: 2.303971  [20000/50000]\n",
            "loss: 2.322528  [30000/50000]\n",
            "loss: 2.306751  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 9.8%, Avg loss: 2.306397 \n",
            "\n",
            "Test: \n",
            " Accuracy: 10.2%, Avg loss: 2.307033 \n",
            "\n",
            "Epoch 11\n",
            "-------------------------------\n",
            "loss: 2.311651  [    0/50000]\n",
            "loss: 2.306987  [10000/50000]\n",
            "loss: 2.303201  [20000/50000]\n",
            "loss: 2.321501  [30000/50000]\n",
            "loss: 2.307612  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 9.9%, Avg loss: 2.306011 \n",
            "\n",
            "Test: \n",
            " Accuracy: 10.2%, Avg loss: 2.306842 \n",
            "\n",
            "Epoch 12\n",
            "-------------------------------\n",
            "loss: 2.310690  [    0/50000]\n",
            "loss: 2.305149  [10000/50000]\n",
            "loss: 2.303295  [20000/50000]\n",
            "loss: 2.320585  [30000/50000]\n",
            "loss: 2.308153  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 9.8%, Avg loss: 2.305674 \n",
            "\n",
            "Test: \n",
            " Accuracy: 10.2%, Avg loss: 2.306548 \n",
            "\n",
            "Epoch 13\n",
            "-------------------------------\n",
            "loss: 2.309719  [    0/50000]\n",
            "loss: 2.303632  [10000/50000]\n",
            "loss: 2.303931  [20000/50000]\n",
            "loss: 2.319835  [30000/50000]\n",
            "loss: 2.308391  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 9.9%, Avg loss: 2.305382 \n",
            "\n",
            "Test: \n",
            " Accuracy: 10.2%, Avg loss: 2.306187 \n",
            "\n",
            "Epoch 14\n",
            "-------------------------------\n",
            "loss: 2.308747  [    0/50000]\n",
            "loss: 2.302425  [10000/50000]\n",
            "loss: 2.304847  [20000/50000]\n",
            "loss: 2.319247  [30000/50000]\n",
            "loss: 2.308400  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 9.8%, Avg loss: 2.305128 \n",
            "\n",
            "Test: \n",
            " Accuracy: 9.9%, Avg loss: 2.305800 \n",
            "\n",
            "Epoch 15\n",
            "-------------------------------\n",
            "loss: 2.307796  [    0/50000]\n",
            "loss: 2.301492  [10000/50000]\n",
            "loss: 2.305843  [20000/50000]\n",
            "loss: 2.318777  [30000/50000]\n",
            "loss: 2.308269  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 9.9%, Avg loss: 2.304906 \n",
            "\n",
            "Test: \n",
            " Accuracy: 9.9%, Avg loss: 2.305423 \n",
            "\n",
            "Epoch 16\n",
            "-------------------------------\n",
            "loss: 2.306891  [    0/50000]\n",
            "loss: 2.300792  [10000/50000]\n",
            "loss: 2.306780  [20000/50000]\n",
            "loss: 2.318374  [30000/50000]\n",
            "loss: 2.308068  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 9.8%, Avg loss: 2.304713 \n",
            "\n",
            "Test: \n",
            " Accuracy: 9.9%, Avg loss: 2.305077 \n",
            "\n",
            "Epoch 17\n",
            "-------------------------------\n",
            "loss: 2.306060  [    0/50000]\n",
            "loss: 2.300287  [10000/50000]\n",
            "loss: 2.307582  [20000/50000]\n",
            "loss: 2.317998  [30000/50000]\n",
            "loss: 2.307840  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 9.8%, Avg loss: 2.304543 \n",
            "\n",
            "Test: \n",
            " Accuracy: 9.9%, Avg loss: 2.304770 \n",
            "\n",
            "Epoch 18\n",
            "-------------------------------\n",
            "loss: 2.305315  [    0/50000]\n",
            "loss: 2.299940  [10000/50000]\n",
            "loss: 2.308216  [20000/50000]\n",
            "loss: 2.317628  [30000/50000]\n",
            "loss: 2.307609  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 9.8%, Avg loss: 2.304393 \n",
            "\n",
            "Test: \n",
            " Accuracy: 9.9%, Avg loss: 2.304506 \n",
            "\n",
            "Epoch 19\n",
            "-------------------------------\n",
            "loss: 2.304659  [    0/50000]\n",
            "loss: 2.299716  [10000/50000]\n",
            "loss: 2.308680  [20000/50000]\n",
            "loss: 2.317256  [30000/50000]\n",
            "loss: 2.307381  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 9.8%, Avg loss: 2.304262 \n",
            "\n",
            "Test: \n",
            " Accuracy: 9.9%, Avg loss: 2.304281 \n",
            "\n",
            "Epoch 20\n",
            "-------------------------------\n",
            "loss: 2.304091  [    0/50000]\n",
            "loss: 2.299585  [10000/50000]\n",
            "loss: 2.308993  [20000/50000]\n",
            "loss: 2.316880  [30000/50000]\n",
            "loss: 2.307159  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 9.8%, Avg loss: 2.304145 \n",
            "\n",
            "Test: \n",
            " Accuracy: 9.9%, Avg loss: 2.304092 \n",
            "\n",
            "Done!\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 학습과정 살펴보기\n",
        "fig, loss_ax = plt.subplots()\n",
        "\n",
        "acc_ax = loss_ax.twinx()\n",
        "\n",
        "loss_ax.plot(loss, 'y', label='train loss')\n",
        "loss_ax.plot(val_loss, 'r', label='val loss')\n",
        "loss_ax.set_ylim([1.5, 2.5])\n",
        "\n",
        "acc_ax.plot(accuracy, 'b', label='train acc')\n",
        "acc_ax.plot(val_accuracy, 'g', label='val acc')\n",
        "acc_ax.set_ylim([0, 100])\n",
        "\n",
        "loss_ax.set_xlabel('epoch')\n",
        "loss_ax.set_ylabel('loss')\n",
        "acc_ax.set_ylabel('accuray')\n",
        "\n",
        "loss_ax.legend(loc='upper left')\n",
        "acc_ax.legend(loc='lower left')\n",
        "\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 683
        },
        "id": "FVhBEtHRAvkk",
        "outputId": "3bb1d706-3308-49f1-8ab1-550adf4f10ef"
      },
      "execution_count": 58,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dTTj4GeCJ8K2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a8ce465c-7ed1-4a62-d3b7-89a56726fa41"
      },
      "source": [
        "# Final evaluation of the model\n",
        "test_loss, test_accuracy = test(test_dataloader, model, loss_fn)\n",
        "print(\"Baseline Error: %.2f%%\" % (100-test_accuracy))"
      ],
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test: \n",
            " Accuracy: 10.0%, Avg loss: 2.303933 \n",
            "\n",
            "Baseline Error: 90.00%\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WYfMo7tFnBed"
      },
      "source": [
        "#STEP 10: 세번째 인공지능 모델의 문제점과 개선"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sOdXlYa0YFTZ"
      },
      "source": [
        "![대체 텍스트](https://image.slidesharecdn.com/usuconference-deeplearning-160418191119/95/introduction-to-deep-learning-7-638.jpg?cb=1461006739)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BDmw6R8_GNed"
      },
      "source": [
        "![대체 텍스트](https://smartstuartkim.files.wordpress.com/2019/02/vanishinggradient-1.png?w=1140&h=492)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class DeepLayerPerceptronRelu(nn.Module):\n",
        "  def __init__(self):\n",
        "    super(DeepLayerPerceptronRelu, self).__init__()\n",
        "    self.flatten = nn.Flatten()\n",
        "    self.mlp = nn.Sequential(\n",
        "        nn.Linear(num_pixels, 256),\n",
        "        nn.ReLU(),\n",
        "        nn.Linear(256, 256),\n",
        "        nn.ReLU(),\n",
        "        nn.Linear(256, 256),\n",
        "        nn.ReLU(),\n",
        "        nn.Linear(256, 256),\n",
        "        nn.ReLU(),\n",
        "        nn.Linear(256, 256),\n",
        "        nn.ReLU(),\n",
        "        nn.Linear(256, num_classes),\n",
        "    )\n",
        "  def forward(self, x):\n",
        "    x = self.flatten(x)\n",
        "    logits = self.mlp(x)\n",
        "    return logits"
      ],
      "metadata": {
        "id": "wHNutVzdBifq"
      },
      "execution_count": 60,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uVgGp42FFocM",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d6a7ab08-4f61-4eea-9149-48aeecd0d3fd"
      },
      "source": [
        "# build the model\n",
        "model = DeepLayerPerceptronRelu().to(device)\n",
        "summary(model, input_size=(1, 28, 28))\n",
        "\n",
        "loss_fn = nn.CrossEntropyLoss()\n",
        "optimizer = torch.optim.SGD(model.parameters(), lr=0.01, weight_decay=1e-6, momentum=0.9, nesterov=True)\n",
        "\n",
        "# fix random seed for reproductibility\n",
        "seed = 7\n",
        "np.random.seed(seed)\n",
        "\n",
        "# Fit the model\n",
        "epochs = 20\n",
        "\n",
        "loss = []\n",
        "accuracy = []\n",
        "val_loss = []\n",
        "val_accuracy = []\n",
        "\n",
        "for t in range(epochs):\n",
        "    print(f\"Epoch {t+1}\\n-------------------------------\")\n",
        "    tmp_loss, tmp_accuracy = train(train_dataloader, model, loss_fn, optimizer)\n",
        "    loss.append(tmp_loss)\n",
        "    accuracy.append(tmp_accuracy)\n",
        "    tmp_loss, tmp_accuracy = test(val_dataloader, model, loss_fn)\n",
        "    val_loss.append(tmp_loss)\n",
        "    val_accuracy.append(tmp_accuracy)\n",
        "print(\"Done!\")"
      ],
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "----------------------------------------------------------------\n",
            "        Layer (type)               Output Shape         Param #\n",
            "================================================================\n",
            "           Flatten-1                  [-1, 784]               0\n",
            "            Linear-2                  [-1, 256]         200,960\n",
            "              ReLU-3                  [-1, 256]               0\n",
            "            Linear-4                  [-1, 256]          65,792\n",
            "              ReLU-5                  [-1, 256]               0\n",
            "            Linear-6                  [-1, 256]          65,792\n",
            "              ReLU-7                  [-1, 256]               0\n",
            "            Linear-8                  [-1, 256]          65,792\n",
            "              ReLU-9                  [-1, 256]               0\n",
            "           Linear-10                  [-1, 256]          65,792\n",
            "             ReLU-11                  [-1, 256]               0\n",
            "           Linear-12                   [-1, 10]           2,570\n",
            "================================================================\n",
            "Total params: 466,698\n",
            "Trainable params: 466,698\n",
            "Non-trainable params: 0\n",
            "----------------------------------------------------------------\n",
            "Input size (MB): 0.00\n",
            "Forward/backward pass size (MB): 0.03\n",
            "Params size (MB): 1.78\n",
            "Estimated Total Size (MB): 1.81\n",
            "----------------------------------------------------------------\n",
            "Epoch 1\n",
            "-------------------------------\n",
            "loss: 2.306039  [    0/50000]\n",
            "loss: 2.290097  [10000/50000]\n",
            "loss: 1.986015  [20000/50000]\n",
            "loss: 1.083272  [30000/50000]\n",
            "loss: 0.981024  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 42.8%, Avg loss: 1.515156 \n",
            "\n",
            "Test: \n",
            " Accuracy: 69.5%, Avg loss: 0.739034 \n",
            "\n",
            "Epoch 2\n",
            "-------------------------------\n",
            "loss: 0.690393  [    0/50000]\n",
            "loss: 0.584905  [10000/50000]\n",
            "loss: 0.670345  [20000/50000]\n",
            "loss: 0.866620  [30000/50000]\n",
            "loss: 0.583158  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 79.1%, Avg loss: 0.581090 \n",
            "\n",
            "Test: \n",
            " Accuracy: 82.6%, Avg loss: 0.484788 \n",
            "\n",
            "Epoch 3\n",
            "-------------------------------\n",
            "loss: 0.434406  [    0/50000]\n",
            "loss: 0.370494  [10000/50000]\n",
            "loss: 0.503696  [20000/50000]\n",
            "loss: 0.727456  [30000/50000]\n",
            "loss: 0.509044  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 84.1%, Avg loss: 0.445189 \n",
            "\n",
            "Test: \n",
            " Accuracy: 84.9%, Avg loss: 0.412572 \n",
            "\n",
            "Epoch 4\n",
            "-------------------------------\n",
            "loss: 0.389876  [    0/50000]\n",
            "loss: 0.305184  [10000/50000]\n",
            "loss: 0.468218  [20000/50000]\n",
            "loss: 0.662817  [30000/50000]\n",
            "loss: 0.443606  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.1%, Avg loss: 0.388258 \n",
            "\n",
            "Test: \n",
            " Accuracy: 86.2%, Avg loss: 0.376132 \n",
            "\n",
            "Epoch 5\n",
            "-------------------------------\n",
            "loss: 0.361439  [    0/50000]\n",
            "loss: 0.287888  [10000/50000]\n",
            "loss: 0.431226  [20000/50000]\n",
            "loss: 0.632450  [30000/50000]\n",
            "loss: 0.398512  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 87.2%, Avg loss: 0.354524 \n",
            "\n",
            "Test: \n",
            " Accuracy: 86.9%, Avg loss: 0.356112 \n",
            "\n",
            "Epoch 6\n",
            "-------------------------------\n",
            "loss: 0.357902  [    0/50000]\n",
            "loss: 0.259472  [10000/50000]\n",
            "loss: 0.414213  [20000/50000]\n",
            "loss: 0.542807  [30000/50000]\n",
            "loss: 0.387557  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 88.0%, Avg loss: 0.331232 \n",
            "\n",
            "Test: \n",
            " Accuracy: 87.1%, Avg loss: 0.354745 \n",
            "\n",
            "Epoch 7\n",
            "-------------------------------\n",
            "loss: 0.356859  [    0/50000]\n",
            "loss: 0.235835  [10000/50000]\n",
            "loss: 0.383108  [20000/50000]\n",
            "loss: 0.477474  [30000/50000]\n",
            "loss: 0.377486  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 88.7%, Avg loss: 0.311883 \n",
            "\n",
            "Test: \n",
            " Accuracy: 87.5%, Avg loss: 0.343175 \n",
            "\n",
            "Epoch 8\n",
            "-------------------------------\n",
            "loss: 0.325207  [    0/50000]\n",
            "loss: 0.239394  [10000/50000]\n",
            "loss: 0.361841  [20000/50000]\n",
            "loss: 0.453579  [30000/50000]\n",
            "loss: 0.345630  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 89.2%, Avg loss: 0.297063 \n",
            "\n",
            "Test: \n",
            " Accuracy: 87.6%, Avg loss: 0.337178 \n",
            "\n",
            "Epoch 9\n",
            "-------------------------------\n",
            "loss: 0.291785  [    0/50000]\n",
            "loss: 0.227766  [10000/50000]\n",
            "loss: 0.320862  [20000/50000]\n",
            "loss: 0.419467  [30000/50000]\n",
            "loss: 0.310567  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 89.7%, Avg loss: 0.283324 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.0%, Avg loss: 0.333212 \n",
            "\n",
            "Epoch 10\n",
            "-------------------------------\n",
            "loss: 0.272954  [    0/50000]\n",
            "loss: 0.232155  [10000/50000]\n",
            "loss: 0.311404  [20000/50000]\n",
            "loss: 0.420107  [30000/50000]\n",
            "loss: 0.306037  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 90.1%, Avg loss: 0.271360 \n",
            "\n",
            "Test: \n",
            " Accuracy: 87.9%, Avg loss: 0.328944 \n",
            "\n",
            "Epoch 11\n",
            "-------------------------------\n",
            "loss: 0.250225  [    0/50000]\n",
            "loss: 0.231612  [10000/50000]\n",
            "loss: 0.297829  [20000/50000]\n",
            "loss: 0.391159  [30000/50000]\n",
            "loss: 0.282035  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 90.5%, Avg loss: 0.262227 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.2%, Avg loss: 0.325683 \n",
            "\n",
            "Epoch 12\n",
            "-------------------------------\n",
            "loss: 0.239978  [    0/50000]\n",
            "loss: 0.201812  [10000/50000]\n",
            "loss: 0.314870  [20000/50000]\n",
            "loss: 0.363202  [30000/50000]\n",
            "loss: 0.250473  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 90.7%, Avg loss: 0.254168 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.1%, Avg loss: 0.328223 \n",
            "\n",
            "Epoch 13\n",
            "-------------------------------\n",
            "loss: 0.241471  [    0/50000]\n",
            "loss: 0.215516  [10000/50000]\n",
            "loss: 0.294815  [20000/50000]\n",
            "loss: 0.365325  [30000/50000]\n",
            "loss: 0.246796  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 91.2%, Avg loss: 0.243846 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.4%, Avg loss: 0.323972 \n",
            "\n",
            "Epoch 14\n",
            "-------------------------------\n",
            "loss: 0.213628  [    0/50000]\n",
            "loss: 0.245447  [10000/50000]\n",
            "loss: 0.297971  [20000/50000]\n",
            "loss: 0.339061  [30000/50000]\n",
            "loss: 0.227159  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 91.3%, Avg loss: 0.236880 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.5%, Avg loss: 0.327256 \n",
            "\n",
            "Epoch 15\n",
            "-------------------------------\n",
            "loss: 0.224054  [    0/50000]\n",
            "loss: 0.213295  [10000/50000]\n",
            "loss: 0.258816  [20000/50000]\n",
            "loss: 0.309880  [30000/50000]\n",
            "loss: 0.196303  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 91.7%, Avg loss: 0.229218 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.5%, Avg loss: 0.326321 \n",
            "\n",
            "Epoch 16\n",
            "-------------------------------\n",
            "loss: 0.206505  [    0/50000]\n",
            "loss: 0.214421  [10000/50000]\n",
            "loss: 0.235696  [20000/50000]\n",
            "loss: 0.339923  [30000/50000]\n",
            "loss: 0.198663  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 92.0%, Avg loss: 0.222092 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.2%, Avg loss: 0.337287 \n",
            "\n",
            "Epoch 17\n",
            "-------------------------------\n",
            "loss: 0.192706  [    0/50000]\n",
            "loss: 0.209779  [10000/50000]\n",
            "loss: 0.238413  [20000/50000]\n",
            "loss: 0.318302  [30000/50000]\n",
            "loss: 0.206663  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 92.2%, Avg loss: 0.214002 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.5%, Avg loss: 0.333257 \n",
            "\n",
            "Epoch 18\n",
            "-------------------------------\n",
            "loss: 0.226926  [    0/50000]\n",
            "loss: 0.172666  [10000/50000]\n",
            "loss: 0.226868  [20000/50000]\n",
            "loss: 0.309923  [30000/50000]\n",
            "loss: 0.215501  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 92.4%, Avg loss: 0.207954 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.3%, Avg loss: 0.338520 \n",
            "\n",
            "Epoch 19\n",
            "-------------------------------\n",
            "loss: 0.207584  [    0/50000]\n",
            "loss: 0.161424  [10000/50000]\n",
            "loss: 0.204951  [20000/50000]\n",
            "loss: 0.271142  [30000/50000]\n",
            "loss: 0.167864  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 92.7%, Avg loss: 0.202459 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.2%, Avg loss: 0.341346 \n",
            "\n",
            "Epoch 20\n",
            "-------------------------------\n",
            "loss: 0.205169  [    0/50000]\n",
            "loss: 0.186967  [10000/50000]\n",
            "loss: 0.221061  [20000/50000]\n",
            "loss: 0.262820  [30000/50000]\n",
            "loss: 0.186823  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 92.8%, Avg loss: 0.197014 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.4%, Avg loss: 0.343547 \n",
            "\n",
            "Done!\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 학습과정 살펴보기\n",
        "fig, loss_ax = plt.subplots()\n",
        "\n",
        "acc_ax = loss_ax.twinx()\n",
        "\n",
        "loss_ax.plot(loss, 'y', label='train loss')\n",
        "loss_ax.plot(val_loss, 'r', label='val loss')\n",
        "loss_ax.set_ylim([0.0, 1.5])\n",
        "\n",
        "acc_ax.plot(accuracy, 'b', label='train acc')\n",
        "acc_ax.plot(val_accuracy, 'g', label='val acc')\n",
        "acc_ax.set_ylim([50, 100])\n",
        "\n",
        "loss_ax.set_xlabel('epoch')\n",
        "loss_ax.set_ylabel('loss')\n",
        "acc_ax.set_ylabel('accuray')\n",
        "\n",
        "loss_ax.legend(loc='upper left')\n",
        "acc_ax.legend(loc='lower left')\n",
        "\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 683
        },
        "id": "9qe-sO0oCFND",
        "outputId": "9ce75d89-08d4-45d7-d6ec-940a16d2ead9"
      },
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cC6n5edlFocR",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a74a7d2c-5898-488b-8fd3-342b519a5069"
      },
      "source": [
        "# Final evaluation of the model\n",
        "test_loss, test_accuracy = test(test_dataloader, model, loss_fn)\n",
        "print(\"Baseline Error: %.2f%%\" % (100-test_accuracy))"
      ],
      "execution_count": 63,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test: \n",
            " Accuracy: 87.4%, Avg loss: 0.381720 \n",
            "\n",
            "Baseline Error: 12.63%\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LLTnR7uKJm5m"
      },
      "source": [
        "![대체 텍스트](https://miro.medium.com/max/1200/1*iWQzxhVlvadk6VAJjsgXgg.png)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class DeepLayerPerceptronReluDropout(nn.Module):\n",
        "  def __init__(self):\n",
        "    super(DeepLayerPerceptronReluDropout, self).__init__()\n",
        "    self.flatten = nn.Flatten()\n",
        "    self.mlp = nn.Sequential(\n",
        "        nn.Linear(num_pixels, 256),\n",
        "        nn.ReLU(),\n",
        "        nn.Dropout(p=0.2),\n",
        "        nn.Linear(256, 256),\n",
        "        nn.ReLU(),\n",
        "        nn.Dropout(p=0.2),\n",
        "        nn.Linear(256, 256),\n",
        "        nn.ReLU(),\n",
        "        nn.Dropout(p=0.2),\n",
        "        nn.Linear(256, 256),\n",
        "        nn.ReLU(),\n",
        "        nn.Dropout(p=0.2),\n",
        "        nn.Linear(256, 256),\n",
        "        nn.ReLU(),\n",
        "        nn.Dropout(p=0.2),\n",
        "        nn.Linear(256, num_classes),\n",
        "    )\n",
        "  def forward(self, x):\n",
        "    x = self.flatten(x)\n",
        "    logits = self.mlp(x)\n",
        "    return logits"
      ],
      "metadata": {
        "id": "rCKCt3dayGjR"
      },
      "execution_count": 64,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "58BeovzhJ59T",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "98ac4625-fc03-4e48-ea02-f7f0fe477a1b"
      },
      "source": [
        "# build the model\n",
        "model = DeepLayerPerceptronReluDropout().to(device)\n",
        "summary(model, input_size=(1, 28, 28))\n",
        "\n",
        "loss_fn = nn.CrossEntropyLoss()\n",
        "optimizer = torch.optim.SGD(model.parameters(), lr=0.01, weight_decay=1e-6, momentum=0.9, nesterov=True)\n",
        "\n",
        "# fix random seed for reproductibility\n",
        "seed = 7\n",
        "np.random.seed(seed)\n",
        "\n",
        "# Fit the model\n",
        "epochs = 20\n",
        "\n",
        "loss = []\n",
        "accuracy = []\n",
        "val_loss = []\n",
        "val_accuracy = []\n",
        "\n",
        "for t in range(epochs):\n",
        "    print(f\"Epoch {t+1}\\n-------------------------------\")\n",
        "    tmp_loss, tmp_accuracy = train(train_dataloader, model, loss_fn, optimizer)\n",
        "    loss.append(tmp_loss)\n",
        "    accuracy.append(tmp_accuracy)\n",
        "    tmp_loss, tmp_accuracy = test(val_dataloader, model, loss_fn)\n",
        "    val_loss.append(tmp_loss)\n",
        "    val_accuracy.append(tmp_accuracy)\n",
        "print(\"Done!\")"
      ],
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "----------------------------------------------------------------\n",
            "        Layer (type)               Output Shape         Param #\n",
            "================================================================\n",
            "           Flatten-1                  [-1, 784]               0\n",
            "            Linear-2                  [-1, 256]         200,960\n",
            "              ReLU-3                  [-1, 256]               0\n",
            "           Dropout-4                  [-1, 256]               0\n",
            "            Linear-5                  [-1, 256]          65,792\n",
            "              ReLU-6                  [-1, 256]               0\n",
            "           Dropout-7                  [-1, 256]               0\n",
            "            Linear-8                  [-1, 256]          65,792\n",
            "              ReLU-9                  [-1, 256]               0\n",
            "          Dropout-10                  [-1, 256]               0\n",
            "           Linear-11                  [-1, 256]          65,792\n",
            "             ReLU-12                  [-1, 256]               0\n",
            "          Dropout-13                  [-1, 256]               0\n",
            "           Linear-14                  [-1, 256]          65,792\n",
            "             ReLU-15                  [-1, 256]               0\n",
            "          Dropout-16                  [-1, 256]               0\n",
            "           Linear-17                   [-1, 10]           2,570\n",
            "================================================================\n",
            "Total params: 466,698\n",
            "Trainable params: 466,698\n",
            "Non-trainable params: 0\n",
            "----------------------------------------------------------------\n",
            "Input size (MB): 0.00\n",
            "Forward/backward pass size (MB): 0.04\n",
            "Params size (MB): 1.78\n",
            "Estimated Total Size (MB): 1.82\n",
            "----------------------------------------------------------------\n",
            "Epoch 1\n",
            "-------------------------------\n",
            "loss: 2.300949  [    0/50000]\n",
            "loss: 2.297657  [10000/50000]\n",
            "loss: 2.274653  [20000/50000]\n",
            "loss: 1.375157  [30000/50000]\n",
            "loss: 1.154408  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 31.9%, Avg loss: 1.750743 \n",
            "\n",
            "Test: \n",
            " Accuracy: 57.4%, Avg loss: 0.925300 \n",
            "\n",
            "Epoch 2\n",
            "-------------------------------\n",
            "loss: 0.911410  [    0/50000]\n",
            "loss: 0.801602  [10000/50000]\n",
            "loss: 0.803188  [20000/50000]\n",
            "loss: 0.894722  [30000/50000]\n",
            "loss: 0.782360  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 70.2%, Avg loss: 0.781624 \n",
            "\n",
            "Test: \n",
            " Accuracy: 78.2%, Avg loss: 0.588764 \n",
            "\n",
            "Epoch 3\n",
            "-------------------------------\n",
            "loss: 0.612095  [    0/50000]\n",
            "loss: 0.514158  [10000/50000]\n",
            "loss: 0.582199  [20000/50000]\n",
            "loss: 0.829076  [30000/50000]\n",
            "loss: 0.616085  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 79.6%, Avg loss: 0.562375 \n",
            "\n",
            "Test: \n",
            " Accuracy: 82.7%, Avg loss: 0.481901 \n",
            "\n",
            "Epoch 4\n",
            "-------------------------------\n",
            "loss: 0.472128  [    0/50000]\n",
            "loss: 0.424991  [10000/50000]\n",
            "loss: 0.440354  [20000/50000]\n",
            "loss: 0.745201  [30000/50000]\n",
            "loss: 0.517627  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 82.8%, Avg loss: 0.481258 \n",
            "\n",
            "Test: \n",
            " Accuracy: 84.6%, Avg loss: 0.424541 \n",
            "\n",
            "Epoch 5\n",
            "-------------------------------\n",
            "loss: 0.371859  [    0/50000]\n",
            "loss: 0.374173  [10000/50000]\n",
            "loss: 0.437518  [20000/50000]\n",
            "loss: 0.627002  [30000/50000]\n",
            "loss: 0.500847  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 84.4%, Avg loss: 0.437904 \n",
            "\n",
            "Test: \n",
            " Accuracy: 85.9%, Avg loss: 0.386493 \n",
            "\n",
            "Epoch 6\n",
            "-------------------------------\n",
            "loss: 0.469838  [    0/50000]\n",
            "loss: 0.333795  [10000/50000]\n",
            "loss: 0.372714  [20000/50000]\n",
            "loss: 0.671073  [30000/50000]\n",
            "loss: 0.462489  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 85.4%, Avg loss: 0.407995 \n",
            "\n",
            "Test: \n",
            " Accuracy: 86.4%, Avg loss: 0.378063 \n",
            "\n",
            "Epoch 7\n",
            "-------------------------------\n",
            "loss: 0.413447  [    0/50000]\n",
            "loss: 0.318850  [10000/50000]\n",
            "loss: 0.372076  [20000/50000]\n",
            "loss: 0.531187  [30000/50000]\n",
            "loss: 0.395649  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.1%, Avg loss: 0.387561 \n",
            "\n",
            "Test: \n",
            " Accuracy: 87.0%, Avg loss: 0.357356 \n",
            "\n",
            "Epoch 8\n",
            "-------------------------------\n",
            "loss: 0.400830  [    0/50000]\n",
            "loss: 0.291232  [10000/50000]\n",
            "loss: 0.399798  [20000/50000]\n",
            "loss: 0.659357  [30000/50000]\n",
            "loss: 0.399702  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.7%, Avg loss: 0.370461 \n",
            "\n",
            "Test: \n",
            " Accuracy: 87.3%, Avg loss: 0.350125 \n",
            "\n",
            "Epoch 9\n",
            "-------------------------------\n",
            "loss: 0.352475  [    0/50000]\n",
            "loss: 0.312454  [10000/50000]\n",
            "loss: 0.361131  [20000/50000]\n",
            "loss: 0.519776  [30000/50000]\n",
            "loss: 0.416627  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 87.2%, Avg loss: 0.358772 \n",
            "\n",
            "Test: \n",
            " Accuracy: 87.6%, Avg loss: 0.343042 \n",
            "\n",
            "Epoch 10\n",
            "-------------------------------\n",
            "loss: 0.402635  [    0/50000]\n",
            "loss: 0.275495  [10000/50000]\n",
            "loss: 0.351298  [20000/50000]\n",
            "loss: 0.543483  [30000/50000]\n",
            "loss: 0.373735  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 87.7%, Avg loss: 0.346683 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.2%, Avg loss: 0.330941 \n",
            "\n",
            "Epoch 11\n",
            "-------------------------------\n",
            "loss: 0.385247  [    0/50000]\n",
            "loss: 0.265827  [10000/50000]\n",
            "loss: 0.345082  [20000/50000]\n",
            "loss: 0.536066  [30000/50000]\n",
            "loss: 0.364821  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 87.9%, Avg loss: 0.336362 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.2%, Avg loss: 0.329227 \n",
            "\n",
            "Epoch 12\n",
            "-------------------------------\n",
            "loss: 0.359950  [    0/50000]\n",
            "loss: 0.284588  [10000/50000]\n",
            "loss: 0.315086  [20000/50000]\n",
            "loss: 0.434352  [30000/50000]\n",
            "loss: 0.317405  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 88.4%, Avg loss: 0.327478 \n",
            "\n",
            "Test: \n",
            " Accuracy: 87.8%, Avg loss: 0.331257 \n",
            "\n",
            "Epoch 13\n",
            "-------------------------------\n",
            "loss: 0.317416  [    0/50000]\n",
            "loss: 0.212102  [10000/50000]\n",
            "loss: 0.344048  [20000/50000]\n",
            "loss: 0.543678  [30000/50000]\n",
            "loss: 0.349522  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 88.4%, Avg loss: 0.321067 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.4%, Avg loss: 0.320653 \n",
            "\n",
            "Epoch 14\n",
            "-------------------------------\n",
            "loss: 0.275492  [    0/50000]\n",
            "loss: 0.298486  [10000/50000]\n",
            "loss: 0.281575  [20000/50000]\n",
            "loss: 0.540925  [30000/50000]\n",
            "loss: 0.273651  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 88.7%, Avg loss: 0.314735 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.2%, Avg loss: 0.328713 \n",
            "\n",
            "Epoch 15\n",
            "-------------------------------\n",
            "loss: 0.260377  [    0/50000]\n",
            "loss: 0.244092  [10000/50000]\n",
            "loss: 0.336014  [20000/50000]\n",
            "loss: 0.515845  [30000/50000]\n",
            "loss: 0.334345  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 88.9%, Avg loss: 0.307530 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.1%, Avg loss: 0.334570 \n",
            "\n",
            "Epoch 16\n",
            "-------------------------------\n",
            "loss: 0.297416  [    0/50000]\n",
            "loss: 0.238936  [10000/50000]\n",
            "loss: 0.266938  [20000/50000]\n",
            "loss: 0.371743  [30000/50000]\n",
            "loss: 0.340087  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 89.2%, Avg loss: 0.299043 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.7%, Avg loss: 0.317451 \n",
            "\n",
            "Epoch 17\n",
            "-------------------------------\n",
            "loss: 0.273187  [    0/50000]\n",
            "loss: 0.266300  [10000/50000]\n",
            "loss: 0.258151  [20000/50000]\n",
            "loss: 0.514752  [30000/50000]\n",
            "loss: 0.288436  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 89.3%, Avg loss: 0.296050 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.5%, Avg loss: 0.313240 \n",
            "\n",
            "Epoch 18\n",
            "-------------------------------\n",
            "loss: 0.330480  [    0/50000]\n",
            "loss: 0.254096  [10000/50000]\n",
            "loss: 0.294427  [20000/50000]\n",
            "loss: 0.449022  [30000/50000]\n",
            "loss: 0.316759  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 89.4%, Avg loss: 0.291019 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.5%, Avg loss: 0.321146 \n",
            "\n",
            "Epoch 19\n",
            "-------------------------------\n",
            "loss: 0.246383  [    0/50000]\n",
            "loss: 0.237124  [10000/50000]\n",
            "loss: 0.271437  [20000/50000]\n",
            "loss: 0.462654  [30000/50000]\n",
            "loss: 0.331273  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 89.7%, Avg loss: 0.282028 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.9%, Avg loss: 0.314373 \n",
            "\n",
            "Epoch 20\n",
            "-------------------------------\n",
            "loss: 0.270035  [    0/50000]\n",
            "loss: 0.178067  [10000/50000]\n",
            "loss: 0.281796  [20000/50000]\n",
            "loss: 0.470827  [30000/50000]\n",
            "loss: 0.311779  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 89.9%, Avg loss: 0.276957 \n",
            "\n",
            "Test: \n",
            " Accuracy: 89.2%, Avg loss: 0.306849 \n",
            "\n",
            "Done!\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 학습과정 살펴보기\n",
        "fig, loss_ax = plt.subplots()\n",
        "\n",
        "acc_ax = loss_ax.twinx()\n",
        "\n",
        "loss_ax.plot(loss, 'y', label='train loss')\n",
        "loss_ax.plot(val_loss, 'r', label='val loss')\n",
        "loss_ax.set_ylim([0.0, 1.5])\n",
        "\n",
        "acc_ax.plot(accuracy, 'b', label='train acc')\n",
        "acc_ax.plot(val_accuracy, 'g', label='val acc')\n",
        "acc_ax.set_ylim([50, 100])\n",
        "\n",
        "loss_ax.set_xlabel('epoch')\n",
        "loss_ax.set_ylabel('loss')\n",
        "acc_ax.set_ylabel('accuray')\n",
        "\n",
        "loss_ax.legend(loc='upper left')\n",
        "acc_ax.legend(loc='lower left')\n",
        "\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 683
        },
        "id": "JU-hmr8PzHRD",
        "outputId": "9c1e4840-a0e0-425f-82f0-f41c262839e3"
      },
      "execution_count": 66,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6cIJuuuENAPP",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6383fff1-d06b-403d-9db2-031d6989c5d5"
      },
      "source": [
        "# Final evaluation of the model\n",
        "test_loss, test_accuracy = test(test_dataloader, model, loss_fn)\n",
        "print(\"Baseline Error: %.2f%%\" % (100-test_accuracy))"
      ],
      "execution_count": 67,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test: \n",
            " Accuracy: 88.3%, Avg loss: 0.338737 \n",
            "\n",
            "Baseline Error: 11.72%\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "E955lLUboIsU"
      },
      "source": [
        "#STEP 11: 네번째 인공지능 모델 (CNN)\n",
        "\n",
        "![대체 텍스트](https://www.mdpi.com/entropy/entropy-19-00242/article_deploy/html/images/entropy-19-00242-g001.png)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CybrZ58MpVGW"
      },
      "source": [
        "\n",
        "# 중요! 입력데이터의 형태가 바뀌어야 한다. \n",
        "# 784 (1D) -> 28x28 (2D)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class SimpleCNN(nn.Module):\n",
        "    def __init__(self):\n",
        "        super(SimpleCNN, self).__init__()\n",
        "        self.conv = nn.Conv2d(1, 32, 5)\n",
        "        self.relu = nn.ReLU()\n",
        "        self.pool = nn.MaxPool2d(2, 2)\n",
        "        self.flatten = nn.Flatten()\n",
        "        self.fc1 = nn.Linear(32*12*12, 128)\n",
        "        self.fc2 = nn.Linear(128, num_classes)\n",
        "\n",
        "    def forward(self, x):\n",
        "        x = self.conv(x)\n",
        "        x = self.relu(x)\n",
        "        x = self.pool(x)\n",
        "        x = self.flatten(x)\n",
        "        x = self.fc1(x)\n",
        "        x = self.fc2(x)\n",
        "        return x\n"
      ],
      "metadata": {
        "id": "6l9_602fzhyS"
      },
      "execution_count": 72,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# build the model\n",
        "model = SimpleCNN().to(device)\n",
        "summary(model, input_size=(1, 28, 28))\n",
        "\n",
        "loss_fn = nn.CrossEntropyLoss()\n",
        "optimizer = torch.optim.SGD(model.parameters(), lr=0.01, weight_decay=1e-6, momentum=0.9, nesterov=True)\n",
        "\n",
        "# fix random seed for reproductibility\n",
        "seed = 7\n",
        "np.random.seed(seed)\n",
        "\n",
        "# Fit the model\n",
        "epochs = 20\n",
        "\n",
        "loss = []\n",
        "accuracy = []\n",
        "val_loss = []\n",
        "val_accuracy = []\n",
        "\n",
        "for t in range(epochs):\n",
        "    print(f\"Epoch {t+1}\\n-------------------------------\")\n",
        "    tmp_loss, tmp_accuracy = train(train_dataloader, model, loss_fn, optimizer)\n",
        "    loss.append(tmp_loss)\n",
        "    accuracy.append(tmp_accuracy)\n",
        "    tmp_loss, tmp_accuracy = test(val_dataloader, model, loss_fn)\n",
        "    val_loss.append(tmp_loss)\n",
        "    val_accuracy.append(tmp_accuracy)\n",
        "print(\"Done!\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JnW5syyDzhzL",
        "outputId": "584467d1-9b4a-46e7-d271-dc5f20bd44e0"
      },
      "execution_count": 73,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "----------------------------------------------------------------\n",
            "        Layer (type)               Output Shape         Param #\n",
            "================================================================\n",
            "            Conv2d-1           [-1, 32, 24, 24]             832\n",
            "              ReLU-2           [-1, 32, 24, 24]               0\n",
            "         MaxPool2d-3           [-1, 32, 12, 12]               0\n",
            "           Flatten-4                 [-1, 4608]               0\n",
            "            Linear-5                  [-1, 128]         589,952\n",
            "            Linear-6                   [-1, 10]           1,290\n",
            "================================================================\n",
            "Total params: 592,074\n",
            "Trainable params: 592,074\n",
            "Non-trainable params: 0\n",
            "----------------------------------------------------------------\n",
            "Input size (MB): 0.00\n",
            "Forward/backward pass size (MB): 0.35\n",
            "Params size (MB): 2.26\n",
            "Estimated Total Size (MB): 2.61\n",
            "----------------------------------------------------------------\n",
            "Epoch 1\n",
            "-------------------------------\n",
            "loss: 2.301720  [    0/50000]\n",
            "loss: 0.527479  [10000/50000]\n",
            "loss: 0.605052  [20000/50000]\n",
            "loss: 0.716180  [30000/50000]\n",
            "loss: 0.460028  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 79.2%, Avg loss: 0.593005 \n",
            "\n",
            "Test: \n",
            " Accuracy: 84.7%, Avg loss: 0.436911 \n",
            "\n",
            "Epoch 2\n",
            "-------------------------------\n",
            "loss: 0.459076  [    0/50000]\n",
            "loss: 0.301469  [10000/50000]\n",
            "loss: 0.463349  [20000/50000]\n",
            "loss: 0.586254  [30000/50000]\n",
            "loss: 0.383935  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.3%, Avg loss: 0.391695 \n",
            "\n",
            "Test: \n",
            " Accuracy: 87.2%, Avg loss: 0.366721 \n",
            "\n",
            "Epoch 3\n",
            "-------------------------------\n",
            "loss: 0.402461  [    0/50000]\n",
            "loss: 0.249696  [10000/50000]\n",
            "loss: 0.397686  [20000/50000]\n",
            "loss: 0.516293  [30000/50000]\n",
            "loss: 0.337064  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 88.2%, Avg loss: 0.337640 \n",
            "\n",
            "Test: \n",
            " Accuracy: 87.9%, Avg loss: 0.337780 \n",
            "\n",
            "Epoch 4\n",
            "-------------------------------\n",
            "loss: 0.369598  [    0/50000]\n",
            "loss: 0.231472  [10000/50000]\n",
            "loss: 0.346154  [20000/50000]\n",
            "loss: 0.465576  [30000/50000]\n",
            "loss: 0.305972  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 89.2%, Avg loss: 0.307076 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.6%, Avg loss: 0.320332 \n",
            "\n",
            "Epoch 5\n",
            "-------------------------------\n",
            "loss: 0.342684  [    0/50000]\n",
            "loss: 0.216271  [10000/50000]\n",
            "loss: 0.314059  [20000/50000]\n",
            "loss: 0.425623  [30000/50000]\n",
            "loss: 0.282956  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 89.9%, Avg loss: 0.285878 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.9%, Avg loss: 0.308988 \n",
            "\n",
            "Epoch 6\n",
            "-------------------------------\n",
            "loss: 0.324991  [    0/50000]\n",
            "loss: 0.201692  [10000/50000]\n",
            "loss: 0.291945  [20000/50000]\n",
            "loss: 0.394233  [30000/50000]\n",
            "loss: 0.265956  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 90.5%, Avg loss: 0.269569 \n",
            "\n",
            "Test: \n",
            " Accuracy: 89.1%, Avg loss: 0.300732 \n",
            "\n",
            "Epoch 7\n",
            "-------------------------------\n",
            "loss: 0.312775  [    0/50000]\n",
            "loss: 0.188104  [10000/50000]\n",
            "loss: 0.274125  [20000/50000]\n",
            "loss: 0.369160  [30000/50000]\n",
            "loss: 0.251655  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 90.9%, Avg loss: 0.256238 \n",
            "\n",
            "Test: \n",
            " Accuracy: 89.5%, Avg loss: 0.293816 \n",
            "\n",
            "Epoch 8\n",
            "-------------------------------\n",
            "loss: 0.300898  [    0/50000]\n",
            "loss: 0.178765  [10000/50000]\n",
            "loss: 0.263987  [20000/50000]\n",
            "loss: 0.346394  [30000/50000]\n",
            "loss: 0.239299  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 91.3%, Avg loss: 0.244846 \n",
            "\n",
            "Test: \n",
            " Accuracy: 89.5%, Avg loss: 0.289622 \n",
            "\n",
            "Epoch 9\n",
            "-------------------------------\n",
            "loss: 0.292997  [    0/50000]\n",
            "loss: 0.170056  [10000/50000]\n",
            "loss: 0.253217  [20000/50000]\n",
            "loss: 0.329275  [30000/50000]\n",
            "loss: 0.228293  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 91.7%, Avg loss: 0.234760 \n",
            "\n",
            "Test: \n",
            " Accuracy: 89.7%, Avg loss: 0.284912 \n",
            "\n",
            "Epoch 10\n",
            "-------------------------------\n",
            "loss: 0.285673  [    0/50000]\n",
            "loss: 0.163089  [10000/50000]\n",
            "loss: 0.248634  [20000/50000]\n",
            "loss: 0.313991  [30000/50000]\n",
            "loss: 0.221315  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 92.0%, Avg loss: 0.225670 \n",
            "\n",
            "Test: \n",
            " Accuracy: 89.8%, Avg loss: 0.282353 \n",
            "\n",
            "Epoch 11\n",
            "-------------------------------\n",
            "loss: 0.277129  [    0/50000]\n",
            "loss: 0.158212  [10000/50000]\n",
            "loss: 0.242844  [20000/50000]\n",
            "loss: 0.297073  [30000/50000]\n",
            "loss: 0.212640  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 92.3%, Avg loss: 0.217542 \n",
            "\n",
            "Test: \n",
            " Accuracy: 89.9%, Avg loss: 0.281454 \n",
            "\n",
            "Epoch 12\n",
            "-------------------------------\n",
            "loss: 0.270648  [    0/50000]\n",
            "loss: 0.153482  [10000/50000]\n",
            "loss: 0.240098  [20000/50000]\n",
            "loss: 0.278293  [30000/50000]\n",
            "loss: 0.205884  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 92.5%, Avg loss: 0.209888 \n",
            "\n",
            "Test: \n",
            " Accuracy: 90.1%, Avg loss: 0.279481 \n",
            "\n",
            "Epoch 13\n",
            "-------------------------------\n",
            "loss: 0.262443  [    0/50000]\n",
            "loss: 0.148968  [10000/50000]\n",
            "loss: 0.238434  [20000/50000]\n",
            "loss: 0.266412  [30000/50000]\n",
            "loss: 0.196440  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 92.8%, Avg loss: 0.202857 \n",
            "\n",
            "Test: \n",
            " Accuracy: 90.0%, Avg loss: 0.280398 \n",
            "\n",
            "Epoch 14\n",
            "-------------------------------\n",
            "loss: 0.258589  [    0/50000]\n",
            "loss: 0.144371  [10000/50000]\n",
            "loss: 0.235004  [20000/50000]\n",
            "loss: 0.251142  [30000/50000]\n",
            "loss: 0.191249  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 93.0%, Avg loss: 0.196194 \n",
            "\n",
            "Test: \n",
            " Accuracy: 90.0%, Avg loss: 0.280643 \n",
            "\n",
            "Epoch 15\n",
            "-------------------------------\n",
            "loss: 0.251428  [    0/50000]\n",
            "loss: 0.141170  [10000/50000]\n",
            "loss: 0.232149  [20000/50000]\n",
            "loss: 0.239827  [30000/50000]\n",
            "loss: 0.190049  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 93.3%, Avg loss: 0.189917 \n",
            "\n",
            "Test: \n",
            " Accuracy: 89.9%, Avg loss: 0.282311 \n",
            "\n",
            "Epoch 16\n",
            "-------------------------------\n",
            "loss: 0.243258  [    0/50000]\n",
            "loss: 0.137305  [10000/50000]\n",
            "loss: 0.228770  [20000/50000]\n",
            "loss: 0.231785  [30000/50000]\n",
            "loss: 0.178222  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 93.5%, Avg loss: 0.184075 \n",
            "\n",
            "Test: \n",
            " Accuracy: 89.9%, Avg loss: 0.284134 \n",
            "\n",
            "Epoch 17\n",
            "-------------------------------\n",
            "loss: 0.241402  [    0/50000]\n",
            "loss: 0.133020  [10000/50000]\n",
            "loss: 0.225559  [20000/50000]\n",
            "loss: 0.220060  [30000/50000]\n",
            "loss: 0.174882  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 93.7%, Avg loss: 0.178408 \n",
            "\n",
            "Test: \n",
            " Accuracy: 89.9%, Avg loss: 0.287707 \n",
            "\n",
            "Epoch 18\n",
            "-------------------------------\n",
            "loss: 0.238775  [    0/50000]\n",
            "loss: 0.131037  [10000/50000]\n",
            "loss: 0.223450  [20000/50000]\n",
            "loss: 0.217827  [30000/50000]\n",
            "loss: 0.170557  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 93.8%, Avg loss: 0.173224 \n",
            "\n",
            "Test: \n",
            " Accuracy: 89.8%, Avg loss: 0.291004 \n",
            "\n",
            "Epoch 19\n",
            "-------------------------------\n",
            "loss: 0.234683  [    0/50000]\n",
            "loss: 0.130660  [10000/50000]\n",
            "loss: 0.221119  [20000/50000]\n",
            "loss: 0.205463  [30000/50000]\n",
            "loss: 0.161142  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 94.0%, Avg loss: 0.168247 \n",
            "\n",
            "Test: \n",
            " Accuracy: 89.9%, Avg loss: 0.291800 \n",
            "\n",
            "Epoch 20\n",
            "-------------------------------\n",
            "loss: 0.227507  [    0/50000]\n",
            "loss: 0.123374  [10000/50000]\n",
            "loss: 0.217563  [20000/50000]\n",
            "loss: 0.199315  [30000/50000]\n",
            "loss: 0.158173  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 94.2%, Avg loss: 0.163267 \n",
            "\n",
            "Test: \n",
            " Accuracy: 89.7%, Avg loss: 0.296466 \n",
            "\n",
            "Done!\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7Y1YC0c-Foca",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 683
        },
        "outputId": "b13d366b-23d6-4112-b619-8f9ec2ebcd63"
      },
      "source": [
        "# 학습과정 살펴보기\n",
        "fig, loss_ax = plt.subplots()\n",
        "\n",
        "acc_ax = loss_ax.twinx()\n",
        "\n",
        "loss_ax.plot(loss, 'y', label='train loss')\n",
        "loss_ax.plot(val_loss, 'r', label='val loss')\n",
        "loss_ax.set_ylim([0.0, 1.5])\n",
        "\n",
        "acc_ax.plot(accuracy, 'b', label='train acc')\n",
        "acc_ax.plot(val_accuracy, 'g', label='val acc')\n",
        "acc_ax.set_ylim([50, 100])\n",
        "\n",
        "loss_ax.set_xlabel('epoch')\n",
        "loss_ax.set_ylabel('loss')\n",
        "acc_ax.set_ylabel('accuray')\n",
        "\n",
        "loss_ax.legend(loc='upper left')\n",
        "acc_ax.legend(loc='lower left')\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 74,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ttpPwLW4Focd",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b2486641-0782-4c65-d782-5ec9e20afef5"
      },
      "source": [
        "# Final evaluation of the model\n",
        "test_loss, test_accuracy = test(test_dataloader, model, loss_fn)\n",
        "print(\"Baseline Error: %.2f%%\" % (100-test_accuracy))"
      ],
      "execution_count": 75,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test: \n",
            " Accuracy: 89.7%, Avg loss: 0.309048 \n",
            "\n",
            "Baseline Error: 10.32%\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-ae0K65Z1IQi"
      },
      "source": [
        "#STEP 12: Convolution kernel 살펴보기 (5x5)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Uf5fPXX_Focf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 733
        },
        "outputId": "217a3baf-dd0a-4d30-ea0a-d6f296a368a9"
      },
      "source": [
        "W1 = model.conv.weight.cpu().detach().numpy()\n",
        "\n",
        "print(W1.shape)\n",
        "\n",
        "plt.figure(figsize=(15, 15), frameon=False)\n",
        "for ind, val in enumerate(W1):\n",
        "    plt.subplot(6, 6, ind + 1)\n",
        "    im = val.reshape((5,5))\n",
        "    plt.axis(\"off\")\n",
        "    plt.imshow(im, cmap='gray',interpolation='nearest')"
      ],
      "execution_count": 82,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(32, 1, 5, 5)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 32 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.conv.weight"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "l2Auzzl883pk",
        "outputId": "306ec7c0-8b6b-4809-f2b2-e2d88b649086"
      },
      "execution_count": 78,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Parameter containing:\n",
              "tensor([[[[-0.0578,  0.2653,  0.5182, -0.1872, -0.3687],\n",
              "          [ 0.3108,  0.4272, -0.0391, -0.2757, -0.2675],\n",
              "          [-0.0685,  0.3351, -0.3866, -0.2090,  0.0220],\n",
              "          [-0.5011, -0.1520, -0.1636,  0.3007,  0.3176],\n",
              "          [-0.4755, -0.2992,  0.0471,  0.5091,  0.2386]]],\n",
              "\n",
              "\n",
              "        [[[ 0.1846, -0.0679, -0.4194,  0.4051,  0.1470],\n",
              "          [-0.0573,  0.0236, -0.4767, -0.0076,  0.5851],\n",
              "          [ 0.0744,  0.2357, -0.4332,  0.0709,  0.1098],\n",
              "          [ 0.0828,  0.1804, -0.1101, -0.4853,  0.4314],\n",
              "          [-0.1997,  0.4782, -0.2237, -0.2349,  0.0054]]],\n",
              "\n",
              "\n",
              "        [[[ 0.3658,  0.1397, -0.0230, -0.2624, -0.1941],\n",
              "          [ 0.3168,  0.2284, -0.0156,  0.1036, -0.1220],\n",
              "          [-0.1300,  0.1320,  0.1425, -0.2209, -0.0899],\n",
              "          [-0.2681, -0.2553, -0.0766, -0.0893, -0.1168],\n",
              "          [ 0.0689,  0.2483,  0.0964, -0.1428, -0.1338]]],\n",
              "\n",
              "\n",
              "        [[[ 0.2571, -0.6671,  0.3897,  0.3211, -0.1028],\n",
              "          [-0.2150, -0.4837,  0.9161, -0.4328, -0.3603],\n",
              "          [-0.5442,  0.2895,  0.4143, -0.4500,  0.3450],\n",
              "          [ 0.2323,  0.4024, -0.1867, -0.3836,  0.2656],\n",
              "          [ 0.2888,  0.1150, -0.0339, -0.0548,  0.0375]]],\n",
              "\n",
              "\n",
              "        [[[ 0.8341, -0.4145, -0.3214, -0.2224, -0.3280],\n",
              "          [ 0.0870,  0.6606, -0.2118, -0.1925,  0.2262],\n",
              "          [-0.4920,  0.2643,  0.3068, -0.0746,  0.1167],\n",
              "          [-0.3119,  0.2050,  0.2075, -0.3282,  0.2341],\n",
              "          [ 0.0741, -0.2632,  0.1464, -0.0545, -0.0651]]],\n",
              "\n",
              "\n",
              "        [[[ 0.1900, -0.2921, -0.1027,  0.2450, -0.2171],\n",
              "          [ 0.2561, -0.2586, -0.1393, -0.3877, -0.0676],\n",
              "          [-0.3953, -0.4073, -0.0806,  0.3744,  0.6714],\n",
              "          [-0.2217,  0.3465,  0.6474,  0.1851,  0.2217],\n",
              "          [-0.0226,  0.5146, -0.1573, -0.2268, -0.5261]]],\n",
              "\n",
              "\n",
              "        [[[ 0.2552,  0.1400,  0.2688, -0.1165, -0.3447],\n",
              "          [-0.1703, -0.0557,  0.4424,  0.1630, -0.4593],\n",
              "          [-0.3040, -0.0041, -0.0295,  0.5483, -0.2580],\n",
              "          [-0.2792,  0.1828, -0.5016,  0.4798,  0.0931],\n",
              "          [ 0.0526,  0.3159, -0.5660, -0.2462,  0.4316]]],\n",
              "\n",
              "\n",
              "        [[[-0.5501,  0.2643, -0.2497,  0.2634,  0.2962],\n",
              "          [ 0.1270,  0.3770, -0.2282, -0.4267,  0.1556],\n",
              "          [-0.2981, -0.0027,  0.4592, -0.1944,  0.0700],\n",
              "          [ 0.1553,  0.2796, -0.3350, -0.5278,  0.4173],\n",
              "          [ 0.4356,  0.2373, -0.3250, -0.3321, -0.0408]]],\n",
              "\n",
              "\n",
              "        [[[-0.1622, -0.0317, -0.1917,  0.2179, -0.1087],\n",
              "          [-0.2941, -0.3837, -0.0223,  0.3617, -0.1351],\n",
              "          [-0.0222, -0.0920,  0.0305, -0.0044,  0.0964],\n",
              "          [ 0.0167,  0.2784, -0.1132, -0.0777,  0.0278],\n",
              "          [ 0.0013,  0.3309,  0.1355,  0.0976, -0.1781]]],\n",
              "\n",
              "\n",
              "        [[[ 0.3030,  0.3018,  0.5431,  0.6499,  0.1341],\n",
              "          [ 0.3047,  0.2190, -0.2499, -0.1305, -0.4221],\n",
              "          [-0.2588,  0.1086, -0.5085, -0.0914, -0.4210],\n",
              "          [-0.1871, -0.2415, -0.6024, -0.4341,  0.1484],\n",
              "          [-0.1889,  0.3108,  0.2392,  0.0866,  0.6497]]],\n",
              "\n",
              "\n",
              "        [[[-0.1543, -0.3369, -0.2616, -0.4378, -0.3308],\n",
              "          [ 0.4658, -0.3348, -0.3359, -0.3573,  0.0543],\n",
              "          [ 0.1440,  0.3307,  0.0678, -0.0142,  0.1360],\n",
              "          [ 0.1837, -0.0923, -0.2029, -0.0952,  0.0682],\n",
              "          [ 0.3772,  0.2633,  0.0362, -0.1300,  0.0339]]],\n",
              "\n",
              "\n",
              "        [[[ 0.2181, -0.1421,  0.1124,  0.0012, -0.0168],\n",
              "          [ 0.0171, -0.1736,  0.1979, -0.1027, -0.1649],\n",
              "          [-0.0039,  0.0531, -0.1303, -0.0961,  0.0141],\n",
              "          [-0.0777, -0.0429,  0.0933, -0.0280,  0.1749],\n",
              "          [-0.1918, -0.0414,  0.1308, -0.0551, -0.1510]]],\n",
              "\n",
              "\n",
              "        [[[ 0.0107,  0.2651,  0.0412, -0.2365, -0.1668],\n",
              "          [ 0.2672, -0.1946,  0.5422, -0.1076, -0.5135],\n",
              "          [ 0.4661, -0.0856, -0.1908,  0.4054,  0.0506],\n",
              "          [ 0.2256, -0.0020, -0.1026,  0.0605, -0.0712],\n",
              "          [ 0.1608, -0.2236, -0.0798,  0.0504, -0.2914]]],\n",
              "\n",
              "\n",
              "        [[[-0.6872, -0.0395,  0.8282, -0.1714, -0.1565],\n",
              "          [ 0.5356,  0.1220, -0.0754,  0.1134, -0.0080],\n",
              "          [-0.2289, -0.0614,  0.0109,  0.1998,  0.1801],\n",
              "          [ 0.0355,  0.3546, -0.2276, -0.3632,  0.0222],\n",
              "          [ 0.2970, -0.3303, -0.5959,  0.2573,  0.0261]]],\n",
              "\n",
              "\n",
              "        [[[-0.4079, -0.2852, -0.4556, -0.5928, -0.3151],\n",
              "          [-0.4161, -0.5703, -0.6146, -0.4668, -0.5507],\n",
              "          [-0.1246, -0.2341, -0.4768, -0.2730,  0.0871],\n",
              "          [ 0.1642,  0.1349, -0.0046,  0.0169, -0.2758],\n",
              "          [ 0.5334,  0.2695, -0.1904, -0.1409, -0.1167]]],\n",
              "\n",
              "\n",
              "        [[[ 0.1998, -0.2877, -0.0107,  0.2730, -0.2117],\n",
              "          [ 0.2120, -0.5598,  0.2971,  0.2172, -0.0877],\n",
              "          [ 0.1275, -0.4115,  0.0110,  0.3710, -0.0968],\n",
              "          [ 0.3288, -0.4447, -0.4117,  0.6874, -0.0885],\n",
              "          [ 0.4146, -0.3206, -0.4630,  0.3075,  0.0061]]],\n",
              "\n",
              "\n",
              "        [[[-0.0070, -0.4057, -0.0626,  0.3514,  0.0409],\n",
              "          [-0.1964,  0.5708, -0.1601, -0.1334,  0.2064],\n",
              "          [-0.4694,  0.4728,  0.4354, -0.2515, -0.2919],\n",
              "          [-0.0052, -0.3506,  0.4062,  0.3625, -0.1913],\n",
              "          [ 0.3714, -0.1585, -0.1871,  0.2069,  0.0821]]],\n",
              "\n",
              "\n",
              "        [[[-0.4268,  0.1935, -0.0978,  0.4188, -0.0774],\n",
              "          [ 0.2860,  0.3210,  0.3370, -0.0970, -0.3431],\n",
              "          [-0.0222, -0.0843, -0.0722, -0.5547, -0.3371],\n",
              "          [-0.3842, -0.3446,  0.3796,  0.3504, -0.1064],\n",
              "          [ 0.1068,  0.4094,  0.3315, -0.3673,  0.1099]]],\n",
              "\n",
              "\n",
              "        [[[-0.1647,  0.0657, -0.0077,  0.3879,  0.1157],\n",
              "          [ 0.1802,  0.0919,  0.0797,  0.2058, -0.4238],\n",
              "          [ 0.0451,  0.0872, -0.0730,  0.2801, -0.3097],\n",
              "          [-0.1936,  0.1499,  0.2400,  0.2719, -0.4543],\n",
              "          [ 0.1440, -0.2981, -0.2352,  0.3459, -0.3967]]],\n",
              "\n",
              "\n",
              "        [[[-0.0756, -0.0080, -0.4591, -0.3325, -0.6801],\n",
              "          [ 0.4516,  0.5238,  0.5381, -0.0179, -0.1884],\n",
              "          [ 0.0808, -0.2290,  0.4692,  0.4711, -0.0427],\n",
              "          [-0.1302, -0.4895, -0.4862,  0.3788,  0.1806],\n",
              "          [ 0.0196, -0.1995, -0.5210, -0.1294,  0.4709]]],\n",
              "\n",
              "\n",
              "        [[[ 0.3420, -0.0651,  0.3430, -0.2833, -0.7730],\n",
              "          [ 0.3469,  0.1922,  0.4206, -0.3688, -0.5612],\n",
              "          [ 0.0156,  0.2690, -0.1457, -0.3271,  0.0687],\n",
              "          [-0.1329, -0.2564, -0.3207, -0.2813, -0.0421],\n",
              "          [-0.0663, -0.1499, -0.0407, -0.1290,  0.4304]]],\n",
              "\n",
              "\n",
              "        [[[-0.5492, -0.2503,  0.0421,  0.1780,  0.1943],\n",
              "          [ 0.5385,  0.5469,  0.2316, -0.0974, -0.1227],\n",
              "          [-0.1197,  0.1921,  0.5693,  0.0667, -0.1767],\n",
              "          [ 0.2546, -0.2521,  0.0298,  0.2227, -0.2228],\n",
              "          [-0.1106, -0.3190, -0.8169, -0.4403,  0.1447]]],\n",
              "\n",
              "\n",
              "        [[[-0.1192, -0.2424,  0.0810, -0.0964,  0.2155],\n",
              "          [-0.3139, -0.4596, -0.1530, -0.2215, -0.0980],\n",
              "          [-0.5376, -0.6153, -0.2895, -0.0995, -0.0752],\n",
              "          [-0.4692, -0.6712, -0.3203,  0.2225,  0.3491],\n",
              "          [-0.2238, -0.2865,  0.0053, -0.1479,  0.4163]]],\n",
              "\n",
              "\n",
              "        [[[-0.5859, -0.2783, -0.5029, -0.6671, -0.3024],\n",
              "          [-0.0494,  0.1757, -0.1202, -0.5558, -0.4461],\n",
              "          [ 0.0467,  0.1212,  0.2869,  0.2674, -0.0706],\n",
              "          [ 0.2120, -0.0314,  0.2080,  0.2384, -0.0437],\n",
              "          [ 0.3830,  0.0263,  0.4561,  0.5878, -0.0077]]],\n",
              "\n",
              "\n",
              "        [[[ 0.0703, -0.1128,  0.1067,  0.1364,  0.0802],\n",
              "          [-0.0627, -0.0488, -0.0899, -0.0745, -0.1209],\n",
              "          [-0.0782,  0.1370, -0.1069,  0.1331,  0.1549],\n",
              "          [-0.1571, -0.0447,  0.0301, -0.1897, -0.1751],\n",
              "          [-0.0191, -0.1646,  0.1197,  0.0895, -0.0017]]],\n",
              "\n",
              "\n",
              "        [[[-0.6195, -0.2047, -0.0333, -0.3111,  0.2797],\n",
              "          [-0.4201,  0.1216,  0.1776, -0.1269,  0.1665],\n",
              "          [ 0.2534, -0.2414,  0.2495,  0.4272, -0.0591],\n",
              "          [-0.1701,  0.1572,  0.0780,  0.2287,  0.0031],\n",
              "          [ 0.2428, -0.0364, -0.0537,  0.1633, -0.2847]]],\n",
              "\n",
              "\n",
              "        [[[ 0.2550,  0.4746, -0.0826, -0.0926,  0.0722],\n",
              "          [ 0.5059, -0.1449, -0.6851,  0.2957,  0.2547],\n",
              "          [ 0.0368, -0.3556, -0.1320,  0.4877, -0.1583],\n",
              "          [-0.4766, -0.4640,  0.3918,  0.2362, -0.1790],\n",
              "          [-0.3716,  0.0437,  0.3791, -0.0909, -0.2668]]],\n",
              "\n",
              "\n",
              "        [[[ 0.1285, -0.1519, -0.0600, -0.0243,  0.0519],\n",
              "          [ 0.0023,  0.2582, -0.1446, -0.3827, -0.2647],\n",
              "          [ 0.1424,  0.2969, -0.1186, -0.1687, -0.3514],\n",
              "          [ 0.4278,  0.0170,  0.0127, -0.0142, -0.2235],\n",
              "          [ 0.2564, -0.0729,  0.1005,  0.0686,  0.0273]]],\n",
              "\n",
              "\n",
              "        [[[ 0.1660,  0.3195, -0.1342, -0.0872,  0.4576],\n",
              "          [-0.4225, -0.1274, -0.4186,  0.1600,  0.4274],\n",
              "          [-0.0295, -0.0829, -0.0746, -0.0303, -0.3100],\n",
              "          [ 0.4101,  0.3228, -0.3500,  0.0726, -0.2244],\n",
              "          [ 0.0841, -0.3971,  0.0438,  0.4507, -0.2290]]],\n",
              "\n",
              "\n",
              "        [[[-0.1176,  0.5792, -0.0539, -0.0162,  0.0424],\n",
              "          [-0.3203,  0.1876, -0.1292, -0.1152, -0.3978],\n",
              "          [ 0.0774, -0.1368, -0.0916,  0.3562,  0.2881],\n",
              "          [ 0.1695,  0.6170,  0.4019, -0.3527,  0.3726],\n",
              "          [-0.6804, -0.5540, -0.2018,  0.0322, -0.1556]]],\n",
              "\n",
              "\n",
              "        [[[ 0.0982,  0.6041, -0.4324, -0.4442,  0.3622],\n",
              "          [ 0.3682,  0.1445,  0.0629, -0.0489, -0.4268],\n",
              "          [ 0.2233, -0.3406,  0.2211,  0.1924,  0.3884],\n",
              "          [-0.3169, -0.0153,  0.0684, -0.1096,  0.3760],\n",
              "          [-0.4585, -0.0685, -0.2625, -0.2314,  0.2349]]],\n",
              "\n",
              "\n",
              "        [[[ 0.5507,  0.4350,  0.2988, -0.0656,  0.4262],\n",
              "          [ 0.1333, -0.2078, -0.0366, -0.3419,  0.3060],\n",
              "          [ 0.1977,  0.0213, -0.0272, -0.3458,  0.0829],\n",
              "          [-0.2795, -0.3978, -0.4744, -0.0624,  0.0653],\n",
              "          [-0.5326, -0.5431, -0.2415,  0.0996,  0.0290]]]], device='cuda:0',\n",
              "       requires_grad=True)"
            ]
          },
          "metadata": {},
          "execution_count": 78
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install torch_intermediate_layer_getter\n",
        "from torch_intermediate_layer_getter import IntermediateLayerGetter as MidGetter"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OWUUr4WnHrDS",
        "outputId": "905b1659-62ba-4b49-d00f-e00be72b9c89"
      },
      "execution_count": 96,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Collecting torch_intermediate_layer_getter\n",
            "  Downloading torch_intermediate_layer_getter-0.1.post1.tar.gz (3.0 kB)\n",
            "Building wheels for collected packages: torch-intermediate-layer-getter\n",
            "  Building wheel for torch-intermediate-layer-getter (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for torch-intermediate-layer-getter: filename=torch_intermediate_layer_getter-0.1.post1-py3-none-any.whl size=3726 sha256=ba6b5c04377c2a535d42833efd6a67b3c68d6437747b257952be765066e66479\n",
            "  Stored in directory: /root/.cache/pip/wheels/59/53/37/8b81c4711686fbae03a35d8d1dac5edd9e4af9221fa8e17f6f\n",
            "Successfully built torch-intermediate-layer-getter\n",
            "Installing collected packages: torch-intermediate-layer-getter\n",
            "Successfully installed torch-intermediate-layer-getter-0.1.post1\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Tdy-keR_Foch",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "fd16d283-eba3-4914-c62e-8253f886c4a3"
      },
      "source": [
        "return_layers = {\n",
        "    'conv':'conv',\n",
        "}\n",
        "\n",
        "test_cases = testX[0:3].to(device)\n",
        "\n",
        "mid_getter = MidGetter(model, return_layers=return_layers, keep_output=True)\n",
        "mid_outputs, model_output = mid_getter(test_cases)\n",
        "\n",
        "#print(mid_outputs['conv'].shape)\n",
        "x_rep = mid_outputs['conv'].cpu().detach().numpy()\n",
        "x_rep = np.squeeze(x_rep)\n",
        "\n",
        "#print(x_rep.shape)\n",
        "\n",
        "for this_x_rep in x_rep:\n",
        "    plt.figure(figsize=(15, 15), frameon=False)\n",
        "    \n",
        "    #print(this_x_rep.shape)\n",
        "\n",
        "    for i in range (this_x_rep.shape[0]):\n",
        "        val = this_x_rep[i,:,:]\n",
        "        plt.subplot(6, 6, i + 1)\n",
        "        plt.axis(\"off\")\n",
        "        plt.imshow(val, cmap='gray',interpolation='nearest')"
      ],
      "execution_count": 107,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 32 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 32 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 32 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "testX[0:3].shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "D2HbAMn491Cv",
        "outputId": "76a51481-4cdd-469b-80a0-6fdbb2152f9b"
      },
      "execution_count": 98,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([3, 1, 28, 28])"
            ]
          },
          "metadata": {},
          "execution_count": 98
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OhErkuAD1laZ"
      },
      "source": [
        "#STEP 13: 마지막 인공지능 모델 (VGG-like CNN)\n",
        "\n",
        "![대체 텍스트](https://neurohive.io/wp-content/uploads/2018/11/vgg16-1-e1542731207177.png)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class VGGLikeCNN(nn.Module):\n",
        "    def __init__(self, num_classes=10):\n",
        "        super(VGGLikeCNN, self).__init__()\n",
        "        self.layer1 = nn.Sequential(\n",
        "            nn.Conv2d(1, 32, kernel_size=3, stride=1, padding=1),\n",
        "            nn.BatchNorm2d(32),\n",
        "            nn.ReLU(),\n",
        "            nn.Conv2d(32, 32, kernel_size=3, stride=1, padding=1),\n",
        "            nn.BatchNorm2d(32),\n",
        "            nn.ReLU(), \n",
        "            nn.MaxPool2d(kernel_size = 2, stride = 2)) \n",
        "        self.layer2 = nn.Sequential(\n",
        "            nn.Conv2d(32, 64, kernel_size=3, stride=1, padding=1),\n",
        "            nn.BatchNorm2d(64),\n",
        "            nn.ReLU(),\n",
        "            nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1),\n",
        "            nn.BatchNorm2d(64),\n",
        "            nn.ReLU(),\n",
        "            nn.MaxPool2d(kernel_size = 2, stride = 2))\n",
        "        self.fc = nn.Sequential(\n",
        "            nn.Dropout(0.5),\n",
        "            nn.Linear(7*7*64, 128),\n",
        "            nn.ReLU())\n",
        "        self.fc1 = nn.Sequential(\n",
        "            nn.Dropout(0.5),\n",
        "            nn.Linear(128, num_classes))\n",
        "        \n",
        "    def forward(self, x):\n",
        "        out = self.layer1(x)\n",
        "        out = self.layer2(out)\n",
        "        out = out.reshape(out.size(0), -1)\n",
        "        out = self.fc(out)\n",
        "        out = self.fc1(out)\n",
        "        return out"
      ],
      "metadata": {
        "id": "G1FKC_zpLThr"
      },
      "execution_count": 108,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3OB5CedfFocl",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c16919f1-069b-40f3-bbfa-a0bdf9c5f88f"
      },
      "source": [
        "# build the model\n",
        "model = VGGLikeCNN().to(device)\n",
        "summary(model, input_size=(1, 28, 28))\n",
        "\n",
        "loss_fn = nn.CrossEntropyLoss()\n",
        "optimizer = torch.optim.SGD(model.parameters(), lr=0.01, weight_decay=1e-6, momentum=0.9, nesterov=True)\n",
        "\n",
        "# fix random seed for reproductibility\n",
        "seed = 7\n",
        "np.random.seed(seed)\n",
        "\n",
        "# Fit the model\n",
        "epochs = 20\n",
        "\n",
        "loss = []\n",
        "accuracy = []\n",
        "val_loss = []\n",
        "val_accuracy = []\n",
        "\n",
        "for t in range(epochs):\n",
        "    print(f\"Epoch {t+1}\\n-------------------------------\")\n",
        "    tmp_loss, tmp_accuracy = train(train_dataloader, model, loss_fn, optimizer)\n",
        "    loss.append(tmp_loss)\n",
        "    accuracy.append(tmp_accuracy)\n",
        "    tmp_loss, tmp_accuracy = test(val_dataloader, model, loss_fn)\n",
        "    val_loss.append(tmp_loss)\n",
        "    val_accuracy.append(tmp_accuracy)\n",
        "print(\"Done!\")"
      ],
      "execution_count": 109,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "----------------------------------------------------------------\n",
            "        Layer (type)               Output Shape         Param #\n",
            "================================================================\n",
            "            Conv2d-1           [-1, 32, 28, 28]             320\n",
            "       BatchNorm2d-2           [-1, 32, 28, 28]              64\n",
            "              ReLU-3           [-1, 32, 28, 28]               0\n",
            "            Conv2d-4           [-1, 32, 28, 28]           9,248\n",
            "       BatchNorm2d-5           [-1, 32, 28, 28]              64\n",
            "              ReLU-6           [-1, 32, 28, 28]               0\n",
            "         MaxPool2d-7           [-1, 32, 14, 14]               0\n",
            "            Conv2d-8           [-1, 64, 14, 14]          18,496\n",
            "       BatchNorm2d-9           [-1, 64, 14, 14]             128\n",
            "             ReLU-10           [-1, 64, 14, 14]               0\n",
            "           Conv2d-11           [-1, 64, 14, 14]          36,928\n",
            "      BatchNorm2d-12           [-1, 64, 14, 14]             128\n",
            "             ReLU-13           [-1, 64, 14, 14]               0\n",
            "        MaxPool2d-14             [-1, 64, 7, 7]               0\n",
            "          Dropout-15                 [-1, 3136]               0\n",
            "           Linear-16                  [-1, 128]         401,536\n",
            "             ReLU-17                  [-1, 128]               0\n",
            "          Dropout-18                  [-1, 128]               0\n",
            "           Linear-19                   [-1, 10]           1,290\n",
            "================================================================\n",
            "Total params: 468,202\n",
            "Trainable params: 468,202\n",
            "Non-trainable params: 0\n",
            "----------------------------------------------------------------\n",
            "Input size (MB): 0.00\n",
            "Forward/backward pass size (MB): 1.82\n",
            "Params size (MB): 1.79\n",
            "Estimated Total Size (MB): 3.61\n",
            "----------------------------------------------------------------\n",
            "Epoch 1\n",
            "-------------------------------\n",
            "loss: 2.455837  [    0/50000]\n",
            "loss: 0.453217  [10000/50000]\n",
            "loss: 0.439213  [20000/50000]\n",
            "loss: 0.637291  [30000/50000]\n",
            "loss: 0.433710  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 80.3%, Avg loss: 0.540260 \n",
            "\n",
            "Test: \n",
            " Accuracy: 88.0%, Avg loss: 0.325800 \n",
            "\n",
            "Epoch 2\n",
            "-------------------------------\n",
            "loss: 0.409117  [    0/50000]\n",
            "loss: 0.355933  [10000/50000]\n",
            "loss: 0.351792  [20000/50000]\n",
            "loss: 0.503609  [30000/50000]\n",
            "loss: 0.371943  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 86.8%, Avg loss: 0.368384 \n",
            "\n",
            "Test: \n",
            " Accuracy: 90.0%, Avg loss: 0.265553 \n",
            "\n",
            "Epoch 3\n",
            "-------------------------------\n",
            "loss: 0.288157  [    0/50000]\n",
            "loss: 0.281952  [10000/50000]\n",
            "loss: 0.329590  [20000/50000]\n",
            "loss: 0.390715  [30000/50000]\n",
            "loss: 0.313035  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 88.7%, Avg loss: 0.319472 \n",
            "\n",
            "Test: \n",
            " Accuracy: 90.6%, Avg loss: 0.244612 \n",
            "\n",
            "Epoch 4\n",
            "-------------------------------\n",
            "loss: 0.334494  [    0/50000]\n",
            "loss: 0.180042  [10000/50000]\n",
            "loss: 0.278338  [20000/50000]\n",
            "loss: 0.439497  [30000/50000]\n",
            "loss: 0.299955  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 89.5%, Avg loss: 0.292153 \n",
            "\n",
            "Test: \n",
            " Accuracy: 91.3%, Avg loss: 0.228449 \n",
            "\n",
            "Epoch 5\n",
            "-------------------------------\n",
            "loss: 0.254521  [    0/50000]\n",
            "loss: 0.291902  [10000/50000]\n",
            "loss: 0.385297  [20000/50000]\n",
            "loss: 0.396223  [30000/50000]\n",
            "loss: 0.394269  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 90.1%, Avg loss: 0.275492 \n",
            "\n",
            "Test: \n",
            " Accuracy: 91.9%, Avg loss: 0.219274 \n",
            "\n",
            "Epoch 6\n",
            "-------------------------------\n",
            "loss: 0.273686  [    0/50000]\n",
            "loss: 0.221731  [10000/50000]\n",
            "loss: 0.301272  [20000/50000]\n",
            "loss: 0.453633  [30000/50000]\n",
            "loss: 0.188158  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 90.8%, Avg loss: 0.255596 \n",
            "\n",
            "Test: \n",
            " Accuracy: 91.5%, Avg loss: 0.225538 \n",
            "\n",
            "Epoch 7\n",
            "-------------------------------\n",
            "loss: 0.294078  [    0/50000]\n",
            "loss: 0.202511  [10000/50000]\n",
            "loss: 0.286774  [20000/50000]\n",
            "loss: 0.422883  [30000/50000]\n",
            "loss: 0.284107  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 91.1%, Avg loss: 0.246583 \n",
            "\n",
            "Test: \n",
            " Accuracy: 92.0%, Avg loss: 0.216013 \n",
            "\n",
            "Epoch 8\n",
            "-------------------------------\n",
            "loss: 0.207121  [    0/50000]\n",
            "loss: 0.143287  [10000/50000]\n",
            "loss: 0.351721  [20000/50000]\n",
            "loss: 0.364016  [30000/50000]\n",
            "loss: 0.188414  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 91.5%, Avg loss: 0.233428 \n",
            "\n",
            "Test: \n",
            " Accuracy: 92.4%, Avg loss: 0.204611 \n",
            "\n",
            "Epoch 9\n",
            "-------------------------------\n",
            "loss: 0.187366  [    0/50000]\n",
            "loss: 0.156379  [10000/50000]\n",
            "loss: 0.273522  [20000/50000]\n",
            "loss: 0.321957  [30000/50000]\n",
            "loss: 0.182102  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 91.9%, Avg loss: 0.226649 \n",
            "\n",
            "Test: \n",
            " Accuracy: 92.6%, Avg loss: 0.200266 \n",
            "\n",
            "Epoch 10\n",
            "-------------------------------\n",
            "loss: 0.219526  [    0/50000]\n",
            "loss: 0.143385  [10000/50000]\n",
            "loss: 0.207971  [20000/50000]\n",
            "loss: 0.259331  [30000/50000]\n",
            "loss: 0.219516  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 92.1%, Avg loss: 0.215190 \n",
            "\n",
            "Test: \n",
            " Accuracy: 92.7%, Avg loss: 0.197883 \n",
            "\n",
            "Epoch 11\n",
            "-------------------------------\n",
            "loss: 0.195074  [    0/50000]\n",
            "loss: 0.161019  [10000/50000]\n",
            "loss: 0.177543  [20000/50000]\n",
            "loss: 0.264082  [30000/50000]\n",
            "loss: 0.175642  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 92.5%, Avg loss: 0.205238 \n",
            "\n",
            "Test: \n",
            " Accuracy: 93.0%, Avg loss: 0.193950 \n",
            "\n",
            "Epoch 12\n",
            "-------------------------------\n",
            "loss: 0.133844  [    0/50000]\n",
            "loss: 0.111514  [10000/50000]\n",
            "loss: 0.229260  [20000/50000]\n",
            "loss: 0.253461  [30000/50000]\n",
            "loss: 0.167205  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 92.8%, Avg loss: 0.199340 \n",
            "\n",
            "Test: \n",
            " Accuracy: 93.0%, Avg loss: 0.190140 \n",
            "\n",
            "Epoch 13\n",
            "-------------------------------\n",
            "loss: 0.170853  [    0/50000]\n",
            "loss: 0.114658  [10000/50000]\n",
            "loss: 0.198662  [20000/50000]\n",
            "loss: 0.331899  [30000/50000]\n",
            "loss: 0.200069  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 92.9%, Avg loss: 0.193533 \n",
            "\n",
            "Test: \n",
            " Accuracy: 93.1%, Avg loss: 0.191672 \n",
            "\n",
            "Epoch 14\n",
            "-------------------------------\n",
            "loss: 0.177888  [    0/50000]\n",
            "loss: 0.136237  [10000/50000]\n",
            "loss: 0.193644  [20000/50000]\n",
            "loss: 0.298129  [30000/50000]\n",
            "loss: 0.194885  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 93.1%, Avg loss: 0.185511 \n",
            "\n",
            "Test: \n",
            " Accuracy: 93.1%, Avg loss: 0.186284 \n",
            "\n",
            "Epoch 15\n",
            "-------------------------------\n",
            "loss: 0.162376  [    0/50000]\n",
            "loss: 0.154641  [10000/50000]\n",
            "loss: 0.205327  [20000/50000]\n",
            "loss: 0.352499  [30000/50000]\n",
            "loss: 0.175597  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 93.4%, Avg loss: 0.182913 \n",
            "\n",
            "Test: \n",
            " Accuracy: 93.3%, Avg loss: 0.185449 \n",
            "\n",
            "Epoch 16\n",
            "-------------------------------\n",
            "loss: 0.212292  [    0/50000]\n",
            "loss: 0.081535  [10000/50000]\n",
            "loss: 0.183463  [20000/50000]\n",
            "loss: 0.222964  [30000/50000]\n",
            "loss: 0.139104  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 93.7%, Avg loss: 0.175959 \n",
            "\n",
            "Test: \n",
            " Accuracy: 93.4%, Avg loss: 0.184094 \n",
            "\n",
            "Epoch 17\n",
            "-------------------------------\n",
            "loss: 0.174077  [    0/50000]\n",
            "loss: 0.117101  [10000/50000]\n",
            "loss: 0.200244  [20000/50000]\n",
            "loss: 0.229081  [30000/50000]\n",
            "loss: 0.167219  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 93.7%, Avg loss: 0.169472 \n",
            "\n",
            "Test: \n",
            " Accuracy: 93.4%, Avg loss: 0.184125 \n",
            "\n",
            "Epoch 18\n",
            "-------------------------------\n",
            "loss: 0.264998  [    0/50000]\n",
            "loss: 0.137866  [10000/50000]\n",
            "loss: 0.245138  [20000/50000]\n",
            "loss: 0.259129  [30000/50000]\n",
            "loss: 0.149050  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 93.8%, Avg loss: 0.167738 \n",
            "\n",
            "Test: \n",
            " Accuracy: 93.2%, Avg loss: 0.186396 \n",
            "\n",
            "Epoch 19\n",
            "-------------------------------\n",
            "loss: 0.156851  [    0/50000]\n",
            "loss: 0.084908  [10000/50000]\n",
            "loss: 0.154475  [20000/50000]\n",
            "loss: 0.162536  [30000/50000]\n",
            "loss: 0.171371  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 93.9%, Avg loss: 0.163433 \n",
            "\n",
            "Test: \n",
            " Accuracy: 93.4%, Avg loss: 0.187592 \n",
            "\n",
            "Epoch 20\n",
            "-------------------------------\n",
            "loss: 0.193167  [    0/50000]\n",
            "loss: 0.101781  [10000/50000]\n",
            "loss: 0.188127  [20000/50000]\n",
            "loss: 0.198248  [30000/50000]\n",
            "loss: 0.096842  [40000/50000]\n",
            "Train: \n",
            " Accuracy: 94.3%, Avg loss: 0.157198 \n",
            "\n",
            "Test: \n",
            " Accuracy: 93.0%, Avg loss: 0.197629 \n",
            "\n",
            "Done!\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 학습과정 살펴보기\n",
        "fig, loss_ax = plt.subplots()\n",
        "\n",
        "acc_ax = loss_ax.twinx()\n",
        "\n",
        "loss_ax.plot(loss, 'y', label='train loss')\n",
        "loss_ax.plot(val_loss, 'r', label='val loss')\n",
        "loss_ax.set_ylim([0.0, 1.5])\n",
        "\n",
        "acc_ax.plot(accuracy, 'b', label='train acc')\n",
        "acc_ax.plot(val_accuracy, 'g', label='val acc')\n",
        "acc_ax.set_ylim([50, 100])\n",
        "\n",
        "loss_ax.set_xlabel('epoch')\n",
        "loss_ax.set_ylabel('loss')\n",
        "acc_ax.set_ylabel('accuray')\n",
        "\n",
        "loss_ax.legend(loc='upper left')\n",
        "acc_ax.legend(loc='lower left')\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 683
        },
        "id": "hAidiz2HPMW8",
        "outputId": "71f96e19-0bae-4a81-ac59-4d48d6d1cf7c"
      },
      "execution_count": 110,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x1080 with 2 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA6AAAANgCAYAAAAyLQUFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXjU9b33/9cnezIJCSRhS9jXCCgqIBUFl6PiBnhccKtr9Wrvnvb4877bep9e59Se+3T3PrX2au1tW3BrBetGFbeqKNSFRQSLEnYCCUsWyDbZZz6/P75kkiEJa+b7nZk8H9c110xmJsl76KJPPt/v52ustQIAAAAAINISvB4AAAAAANA3EKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAABBHjDGLjDHlxphNnZ4bYIz5mzFm25H7/keeN8aYx4wx240xnxtjzonkbAQoAAAAAMSXJyXNPeq5hyS9a60dJ+ndI19L0pWSxh253S/p8UgORoACAAAAQByx1q6UdOiop+dLeurI46ckLej0/NPW8YmkHGPMkEjNlhSpHxwpCQkJNj093esxAAAAAMATDQ0NVtL6Tk89Ya194jjfNshau//I4wOSBh15XCBpb6f3lR55br8iIOYCND09XX6/3+sxAAAAAMATxphGa+20U/1+a601xtjenOlEcQguAAAAAMS/g+2H1h65Lz/yfJmkYZ3eV3jkuYggQAEAAAAg/v1V0p1HHt8paVmn5+84shvuTEk1nQ7V7XUxdwguAAAAAKBnxpjnJF0kKc8YUyrpB5J+Kul5Y8y9kkok3XTk7a9LukrSdkkNku6O6GzWenLo7ynz+Xz26HNAW1tbVVpaqqamJo+min1paWkqLCxUcnKy16MAAAAAOAZjTIO11uf1HKciLlZAS0tLlZWVpZEjR8oY4/U4Mcdaq6qqKpWWlmrUqFFejwMAAAAgTsXFOaBNTU3Kzc0lPk+RMUa5ubmsIAMAAACIqLgIUEnE52nizw8AAABApMVNgAIAAAAAohsB2guqq6v129/+9pS+96qrrlJ1dfUJv//hhx/WI488ckq/CwAAAAC8RID2gmMFaFtb2zG/9/XXX1dOTk4kxgIAAACAqEKA9oKHHnpIO3bs0NSpU/Wd73xH77//vi688ELNmzdPZ5xxhiRpwYIFOvfcczVp0iQ98cQToe8dOXKkKisrtXv3bhUVFem+++7TpEmTdPnll6uxsfGYv3fDhg2aOXOmzjzzTF133XU6fPiwJOmxxx7TGWecoTPPPFM333yzJOmDDz7Q1KlTNXXqVJ199tmqq6uL0J8GAAAAAHQvLi7D0tm2bQ+ovn5Dr/7MzMypGjfu0R5f/+lPf6pNmzZpwwbn977//vtav369Nm3aFLqsyaJFizRgwAA1NjZq+vTpuv7665Wbm3vU7Nv03HPP6fe//71uuukmvfjii7r99tt7/L133HGHfv3rX2vOnDn6j//4D/3whz/Uo48+qp/+9KfatWuXUlNTQ4f3PvLII/rNb36jWbNmqb6+Xmlpaaf7xwIAAAAAJ4UV0AiZMWNG2DU1H3vsMZ111lmaOXOm9u7dq23btnX5nlGjRmnq1KmSpHPPPVe7d+/u8efX1NSourpac+bMkSTdeeedWrlypSTpzDPP1G233aZnn31WSUnO3zHMmjVLDz74oB577DFVV1eHngcAAAAAt8RdhRxrpdJNPp8v9Pj999/XO++8o48//lgZGRm66KKLur3mZmpqauhxYmLicQ/B7cny5cu1cuVKvfrqq/rRj36kf/zjH3rooYd09dVX6/XXX9esWbP01ltvaeLEiaf08wEAAADgVLAC2guysrKOeU5lTU2N+vfvr4yMDBUXF+uTTz457d+ZnZ2t/v37a9WqVZKkZ555RnPmzFEwGNTevXt18cUX62c/+5lqampUX1+vHTt2aMqUKfre976n6dOnq7i4+LRnAAAAAICTEXcroF7Izc3VrFmzNHnyZF155ZW6+uqrw16fO3eufve736moqEgTJkzQzJkze+X3PvXUU/r617+uhoYGjR49WosXL1YgENDtt9+umpoaWWv17W9/Wzk5Ofr3f/93rVixQgkJCZo0aZKuvPLKXpkBAAAAAE6UsdZ6PcNJ8fl81u/3hz23efNmFRUVeTRR/ODPEQAAAIh+xpgGa63v+O+MPhyCCwAAAABwBQEKAAAAAHAFAQoAAAAAcAUBCgAAAABwBQEKAAAAAHAFAQoAAAAAcAUB6pHMzMyTeh4AAAAAYh0BCgAAAABwBQHaCx566CH95je/CX398MMP65FHHlF9fb0uvfRSnXPOOZoyZYqWLVt2wj/TWqvvfOc7mjx5sqZMmaKlS5dKkvbv36/Zs2dr6tSpmjx5slatWqVAIKC77ror9N5f/vKXvf4ZAQAAAOB0JXk9QK974AFpw4be/ZlTp0qPPtrjywsXLtQDDzygb37zm5Kk559/Xm+99ZbS0tL08ssvq1+/fqqsrNTMmTM1b948GWOO+ytfeuklbdiwQRs3blRlZaWmT5+u2bNn689//rOuuOIKff/731cgEFBDQ4M2bNigsrIybdq0SZJUXV3dO58bAAAAAHpR/AWoB84++2yVl5dr3759qqioUP/+/TVs2DC1trbq3/7t37Ry5UolJCSorKxMBw8e1ODBg4/7M//+97/rlltuUWJiogYNGqQ5c+Zo7dq1mj59uu655x61trZqwYIFmjp1qkaPHq2dO3fqW9/6lq6++mpdfvnlLnxqAAAAADg58Regx1ipjKQbb7xRL7zwgg4cOKCFCxdKkv70pz+poqJCn376qZKTkzVy5Eg1NTWd1u+ZPXu2Vq5cqeXLl+uuu+7Sgw8+qDvuuEMbN27UW2+9pd/97nd6/vnntWjRot74WAAAAADQazgHtJcsXLhQS5Ys0QsvvKAbb7xRklRTU6OBAwcqOTlZK1asUElJyQn/vAsvvFBLly5VIBBQRUWFVq5cqRkzZqikpESDBg3Sfffdp6997Wtav369KisrFQwGdf311+u//uu/tH79+kh9TAAAAAA4ZfG3AuqRSZMmqa6uTgUFBRoyZIgk6bbbbtO1116rKVOmaNq0aZo4ceIJ/7zrrrtOH3/8sc466ywZY/Tzn/9cgwcP1lNPPaVf/OIXSk5OVmZmpp5++mmVlZXp7rvvVjAYlCT95Cc/ichnBAAAAIDTYay1Xs9wUnw+n/X7/WHPbd68WUVFRR5NFD/4cwQAAACinzGmwVrr83qOU8EhuAAAAAAAVxCgAAAAAABXxE2AxtqhxNGGPz8AAAAAkRYXAZqWlqaqqioi6hRZa1VVVaW0tDSvRwEAAAAQx+JiF9zCwkKVlpaqoqLC61FiVlpamgoLC70eAwAAAEAci4tdcAEAAACgr2AXXAAAAAAAjoMABQAAAAC4ggAFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALgiYgFqjFlkjCk3xmw6zvumG2PajDE3RGoWAAAAAID3IrkC+qSkucd6gzEmUdLPJL0dwTkAAAAAAFEgYgFqrV0p6dBx3vYtSS9KKo/UHAAAAACA6ODZOaDGmAJJ10l6/ATee78xZp0xZl1bW1vkhwMAAAAA9DovNyF6VNL3rLXB473RWvuEtXaatXZaUlKSC6MBAAAAAHqblzU3TdISY4wk5Um6yhjTZq19xcOZAAAAAAAR4lmAWmtHtT82xjwp6TXiEwAAAADiV8QC1BjznKSLJOUZY0ol/UBSsiRZa38Xqd8LAAAAAIhOxlrr9QwnxefzWb/f7/UYAAAAAOAJY0yDtdbn9RynwstNiAAAAAAAfQgBCgAAAABwBQEKAAAAAHAFAQoAAAAAcAUBCgAAAABwBQEKAAAAAHAFAQoAAAAAcAUBCgAAAABwBQEKAAAAAHAFAQoAAAAAcAUBCgAAAABwBQEKAAAAAHAFAQoAAAAAcAUBCgAAAABwBQEKAAAAAHAFAQoAAAAAcAUBCgAAAABwBQEKAAAAAHAFAQoAAAAAcAUBCgAAAABwBQEKAAAAAHAFAQoAAAAAcAUBCgAAAABwBQEKAAAAAHAFAQoAAAAAcAUBCgAAAABwBQEKAAAAAHAFAQoAAAAAcAUBCgAAAABwBQEKAAAAAHAFAQoAAAAAcAUBCgAAAABwRZLXAwAAAACAl6yVqqqkXbuknTs77gMB6Y9/9Hq6+EKAAgAAAIh7jY3S7t1dI7P9cV1d+Pvz86VJkzwZNa4RoAAAAABiXjAo7dsXHped7/ftC39/ero0apQ0erQ0Z45z3/71qFFSZqY3nyPeEaAAAAAAYkJNTferlzt3OqubLS0d7zVGGjbMickrrgiPy9GjpUGDnPfAXcZa6/UMJ8Xn81m/3+/1GAAAAAB6WUuLtGdP96uYO3dKhw+Hv79//64rl6NHO7fhw6WUFG8+R6QZYxqstT6v5zgVrIACAAAA6HXBoOT3O+dW1tc79+239q/Ly8Mjs7TU+b52KSnSyJFOUM6Y0RGXo0Y5t5wczz4eThEBCgAAAECBgBOGnWOxp3A83mt1dU58noihQ52YnDMnfAVz1CjntQQuHBlXCFAAAAAgxlgrNTV1BGNPt5MJyYaGE//9mZlSVpZza388dGj410c/7u7r/v2dzYDQdxCgAAAAQARZ68Td8WKxPQRP5H319c6K5YkwpvsAHDas5zg8VjhmZLAqiVNHgAIAACDqWeucG9jW5oRXW1vHzcuvT2QVsr7emf9EJCQ4sdcefO2PBw/ueNz51vk9Pd0yMtjtFdGDAAUAAIBnGhqkkhLnEhrd3WpqOoIvmiQnS4mJUmpqeARmZUmFhcePwp7CMS2NWER8I0ABAAAQMX5/z4FZUuLsgtpZcrI0YoSz8+m8ec45gsnJUlKSE3xJSR03r77m8FPg1BGgAAAAOGX19U5I9hSZFRXh709J6QjMqVOd+863wYMJPCCeEaAAAADoUXtg9nSIbGVl+PtTUzsC8+yzw+NyxAgCE+jrCFAAAIA+rK7u2IfIdheY7UF57rkdYdn+3KBBpx+YQRtUc1uzWgItCtrg6f2wCEtPTldaUprXY8SklkCLJCklMcXjSeAmY090S64o4fP5rP9Er2oLAADQR1krHT4slZVJpaVd70vLrErLa1XdUC8ltoRuKRnNGlzQokEFLRo4uEW5g5o1IL9F/fNalDOgRSm+ZrUFW9QScG7todgSaFFzoONxt1+3ndjrAXuC1xeJElkpWRroG3jMW35Gvgb6Bio3I1dJCfG5BhS0QR1qPKRyf3nYrcJf4TxuCH++uqlaCSZBo/uPVlFekSbmTVRRXpGK8p3HOWk5Xn+kqGWMabDW+rye41QQoAAAADEmEJAOHuwIyr17rXbtq9OuA1Xae6hS+2sqVdlQqdakSimjUsqoOnJfqaR+zn0gtUrW9M7WsokmUalJqUpJTFFKYopSEzsepySmhL129OvHe2+iSeyVGSPF3+rvGlwNFarwV3Qb0kZGuRm54YGacSRSffldwjU7NVvGo21xrbWqb6nv8vlCt4ZOcXnkc3e3Yn2sz9wSaFFxVbGKK4u1tWpraFVUkgZnDu42TAuyCjz7M4kWBKiLCFAAABDPGhuttpX4tbmkUtvKKrW7vFJ7q6p0oNaJyprWSvltpZTeEZXKqJQSW7v9eQlKVHZKrnLTczW4X57yfXnKy3Buuem5ykrNOqEY7Om15IRkJSZEdyR6IWiDOtx4uMd4q2ioCPv6cNPhbn9OckJy+EqqLz8Ub93d0pPTjzlXU1tTKBqPnqG7GZvamrr9Of1S+3W7wtvdLTc994T+O9IWbNOuw7tUXFmszZWbtblys/O4YrNqmmtC78tKydLEvIldwnRM/zFKTkw+7u+JBwSoiwhQAAD6Fmutmtqa5G/1y9/iV31LvfytR+5b/AraoBJMwjFviQmJx31Pl+8xJ/89nX+XkbNC09DaoMqGSlU2VKmkolLb91Vq18FKlR2u1IHaKlU1VqqmpVJ+Vao5sVI2rUpKau7hDyNBqcEBykzIU05KnvIz8jQkO08FubkantsRl7npuaHIzE7LVoJh159o1xJoUWVDZfeHrh51+OrB+oNqbGvs9udkpmSGxWBiQmLYz6xtru32+1ITU0/oEOL2CHbzvFdrrQ76D2pzRacordyszRWbVVZXFnpfUkKSxg0YFwrTiXkTQ3GamZLp2rxuIEBdRIACABCdgjaohtaGUBgeHYo9fX0i74n2jWhOiTVS4wAlNucpLZinzIRc9U/NU54vT0Oz8zQsN0+jBudqfGGeRuQ7MZmTlkNMQpLkb/EfdxXzoP+ggjaoQb5Bx105zUzJjMnDWuua60JB2vl++6Htagt2HGJe2K/QWS3tFKZFeUUa6BsYk5+bAHURAQoAiCdtwTY1tzWrLdimgA0oEAz0eH+897QF2yLy/Y1tjScUkz2tyPQkPSldvhSffMk+ZaZkypdy5D75qPujnz/qexJNooI2eMxbwAaO+572W2NTUNU1QdXUBlVdE1BNXVC1tUHV1gdVWxdUXb1zq/c794FAUDKdbgkByQSVlBKUzxdUdoZP+e1RmZenMUPyNK4gV0Uj+2tYYaJSUyP0Xy6gD2sJtGjHoR1dDuctrixWfUt96H05aTndhunInJFRfWg5AeoiAhQAECua2pq0r26f9tbsVWltacetzrnfW7NXB/0HvR7zmBJNotKT048dhD2E4bG+JyM5w7V/uWtokCoqnFt5+fHvG3vo6LQ0aeBA55af3/N9+80Xk/9qCMQ3a61Ka0vDDuMtrnLOM+38/8epiakanzteUwZN0bPXPRt1q6QEqIsIUABANGhobVBZbVlYWO6tDQ/NioaKLt+XnZqtwn6FGpY9TIVZhSroV+DEmElUUkKSEhMSlWgSu71PSkjq8bVIfH80HerZ2ir5/VJ9fcd9++3QoWMHZU//2pCaemIx2f7Y55Oi7N9BAfSiw42HuxzO2xpo1Zu3v+n1aF0QoC4iQAEAkVbfUh++YtlNZB5qPNTl+wakD1Bhv0LnlnUkMtu/7leogqwCZaVmefCJ3NPW1nModv76ZB+3tBz/d6ekHDsmj47KzEyCEkBsIkBdRIACAE5HbXNtR0x2PjS2riMyq5uqu3xffkZ+WEx2vg3rNyy0khnrgkHn+pJ79nTcqqtPPBSbe9i8tTsJCU4EZmY6q4vdPT7Waz6fc8vNdaKyXz+CEkDfQIC6iAAFAARtUP4Wv2qaa1TbXKuappqwx7XNtapprgk9v69uXygu61rquvy8Qb5BYTF5dGAW9Ctw9ZIDkdTQIO3d2xGXJSXhsbl3b9fVRmNOLAZPNiBTUwlGADgVBKiLCFAAiG3Nbc2hQDxePNa2dPNcc61qm2tldex/fhkZZaVmKTs1W0OyhoQOiw2df3kkLodmDVVKYopLnz6yrHXOeewpLktKpMrK8O9JSJCGDpWGD3duI0Z0PB4+XBo2TMrJIRQBIJrEcoAmeT0AACC2tAZaVdlQqYqGCufeXxEWhzXNxwjK5lo1B45/jGZaUpr6pfZTdmq2c5+WrXG+cV2e6/z46NcyUzKjahOd3tDU5KxQHh2W7XG5d2/XQ2B9vo6onDYtPC5HjHDiMznZm88DAOh7WAEFgD6uobVBFf6KUFRW+CvC7zuFZkVDRbfnR7ZLMAldQjDs69TsbmPx6NfiZUXyZFjrrE72FJd79jirm50ZIw0Z0v3KZftzrF4CQPyJ5RVQAhQA4oi1VrXNtd2HpL9ClY2VXZ5vaG3o9mclJSQpLyNP+Rn5yvflO/edH/vyQ6/npOUoOy1bvmRf1F0rLZo0NUm7dknbtzu3bduc+/bAbGoKf39GRvdx2f51QYGz8ysAoG8hQF1EgALoaw41HtKB+gPhIdm+Wnl0YDZUqjXY2u3PSU9KD4vH9qDMy8jr+rwvX9mp2cTkKWhslHbu7BqZ27c7kdn5H7s5OdLYsdKoUd2fgzlgAKuXAICuCFAXEaAA4lHQBlVSXaLNlZu1uWJz2IWwqxqruv2e7NTsUDB2Wak86j4vI0++lJj851RUamiQduzoGpjbt0ulpeGRmZvrRGb7bdy4jse5ud59BgBA7CJAXUSAAohlTW1N2la1rSM0q4q1uWKztlRtUVNbx/GXeRl5KsorUlFekSbkTdCQzCFdDn3ti+dJuqm+vufILCsLf29+fs+R2b+/N/MDAOIXAeoiAhRALDjceDi0itk5NHdV71LQBiU5lwkZmTNSE/MmqiivyLnPd+7zMvI8/gR9Q11d94fKbt8u7d8f/t5Bg7oPzDFjnENpAQBwCwHqIgIUQLSw1qq0trTb0DzoPxh6X2piqsbnjg+FZntkjs8dr4zkDA8/QfyzVqqp6Xkl8+DB8PcPGdJzZPbr581nAADgaASoiwhQAG5rDbRq+6HtoXMy2++LK4tV31Ifel9OWk7HSman0ByVM0qJCYkefoL4FAw6ly0pK3POu2y/Hf310f/IKCjoOTIzM735LAAAnAwC1EUEKIBIqWuuCwvM9lXNHYd3qC3YFnpfYb/C0PmZ7YfNFuUVaaBvILvG9pK2NunAgZ6jsqzMubW0hH9fYqI0dKhUWOiEZvv9mDFOZI4eLfli8h/XAAB0IEBdRIACOF3th86u3bdWa8vW6tP9n+rLii9VVtexs0xSQpLGDhjbJTQn5E5QVmqWh9PHvqYmad++Y69aHjjgrHB2lpbWEZWdA7Pz14MGOREKAEA8i+UATfJ6AACItKqGqlBsrtm3RmvL1obO0UxKSNKUgVN0yahLwkJzTP8xSk5M9njy2FNX1/OKZfvjysqu39evX0dETprUfWRyTUwAAGIfK6BADAraoIori7WmbI3WlK3R1qqtYbupFuUXaUT2iD553mF9S73W71+vNWVrQtG5q3qXJGfX2Ql5EzSjYIamD52u6UOn66zBZyktKc3jqWOHtc7qZXGxtGWLc791q7RnjxOZtbVdvycv79irlgUFbPADAMDJiOUVUAIUiAFltWWh2GxfwatrqZMk9Uvtp/G547WnZo/K/eWh70lLStP43PFdNsUZN2Cc0pPTvfoovaq5rVmfH/zcCc0jsbm5cnPoMicjskdoesH0UGyeO/Rc9UuldE5EU5OzY2xxcXhsbtniXB+zXWamNH68NHJk94E5dKhz6CwAAOg9BKiLCFDEu9rmWq3bty4UnKvLVmtf3T5JUnJCss4afJZmDJ2hGQXObULeBCWYBEnSocZDzqVA2jfQObKZzq7Du2Tl/G/dyGhU/1Hd7tY6IH2AZ5/7eALBgIori0OhuXbfWm08uFEtAWcXmvyM/LDYnF4wXQN9Az2eOrpZ61yGpHNgtj/evdt5vd3w4dLEidKECc59++OhQzksFgAAtxGgLiJAEU9aAi36x8F/hFY2V5euVnFlcSgWxw0YFwrNGQUzNHXw1FM6XLSxtVFbq7Z2CdMtlVvUHGgOvW+gb2C3YTqs3zBXd3e11mp39e6w8zbX718fuuRJVkqWzh16rqYPnR46nHZ49nB2oO1Bc7NzHczuVjNrajrel5HhRGV7ZLbfjxvHzrEAAEQTAtRFBChilbVWOw7vCFvZ/Gz/Z6EAzM/I13mF54VWN6cXTI/4imQgGNDu6t1hlxwprirW5orNOtx0OPQ+X7JPE/MmdgnTsQPGKiUx5bTnOFh/UGv3rQ2dt7lu3zpVNjg71aQkpujswWeHVjWnD50etuoLh7XO5j7drWbu3Bm+o2xBQfgqZvt9YaGUwB8rAABRjwB1EQGKWFHhrwiLzbX71upQ4yFJUkZyhs4dcm7Y6uaI7BFRs4JnrVW5v7zbMN1buzf0vkSTqLEDxnYJ04l5E3s817KmqUbr9q0LO2+z/WcmmARNyp8UFptTBk3plciNF62tzmpm51XM9tg83PF3BkpLc87NPDoyx4+XsriKDAAAMY0AdREBimjU0NoQ2nl1ddlqrSlbo93VuyU5UTV54OSw8zYnDZykpITYvApSfUu9tlRu6RKm2w5tU1uwLfS+gqyCUJgW9CvQFxVfaG3ZWm2p2hJ6z5j+Y8LO2zxnyDnypcTk/5f2qvbVzC1bnB1mO8fmjh1SW8cfs4YM6Xpe5sSJzjmbrGYCABCfCFAXEaDwWiAY0JcVX4atbm4q36SADUhydl7tvLJ57pBz+0RUtQZatfPwzi5hWlxZrLqWOg3JHBIWm9OGTlNuRq7XY3uqqUnavt0Jy/Zbe3B2Xs1MSXHOwzx6E6Dx46XsbO/mBwAA3iBAXUSAwm2HGg/pvV3vhYJz3b518rc6/x3MSctxQrPTeZuDMwd7PHF0sdaqtrlW2Wl9s5SslUpLO8Ky862kJHyn2YKCjsNk2zcDmjBBGjFCSux7l3QFAAA9IEBdRIDCDburd2tZ8TIt27JMK0tWKmADoc1wOq9ujh0wls1wIEmqre2IzM6xuXWr1NDQ8b7262a2x2X74/HjndcAAACOhwB1EQGKSLDW6rMDn4Wic+PBjZKkorwiLZi4QNeMv0bnDjlXqUmpHk8KL7W1OdfHPHolc+tWaf/+jvclJEgjR4avYrZHJtfNBAAAp4sAdREBit7SGmjVByUfhKJzb+1eGRnNGj5L8yfM1/wJ8zUud5zXY8JlnTcAOno1c8cOZxfadrm53R8yO2aMlMrfVQAAgAghQF1EgOJ01DbX6o1tb2jZlmV6fdvrqmmuUXpSui4bc5kWTHBWOvN9+V6PCRdUVztBuWNHx0ZAPW0ANHZs10NmJ0xwAhQAAMBtBKiLCFCcrLLaMv11y1+1bMsyvbfrPbUGW5WXkadrx1+r+RPm67IxlykjOcPrMdHLrJUOHOgIzPbYbP/60KHw9xcUdF3JZAMgAAAQjQhQFxGgOB5rrb6o+CJ0aO3afWslOdecXDBxgeZPmK/zh52vxASqIta1tUl79nQfmDt3hm/+k5DgxOSYMR23sWOd+9Gj2QAIAADEDgLURQQoutMWbNNHez/SsuJlemXLK9p5eKckaUbBjND5nGfknyHD7i8xp6HBicnOcdn+uJg+sYcAACAASURBVKTEidB2aWlOTLaHZefbiBHO4bQAAACxjgB1EQGKdv4Wv97e8baWbVmm17a+pqrGKqUkpuiSUZdowYQFunbCtRqaNdTrMXECDh3qPjB37JD27Qt/b05O18Bs/3rIEGelEwAAIJ7FcoAmeT0AcDLK/eV6dcurWrZlmf62829qamtSTlqOrh53teZPmK+5Y+cqKzXL6zFxlGDQuUzJ0XHZHpzV1eHvHzrUCcrLL+8amwMGePMZAAAAcPpYAUXU21q1NXRo7cd7P5aV1fDs4aFDa2ePmK3kxGSvx8RR9u6Vli93bitWSJ3/Z5uY6Fwns/PqZftt9Ggpgz2hAAAAehTLK6AEKKJO0Aa1pmyNXil+Rcu2LFNxZbEkaergqZo/Yb4WTFygswadxfmcUSYQkNaskV57zYnOjRud50eNkq68Upo8uSM4hw+Xkjj+AgAA4JQQoC4iQONTU1uT3t35rpZtWaZXt76qA/UHlGgSddHIizR/wnzNmzBPI3JGeD0mjlJTI731lhOcr78uVVY6q5sXXCBdc4109dXSxIkSf1cAAADQe2I5QFmDgCea25q1dt9arSxZqVV7VmlVySr5W/3KTMnUlWOv1PwJ83XVuKvUP72/16OiE2ulrVs7VjlXrXJ2oR0wQLrqKic4r7hC6s9/bAAAAOgGK6BwRV1znT4u/TgUnKtLV6s50CxJOiP/DM0ZMUfXjr9Wl4y6RKlJqR5Pi85aWqSVKzuic/t25/kpUzpWOWfOdFY+AQAAEHmxvAJKgCIiKhsq9fc9fw8F52f7P1PABpRoEnXOkHN04fALNXvEbM0aPkt5GXlej4ujHDzoHFL72mvS3/4m1dVJqanSpZc60XnVVc51NQEAAOA+AtRFBGh02luzV6v2rAoF55cVX0qSUhNTdV7heZo9fLZmj5itmYUzuUxKFLJW+uwzJzhfe01au9Z5vqCgY5XzkkskX0z+3xwAAEB8IUBdRIB6z1qrrVVbw4Jzd/VuSVJWSpZmDZ+l2cNn68IRF2r60OkcUhul6uuld97puFTK/v3OZkHnndcRnWedxQZCAAAA0YYAdREB6r5AMKDPD34eFpzl/nJJUn5GvmaPmB06pPbMQWcqMYGTAaPVrl0d53KuWOGc39mvn7Nx0NVXO5dLGTjQ6ykBAABwLLEcoOyCiy6a25q1bt+6UHB+uPdD1TbXSpJGZI/QFWOuCAXn+NzxXI8zirW1SR991BGdXzpHRmv8eOlf/sWJzgsukFJSvJ0TAAAAfUPEVkCNMYskXSOp3Fo7uZvXb5P0PUlGUp2kb1hrNx7v57IC2vvqW+r18d6PQ8G5umy1mtqaJElFeUWhFc4LR1yo4dnDPZ4Wx1NVJb35phOcb7whVVdLSUnSnDkdh9aOG+f1lAAAADhVJ7ICaoz5V0n3yemt31trHzXGDJC0VNJISbsl3WStPRzhccPnimCAzpZUL+npHgL0fEmbrbWHjTFXSnrYWnve8X4uAXr6DjUeCtuh9tN9nypgA0owCTp78Nmh4Lxg+AXK9+V7PS6Ow1pp06aOczk/+kgKBqX8fCc2r7lGuuwy51BbAAAAxL7jBagxZrKkJZJmSGqR9Kakr0u6X9Iha+1PjTEPSepvrf2eGzOHZovkOaDGmJGSXusuQI96X39Jm6y1Bcf7mQToyWtsbdSyLctCwbmpfJMkZ4faGQUzQsH5lWFfUb9UKiXaNTZK69ZJH37o3D7+2Fn1lKSzz3aC85prpGnTpIQEb2cFAABA7zuBAL1R0lxr7b1Hvv53Sc2S7pV0kbV2vzFmiKT3rbUTXBn6iGg5B/ReSW/09KIx5n45ta4UTlY7KSXVJVqwdIE2HNigrJQsnT/sfN0y+RZdOPxCTS+YrrSkNK9HxHEcONARmx99JK1fL7W2Oq9NmCDNm+ecx3nFFc5lUwAAABD3kowx6zp9/YS19olOX2+S9CNjTK6kRklXSVonaZC1dv+R9xyQNMiVaTvxPECNMRfLCdALenrPkT/MJyRnBdSl0WLeB7s/0A1/uUGtgVa9dNNLunbCtUpK8Pw/chxDICB98UVHbH74obNzrSSlpkrTp0sPPiidf75zy8vzdl4AAAB4os1aO62nF621m40xP5P0tiS/pA2SAke9xxpjXG8rT2vEGHOmpD9IutJaW+XlLPHEWqvH1z2uf33zXzV2wFgtu3mZxueO93osdKOuTlq9uiM2P/lEqnU2HNagQdKsWc5uteefL51zDrvVAgAA4MRYa/8o6Y+SZIz5saRSSQeNMUM6HYJb7vZcngWoMWa4pJckfdVau9WrOeJNc1uz/uX1f9EfPvuDrhl/jZ697lllp2V7PRbkbBa0Z09HbH74ofT5586GQcZIkydLt97qxOasWdKoUc7zAAAAwMkyxgy01pYf6a5/ljRT0ihJd0r66ZH7Za7PFcFdcJ+TdJGkPEkHJf1AUrIkWWt/Z4z5g6TrJZUc+ZZjLiO3YxOinh2oP6Drn79eH+39SN+/8Pv6z4v/UwmGXWi80toqbdwYfv5mWZnzms8nzZzphOb55zuPs/l7AgAAAJyAE7wMyypJuZJaJT1orX33yDmhz0saLqfDbrLWHor4wJ3niuQuuJFAgHZvbdlaXbf0Oh1uOqwn5z+pGyfd6PVIfc7hw86OtO2xuWaN1NDgvDZ8eEdszpolTZniXJsTAAAAOFknEqDRin8FjgNPb3xa9796v4ZkDdFH93ykswaf5fVIcc9aafv28M2CvvzSeS0x0bkcyte+1hGdhYXezgsAAABEAwI0hrUF2/Tdv31Xv/zkl7p45MV6/sbnlZfBtqiR0NQkffppR2x+9JFUUeG8lpMjfeUrHedvzpjhHGILAAAAIBwBGqMONR7SwhcW6p2d7+jbM76tRy5/RMmJyV6PFVfq6qS//lVaulR6+22pudl5ftw46aqrOlY3i4qkBE61BQAAAI6LAI1Bm8o3af6S+SqtLdWieYt099l3ez1S3GhokJYvd6Jz+XJn5bOwUPr616WLLnKCc+BAr6cEAAAAYhMBGmNe3vyyvvryV9UvtZ8+uOsDzSyc6fVIMa+5WXrzTSc6//pXye93rsH5ta9JCxc60ckKJwAAAHD6CNAYEbRB/ecH/6kffvBDnVdwnl5a+JKGZg31eqyY1doqvfOOE52vvCLV1Ei5udJttznROWeOs5kQAAAAgN5DgMaAuuY63fHKHXql+BXdNfUuPX7140pLSvN6rJgTCEjvv+9E54svSocOOdfevO46JzovvVRK5jRaAAAAIGII0Ci3/dB2LViyQMWVxfrV3F/pWzO+JWOM12PFjGDQ2bV26VLphRekgwedHWrnz3ei84orpNRUr6cEAAAA+gYCNIq9veNtLXxhoRJMgt7+6tu6ZNQlXo8UE6yV1q6VliyRnn9eKiuT0tKka65xovOqq6SMDK+nBAAAAPoeAjQKWWv13x//t777znc1eeBkvbLwFY3qP8rrsaKatdLGjR3RuWuXczjtlVdKP/+5dO21UlaW11MCAAAAfRsBGmUaWxt1/2v369nPn9X1RdfryQVPKjMl0+uxotaXXzqH1y5ZIm3d6mwcdNll0n/8h7RggZST4/WEAAAAANoRoFFkb81eXbf0On26/1P9n4v/j75/4fc537Mb27Y50bl0qbRpk3OJlIsukv7n/5T++Z+lvDyvJwQAAADQHQI0Svx9z991/fPXq7G1UctuXqZ5E+Z5PVJUKSnpiM71653nLrhA+vWvpRtukAYP9nY+AAAAAMdHgEaB33/6e33z9W9qZM5IvX/n+yrKL/J6pKiwb5/0l784h9d+8onz3PTp0v/9v9KNN0rDhnk7HwAAAICTQ4B6qCXQogfefECPr3tcc8fO1XPXP6ectL590mJ5uXONziVLpFWrnM2FzjpL+slPpJtukkaP9npCAAAAAKeKAPVIub9cN/7lRq0sWanvnv9d/fjSHysxIdHrsTzR1CT96U/O4bXvvutcu7OoSHr4YeeyKRMmeD0hAAAAgN5AgHpg/f71WrBkgSoaKvTnf/6zbplyi9cjeWb7dudw2g0bpDFjpP/9v53onDxZYv8lAAAAIL4QoC577h/P6d6/3qu8jDx9eM+HOmfIOV6P5JkXX5TuvltKSpKWLXOu1Ul0AgAAAPErwesB+opAMKDv/e17uvWlWzVt6DStu39dn43PlhbpgQec3WvPOEP67DNp3jziEwAAAIh3rIC64HDjYd360q16c/ub+sa0b+jRuY8qJTHF67E8sWePc4jtJ59I3/629ItfSCl9848CAAAA6HMI0AjbXLFZ85fM1+7q3fp/1/w/3X/u/V6P5Jk33pBuv11qbXUur3LDDV5PBAAAAMBNHIIbQa9ueVXn/eE81TTX6L073+uz8dnWJn3/+9JVV0mFhdKnnxKfAAAAQF9EgEaAtVY/WvkjzV8yX+Nzx2vdfet0wfALvB7LEwcOSJddJv34x9LXvuYcejtunNdTAQAAAPACh+D2svqWet297G698OULuv3M2/XENU8oPTnd67E88f770s03S7W10pNPSnfe6fVEAAAAALzECmgv2nV4l2YtmqWXNr+kRy57RE8veLpPxmcw6Kx4XnqplJMjrVlDfAIAAABgBbTXvLfrPd30l5sUsAG9cdsbunzM5V6P5ImqKumrX3U2HLr5ZumJJ6SsLK+nAgAAABANWAHtBU98+oQuf+ZyDcocpLX3re2z8fnJJ9LZZ0vvviv99rfSn/9MfAIAAADoQID2ghHZIzR/4nx9cu8nGjtgrNfjuM5a6Ve/kmbPlhITpQ8/lL7xDckYrycDAAAAEE2MtdbrGU6Kz+ezfr/f6zFwRE2NdO+90osvSvPmOZsN9e/v9VQAAABA/DLGNFhrfV7PcSpYAcUp27BBmjZNeuUV6Re/cO6JTwAAAAA9IUBx0qyV/vAHaeZMqaHBudzK//pfHHILAAAA4NgIUJwUv9+5pMp990kXXih99pl0wQVeTwUAAAAgFhCgOGGbN0vnnSc9+6z08MPSm29KAwd6PRUAAACAWMF1QHFCnnvOWfXMyJDeeku67DKvJwIAAAAQa1gBxTE1NUn/439It94qTZ3qHHJLfAIAAAA4FQQoerRzpzRrlvT449J3viOtWCEVFHg9FQAAAIBYxSG46NayZc5mQ8Y4j+fN83oiAAAAALGOFVCEaW11LqmyYIE0dqy0fj3xCQAAAKB3sAKKkNJS6eabpQ8/dM77/O//llJTvZ4KAAAAQLwgQCFJevtt6bbbpMZG6c9/lm65xeuJAAAAAMQbDsHt4wIB6Qc/kObOlQYPltatIz4BAAAARAYroH1Yebmz6vnOO86GQ7/9rXOdTwAAAACIBAK0j1q1Slq4UDp8WPrjH6W773Z2vAUAAACASOEQ3D4mGJR+/nPp4osln0/65BPpnnuITwAAAACRxwpoH3LokHTXXdKrr0o33OCsfPbr5/VUAAAAAPoKArSPWLtWuvFGad8+6Ve/kr71LVY9AQAAALiLAO0D3nlHuvpqZ5fbVauk887zeiIAAAAAfREB2gf87GdOfK5fL+Xmej0NAAAAgL6KTYjiXEmJ9O67zkZDxCcAAAAALxGgce6pp5z7u+7ydAwAAAAAkLHWej3DSfH5fNbv93s9RkwIBqWxY6XRo53zQAEAAADEPmNMg7XW5/Ucp4IV0Dj2wQfSrl3O4bcAAAAA4DUCNI4tXixlZ0vXXef1JAAAAABAgMatmhrphRekW26R0tO9ngYAAAAACNC4tXSp1Ngo3X2315MAAAAAgINNiOLUV74i1dVJ//iHZIzX0wAAAADoLWxChKiyebP0ySfO5kPEJwAAAIBoQYDGocWLpaQk6fbbvZ4EAAAAADoQoHGmtVV6+mnpmmukgQO9ngYAAAAAOhCgcebNN6WDB9l8CAAAAED0IUDjzKJF0qBB0pVXej0JAAAAAIQjQONIebn02mvSV78qJSd7PQ0AAAAAhCNA48izz0ptbRx+CwAAACA6cR3QOGGtNGWKlJnpXIIFAAAAQHziOqDw3Lp10hdfONf+BAAAAIBoRIDGiUWLpPR0aeFCrycBAAAAgO4RoHGgsVF67jnp+uul7GyvpwEAAACA7hGgceDll6WaGg6/BQAAABDd2IQoDlx2mbR9u7Rjh5TAXykAAAAAcY1NiOCZ3buld991Lr1CfAIAAACIZiRLjHvqKef+zju9nQMAAAAAjodDcGNYMCiNGSONHSv97W9eTwMAAADADRyCC0+8/75zCO7dd3s9CQAAAAAcHwEawxYvdi67ct11Xk8CAAAAAMdHgMaomhrphRekW2+V0tO9ngYAAAAAjo8AjVFLl0pNTRx+CwAAACB2sAlRjJo5U/L7pc8/l4zxehoAAAAAbmETIrjqyy+l1aud1U/iEwAAAECsIEBj0OLFUlKSdPvtXk8CAAAAACeOAI0xra3S009L11wjDRzo9TQAAAAAcOII0BjzxhtSebl0zz1eTwIAAAAAJ4cAjTGLFkmDBklXXun1JAAAAABwcgjQGHLwoLR8uXTHHc45oAAAAAAQSwjQGPLss1JbG9f+BAAAABCbuA5ojLBWmjxZ6tdP+vhjr6cBAAAA4BWuA4qIW7vWuf4nmw8BAAAAiFUEaIxYtEhKT5cWLvR6EgAAAAA4NQRoDGhokJ57TrrhBucQXAAAAACIRQRoDHj5Zam2ls2HAAAAAMQ2NiGKAf/0T9LOndL27VICf2UAAAAA9GlsQoSI2b1bevddZ/WT+AQAAAAQy0iaKPfUU5Ix0p13ej0JAAAAAJweDsGNYsGgNHq0NH689PbbXk8DAAAAIBpwCC4iYsUKqaSEzYcAAAAAxAcCNIotXizl5EgLFng9CQAAAACcPgI0SlVXSy++KN1yi5Se7vU0AAAAAHD6CNAotXSp1NQk3XOP15MAAAAAQO9gE6Iodd55UkOD9Pnnzi64AAAAACCxCRF62RdfSGvWOKufxCcAAACAeEGARqHFi6WkJOn2272eBAAAAAB6T8QC1BizyBhTbozZ1MPrxhjzmDFmuzHmc2PMOZGaJZa0tkrPPCNde62Un+/1NAAAAADQeyK5AvqkpLnHeP1KSeOO3O6X9HgEZ4kZr78ulZez+RAAAACA+BOxALXWrpR06BhvmS/paev4RFKOMWZIpOaJFYsWSYMHS3OPle4AAAAAEIO8PAe0QNLeTl+XHnmuC2PM/caYdcaYdW1tba4M54UDB6Tly6U77nDOAQUAAACAeBITmxBZa5+w1k6z1k5LiuMye/ZZKRCQ7r7b60kAAAAAoPd5GaBlkoZ1+rrwyHN9krXO7rdf+Yo0caLX0wAAAABA7/MyQP8q6Y4ju+HOlFRjrd3v4TyeWrNG+vJLVj8BAAAAxK+IHc9qjHlO0kWS8owxpZJ+IClZkqy1v5P0uqSrJG2X1CCpT6fX4sVSerq0cKHXkwAAAABAZBhrrdcznBSfz2f9fr/XY/SqhgZpyBBpwQLpqae8ngYAAABANDPGNFhrfV7PcSpiYhOiePfSS1JtLYffAgAAAIhvrIBGgUsvlXbvlrZtkxL4KwEAAAAAx8AKKE7Zrl3Se+9Jd91FfAIAAACIbySPx556SjJGuvNOrycBAAAAgMjiEFwPBYPS6NHS+PHS2297PQ0AAACAWMAhuDglK1ZIJSXSPfd4PQkAAAAARB4B6qFFi6ScHOfyKwAAAAAQ7whQj1RXO5dfufVWKS3N62kAAAAAIPIIUI8sWSI1NXH4LQAAAIC+g02IPDJjhhOgGzc6u+ACAAAAwIlgEyKclE2bpLVrndVP4hMAAABAX0GAemDxYikpSbrtNq8nAQAAAAD3EKAua22VnnlGmjdPys/3ehoAAAAAcA8B6rLly6WKCunuu72eBAAAAADcRYC6bPFiafBgae5crycBAAAAAHcRoC46cMBZAb3jDuccUAAAAADoSwhQFz37rBQIcPgtAAAAgL6J64C6xFpp0iSpf3/pww+9ngYAAABArOI6oDiu1aulzZtZ/QQAAADQdxGgLlm8WMrIkG66yetJAAAAAMAbBKgLGhqk556TbrhB6tfP62kAAAAAwBsEqAteekmqq5PuucfrSQAAAADAO2xC5IJLLpFKSqTt2yVjvJ4GAAAAQCxjEyL0aNcuacUKZ/Mh4hMAAABAX0aARtiTTzrheeedXk8CAAAAAN7iENwICgalUaOkiROlt97yehoAAAAA8YBDcNGt996T9uxh8yEAAAAAkAjQiFq0SMrJkebP93oSAAAAAPAeARohhw87l1+57TYpLc3raQAAAADAewRohCxZIjU3O7vfAgAAAADYhChiZsxwAnTDBi6/AgAAAKD3sAkRwmzaJK1dy7U/AQAAAKAzAjQCFi+WkpOd8z8BAAAAAA4CtJe1tEjPPCNde62Un+/1NAAAAAAQPQjQXrZ8uVRRwbU/AQAAAOBoBGgvW7xYGjJEuuIKrycBAAAAgOhCgPaiAwek11+X7rhDSkryehoAAAAAiC4EaC965hkpEODanwAAAADQHa4D2kuslc44QxowQPrwQ6+nAQAAABCvuA4otHq1VFzM5kMAAAAA0BMCtJcsWiRlZEg33eT1JAAAAAAQnQjQXuD3S0uWSDfeKGVleT0NAAAAAEQnArQXvPSSVFfH5kMAAAAAcCwEaC9oaZEuuECaPdvrSQAAAAAgerELLgAAAADEEHbBBQAAAABEDWPM/2eM+cIYs8kY85wxJs0YM8oYs9oYs90Ys9QYk+L2XAQoAAAAAMQRY0yBpG9LmmatnSwpUdLNkn4m6ZfW2rGSDku61+3ZCFAAAAAAiD9JktKNMUmSMiTtl3SJpBeOvP6UpAVuD0WAAgAAAEAcsdaWSXpE0h454Vkj6VNJ1dbatiNvK5VU4PZsBCgAAAAAxJYkY8y6Trf7O79ojOkvab6kUZKGSvJJmuvBnF0keT0AAAAAAOCktFlrpx3j9X+StMtaWyFJxpiXJM2SlGOMSTqyClooqSzyo4ZjBRQAAAAA4sseSTONMRnGGCPpUklfSloh6YYj77lT0jK3ByNAAQAAACCOWGtXy9lsaL2kf8jpvickfU/Sg8aY7ZJyJf3R7dmMtdbt33lafD6f9fv9Xo8BAAAAAJ4wxjRYa31ez3EqWAEFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALiCAAUAAAAAuIIABQAAAAC4ggAFAAAAALgiogFqjJlrjNlijNlujHmom9eHG2NWGPP/s3fv8XHVBf7/35+55X6/NKUtJAVaQJBytSAiolXAFdhFrHdYLzz2t4qw3324i6yuXdHVddeviqLSVfaHflFkuYh+fytIWbCsUKFgoQVaSm8ktLm0SZPmMpnb5/fHmSQzySSdhMw5M5PX8/E4j7nknMknAbGvfs45H/MnY8wLxpjLcjkeAAAAAIB3chagxhi/pNskXSrpFEkfMsacMmm3L0q6x1p7hqQPSvpBrsYDAAAAAPBWLmdAz5X0qrV2t7U2IuluSVdM2sdKqk4+r5G0P4fjAQAAAAB4KJDDz14iqT3ldYekt0zaZ52k3xljrpdUIeldmT7IGHOdpOskKRQKzftAAQAAAAC55/VNiD4k6f+11i6VdJmknxljpozJWrveWnu2tfbsQCCXzQwAAAAAyJVcBujrkpalvF6afC/VJyXdI0nW2qcklUpqzOGYAAAAAAAeyWWAPiPpRGNMmzEmJOcmQ7+etM9rkt4pScaYk+UEaE8OxwQAAAAA8EjOAtRaG5P0WUkPS3pZzt1uXzTGfMUYc3lyt7+V9GljzPOSfiHpWmutzdWYAAAAAADeMYXWexUVFXZoaMjrYQAAAACAJ4wxw9baCq/HMRde34QIAAAAALBAEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAgKwZY06b67EEKAAAAABgNn5gjHnaGPPXxpia2RxIgAIAAAAAsmatfZukj0haJulZY8zPjTFrsjmWdUABAAAAoIDkyzqgxhi/pCsl3SppQJKRdLO19v7pjmEGFAAAAACQNWPMm40x35b0sqSLJb3PWnty8vm3Zzo24ML4AAAAAADF43uSfixntnNk7E1r7X5jzBdnOpBTcAEAAACggOTLKbhzwQwoAAAAACBrxpgTJX1d0imSSsfet9YuP9qxXAMKAAAAAJiN/5D0Q0kxSe+Q9FNJ/yebAwlQAAAAAMBslFlrH5VzSec+a+06Se/N5kBOwQUAAAAAzMaoMcYnaacx5rOSXpdUmc2BWc2AGmNuMMZUG8dPjDHPGWPe/QYGDAAAAAAoTDdIKpf0OUlnSfqopGuyOTDbU3A/Ya0dkPRuSXWSPibpG7MfJwAAAACgUBlj/JLWWmsHrbUd1tq/tNZeZa3dlM3x2QaoST5eJuln1toXU94DAAAAACwA1tq4pAvmeny214A+a4z5naQ2SV8wxlRJSsz1mwIAAAAACtafjDG/lvSfkobG3rTW3n+0A7MN0E9KWiVpt7V22BhTL+kv5zJSAAAAAEBBK5V0xcaLNwAAIABJREFUSNLFKe9ZSfMWoOdJ2mKtHTLGfFTSmZK+O9tRAgAAAAAKm7V2zpOR2QboDyWdbow5XdLfSvqxnMVG3z7XbwwAAAAAKDzGmP+QM+OZxlr7iaMdm22Axqy11hhzhaTvW2t/Yoz55CzHCQAAAAAofP835XmppD+XtD+bA7MN0CPGmC/IWX7lbclFR4OzGiIAAAAAoOBZa+9LfW2M+YWk/8nm2GyXYVkraVTOeqCdkpZK+tfZDBIAAAAAUJROlNSczY7G2imn7mbe0ZhFks5JvnzaWts9t7G9MRUVFXZoaOjoOwIAAABAETLGDFtrKzz8/keUfg1op6QvTJ4ZzSSrU3CNMR+QM+P5uCQj6XvGmM9ba++d/XABAAAAAIXKWls112OzvQb0HySdMzbraYxpkrRBEgEKAAAAAAuIMebPJf23tbY/+bpW0kXW2l8d7dhsrwH1TTrl9tAsjgUAAAAAFI8vj8WnJFlrD0v6cjYHZjsD+pAx5mFJv0i+Xivpv2Y1RAAAAABAMcg0GZnd5Z2zuAnRVZLemnz5hLX2gezGNr+4CREAAACAhSwPbkJ0h6TDkm5LvvUZSfXW2muPemy2AZovCFAAAAAAC1keBGiFpC9Jepecu+E+Iulr1tqjhtqMAZrh9rrjX5JkrbXVcxrxG0CAAgAAAFjIvA7QN2LGGwlZa6ustdUZtiov4hMAAAAA4C1jzCPJO9+Ova5L3jPoqLiTLQAAAABgNhqTd76VJFlr+yQ1Z3MgAQoAAAAAmI2EMebYsRfGmFZlvnRzimyXYQEAAAAAQJL+QdL/GGN+L+f+QG+TdF02B3IXXAAAAAAoIPlwEyJjTLOc6PyTpDJJ3dbajUc7jhlQAAAAAEDWjDGfknSDpKWStkhaLekpSRcf7ViuAQUAAAAAzMYNks6RtM9a+w5JZ0g6PPMhDgIUAAAAADAbYWttWJKMMSXW2u2SVmZzIKfgAgAAAABmoyO5DuivJD1ijOmTtC+bA7kJEQAAAAAUkHy4CdEYY8zbJdVIeshaGzna/syAAgAAAADmxFr7+9nszzWgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABXEKAAAAAAAFcQoAAAAAAAVxCgAAAAAABX5DRAjTGXGGN2GGNeNcbcNM0+HzDGvGSMedEY8/NcjgcAAAAAip0xZqUxZkvKNmCMudEYU2+MecQYszP5WOf62Ky1uflgY/ySXpG0RlKHpGckfcha+1LKPidKukfSxdbaPmNMs7W2e6bPraiosENDQzkZMwAAAADkO2PMsLW2Ist9/ZJel/QWSZ+R1Gut/UZygrDOWvv3ORzqFLmcAT1X0qvW2t3W2oikuyVdMWmfT0u6zVrbJ0lHi08AAAAAwKy8U9Iua+0+OT12Z/L9OyVd6fZgchmgSyS1p7zuSL6XaoWkFcaYPxhjNhljLsnheAAAAABgofmgpF8kny+y1h5IPu+UtMjtwQTc/oYZvv+Jki6StFTSRmPMadbaw6k7GWOuk3SdJIVCIbfHCAAAAAD5JGCM2Zzyer21dv3knYwxIUmXS/rC5K9Za60xJjfXY84glwH6uqRlKa+XJt9L1SHpj9baqKQ9xphX5ATpM6k7JX+Z6yXnGtCcjRgAAAAA8l/MWnt2FvtdKuk5a21X8nWXMWaxtfaAMWaxJNcvgczlKbjPSDrRGNOWLO8PSvr1pH1+JWf2U8aYRjmn5O7O4ZgAAAAAYKH4kCZOv5WcHrsm+fwaSQ+6PaCcBai1Nibps5IelvSypHustS8aY75ijLk8udvDkg4ZY16S9Jikz1trD+VqTAAAAACwEBhjKuSsSHJ/ytvfkLTGGLNT0ruSr90dV66WYckVlmEBAAAAsJDNZhmWfJPLU3ABAAAAABhHgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAzoODBx/Uyy9/XNZar4cCAAAAAHmLAJ0Ho6MH1NX1Mw0MPOX1UAAAAAAgbxGg86Cl5WPy+2vU0XGr10MBAAAAgLxFgM4Dv79Cixd/Sj0992p09HWvhwMAAAAAeYkAnSdLlvy1pIT27/+R10MBAAAAgLxEgM6TsrLlamh4n/bvv13xeNjr4QAAAABA3iFA59HSpZ9TNNqjnp5fej0UAAAAAMg7BOg8qq29WOXlp6ij43ssyQIAAAAAkxCg88gYoyVLrtfg4LMsyQIAAAAAkxCg84wlWQAAAAAgMwJ0nrEkCwAAAABkRoDmAEuyAAAAAMBUBGgOsCQLAAAAAExFgOYIS7IAAAAAQDoCNEdYkgUAAAAA0hGgOcKSLAAAAACQjgDNIZZkAQAAAIAJBGgOsSQLAAAAAEwgQHOMJVkAAAAAwEGA5hhLsgAAAACAgwB1AUuyAAAAAAAB6oqJJVluZUkWAAAAAAsWAeqCiSVZnmNJFgAAAAALFgHqEpZkAQAAALDQEaAuYUkWAAAAAAsdAeqiJUs+I5ZkAQAAALBQEaAuKitrY0kWAAAAAAsWAeoylmQBAAAAsFARoC5jSRYAAAAACxUB6jKWZAEAAACwUBGgHmBJFgAAAAALEQHqgdQlWcLhDq+HAwAAAACuIEA9wpIsAAAAABYaAtQjY0uyHDiwniVZAAAAACwIBKiHWJIFAAAAwEJCgHqIJVkAAAAALCQEqIeMMVq69HMsyQIAAABgQSBAPbZo0UcVCNSyJAsAAACAokeAeszvr1BLyydZkgUAAABA0SNA8wBLsgAAAABYCAjQPMCSLAAAAAAWAgI0T7AkCwAAAIBiR4DmCZZkAQAAAFDsCNA8wZIsAAAAAIodAZpHWJIFAAAAQDEjQPMIS7IAAAAAKGYEaJ5hSRYAAAAAxYoAzTPOkiyXsyQLAAAAgKJDgOahpUuvZ0kWAAAAAEWHAM1DE0uyfJclWQAAAAAUDQI0D00syfInDQw86fVwAAAAAGBeEKB5amJJlu95PRQAAAAAmBcEaJ5iSRYAAAAAxYYAzWMsyQIAAACgmBCgeYwlWQAAAAAUEwI0z7EkCwAAAIBiQYDmOZZkAQAAAFAsCNA8x5IsAAAAAIoFAVoAWJIFAAAAQDEgQAsAS7IAAAAAKAYEaIFgSRYAAAAAhY4ALRAsyQIAAACg0BGgBYQlWQAAAAAUMgK0gLAkCwAAAIBCRoAWEJZkAQAAAFDICNACw5IsAAAAAAoVAVpgWJIFAAAAQKEiQAsQS7IAAAAAKEQEaAGaWJLldpZkAQAAAFAwCNAC5SzJcpAlWQAAAAAUDAK0QLEkCwAAAIBCQ4AWKJZkAQAAAFBoCNACxpIsAAAAAAoJAVrAWJIFAAAAQCHJaYAaYy4xxuwwxrxqjLlphv2uMsZYY8zZuRxPMWJJFgAAAACFImcBaozxS7pN0qWSTpH0IWPMKRn2q5J0g6Q/5mosxYwlWQAAAAAUilzOgJ4r6VVr7W5rbUTS3ZKuyLDfLZL+RRL1NEdLl36OJVkAAAAA5L1cBugSSe0przuS740zxpwpaZm19v+b6YOMMdcZYzYbYzbHYrH5H2mBq619h8rL38SSLAAAAADymmc3ITLG+CT9b0l/e7R9rbXrrbVnW2vPDgQCuR9cgXGWZLmeJVkAAAAA5LVcBujrkpalvF6afG9MlaRTJT1ujNkrabWkX3MjormZWJLlVq+HAgAAAAAZ5TJAn5F0ojGmzRgTkvRBSb8e+6K1tt9a22itbbXWtkraJOlya+3mHI6paE0syXIfS7IAAAAAyEs5C1BrbUzSZyU9LOllSfdYa180xnzFGHN5rr7vQsaSLAAAAADymSm0m9ZUVFTYoaEhr4eRt7ZuvVIDA3/Q6tXt8vtLvR4OAAAAgHlmjBm21lZ4PY658OwmRMgNlmQBAAAAkK8I0CLDkiwAAAAA8hUBWmRYkgUAAABAviJAixBLsgAAAADIRwRoEfL7K7R48adYkgUAAABAXiFAi9Qxx/y1WJIFAAAAQD4hQItUWVmbGhou14EDtyseD3s9HAAAAAAgQIsZS7IAAAAAyCcEaBFjSRYAAAAA+YQALWIsyQIAAAAgnxCgRY4lWQAAAADkCwK0yDlLsnxaPT3/qVdf/V+Kx4e9HhIAAACABSrg9QCQe62t6xSPD6mj49s6dOj/auXKO1Rbe4HXwwIAAACwwDADugD4/eVaseI2nX76f8vaqLZsuVA7d97IbCgAAAAAV5lCuztqRUWFHRoa8noYBSsWG9Tu3Tdp//7bVFp6vE466Q7V1l7o9bAAAAAAZMkYM2ytrfB6HHPBDOgCEwhUasWK7+v00x+TZLVly9u1c+fnFI8T9QAAAAByixnQBSweH9Lu3V/Q669/T6Wly7Vy5U9UV3eR18MCAAAAMANmQFGQ/P4KnXjirVq16veSjJ5//h165ZXPKhYb9HpoAAAAAIoQM6CQJMXjw9qz5x/U0fFdlZYep5Ur71Bd3Tu8HhYAAACASZgBRcHz+8t1wgnf1qpVG2VMUM8/f7FeeeWvFYsd8XpoAAAAAIoEM6CYwpkN/ZI6Or6tkpJjddJJP1Fd3Tu9HhYAAAAAMQOKIuPMhn5LZ5zxP/L5SvT88+/Sjh1/pVhswOuhAQAAAChgzIBiRvH4iPbu/Ue1t39LJSXLtHLlj1Vfv8brYQEAAAALFjOgKFp+f5mOP/5fdcYZf5DPV6YXXni3duy4TrFYv9dDAwAAAFBgmAFF1pzZ0HVqb/83lZQck5wNfY/XwwIAAAAWFGZAsSA4s6H/ojPPfFJ+f5VeeOESbd/+KWZDAQAAAGSFGVDMSTwe1r59/6TXXvumQqHFWrny39XQcKnXwwIAAACKHjOgWHD8/lItX/51nXnmJgUCNdq69TJt3/4JRaOHvR4aAAAAgDzFDCjesERiVHv3fkWvvfYvCoUWaeXK9WpoeK/XwwIAAACKEjOgWNB8vhItX/41nXnmJgWD9dq69c/08svXKBrt83poAAAAAPIIM6CYV4nEqPbt+6r27fu6QqFmrVhxuxob3+f1sAAAAICiwQwokOTzlait7RadddbTCgYbtW3b5Xr55Y8pGu31emgAAAAAPMYMKHImkYho376v6bXX/lmBQINWrrxdjY1XeD0sAAAAoKAxA7rQxWLSvn1ejyLv+HwhtbX9k84882mFQi3atu1KvfTSRxSNHvJ6aAAAAAA8QIDOh3/8R+mMM6QNG7weSV6qqjpDZ531tFpb16mn5x49/fQp6ul5wOthAQAAAHAZATofPvUpackS6ZJLpFtvlQrstGY3+HwhtbZ+WWedtVklJcfoxRf/Qi+99CFFIge9HhoAAAAAlxCg82H5cunJJ6X3vU+64QYnSEdHvR5VXqqsPF1nnvm0Wlu/op6e+/TMM6eou/ter4cFAAAAwAXchGg+JRLSunXSLbdI558v3Xef1NLi9ajy1uDgC9q+/S81OPicGhv/XEuX/o1qai6QMcbroQEAAAB5q5BvQkSA5sJ//qd07bVSfb30q19JZ53l9YjyViIRVXv7N7Vv39eVSAyptLRNixZ9XC0tH1dZ2XKvhwcAAADkHQLURQURoJK0ZYt0xRVSd7f0H/8hffCDXo8or8Vigzp48H51dv5Uhw//tySrmpq3qaXlGjU1Xa1AoNrrIQIAAAB5gQB1UcEEqOTE5/vfLz3xhPSFL0hf/ark47LbowmHX1NX1/9RZ+edGhl5RT5fmRob/1wtLdeoru6dMsbv9RABAAAAzxCgLiqoAJWkSES6/npp/Xrpz/5MuusuqZrZvGxYazUw8Ed1dd2p7u67FYsdVii0RIsWfVQtLdeoouJkr4cIAAAAuI4AdVHBBajkLMvywx9Kn/uctGKF9OCD0oknej2qghKPh3Xo0G/U2XmnensfkhRXVdU5amm5Rs3NH1Qw2OD1EAEAAABXEKAuKsgAHfPYY9LVV0vxuHTPPdKaNV6PqCCNjnaqu/vn6uy8U0NDL8iYoBoa3qeWlmtUX3+pfL6g10MEAAAAcoYAdVFBB6gk7dkjXX659NJL0re+5awbyrIjc3bkyBZ1dd2prq67FI32KBhsUnPzh9XSco0qK1expAsAAACKDgHqooIPUEk6ckT6+MedJVquvVb60Y+kkhKvR1XQEomoensfVlfXnTp48NeyNqKKitOSp+h+WCUli70eIgAAADAvCFAXFUWASlIiIf3TP0lf+Yp03nnS/fdLLS1ej6ooRKO96u7+pTo779SRI3+U5FN9/XvU0nKNGhqukN9f6vUQAQAAgDnLJkCNMbWSfizpVElW0ick7ZD0S0mtkvZK+oC1ti+ng508LgLUY/feK11zjVRX58yInn221yMqKkND29XV9VN1df1Mo6Md8vtr1Ny8Vi0t16i6+jxO0QUAAEDByTJA75T0hLX2x8aYkKRySTdL6rXWfsMYc5OkOmvt37sw5IlxEaB5YMsW6YornHVDf/IT6cMf9npERcfauPr6HlNX153q6blficSwyspO1KJFH1dLy8dUWnqc10MEAAAAsnK0ADXG1EjaImm5TQk+Y8wOSRdZaw8YYxZLetxauzL3I04ZGwGaJ7q7pfe/X3riCenv/1762tckv9/rURWlWOyIenruU1fXnTp8+HFJUm3tO9TSco0aG69SIFDp7QABAACAGRhjIpK2pry13lq7PuXrqyStl/SSpNMlPSvpBkmvW2trk/sYSX1jr91CgOaTSMRZK/T226X3vle66y6ppsbrURW1kZG96ur6mTo771Q4vEs+X4Wamq5SS8s1qq29SMb4vB4iAAAAkCaLGdCzJW2S9FZr7R+NMd+VNCDp+tTgNMb0WWvrcj/ilLERoHnohz90QvSEE6Rf/1o68USvR1T0rLUaGHhSnZ13qrv7l4rHB1RScqwWLfqYWlo+rvLyFV4PEQAAAJCUVYC2SNpkrW1Nvn6bpJsknSBOwZ2dBRGgkvT4484pufG49MtfSu9+t9cjWjDi8REdPPigurruVG/v7yQlVFV1rmprL1JNzfmqrj5PoVCz18MEAADAApXlTYiekPQpa+0OY8w6SWP7H0q5CVG9tfbvcjzc9HERoHlszx7n5kQvvij9279JN94ocddWV42OHlBX113q6blXg4PPydqoJKms7ARVV5+fDNLzVVFxiozhml0AAADkXpYBukrOMiwhSbsl/aUkn6R7JB0raZ+cZVh6czzc9HERoHlucFD6+MelBx6Qrr1W+tGPpJISr0e1IMXjIxocfE79/U9qYOBJ9fc/qWi0W5Lk91erunq1qqvPS0bpWxQIcP0uAAAA5l82AZqvCNBCkEhIt9wirVsnrV4t3X+/tHix16Na8Ky1Cod3pwXp0NBWOev8GlVUnJo2S1pWdjzrjgIAAOANI0BdtCADdMx99zmzoXV1zozoOed4PSJMEosNaGDg6fEgHRh4SvH4gCQpGGxKBul5qq4+X1VVZ8vvL/N4xAAAACg0BKiLFnSAStLzzzvXhXZ1ST/5ifThD3s9IszA2oSGhl5KC9KRkVckScYEVFl55vgMaXX1eSotXerxiAEAAJDvCFAXLfgAlaSeHunqq6Xf/176u7+T/vmfJT83wCkUkUiPBgY2jUfpkSNPK5EIS5JKSpalnbZbWXm6fL6gxyMGAABAPiFAXUSAJkWj0g03OGuGXnaZ9POfSzXc9KYQJRIRDQ4+nwzSpzQw8KRGR9slST5fmaqqzk2ZJV2tUKjR4xEDAADASwSoiwjQSX70I+n666UTTpAefFBascLrEWEehMPtGhh4avwGR4ODf5K1MUlSWdmK8SCtqTlf5eUnyxifxyMGAACAWwhQFxGgGfz+99L73+/Miv7yl9J73uP1iDDP4vFhHTmyOS1Ko9GDkiS/v0Y1Neeptvadqq9fo4qKN3O3XQAAgCJGgLqIAJ3G3r3OzYm2bZP+9V+lv/kbiQgpWtZajYy8On4daX//ExoeflmSFAwuUl3du1Rfv0Z1dWtUUnKMx6MFAADAfCJAXUSAzmBwULr22onlWm6/XSot9XpUcEk43KG+vg3q63tEfX0bFI12S5LKy09RXd0a1de/W7W1b5ffX5D/rQIAAEASAeoiAvQoEgnpllukdeukt7xFuv9+6RhmwBYaZ/mXrert/Z36+h5Rf/8TSiTCMiao6urzx2dHq6rOkjHcQRkAAKCQEKAuIkCzdP/9zixoTY30q19J55zj9YjgoXg8rP7+/0nOjj6iwcE/SZICgTrV1l6s+vp3q65ujcrK2jweKQAAAI6GAHURAToLL7zgXBd64ID0/e9LH/mIVFbm9aiQByKRbvX1PToepKOjHZKk0tLjx2dHa2svVjBY6/FIAQAAMBkB6iICdJYOHnTukPv73zvxedFF0qWXOtsJJ3g9OuQBa62Gh3eMx+jhw48pHh+U5FN19bmqq3OCtLp6tXy+oNfDBQAAWPAIUBcRoHMQi0mPPCL99rfSQw9JO3c67x9//ESMXnSRVF7u6TCRHxKJqAYGNo0H6cDA05IS8vsrVVt70XiQlpefxHIvAAAAHiBAXUSAzoNdu5wY/e1vpccek0ZGpJIS6e1vnwjSFStYxgWSpGj0sA4ffkx9fY+ot/d3Cod3SZJKSpaOx2hd3bsUCjV5PFIAAICFgQB1EQE6z8JhaePGiSDdscN5v63NCdFLLpEuvliqKMh/v5EDIyN7xmdH+/oeVSzWJ0mqrFyVjNF3q6bmAvn9LAEEAACQCwSoiwjQHNuzxzlN97e/lR59VBoelkIh6cILJ4L05JOZHYUkydq4jhx5Tn19v1Nv7yMaGHhS1kbl85WqpuZt4zOklZVvljE+r4cLAABQFAhQFxGgLhodlZ54YuLa0Zdect4/7jgnRC+91JkdrarydpzIG7HYoPr7NyZP131Ew8MvSpICgQaVl5+o0tJWlZQcp9LS41Ra2pp8PE5+f0H+9xMAAMATBKiLCFAP7duXPjs6OCgFg9IFF0xcO/qmNzE7inGjo/vV17dB/f1PaGRkj8LhvRodfU3WRtP2CwYbM4bpWLCyHAwAAMAEAtRFBGieiESkP/xh4trRbduc95cunZgdfde7pOpqb8eJvGNtQpHIAYXD+1K2vRodnXieSIykHeP3V6fFqROrE6EaDDZxR14AALBgEKAuIkDzVHu79PDDTow+8oh05IgUCEhvfetEkL75zcyO4qistYpGD2YM07HHeHwg7RifryxjmI49D4UWcw0qAAAoGgSoiwjQAhCNSk8+OXG67vPPO+8fc0z67Ggtp1VibqLRwxnCdCJYo9GDafsbE1RJybKMp/c64bpUPl/Qo58GAABgdghQFxGgBWj//okYfeQRqb9f8vul886buHZ01SpmRzFv4vGhGU/xjUQOTDrCr/LyFaqoeLMqK9+cfDxNJSXHcmovAADIOwSoiwjQAheLSZs2TVw7+qc/Oe83NztBunq183j22aw9ipxJJEYVDrePh+nIyG4NDW3T0NALCof3ju/n99eosvI0VVSclhKnpyoQ4NpmAADgHQLURQRokensdK4dffRRJ0x37nTe9/ul005zgnRsW7GCWVLkXCw2oKGhbRocfEFDQy9oaGirBgdfSLvutLS0NW22tKLiNJWVnSCfL+DhyAEAwEJBgLqIAC1yBw9KTz8tPfWUE6R//KNzQyNJqqtLD9Jzz+U6UrjCWqvR0dc0OLhVQ0MvjMfp8PArkuKSJJ+vVOXlb0rOmI7F6WkKhZq9HTwAACg6BKiLCNAFJh6Xtm93YnRse/FFaezf25NPnjh1d/Vq6ZRTnNlTwAXxeFjDwy8no3QiTqPRrvF9gsFFadeVVlS8WeXlJ8vvL/Vw5AAAoJARoC4iQKH+fumZZ9Kj9NAh52uVlc7M6Ni1pG95i9TU5O14seBEIt3jp+6Oxenw8ItKJMLJPSbf9Og0VVa+mZseAQCArBCgLiJAMYW10q5dE6ftbtrkLP0Sd06N1PHHp5+6e/rpUpAlN+CuRCKmkZFXNTSUfhpv+k2PqseDlJseAQCA6RCgLiJAkZXhYenZZ50YfeopZ+vsdL5WWurcZTc1Spcs8Xa8WLDSb3o0EaepNz1y1itdpmBwkUKhFoVCUx+DwUWc1gsAwAJBgLqIAMWcWCu1t6eftvvss1Ik4nx96dKJ03ZXr5bOPNMJVcADk296NDT0oiKR/YpEuhSJdCoW68t4XCBQO2OkOo8tCgab5fNxFgAAAIWKAHURAYp5MzoqbdmSHqV79zpfCwalVavS77iJlOZZAAAgAElEQVR73HGcuou8kEiMKhLpViTSOR6l0WhX2uux56kzqakCgYajROqi5GMjy8sAAJBnCFAXEaDIqc5OZ+mXsVN3n3nGOZ1Xknw+Z6a0tdWJ0dbW9G3ZMgIVeSceH8kQqamhOvFeIjGc4ROMgsGmKZE6MdPqbGVly+X3l7v+8wEAsBARoC4iQOGqWEzatk167jlndjR1e/11KZGY2Nfnc64lnSlQQyHXfwTXWCsNDkrd3VJPj/M4tg0MSIsXS8uXS21tzlZRkP/NLGqx2OA0s6lTw9Xa0SnHl5Qcq/Lyk6ZsoVALd/cFAGAeEaAuIkCRN6JRqaNjapiObR0d6YFqzNEDtaTE5R/iKMLhqTE50+twOPPnBAJOzKdqapqI0dRt+XLp2GOZTc5j1lrF4wPjQTo6ul8jIzs1PLw9ue1QIjHx32m/vypjmJaVHS+fL8/+nQcAoAAQoC4iQFEwxgJ1377MgdrePjVQjzlm+kA99tg3HqixmHTwYHYx2d0tHTmS+XNKSqTmZicim5vTt8nvNTU5N3Tq6ZH27JnYdu+eeP7aa+mBOna6c6ZAbWtzZlN9vjf2u0DOODdRej0lSJ1tZGSHRkc7Uvb0qaxseYYwXalQqNGz8QMAkO8IUBcRoCga0ahzGu9MgTq2lumY6QK1pUU6fHjmmOzulnp7M4/F75caG7OLyeZmqarKCeb5Eos5v4vUQE3d9u9P37+kxPm5pwvUurr5HR/mTSx2RCMjr0yK0x0aHn4l7bTeQKAhw6zpSpWWtnFTJADAgkeAuogAxYIxFmV792aO1Ndemxqoqerrs4vJ5mYn2PJ5RjEcdn7m6QK1b9KyJNXVE6fzTo7T1lapnJvl5Btr4wqH902KUud5NNo9vp8xQZWVnZgxTgOBag9/AgAA3EOAuogABZJiMWdmcO9eqavLicixoGxoWFjXUPb3Zw7T3bud38/ISPr+ixZNve505Upna2xk9jTPRKO9KUG6I+WU3lclTfwlTCi0eEqUlpefpJKSZTImj/+CBQCAWSJAXUSAApgVa51An272dPJMcn29E6InneRsY8+XL19YUV8AEomIRkZ2j19fmnpabyx2eHw/n69MZWXHq7T0eJWVLU8+H3tslc9XxHenBgAUJQLURQQogHkVizkR+sor0vbtzrZjh/PY2TmxXyAgHX/81DBdudKJVuQNa62i0e5Js6W7NDKyS+HwbiUSqTPiRiUly9KitKxseTJWj1cwWOfZzwEAwHQIUBcRoABc098/EaNjj9u3S6++KkUiE/s1NaVH6djz1lYnXJE3rLWKRDoVDu9ORuluhcO7xp9Ho11p+wcCtZNmTieel5QslTF+j34SAMBCRoC6iAAF4LlYzLm2dHKY7tjh3H14TCgknXBC5lnTmhrPho/pxWKDCof3jM+Wpj/ulbXR8X2NCaq0tDVl5jR9FtXvL8g/FwAACgAB6iICFEBe6+2dGqXbt0u7dqWvddrSknnW9NhjnWVxkHesjWt0tGPSzOlYnO5Ku+5UkoLBRZNO6Z24/jQUapHhZlcAgDkiQF1EgAIoSNGoc1feyWG6fXv6MjKlpdKKFemzpStWOKf51tQ4S8wQqHkpGu3LOHM6MrJbo6PtkhLj+/p85ckwdaI0FFqsYLAxuTUpGGxUKNQkv7+aUAUATEGATvfhxlwi6buS/JJ+bK39xqSv/y9Jn5IUk9Qj6RPW2n0zfSYBCqCoWCsdPJj5dN7du6VEYuoxlZVOjKZutbVT38u01dZKVVVErMsSiYjC4X3js6UTM6e7NTKyW4lE5v9fMyY4JUxTA3Xq+w3c1RcAFgACNNMHO3dmeEXSGkkdkp6R9CFr7Usp+7xD0h+ttcPGmP9H0kXW2rUzfS4BCmDBGB11bni0c6dzam9/f3bb6OjRP7uq6uihOtPXq6slH2trzgdrrRKJYUUiPYpGDyoazfyY+vVYrHfaz/P7a2YI1KkByywrABQeAjTTBxtznqR11tr3JF9/QZKstV+fZv8zJH3fWvvWmT6XAAWAoxgddUL08OHso3VsGzsm9S6/00mN2PJyqazMOYV4bJvN62z2ZdZ2XCIRUyzWmzFOp4tXazP/xcTELOvkUM0Usc7GLCsAeKuQAzSX6wMskdSe8rpD0ltm2P+Tkn6b6QvGmOskXSdJoRD/pwcAMyopkZqbnW2uwuHZhevIiHPMwIDU3e08H3tvbMtmZnYmgcDc47Wy0pm1ra52wjnT84qKgpnV9fkCCoWaFQpl98/YWqt4fCir2dXBwT8lZ1n7pv08v786LUinnibcmBaxgUCdjCmM3y0AaGTEuQxm1y7pyBHpIx/xekRFJS8WqDPGfFTS2ZLenunr1tr1ktZLzgyoi0MDgIVpLN4WLZq/z0wknAidHKeTQ3WuX+vvl7q6Mu8XjR59fMZMhOp0kZr6fKavlZY6n5cnjDEKBCoVCFSqrKwtq2OcWdZDKXGaaetRJNKpoaFtikYPKpEYnubTfAoG648aq6nv+/2VnBoMIDeslQ4dcgJzbBsLzl27pP37J/atqyNA51kuA/R1SctSXi9NvpfGGPMuSf8g6e12uvODAACFz+dzZiPLypz/Q3fT6Kjzt9gDAxOP2T7v7Ex/Px4/+vfz+48eqZOfj22TX3sUs84s6yKFQtn/JUQ8PjxjrI49Hxl5VQMDmxSNHpS1sYyfZUwoq2ANhZoUCrUoGGxilhXAhHhcam+fGpdj28BA+v7HHCMtXy6tWSMdf3z6hnmVy2tAA3JuQvROOeH5jKQPW2tfTNnnDEn3SrrEWrszm8/lGlAAgGesdWZU5xKyk58fOZLd9wwEZg7Uo71OfS/PTjO21ioW658xVse3cLfiIwcVDx+WiUm+mGSSW6JUitUEFKpcqpKSzFsotEShUIt8vrw4+QvAfBgenojLyZG5d2/62S/BoNTW5kTm5MBsa3PuZVBACvka0Fwvw3KZpO/IWYblDmvt14wxX5G02Vr7a2PMBkmnSTqQPOQ1a+3lM30mAQoAKAqJhDQ4mB6kY9vk97J5nWnJnsmMcSJ0thErOX+Qi0adG1TN5vlcjsn0PIs/ryTKAorV+BWttopURhWttopWS9EaKVYlRauN1FAv07BIvqal8je3Kth0vErKj1VJyZJkrB4jn6/kDf7DBTAvxpYqm24W88CB9P1rapygzBSZS5cW1c3sCFAXEaAAAEwyNjP7RgI29fVcbhjl8zkzDKGQ8zjT82z3y/azhoac67l6e53HQ4dke3ulg92yhw7JHB6QmSbQrZFilVK0WopVO4/x2hLZumqpoV5qaJav8Rj5m49VoHm5gi0rFFp8ivw1i/LqOl+gYMVizqmymQJz7CZAqZYsmT4y6+sXzP8uCVAXEaAAAORYJJIepNLRIzCfZxYSCecmVck4TQ3VRM9+xXvalTh4QDrYLfUdlukdkP9wWP6hzNenSlIiIMWrA4rXlipRXynV1UgNTTJNLfI3LZO/qVW+piUyjY0Tywj5/U6ojz2faZtuvzw6hRp5IB5Pvxlb6l3Hx56PnUUQi6Vvmd6b7fvz8Rn9/emnyoZCzimxmSKzrc353xMIUDcRoAAAwBWRiNTXp3h3h6KdOxTr3q14914levbLHuqSDh2S6e2Xr29I/oGoggNSsF/yZXHT5TfCpgSpeaNB6/c71xkHgxOPk5/P5mvz9Tmpz7OJ7mz/PDvf+0lORGWKwJmicL72iU3/lyRvmM/n/HOYvI3983kj76e+V12dHplLluT3X2jlCQLURQQoAADIN4lERKOj+zUablekf49iXTsV69qtRE+7EsNHZOOjstFR2VhENj4qxSNKxCJSLCLF4zIJySQkxTXx3Eom9XVi6muf9cvYgHyJgIz1y2f98iX8Mgm/8zrhk7G+5HMjk/DJJMykzSdf3C8TN/LFjEwsPjFbNTZjlXpN79jrbK47xswCAWft5tS1izNtb2SfyZGfbTgy257XCFAXEaAAAKCYWBtXPD6iRGJiy93rYSUS4RnH4/NVKBRqUSjUopKSxePPJ29Bf6N8CZM5TmcK19num+2fVbO99m821whms6/f75wWOttwLClxQg+YAwLURQQoAADA3FlrlUiEx+M0Gu1VJNKZsh2Y9LpTsVhfhk8yybVYJ8fp1GgNBGplFsjNYQA3EKAuIkABAADcFY+HFY12TQnTsW10dCJarZ16F2VjSqadSU2dZQ0GF8nvL/XgJwQKCwHqIgIUAAAgP1lrFYv1zzibOrZFoz2Spv45NBConTSbukjBYHPa49hzYhULFQHqIgIUAACg8CUSUUWjPRni9EDazGo02q14/EjGz/D7qxUKNSsYXKRQqHmaWHW+HgjUcBowigYB6iICFAAAYGGJx4cViXQrGu1WJNI1/ui815X2PBo9pEwzq8aEso7VYLBRPh83CEL+IkBdRIACAABgOolETNHowQyxmilcu2VtJMOnGAWDDVnPrPp8IRkTkjEBZlnhCgLURQQoAAAA5sPYNauZIjVTrMbjA0f9TGNC40Ga/WPJHI87+qPPV6ZAoEZ+f6WMYW3PYkGAuogABQAAgBfi8ZFkmI5FapdisSOyNqJEIpKTR2uj8zR6I7+/WoFAzfjm99fM8nUVEZsnCFAXEaAAAABYKKy1sjY654CNx4cUi/UrHu9XLDaxZXptbewoozHy+6syBmr2cVstY/yu/O6KWSEHKFdXAwAAAHnKGDN+Om0uWWuVSIxkDNR4fGDaeI1EujQy8sr468zX1Kbz+yvHgzQYrFcw2JS8trZZwWBT8jH1eQPRWkSYAQUAAAAwL+LxcFazrROvexWJ9Cga7U7ewTiR4VPHbgo1EakzBWsgUFv0pwoX8gwoAQoAAADAc9bGFY32Jq+z7Rm/3jYaTX8+9hiL9Wb8HGMCCgYbsw5W59rWwrp7MQHqIgIUAAAAQCIRTS650zO+Tmzq88kRO91djI0pUSjUlDFYQ6Fj1NLyUZd/sqMjQF1EgAIAAACYrXg8nJxNzRSsk9/rUiIxolDoGJ1//uteD32KQg5QbkIEAAAAoOj5/aXy+5eptHRZVvuP3UEY84sABQAAAIBJ/P4K+f0FOcmY14r79lAAAAAAgLxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcQYACAAAAAFxBgAIAAAAAXEGAAgAAAABcEfB6APMhGo2qo6ND4XDY66EUnNLSUi1dulTBYNDroQAAAAAockURoB0dHaqqqlJra6uMMV4Pp2BYa3Xo0CF1dHSora3N6+EAAAAAKHJFcQpuOBxWQ0MD8TlLxhg1NDQwcwwAAADAFUURoJKIzzni9wYAAADALUUToAAAAACA/EaAzoPDhw/rBz/4wZyOveyyy3T48OF5HhEAAAAA5B8CdB7MFKCxWGzGY//rv/5LtbW1uRgWAAAAAOSVorgLbqobb5S2bJnfz1y1SvrOd6b/+k033aRdu3Zp1apVWrNmjd773vfqS1/6kurq6rR9+3a98soruvLKK9Xe3q5wOKwbbrhB1113nSSptbVVmzdv1uDgoC699FJdcMEFevLJJ7VkyRI9+OCDKisrS/tev/nNb/TVr35VkUhEDQ0Nuuuuu7Ro0SINDg7q+uuv1+bNm2WM0Ze//GVdddVVeuihh3TzzTcrHo+rsbFRjz766Pz+cgAAAAAgS0UXoF74xje+oW3btmlLsnwff/xxPffcc9q2bdv48iZ33HGH6uvrNTIyonPOOUdXXXWVGhoa0j5n586d+sUvfqF///d/1wc+8AHdd999+uhHP5q2zwUXXKBNmzbJGKMf//jH+uY3v6lvfetbuuWWW1RTU6OtW7dKkvr6+tTT06NPf/rT2rhxo9ra2tTb2+vCbwMAAAAAMiu6AJ1pptJN5557btramrfeeqseeOABSVJ7e7t27tw5JUDb2tq0atUqSdJZZ52lvXv3Tvncjo4OrV27VgcOHFAkEhn/Hhs2bNDdd989vl9dXZ1+85vf6MILLxzfp76+fl5/RgAAAACYDa4BzZGKiorx548//rg2bNigp556Ss8//7zOOOOMjGtvlpSUjD/3+/0Zrx+9/vrr9dnPflZbt27V7bffzhqeAAAAAAoGAToPqqqqdOTIkWm/3t/fr7q6OpWXl2v79u3atGnTnL9Xf3+/lixZIkm68847x99fs2aNbrvttvHXfX19Wr16tTZu3Kg9e/ZIEqfgAgAAAPAUAToPGhoa9Na3vlWnnnqqPv/5z0/5+iWXXKJYLKaTTz5ZN910k1avXj3n77Vu3TpdffXVOuuss9TY2Dj+/he/+EX19fXp1FNP1emnn67HHntMTU1NWr9+vf7iL/5Cp59+utauXTvn7wsAAAAAb5Sx1no9hlmpqKiwQ0NDae+9/PLLOvnkkz0aUeHj9wcAAAAUDmPMsLW24uh75h9mQAEAAAAAriBAAQAAAACuIEABAAAAAK4gQAEAAAAAriBAAQAAAACuIEABAAAAAK4gQD1SWVnp9RAAAAAAwFUEKAAAAADAFQGvBzDfbnzoRm3p3DKvn7mqZZW+c8l3pv36TTfdpGXLlukzn/mMJGndunWqrKzUX/3VX+mKK65QX1+fotGovvrVr+qKK66Y8XtdeeWVam9vVzgc1g033KDrrrtOkvTQQw/p5ptvVjweV2Njox599FENDg7q+uuv1+bNm2WM0Ze//GVdddVV8/eDAwAAAMA8KroA9cLatWt14403jgfoPffco4cfflilpaV64IEHVF1drYMHD2r16tW6/PLLZYyZ9rPuuOMO1dfXa2RkROecc46uuuoqJRIJffrTn9bGjRvV1tam3t5eSdItt9yimpoabd26VZLU19eX+x8WAAAAAOao6AJ0ppnKXDnjjDPU3d2t/fv3q6enR3V1dVq2bJmi0ahuvvlmbdy4UT6fT6+//rq6urrU0tIy7WfdeuuteuCBByRJ7e3t2rlzp3p6enThhReqra1NklRfXy9J2rBhg+6+++7xY+vq6nL4UwIAAADAG1N0AeqVq6++Wvfee686Ozu1du1aSdJdd92lnp4ePfvsswoGg2ptbVU4HJ72Mx5//HFt2LBBTz31lMrLy3XRRRfNuD8AAAAAFBJuQjRP1q5dq7vvvlv33nuvrr76aklSf3+/mpubFQwG9dhjj2nfvn0zfkZ/f7/q6upUXl6u7du3a9OmTZKk1atXa+PGjdqzZ48kjZ+Cu2bNGt12223jx3MKLgAAAIB8RoDOkze96U06cuSIlixZosWLF0uSPvKRj2jz5s067bTT9NOf/lQnnXTSjJ9xySWXKBaL6eSTT9ZNN92k1atXS5Kampq0fv36/7+9+4+5s6zvOP7+jCp1xUDR2TnKxClZQGKrEtiGLl2YHRAjSJiWOdY5FsYmy/hjmbixabp/dJsz22ImbpBUh4qi1WbxBx0jLP4BUpoqP4qjIyU+TaVYmmLBzRW/++NcbQ7PzinPAz33Oc9z3q+kOee+r+s8+Z7z7XWf+3vu61yHSy+9lFWrVh25wnr99dezf/9+zjrrLFatWsUdd9wx2icpSZIkSS9AqmrcMczLsmXL6qmnnnrWvh07dnDGGWeMKaKFz9dPkiRJWjiSPF1Vy8Ydx/PhFVBJkiRJUicsQCVJkiRJnVg0BehCm0o8KXzdJEmSJHVlURSgS5cuZd++fRZT81RV7Nu3j6VLl447FEmSJElTYFH8DujKlSuZmZnh8ccfH3coC87SpUtZuXLluMOQJEmSNAUWxSq4kiRJkjQtXAV3iCQXJPlOkp1JrhvQfnySW1r73UlOG2U8kiRJkjQNkuxKcl+S7Um2tn0nJ9mS5OF2u7zruEZWgCY5DvgYcCFwJnB5kjNndbsS2F9VrwU+Cnx4VPFIkiRJ0pT5lapaXVVnt+3rgNur6nTg9rbdqVFeAT0H2FlVj1TVj4DPAhfP6nMxsLHdvxU4P0lGGJMkSZIkTav++msjcEnXAYxyEaJTgO/2bc8A5w7rU1WHkhwAXgZ8v79TkquAq9pmJfnhSCJ+YZYAh8YdhI4wH5PFfEweczJZzMdkMR+TxXxMFvMxGV5yeFpt84mq+sSsPgXclqSAG1r7iqra09q/B6zoINZnWRCr4LYXa/YLOlGSbO27tK0xMx+TxXxMHnMyWczHZDEfk8V8TBbzsaC8uap2J3kFsCXJQ/2NVVWtOO3UKKfg7gZO7dte2fYN7JNkCXAisG+EMUmSJEnSoldVu9vtXmATva9IPpbklQDtdm/XcY2yAL0HOD3Jq5O8GFgHbJ7VZzOwvt2/DPj3Wmi/CyNJkiRJEyTJsiQvPXwfWAvcz7Prr/XAl7uObWRTcNt3Oq8Bvg4cB9xUVQ8k2QBsrarNwI3Ap5LsBJ6gV6QuVBM9RXgKmY/JYj4mjzmZLOZjspiPyWI+Jov5WBhWAJva+q5LgE9X1deS3AN8LsmVwKPAO7sOLF5wlCRJkiR1YZRTcCVJkiRJOsICVJIkSZLUCQvQeUpyQZLvJNmZ5LoB7ccnuaW1353ktO6jnA5JTk1yR5IHkzyQ5I8G9FmT5ECS7e3fX4wj1mmRZFeS+9prvXVAe5L8fRsf307yxnHEOQ2S/Hzf//vtSZ5Mcu2sPo6PEUtyU5K9Se7v23dyki1JHm63y4c8dn3r83CS9YP6aH6G5OOvkzzUjkmbkpw05LFHPb5p/obk44NJdvcdly4a8tijno9p/obk45a+XOxKsn3IYx0fmjO/AzoPSY4D/hN4KzBDb6Xfy6vqwb4+fwC8vqquTrIOeEdVvWssAS9ybenoV1bVtrbK173AJbPysQb446p625jCnCpJdgFnV9X3h7RfBPwhcBFwLvB3VXVudxFOp3bs2g2cW1WP9u1fg+NjpJL8MnAQ+GRVndX2/RXwRFV9qJ04L6+q98163MnAVuBsej8kfi/wpqra3+kTWGSG5GMtvVX4DyX5MMDsfLR+uzjK8U3zNyQfHwQOVtXfHOVxz3k+pvkblI9Z7R8BDlTVhgFtu3B8aI68Ajo/5wA7q+qRqvoR8Fng4ll9LgY2tvu3AuenLT+lY6uq9lTVtnb/B8AO4JTxRqXncDG9N7aqqruAk9oHCRqt84H/6i8+1Y2q+g96q7z363+f2AhcMuChvwZsqaonWtG5BbhgZIFOiUH5qKrbqupQ27yL3u+WqwNDxsdczOV8TPN0tHy0c9l3Ap/pNCgtShag83MK8N2+7Rn+f8FzpE97QzsAvKyT6KZYm+r8BuDuAc2/mORbSb6a5HWdBjZ9Crgtyb1JrhrQPpcxpGNvHcNPGhwf3VtRVXva/e/RWyp/NsfKePwO8NUhbc91fNOxc02bEn3TkCnqjo/uvQV4rKoeHtLu+NCcWYBqwUtyAvAF4NqqenJW8zbgVVW1CvgH4Etdxzdl3lxVbwQuBN7bpvNojJK8GHg78PkBzY6PMave92D8LswESPJnwCHg5iFdPL514x+B1wCrgT3AR8YbjprLOfrVT8eH5swCdH52A6f2ba9s+wb2SbIEOBHY10l0UyjJi+gVnzdX1Rdnt1fVk1V1sN3/CvCiJC/vOMypUVW72+1eYBO9aVL95jKGdGxdCGyrqsdmNzg+xuaxw1PP2+3eAX0cKx1K8tvA24B315DFMeZwfNMxUFWPVdUzVfVj4J8Y/Do7PjrUzmcvBW4Z1sfxofmwAJ2fe4DTk7y6XVVYB2ye1WczcHi1wsvoLWzgp9sj0L6PcCOwo6r+dkifnz78Hdwk59D7P+8HAiOQZFlbDIoky4C1wP2zum0Gfis9v0BvMYM9aJSGfmrt+Bib/veJ9cCXB/T5OrA2yfI2BXFt26djLMkFwJ8Ab6+qp4f0mcvxTcfArHUB3sHg13ku52M6dn4VeKiqZgY1Oj40X0vGHcBC0lbIu4beScBxwE1V9UCSDcDWqtpMryD6VJKd9L7IvW58ES965wFXAPf1LQv+p8DPAlTVx+l9CPD7SQ4BPwTW+YHAyKwANrV6Zgnw6ar6WpKr4Ug+vkJvBdydwNPAe8YU61RoJwJvBX6vb19/PhwfI5bkM8Aa4OVJZoAPAB8CPpfkSuBRegt7kORs4Oqq+t2qeiLJX9I70QbYUFXPZ7EW9RmSj/cDxwNb2vHrrraS/c8A/1xVFzHk+DaGp7CoDMnHmiSr6U1N30U7fvXnY9j52BiewqIyKB9VdSMD1hFwfOiF8GdYJEmSJEmdcAquJEmSJKkTFqCSJEmSpE5YgEqSJEmSOmEBKkmSJEnqhAWoJEmSJKkTFqCSJB1jSdYk+ddxxyFJ0qSxAJUkSZIkdcICVJI0tZL8ZpJvJtme5IYkxyU5mOSjSR5IcnuSn2p9Vye5K8m3k2xKsrztf22Sf0vyrSTbkrym/fkTktya5KEkN6f9SrskSdPMAlSSNJWSnAG8CzivqlYDzwDvBpYBW6vqdcCdwAfaQz4JvK+qXg/c17f/ZuBjVbUK+CVgT9v/BuBa4Ezg54DzRv6kJEmacEvGHYAkSWNyPvAm4J52cfIlwF7gx8Atrc+/AF9MciJwUlXd2fZvBD6f5KXAKVW1CaCq/hug/b1vVtVM253xYd4AAAD3SURBVN4OnAZ8Y/RPS5KkyWUBKkmaVgE2VtX7n7Uz+fNZ/ep5/v3/6bv/DL7nSpLkFFxJ0tS6HbgsySsAkpyc5FX03hsva31+A/hGVR0A9id5S9t/BXBnVf0AmElySfsbxyf5yU6fhSRJC4ifxkqSplJVPZjkeuC2JD8B/C/wXuAp4JzWtpfe90QB1gMfbwXmI8B72v4rgBuSbGh/49c7fBqSJC0oqXq+M4skSVp8khysqhPGHYckSYuRU3AlSZIkSZ3wCqgkSZIkqRNeAZUkSZIkdcICVJIkSZLUCQtQSZIkSVInLEAlSZIkSZ2wAJUkSZIkdeL/APcMBrPVpHbDAAAAAElFTkSuQmCC\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FR3ZmGKkFocn",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "1494b84a-98a5-40fb-d26b-0b427dedcf8e"
      },
      "source": [
        "# Final evaluation of the model\n",
        "test_loss, test_accuracy = test(test_dataloader, model, loss_fn)\n",
        "print(\"Baseline Error: %.2f%%\" % (100-test_accuracy))"
      ],
      "execution_count": 111,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test: \n",
            " Accuracy: 92.8%, Avg loss: 0.208501 \n",
            "\n",
            "Baseline Error: 7.17%\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BM17JvSG1_y_"
      },
      "source": [
        "#STEP 14: 결과 확인하기 (틀린 것 들만)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yMnR251GFocp",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 727
        },
        "outputId": "20199556-b63c-4f04-ccbd-88d320c111df"
      },
      "source": [
        "# 7. 모델 사용하기\n",
        "plt_row = 5\n",
        "plt_col = 5\n",
        "\n",
        "test_dataloader = DataLoader(test_data, batch_size=1)\n",
        "\n",
        "failed_cases = []\n",
        "failed_labels = []\n",
        "real_labels = []\n",
        "cnt = 0\n",
        "\n",
        "while cnt < (plt_row*plt_col):\n",
        "  with torch.no_grad():\n",
        "      for X, y in test_dataloader:\n",
        "          X, y = X.to(device), y.to(device)\n",
        "          pred = model(X)\n",
        "          X = X.cpu()\n",
        "          y = y.cpu()\n",
        "          pred = pred.cpu()\n",
        "          if y != np.argmax(pred):\n",
        "            failed_cases.append(X)\n",
        "            failed_labels.append(np.argmax(pred))\n",
        "            real_labels.append(y)\n",
        "            cnt += 1        \n",
        "\n",
        "\n",
        "plt.rcParams[\"figure.figsize\"] = (20,20)\n",
        "\n",
        "f, axarr = plt.subplots(plt_row, plt_col)\n",
        "\n",
        "i = 0\n",
        "while i < (plt_row*plt_col):\n",
        "    \n",
        "    sub_plt = axarr[(int)(i/plt_row), i%plt_col]\n",
        "    sub_plt.axis('off')\n",
        "    sub_plt.imshow(failed_cases[i].reshape(width, height), cmap='gray')\n",
        "    sub_plt_title = 'R: ' + labelNames[real_labels[i]] + ' P: ' + labelNames[failed_labels[i]]\n",
        "    sub_plt.set_title(sub_plt_title)\n",
        "\n",
        "    i += 1    \n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 125,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1440x1440 with 25 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}