Notebook: CNN Img Classification.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "provenance": []
    },
    "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": "3.7.6"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "b354718abe934b8a9f777444dd3a874c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_f544900bb96542b28ba9139367b46eae",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_c4eb5dfaf35e40f68d306ceff9909a0d",
              "IPY_MODEL_cd8d3bf9f23a4e82acccde7de649468c"
            ]
          },
          "model_module_version": "1.5.0"
        },
        "f544900bb96542b28ba9139367b46eae": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "c4eb5dfaf35e40f68d306ceff9909a0d": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_df1504e247324fc9aec3a3e673847e4b",
            "_dom_classes": [],
            "description": "",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 26421880,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 26421880,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_9b927b6f4b304097bd2cb1f93cced1d4"
          },
          "model_module_version": "1.5.0"
        },
        "cd8d3bf9f23a4e82acccde7de649468c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_46240f5f307546fcadd66333a0f230f5",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 26422272/? [00:04&lt;00:00, 6216487.65it/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_eeff40836fdd46799eda6affecb4d00b"
          },
          "model_module_version": "1.5.0"
        },
        "df1504e247324fc9aec3a3e673847e4b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          },
          "model_module_version": "1.5.0"
        },
        "9b927b6f4b304097bd2cb1f93cced1d4": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "46240f5f307546fcadd66333a0f230f5": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          },
          "model_module_version": "1.5.0"
        },
        "eeff40836fdd46799eda6affecb4d00b": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "fa1b3583c0244f0187600e2856bbda1b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_0e83da34f6314833af0afd0613c8d38d",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_6ec41ca546734a68ae902712cf6a7e23",
              "IPY_MODEL_3abdbd007cf147b9bd3e88535f89a53d"
            ]
          },
          "model_module_version": "1.5.0"
        },
        "0e83da34f6314833af0afd0613c8d38d": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "6ec41ca546734a68ae902712cf6a7e23": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_d8390ea3d7ac432a8b57a33a641492e8",
            "_dom_classes": [],
            "description": "",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 29515,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 29515,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_64f5cdb241ed4b928ce28e78db18270c"
          },
          "model_module_version": "1.5.0"
        },
        "3abdbd007cf147b9bd3e88535f89a53d": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_f2a7fe200449434aa80d28290ef7e226",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 29696/? [00:01&lt;00:00, 15224.65it/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_0e10c176df93404f98d558e9f242e66f"
          },
          "model_module_version": "1.5.0"
        },
        "d8390ea3d7ac432a8b57a33a641492e8": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          },
          "model_module_version": "1.5.0"
        },
        "64f5cdb241ed4b928ce28e78db18270c": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "f2a7fe200449434aa80d28290ef7e226": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          },
          "model_module_version": "1.5.0"
        },
        "0e10c176df93404f98d558e9f242e66f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "846871931bed4621a82e2e39d337c5de": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_b0bc5fcea2014166ad5f871e337e9b26",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_ccca4ef3753b4e78856d5f5a122f71b3",
              "IPY_MODEL_fb14fe7e0b09489b8470d2e525ac00e3"
            ]
          },
          "model_module_version": "1.5.0"
        },
        "b0bc5fcea2014166ad5f871e337e9b26": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "ccca4ef3753b4e78856d5f5a122f71b3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_58328502aa5540b39263dcdd370f92fb",
            "_dom_classes": [],
            "description": "",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 4422102,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 4422102,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_42dab2b43021447990cf3c05e94ef049"
          },
          "model_module_version": "1.5.0"
        },
        "fb14fe7e0b09489b8470d2e525ac00e3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_224d4a1f603449fcaf86622170c4dec3",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 4422656/? [00:01&lt;00:00, 3064433.64it/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_27fe80fd4bd04c829615c6a9178a9edb"
          },
          "model_module_version": "1.5.0"
        },
        "58328502aa5540b39263dcdd370f92fb": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          },
          "model_module_version": "1.5.0"
        },
        "42dab2b43021447990cf3c05e94ef049": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "224d4a1f603449fcaf86622170c4dec3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          },
          "model_module_version": "1.5.0"
        },
        "27fe80fd4bd04c829615c6a9178a9edb": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "e4dbb92d29d64c9d8f2ac2a7d6db2a8b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_46d9b456046b4fc985d7a47d6b70fb34",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_560b497d2d084e97995cf6a663285667",
              "IPY_MODEL_aff888172503487e99bc40ae66890f33"
            ]
          },
          "model_module_version": "1.5.0"
        },
        "46d9b456046b4fc985d7a47d6b70fb34": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "560b497d2d084e97995cf6a663285667": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_bf20de7ac71a439aa3d17a2cb715daf7",
            "_dom_classes": [],
            "description": "",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 5148,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 5148,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_e4071b1fc3554e1186d1cd4b9fb8aea7"
          },
          "model_module_version": "1.5.0"
        },
        "aff888172503487e99bc40ae66890f33": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_2b07a4e5d756414bb8aab9332640e94b",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 6144/? [00:00&lt;00:00, 24090.86it/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_d0ecf7bda9b142b185c5806973c6eebe"
          },
          "model_module_version": "1.5.0"
        },
        "bf20de7ac71a439aa3d17a2cb715daf7": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          },
          "model_module_version": "1.5.0"
        },
        "e4071b1fc3554e1186d1cd4b9fb8aea7": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "2b07a4e5d756414bb8aab9332640e94b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          },
          "model_module_version": "1.5.0"
        },
        "d0ecf7bda9b142b185c5806973c6eebe": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "deletable": false,
        "editable": false,
        "id": "64bIrOv5rEYJ",
        "nbgrader": {
          "cell_type": "markdown",
          "checksum": "2e176d1ab1160888bd08953c1fec6598",
          "grade": false,
          "grade_id": "cell-559e06fab022e928",
          "locked": true,
          "schema_version": 3,
          "solution": false
        }
      },
      "source": [
        "## Convolutional Neural Networks\n",
        "\n",
        "This Notebook will create a CNN to classify images of fashion items"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "deletable": false,
        "editable": false,
        "id": "QNjuVk66rEYO",
        "nbgrader": {
          "cell_type": "markdown",
          "checksum": "6da914e441f04a8a66e6dceec29a35c4",
          "grade": false,
          "grade_id": "cell-98105dbd82868b57",
          "locked": true,
          "schema_version": 3,
          "solution": false,
          "task": false
        }
      },
      "source": [
        "We first import the neccesary packages required to build the model. We also fix a seed through out the process to have a deterministic behaviour (reproduce same results) of the model. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "deletable": false,
        "editable": false,
        "id": "9v4EAhu1rEYO",
        "nbgrader": {
          "cell_type": "code",
          "checksum": "19fe0393c09e9a63890d67116c2c72f2",
          "grade": false,
          "grade_id": "cell-675473914c4b9f68",
          "locked": true,
          "schema_version": 3,
          "solution": false
        },
        "outputId": "c7c3f2bc-76a3-4183-e550-1178f3ecde9e"
      },
      "source": [
        "## Imports\n",
        "import os\n",
        "import time\n",
        "import numpy as np\n",
        "import torch\n",
        "import random\n",
        "import torchvision\n",
        "import torchvision.transforms as transforms\n",
        "from torch.utils.data.sampler import SubsetRandomSampler\n",
        "import numpy.testing as npt\n",
        "#from torchsummary import summary\n",
        "# from tqdm import trange\n",
        "\n",
        "# Checks for the availability of GPU \n",
        "is_cuda = torch.cuda.is_available()\n",
        "if torch.cuda.is_available():\n",
        "    print(\"working on gpu!\")\n",
        "else:\n",
        "    print(\"No gpu! only cpu ;)\")\n",
        "    \n",
        "## The following random seeds are just for deterministic behaviour of the code and evaluation\n",
        "\n",
        "random.seed(0)\n",
        "np.random.seed(0)\n",
        "torch.manual_seed(0)\n",
        "torch.cuda.manual_seed_all(0)\n",
        "torch.backends.cudnn.deterministic = True\n",
        "torch.backends.cudnn.benchmark = False\n",
        "os.environ['PYTHONHASHSEED'] = '0'\n",
        "\n",
        "############################################################################### \n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "working on gpu!\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "deletable": false,
        "editable": false,
        "id": "dBP8tQblrEYQ",
        "nbgrader": {
          "cell_type": "markdown",
          "checksum": "7eb19f0b795a90d9575c1139d3a87e7c",
          "grade": false,
          "grade_id": "cell-9c87add622617c1e",
          "locked": true,
          "schema_version": 3,
          "solution": false
        }
      },
      "source": [
        "### Setting up the DataLoaders\n",
        "\n",
        "In the following cell we will first download and arrange the data. FashionMNIST dataset is already available in the official PyTorch repository. Hence, the following cell checks for the availability of FashionMNIST data and downloads if the data is not available.   \n",
        "\n",
        "The following parts are already written for you to handle the data.\n",
        "- import neccesary pytorch packages for data handling.\n",
        "- We then move the data onto PyTorch tensors.  \n",
        "- Next we define the parameters like batch_size for data handling. A different batch_size for test data is used to     make sure that number of samples in the test data are perfectly divisible.\n",
        "- create dataloaders for training and testing data to iterate."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "deletable": false,
        "editable": false,
        "id": "xueQ11uDrEYQ",
        "nbgrader": {
          "cell_type": "code",
          "checksum": "f9026479c9b2db3497126f8ec4a5868f",
          "grade": false,
          "grade_id": "cell-e518bb36bae17b30",
          "locked": true,
          "schema_version": 3,
          "solution": false
        },
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 571,
          "referenced_widgets": [
            "b354718abe934b8a9f777444dd3a874c",
            "f544900bb96542b28ba9139367b46eae",
            "c4eb5dfaf35e40f68d306ceff9909a0d",
            "cd8d3bf9f23a4e82acccde7de649468c",
            "df1504e247324fc9aec3a3e673847e4b",
            "9b927b6f4b304097bd2cb1f93cced1d4",
            "46240f5f307546fcadd66333a0f230f5",
            "eeff40836fdd46799eda6affecb4d00b",
            "fa1b3583c0244f0187600e2856bbda1b",
            "0e83da34f6314833af0afd0613c8d38d",
            "6ec41ca546734a68ae902712cf6a7e23",
            "3abdbd007cf147b9bd3e88535f89a53d",
            "d8390ea3d7ac432a8b57a33a641492e8",
            "64f5cdb241ed4b928ce28e78db18270c",
            "f2a7fe200449434aa80d28290ef7e226",
            "0e10c176df93404f98d558e9f242e66f",
            "846871931bed4621a82e2e39d337c5de",
            "b0bc5fcea2014166ad5f871e337e9b26",
            "ccca4ef3753b4e78856d5f5a122f71b3",
            "fb14fe7e0b09489b8470d2e525ac00e3",
            "58328502aa5540b39263dcdd370f92fb",
            "42dab2b43021447990cf3c05e94ef049",
            "224d4a1f603449fcaf86622170c4dec3",
            "27fe80fd4bd04c829615c6a9178a9edb",
            "e4dbb92d29d64c9d8f2ac2a7d6db2a8b",
            "46d9b456046b4fc985d7a47d6b70fb34",
            "560b497d2d084e97995cf6a663285667",
            "aff888172503487e99bc40ae66890f33",
            "bf20de7ac71a439aa3d17a2cb715daf7",
            "e4071b1fc3554e1186d1cd4b9fb8aea7",
            "2b07a4e5d756414bb8aab9332640e94b",
            "d0ecf7bda9b142b185c5806973c6eebe"
          ]
        },
        "outputId": "091f9ced-d493-4272-ea29-ab8d93ff499d"
      },
      "source": [
        "import torchvision\n",
        "import torchvision.transforms as transforms\n",
        "import os\n",
        "\n",
        "if not os.path.isdir('./data'):\n",
        "    os.mkdir('./data')\n",
        "root = './data/'\n",
        "\n",
        "# List of transformation on the data - here we will normalize the image data to (-1,1)\n",
        "transform = transforms.Compose([transforms.ToTensor(),transforms.Normalize((0.5), (0.5)),])\n",
        "# Geta  handle to Load the data\n",
        "training_data = torchvision.datasets.FashionMNIST(root, train=True, transform=transform,download=True)\n",
        "testing_data = torchvision.datasets.FashionMNIST(root, train=False, transform=transform,download=True)\n",
        "\n",
        "num_train = len(training_data)\n",
        "indices = list(range(num_train))\n",
        "np.random.shuffle(indices)\n",
        "split = int(np.floor(0.2 * num_train))\n",
        "train_idx, valid_idx = indices[split:], indices[:split]\n",
        "\n",
        "train_bs = 60\n",
        "test_bs = 50\n",
        "\n",
        "train_sampler = SubsetRandomSampler(train_idx)\n",
        "valid_sampler = SubsetRandomSampler(valid_idx)\n",
        "\n",
        "# Create Data loaders which we will use to extract minibatches of data to input to the network for training\n",
        "train_loader = torch.utils.data.DataLoader(training_data, batch_size=train_bs,\n",
        "    sampler=train_sampler, drop_last=False)\n",
        "valid_loader = torch.utils.data.DataLoader(training_data, batch_size=train_bs, \n",
        "    sampler=valid_sampler, drop_last=False)\n",
        "test_loader = torch.utils.data.DataLoader(testing_data, batch_size=test_bs, \n",
        "    drop_last=False)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "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"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "b354718abe934b8a9f777444dd3a874c",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, max=26421880.0), HTML(value='')))"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n",
            "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"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "fa1b3583c0244f0187600e2856bbda1b",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, max=29515.0), HTML(value='')))"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n",
            "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"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "846871931bed4621a82e2e39d337c5de",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, max=4422102.0), HTML(value='')))"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n",
            "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"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "e4dbb92d29d64c9d8f2ac2a7d6db2a8b",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, max=5148.0), HTML(value='')))"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n",
            "Extracting ./data/FashionMNIST/raw/t10k-labels-idx1-ubyte.gz to ./data/FashionMNIST/raw\n",
            "\n",
            "Processing...\n",
            "Done!\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/torchvision/datasets/mnist.py:502: UserWarning: The given NumPy array is not writeable, and PyTorch does not support non-writeable tensors. This means you can write to the underlying (supposedly non-writeable) NumPy array using the tensor. You may want to copy the array to protect its data or make it writeable before converting it to a tensor. This type of warning will be suppressed for the rest of this program. (Triggered internally at  /pytorch/torch/csrc/utils/tensor_numpy.cpp:143.)\n",
            "  return torch.from_numpy(parsed.astype(m[2], copy=False)).view(*s)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "deletable": false,
        "editable": false,
        "id": "3CQY7vi4rEYQ",
        "nbgrader": {
          "cell_type": "markdown",
          "checksum": "6f5531a49684a81df0e0f59102be6227",
          "grade": false,
          "grade_id": "cell-f9b382eeb7593ce2",
          "locked": true,
          "schema_version": 3,
          "solution": false,
          "task": false
        }
      },
      "source": [
        "### Visualize a Few Data Samples\n",
        "\n",
        "In the following cell we first peek into a random batch of images together with labels and visualize them. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 484
        },
        "deletable": false,
        "editable": false,
        "id": "3aQM8-KfrEYR",
        "nbgrader": {
          "cell_type": "code",
          "checksum": "87e6c7dfdb9d4d57ab2421fad9e63cb9",
          "grade": false,
          "grade_id": "cell-1168a8b4140174bf",
          "locked": true,
          "schema_version": 3,
          "solution": false
        },
        "outputId": "8ac52134-ed2c-4000-85a1-9bf9038c68e9"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "\n",
        "## get a batch of data\n",
        "images, labels = iter(train_loader).next()\n",
        "\n",
        "\n",
        "image_dict = {0:'T-shirt/Top', 1:'Trouser', 2:'Pullover', 3:'Dress',\n",
        "              4:'Coat', 5:'Sandal', 6:'Shirt', 7:'Sneaker',\n",
        "              8:'Bag', 9:'Ankle Boot'}\n",
        "\n",
        "fig = plt.figure(figsize=(8,8))\n",
        "\n",
        "print(images.size())\n",
        "\n",
        "for i in np.arange(1, 13):\n",
        "    ax = fig.add_subplot(3,4,i, frameon=False)\n",
        "    img = images[i][0]\n",
        "    ax.set_title(image_dict[labels[i].item()])\n",
        "    plt.imshow(img, cmap='gray')\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "torch.Size([60, 1, 28, 28])\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 12 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "deletable": false,
        "editable": false,
        "id": "shsQOK9SrEYR",
        "nbgrader": {
          "cell_type": "markdown",
          "checksum": "6dcbe042b4711d11eafae3ec791e14b6",
          "grade": false,
          "grade_id": "cell-f48cb922457e6d1e",
          "locked": true,
          "schema_version": 3,
          "solution": false,
          "task": false
        }
      },
      "source": [
        "### Architecture\n",
        "\n",
        "We implement a Convolutional Neural Network as our model. We make use of the following layers in our model.\n",
        "- a convolution layer for extracting features.\n",
        "- batchnorm layer for normalizing the weights in the hidden layers.\n",
        "- ReLU activation function for the non-linearity between layers.\n",
        "- Finally fully connected layers in the end.\n",
        "\n",
        "#### Model:\n",
        "\n",
        "we make use of the following convolutional neural network architecture for our dataset. \n",
        "\n",
        "- convolution layer output_channels-16 kernel_size=3 stride=1 padding-1\n",
        "- batchnormalization layer\n",
        "- ReLU activation layer\n",
        "- maxpool layer kernel_size=2 stride=2\n",
        "- convolution layer output_channels-32 kernel_size=3 stride=1 padding-1\n",
        "- batchnormalization layer\n",
        "- ReLU activation layer\n",
        "- maxpool layer kernel_size=2 stride=2\n",
        "- convolution layer output_channels-64 kernel_size=5 stride=1 padding-2\n",
        "- batchnormalization layer\n",
        "- ReLU activation layer\n",
        "- maxpool layer kernel_size=2 stride=2\n",
        "- fully connected layer - number_of_classes\n",
        "\n",
        "### Build the model\n",
        "\n",
        "- We first define a class called Model nheriting from Pytorch's nn.Module.\n",
        "- In init(constructor), we define all the layers that are used to build the model\n",
        "- Define a forward function for a sequential model that takes in images as input and returns the predictions as output.\n",
        "\n",
        "All the functions are available in the PyTorch package. Read the documentation/source code for a better understanding.\n",
        "\n",
        "- Convolutional layer: https://pytorch.org/docs/stable/nn.html#convolution-layers\n",
        "- Batchnorm layer: https://pytorch.org/docs/stable/nn.html#normalization-layers\n",
        "- Activation ReLU: https://pytorch.org/docs/stable/nn.html#non-linear-activations-weighted-sum-nonlinearity\n",
        "- Maxpooling layer: https://pytorch.org/docs/stable/nn.html#pooling-layers\n",
        "- Fully connected layer: https://pytorch.org/docs/stable/nn.html#linear-layers\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "deletable": false,
        "id": "BIJgXe8grEYS",
        "nbgrader": {
          "cell_type": "code",
          "checksum": "517fa5c8accc5740302af5e2cfbcc415",
          "grade": false,
          "grade_id": "network_model_soln",
          "locked": false,
          "schema_version": 3,
          "solution": true
        }
      },
      "source": [
        "import torch.nn as nn\n",
        "\n",
        "class Model(nn.Module):\n",
        "    ## init function is the constructor and we define all the layers used in our model. \n",
        "    def __init__(self, num_classes=10):\n",
        "        super(Model, self).__init__()\n",
        "        ''' \n",
        "        REST OF THE MODEL HERE\n",
        "        \n",
        "        # define a convolutional layer with 16 channels, kernel_size=3, stride=1 and padding=1\n",
        "        \n",
        "        # define a batchnormalization layer \n",
        "        \n",
        "        # define a relu layer\n",
        "        \n",
        "        # define a maxpool layer with kernel_size=2, stride=2\n",
        "        \n",
        "        # define a convolutional layer with 32 channels, kernel_size=3, stride=1 and padding=1\n",
        "        \n",
        "        # define a batchnormalization layer \n",
        "        \n",
        "        # define a relu layer\n",
        "        \n",
        "        # define a maxpool layer with kernel_size=2, stride=2\n",
        "        \n",
        "        # define a convolution layer with 64 channels, kernel_size=5, stride=1 and padding=2\n",
        "        \n",
        "        # define a batchnorm layer\n",
        "        \n",
        "        # define a relu layer\n",
        "        \n",
        "        # define a maxpool layer with kernel_size=2, stride=2\n",
        "        \n",
        "        # define a fully connected layer from resulting dimension -> number of classes\n",
        "        \n",
        "        '''\n",
        "        ###\n",
        "        self.conv1 = nn.Conv2d(in_channels=1, out_channels=16, kernel_size=3, stride=1, padding=1)  #28 x 28\n",
        "\n",
        "        self.bn1 = nn.BatchNorm2d(16)\n",
        "\n",
        "        self.relu1 = nn.ReLU(inplace=False)\n",
        "\n",
        "        self.maxp1 = nn.MaxPool2d(kernel_size = 2,stride = 2)\n",
        "\n",
        "        self.conv2 = nn.Conv2d(in_channels=16, out_channels=32, kernel_size=3, stride=1, padding=1)\n",
        "\n",
        "        self.bn2 = nn.BatchNorm2d(32)\n",
        "\n",
        "        self.relu2 = nn.ReLU(inplace=False)\n",
        "\n",
        "        self.maxp2 = nn.MaxPool2d(kernel_size = 2,stride = 2)\n",
        "\n",
        "        self.conv3 = nn.Conv2d(in_channels=32, out_channels=64, kernel_size=5, stride=1, padding=2)\n",
        "\n",
        "        self.bn3 = nn.BatchNorm2d(64)\n",
        "\n",
        "        self.relu3 = nn.ReLU(inplace=False)\n",
        "\n",
        "        self.maxp3 = nn.MaxPool2d(kernel_size = 2,stride = 2)\n",
        "\n",
        "        self.fc = nn.Linear(576,num_classes)\n",
        "\n",
        "\n",
        "        \n",
        "    \n",
        "    def forward(self, x):\n",
        "        # We will start with feeding the data to the first layer. \n",
        "        # We take the output x and feed it back to the next layer \n",
        "        x = self.conv1(x)\n",
        "        x = self.bn1(x)\n",
        "        # Continue in ths manner to get the output of the final layer. \n",
        "        \n",
        "        ###\n",
        "        x = self.relu1(x)\n",
        "        x = self.maxp1(x)\n",
        "        x = self.conv2(x)\n",
        "        x = self.bn2(x)\n",
        "        x = self.relu2(x)\n",
        "        x = self.maxp2(x)\n",
        "        x = self.conv3(x)\n",
        "        x = self.bn3(x)\n",
        "        x = self.relu3(x)\n",
        "        x = self.maxp3(x)\n",
        "        x = self.flatten(x)        \n",
        "        x = self.fc(x)\n",
        "        \n",
        "        return x\n",
        " \n",
        "    # Note that before the fully connected(fc) layer, the output is a feature map with size (N,C,H,W)\n",
        "    # but a fully connected layers expects a input of size (N, dimension). Therefore, before passing the output of\n",
        "    # maxpool layer to the fc layer, we must first flatten the output of previous layer to a size (N,C*H*W) \n",
        "    # and then pass it to the fully connected layer, like x = x.flatten(x)\n",
        "    # Define the flatten function below, use the x = x.view() function\n",
        "    def flatten(self, x):\n",
        "        N, C, H, W = x.size()\n",
        "        #reshape x to (N, C*H*W) \n",
        "        \n",
        "        # your code here\n",
        "        x = x.view(N,C*H*W)\n",
        "        \n",
        "        \n",
        "        return x\n",
        "        \n",
        "        \n",
        "    "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "deletable": false,
        "editable": false,
        "id": "prxsnKgirEYS",
        "nbgrader": {
          "cell_type": "code",
          "checksum": "e8b0566eda978f395e9851885f70d7db",
          "grade": true,
          "grade_id": "network_model_soln_test",
          "locked": true,
          "points": 30,
          "schema_version": 3,
          "solution": false
        }
      },
      "source": [
        "## Run the cell to check the implementation of your model\n",
        "# The testcase only tests the input and output dimensions of your architecture. \n",
        "# The only constraints you Model needs to satisfy are: \n",
        "# The Model object is initialized by providing num_classes as input \n",
        "# The network takes input Tensors of dimensions (B,1,28,28), where B is arbitrary batch_size, \n",
        "#    1 is the number of channels in the grayscale image and 28 is image size\n",
        "# The output of the network is Tensor of dimensions (B,10) where 10 is the num_classes\n",
        "\n",
        "model = Model(num_classes=10)\n",
        "test_input1 = torch.randn(16,1,28,28)\n",
        "out1 = model(test_input1)\n",
        "test_input2 = torch.rand(20,1,28,28)\n",
        "out2 = model(test_input2)\n",
        "\n",
        "#hidden tests follow\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_2Vu_5WjcYbu"
      },
      "source": [
        "#Extraneous Hidden tests follow        ................      Extraneous Hidden Tests, run to further confirm everything is set up correctly\n",
        "# HIDDEN TESTS\n",
        "npt.assert_equal(out1.shape, (16,10))\n",
        "npt.assert_equal(out2.shape, (20,10))\n",
        "# HIDDEN TESTS"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IvXGBtROrEYT"
      },
      "source": [
        "# Setting up a few learning parameters\n",
        "learning_rate = 1e-2\n",
        "decayRate = 0.999\n",
        "epochs = 5\n",
        "number_of_classes = 10"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XjHis8E-rEYT"
      },
      "source": [
        "### Initialize the CNN Model\n",
        "\n",
        "Define a loss criterion, using Cross-entropy Loss between the predictions and ground truth to estimate the loss. \n",
        "- CrossEntropyLoss - https://pytorch.org/docs/stable/nn.html#crossentropyloss\n",
        "\n",
        "We also define a optimization strategy to update the weights. - use the most commonly used Adam optimizer from the PyTorch package.\n",
        "\n",
        "- Adam - https://pytorch.org/docs/stable/optim.html#algorithms "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "deletable": false,
        "id": "jqDuph8prEYT",
        "nbgrader": {
          "cell_type": "code",
          "checksum": "dff0b3cd47411f8bf3dc16268a3a7c6e",
          "grade": false,
          "grade_id": "optim_soln",
          "locked": false,
          "schema_version": 3,
          "solution": true
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "70652ab7-6413-4ef2-d9cc-056687933d09"
      },
      "source": [
        "## First we will define an instance of the model to train\n",
        "model = Model(num_classes=number_of_classes)\n",
        "print(model)\n",
        "\n",
        "#Move the model to the gpu if is_cuda\n",
        "if is_cuda:\n",
        "  model = model.cuda()\n",
        "\n",
        "# define the loss 'criterion' as nn.CrossEntropyLoss() object\n",
        "###\n",
        "criterion = nn.CrossEntropyLoss()\n",
        "\n",
        "\n",
        "# Initialize the Adam optimizer for the model.parameters() using the learning_rate\n",
        "###\n",
        "optimizer = torch.optim.Adam(model.parameters(),lr = learning_rate)\n",
        "\n",
        "# This is the learning rate scheduler. It decreases the learning rate as we approach convergence\n",
        "my_lr_scheduler = torch.optim.lr_scheduler.ExponentialLR(optimizer=optimizer, gamma=decayRate)\n",
        "\n",
        "# optimizer = None"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model(\n",
            "  (conv1): Conv2d(1, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
            "  (bn1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
            "  (relu1): ReLU()\n",
            "  (maxp1): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
            "  (conv2): Conv2d(16, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
            "  (bn2): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
            "  (relu2): ReLU()\n",
            "  (maxp2): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
            "  (conv3): Conv2d(32, 64, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))\n",
            "  (bn3): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
            "  (relu3): ReLU()\n",
            "  (maxp3): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
            "  (fc): Linear(in_features=576, out_features=10, bias=True)\n",
            ")\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "deletable": false,
        "editable": false,
        "id": "mklkuII0rEYT",
        "nbgrader": {
          "cell_type": "code",
          "checksum": "7c3ddc62849e994cda0149bd7085e6ee",
          "grade": true,
          "grade_id": "optim_soln_test",
          "locked": true,
          "points": 10,
          "schema_version": 3,
          "solution": false
        }
      },
      "source": [
        "out = torch.FloatTensor([[0.1,0.8,0.05,0.05]])\n",
        "true = torch.LongTensor([1])\n",
        "assert criterion(out, true), 0.8925"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "deletable": false,
        "editable": false,
        "id": "wqD65PCSrEYU",
        "nbgrader": {
          "cell_type": "markdown",
          "checksum": "4e52cd64b766001f7527c8f64e0a0870",
          "grade": false,
          "grade_id": "cell-1e2e5baffbc2c303",
          "locked": true,
          "schema_version": 3,
          "solution": false
        }
      },
      "source": [
        "### Training the Model\n",
        "\n",
        "The training loop is setup in the following way:\n",
        "\n",
        "For every batch in the defined number of epochs\n",
        "\n",
        "- Move the images and labels to the gpu by checking is_cuda\n",
        "- Extract output by passing images through the model \n",
        "- pass the output and ground truth to the loss criterion for batch loss\n",
        "- clear the gradients \n",
        "- backpropagate (compute gradients w.r.t the parameters) using backward() \n",
        "- update the parameters with a single optimization step\n",
        "- update the training loss for plots\n",
        "\n",
        "repeat"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "deletable": false,
        "id": "t3CtvtqQrEYU",
        "nbgrader": {
          "cell_type": "code",
          "checksum": "06673cae7788c67677e8deda3b73c167",
          "grade": false,
          "grade_id": "train_model_soln",
          "locked": false,
          "schema_version": 3,
          "solution": true
        }
      },
      "source": [
        "## training loop \n",
        "\n",
        "## Number of epochs the model runs\n",
        "\n",
        "def train_model(epochs=25, validate=True):\n",
        "    '''\n",
        "    A function to train the model on the dataset and returns the trained model, training loss and\n",
        "    validation loss for every epoch.\n",
        "    \n",
        "    Inputs:\n",
        "        epochs: Number of times the model should be trained on the whole data.\n",
        "        validate: A boolean parameter that validates on validation data.\n",
        "        \n",
        "    Outputs:\n",
        "        model: The model trained for specified number of epochs\n",
        "        training loss: A list of training losses computed for every epoch.\n",
        "        validation loss: A list of validation losses computed for every epoch.\n",
        "    \n",
        "    '''\n",
        "    train_loss = []\n",
        "    val_loss = []\n",
        "\n",
        "    for epoch in range(epochs):\n",
        "        # Iterate through the batches in the data\n",
        "        training_loss = 0.0\n",
        "        validation_loss = 0.0\n",
        "        \n",
        "        # Move the images and labels to the gpu by checking is_cuda\n",
        "        # Extract outputs by passing images through the model \n",
        "        # estimate loss using criterion(.) with 'outputs' and 'labels'\n",
        "        # clear the gradients with .zero_grad()\n",
        "        # Estimate gradients using .backward()\n",
        "        # update model with .step()\n",
        "        # update learning_rate with with a .step using my_lr_scheduler\n",
        "        # add loss to training_loss\n",
        "        model.train()\n",
        "        itr = 0\n",
        "        for (images,labels)  in train_loader:\n",
        "            ###\n",
        "            if is_cuda:\n",
        "                images = images.cuda()\n",
        "                labels = labels.cuda()\n",
        "\n",
        "            outputs = model(images)\n",
        "            \n",
        "            loss = criterion(outputs,labels)\n",
        "            \n",
        "            optimizer.zero_grad()\n",
        "            \n",
        "            loss.backward()\n",
        "\n",
        "            optimizer.step()\n",
        "\n",
        "            my_lr_scheduler.step(epoch)\n",
        "\n",
        "            training_loss += loss\n",
        "\n",
        "            if itr%100 == 0:\n",
        "                print('Epoch %d/%d, itr = %d, Train Loss = %.3f, LR = %.3E'\\\n",
        "                      %(epoch, epochs, itr, loss.item(),optimizer.param_groups[0]['lr']))\n",
        "            itr += 1\n",
        "        train_loss.append(training_loss/len(train_loader))\n",
        "        print('------------------------------------------------')\n",
        "        \n",
        "        # During validation we do not update gradients. We set model to eval()\n",
        "        # Move images and labels to gpu if is_cuda\n",
        "        # get 'outputs' using model and images\n",
        "        # estimate loss using criterion(.) with 'outputs' and 'labels'\n",
        "        # add loss to 'validation_loss'\n",
        "        if validate:\n",
        "            model.eval()\n",
        "            with torch.no_grad():\n",
        "                itr = 0\n",
        "                for (images,labels)  in valid_loader:\n",
        "                    ###\n",
        "                    if is_cuda:\n",
        "                      images = images.cuda()\n",
        "                      labels = labels.cuda()\n",
        "\n",
        "                    outputs = model(images)\n",
        "                    loss = criterion(outputs,labels)\n",
        "                    validation_loss += loss\n",
        "\n",
        "\n",
        "                    \n",
        "                    if itr%100 == 0:\n",
        "                        print('Epoch %d/%d, itr = %d, Val Loss = %.3f, LR = %.3E'\\\n",
        "                              %(epoch, epochs, itr, loss.item(),optimizer.param_groups[0]['lr']))\n",
        "                    itr += 1\n",
        "                val_loss.append(validation_loss/len(valid_loader))\n",
        "                print('################################################')\n",
        "                \n",
        "    return model, train_loss, val_loss\n",
        "                "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "deletable": false,
        "editable": false,
        "id": "FP0Z9aPyrEYU",
        "nbgrader": {
          "cell_type": "code",
          "checksum": "9850cf853a8d0e7a84387bbb72f8c10f",
          "grade": false,
          "grade_id": "cell-3ef122427f827977",
          "locked": true,
          "schema_version": 3,
          "solution": false
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "cf22703f-f591-49c5-d1ba-ce69962f41a7"
      },
      "source": [
        "start = time.time()\n",
        "trained_model, train_loss, val_loss = train_model(epochs, validate=True)\n",
        "end = time.time()\n",
        "print('Time to train in seconds ',(end - start))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/torch/optim/lr_scheduler.py:154: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.\n",
            "  warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 0/5, itr = 0, Train Loss = 2.486, LR = 1.000E-02\n",
            "Epoch 0/5, itr = 100, Train Loss = 0.858, LR = 1.000E-02\n",
            "Epoch 0/5, itr = 200, Train Loss = 0.464, LR = 1.000E-02\n",
            "Epoch 0/5, itr = 300, Train Loss = 0.311, LR = 1.000E-02\n",
            "Epoch 0/5, itr = 400, Train Loss = 0.419, LR = 1.000E-02\n",
            "Epoch 0/5, itr = 500, Train Loss = 0.358, LR = 1.000E-02\n",
            "Epoch 0/5, itr = 600, Train Loss = 0.312, LR = 1.000E-02\n",
            "Epoch 0/5, itr = 700, Train Loss = 0.402, LR = 1.000E-02\n",
            "------------------------------------------------\n",
            "Epoch 0/5, itr = 0, Val Loss = 0.347, LR = 1.000E-02\n",
            "Epoch 0/5, itr = 100, Val Loss = 0.211, LR = 1.000E-02\n",
            "################################################\n",
            "Epoch 1/5, itr = 0, Train Loss = 0.433, LR = 9.990E-03\n",
            "Epoch 1/5, itr = 100, Train Loss = 0.343, LR = 9.990E-03\n",
            "Epoch 1/5, itr = 200, Train Loss = 0.358, LR = 9.990E-03\n",
            "Epoch 1/5, itr = 300, Train Loss = 0.258, LR = 9.990E-03\n",
            "Epoch 1/5, itr = 400, Train Loss = 0.359, LR = 9.990E-03\n",
            "Epoch 1/5, itr = 500, Train Loss = 0.280, LR = 9.990E-03\n",
            "Epoch 1/5, itr = 600, Train Loss = 0.493, LR = 9.990E-03\n",
            "Epoch 1/5, itr = 700, Train Loss = 0.203, LR = 9.990E-03\n",
            "------------------------------------------------\n",
            "Epoch 1/5, itr = 0, Val Loss = 0.233, LR = 9.990E-03\n",
            "Epoch 1/5, itr = 100, Val Loss = 0.188, LR = 9.990E-03\n",
            "################################################\n",
            "Epoch 2/5, itr = 0, Train Loss = 0.123, LR = 9.980E-03\n",
            "Epoch 2/5, itr = 100, Train Loss = 0.225, LR = 9.980E-03\n",
            "Epoch 2/5, itr = 200, Train Loss = 0.196, LR = 9.980E-03\n",
            "Epoch 2/5, itr = 300, Train Loss = 0.334, LR = 9.980E-03\n",
            "Epoch 2/5, itr = 400, Train Loss = 0.244, LR = 9.980E-03\n",
            "Epoch 2/5, itr = 500, Train Loss = 0.372, LR = 9.980E-03\n",
            "Epoch 2/5, itr = 600, Train Loss = 0.199, LR = 9.980E-03\n",
            "Epoch 2/5, itr = 700, Train Loss = 0.325, LR = 9.980E-03\n",
            "------------------------------------------------\n",
            "Epoch 2/5, itr = 0, Val Loss = 0.229, LR = 9.980E-03\n",
            "Epoch 2/5, itr = 100, Val Loss = 0.349, LR = 9.980E-03\n",
            "################################################\n",
            "Epoch 3/5, itr = 0, Train Loss = 0.176, LR = 9.970E-03\n",
            "Epoch 3/5, itr = 100, Train Loss = 0.152, LR = 9.970E-03\n",
            "Epoch 3/5, itr = 200, Train Loss = 0.218, LR = 9.970E-03\n",
            "Epoch 3/5, itr = 300, Train Loss = 0.217, LR = 9.970E-03\n",
            "Epoch 3/5, itr = 400, Train Loss = 0.351, LR = 9.970E-03\n",
            "Epoch 3/5, itr = 500, Train Loss = 0.228, LR = 9.970E-03\n",
            "Epoch 3/5, itr = 600, Train Loss = 0.210, LR = 9.970E-03\n",
            "Epoch 3/5, itr = 700, Train Loss = 0.193, LR = 9.970E-03\n",
            "------------------------------------------------\n",
            "Epoch 3/5, itr = 0, Val Loss = 0.300, LR = 9.970E-03\n",
            "Epoch 3/5, itr = 100, Val Loss = 0.336, LR = 9.970E-03\n",
            "################################################\n",
            "Epoch 4/5, itr = 0, Train Loss = 0.125, LR = 9.960E-03\n",
            "Epoch 4/5, itr = 100, Train Loss = 0.204, LR = 9.960E-03\n",
            "Epoch 4/5, itr = 200, Train Loss = 0.426, LR = 9.960E-03\n",
            "Epoch 4/5, itr = 300, Train Loss = 0.377, LR = 9.960E-03\n",
            "Epoch 4/5, itr = 400, Train Loss = 0.149, LR = 9.960E-03\n",
            "Epoch 4/5, itr = 500, Train Loss = 0.211, LR = 9.960E-03\n",
            "Epoch 4/5, itr = 600, Train Loss = 0.152, LR = 9.960E-03\n",
            "Epoch 4/5, itr = 700, Train Loss = 0.215, LR = 9.960E-03\n",
            "------------------------------------------------\n",
            "Epoch 4/5, itr = 0, Val Loss = 0.153, LR = 9.960E-03\n",
            "Epoch 4/5, itr = 100, Val Loss = 0.251, LR = 9.960E-03\n",
            "################################################\n",
            "Time to train in seconds  50.93747425079346\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "deletable": false,
        "editable": false,
        "id": "6wKIxmkHrEYU",
        "nbgrader": {
          "cell_type": "code",
          "checksum": "55d34f1ee711a47be6689c5ecdf6a9af",
          "grade": true,
          "grade_id": "train_model_soln_test",
          "locked": true,
          "points": 30,
          "schema_version": 3,
          "solution": false
        }
      },
      "source": [
        "# Test cases to ensure loss is reducing every epoch\n",
        "# hidden test cases follow "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "brKsFsDh83D1"
      },
      "source": [
        "# Test cases to ensure loss is reducing every epoch           ... extraneous Tests              ..................................................\n",
        "# hidden test cases follow \n",
        "# HIDDEN TESTS ###\n",
        "assert np.all(np.diff(train_loss) <= 0)\n",
        "assert np.all(np.diff(val_loss) <= 0)\n",
        "# HIDDEN TESTS ###"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "deletable": false,
        "editable": false,
        "id": "Z5Yy9QCurEYV",
        "nbgrader": {
          "cell_type": "code",
          "checksum": "8e12aa54daa529184ce87590889347cc",
          "grade": false,
          "grade_id": "cell-35244d9978da3c81",
          "locked": true,
          "schema_version": 3,
          "solution": false
        },
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        },
        "outputId": "a938c31a-e97d-4a4c-fa93-8010c852cd4e"
      },
      "source": [
        "# Plot the losses\n",
        "it = np.arange(epochs)\n",
        "plt.plot(it, train_loss, label='training loss')\n",
        "plt.plot(it, val_loss, label='validation loss')\n",
        "plt.xlabel('epochs')\n",
        "plt.ylabel('losses')\n",
        "plt.legend(loc='upper right')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "deletable": false,
        "editable": false,
        "id": "nH_GcUV3rEYV",
        "nbgrader": {
          "cell_type": "markdown",
          "checksum": "bfa73d763cb1ddbd42d8ff1548d42e1c",
          "grade": false,
          "grade_id": "cell-fe85b51930497822",
          "locked": true,
          "schema_version": 3,
          "solution": false
        }
      },
      "source": [
        "### Testing the Classsification\n",
        "\n",
        "In the testing loop we don't update the weights. The trained model is tested for all the samples in test data to compute the accuracy and observe how well the model is generalizing to the unseen data. \n",
        "\n",
        "The testing loop is setup in the following way: \n",
        "\n",
        "For every batch in the testing data\n",
        "\n",
        "- Put the model in the evaluation mode and turn off the gradients\n",
        "- Move the images and labels to the device available\n",
        "- extract output from the model for the input\n",
        "- compute the prediction class by choosing the one with maximum probability in the predictions.\n",
        "- Compare the prediction classes with true classes.\n",
        "- calculate accuracy\n",
        "- update test_loss for plots\n",
        "\n",
        "repeat\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "deletable": false,
        "id": "VwErQNu3rEYV",
        "nbgrader": {
          "cell_type": "code",
          "checksum": "6f32a3ca4f639cbfd8e5c64aa8775d11",
          "grade": false,
          "grade_id": "test_model_soln",
          "locked": false,
          "schema_version": 3,
          "solution": true
        }
      },
      "source": [
        "## Testing Loop\n",
        "\n",
        "def evaluate_model(model, loader):\n",
        "    '''\n",
        "    A function to test the trained model on the dataset and print the accuracy on the testset.\n",
        "    \n",
        "    Inputs:\n",
        "        model: Trained model\n",
        "        loader: train_loader or test_loader\n",
        "        \n",
        "    outputs:\n",
        "        accuracy. returns the accuracy of prediction\n",
        "    '''\n",
        "    model.eval()\n",
        "    with torch.no_grad():\n",
        "        correct = 0\n",
        "        total_samples = 0\n",
        "        for images, labels in loader:\n",
        "            ## Move the images and labels to gpu if is_cuda \n",
        "            ## Get the output of the model by passing images as input to the model\n",
        "            ## estimate the index of the highest output in each row, that is the predicted label \n",
        "            ## There is no need for softmax if the goal is to merely identify the predicted category\n",
        "            ## compare predictions with ground truth for number of correct samples\n",
        "            ###\n",
        "            if is_cuda:\n",
        "              images = images.cuda()\n",
        "              labels = labels.cuda()\n",
        "\n",
        "            #Predict classes using images from the test set\n",
        "            outputs = model(images)\n",
        "            _,prediction = torch.max(outputs.data, 1)\n",
        "           # prediction = prediction.cpu().numpy()\n",
        "            correct += torch.sum(prediction == labels.data)\n",
        "              \n",
        "            total_samples += labels.size(0)\n",
        "        \n",
        "        accuracy = correct/total_samples*100\n",
        "        print(\"Total Accuracy on the Input set: {} %\".format(accuracy))\n",
        "        return accuracy"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "deletable": false,
        "editable": false,
        "nbgrader": {
          "cell_type": "code",
          "checksum": "d0809b359b485a5d3c607c767fd2648e",
          "grade": false,
          "grade_id": "cell-99bb2d4e6a095086",
          "locked": true,
          "schema_version": 3,
          "solution": false,
          "task": false
        },
        "id": "NjrAl2Jj_PhX",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "19530fc3-5fa3-42f5-867d-c213171d60fb"
      },
      "source": [
        "# With these settings, obtained 95% train and 91% test accuracy\n",
        "tr_acc = evaluate_model(model, train_loader)\n",
        "ts_acc = evaluate_model(model, test_loader)\n",
        "print('Train Accuracy = %.3f'%(tr_acc))\n",
        "print('Test Accuracy = %.3f'%(ts_acc))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Total Accuracy on the Input set: 94.40833282470703 %\n",
            "Total Accuracy on the Input set: 91.05999755859375 %\n",
            "Train Accuracy = 94.408\n",
            "Test Accuracy = 91.060\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "deletable": false,
        "editable": false,
        "id": "TrcBxwqCrEYV",
        "nbgrader": {
          "cell_type": "code",
          "checksum": "ab282ae6551e71d844f7281b7ee39aff",
          "grade": true,
          "grade_id": "test_model_soln_test",
          "locked": true,
          "points": 30,
          "schema_version": 3,
          "solution": false
        }
      },
      "source": [
        "# test cases for test accuracy > 90%\n",
        "# hidden tests follow "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "deletable": false,
        "editable": false,
        "id": "9_MkVg6ErEYW",
        "nbgrader": {
          "cell_type": "code",
          "checksum": "b4ece53bc914981efc7566f454d2c571",
          "grade": false,
          "grade_id": "cell-06766329c9d4e874",
          "locked": true,
          "schema_version": 3,
          "solution": false
        },
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 879
        },
        "outputId": "2342fd87-1172-4db1-9c29-45e06220cae8"
      },
      "source": [
        "## Visualize the test samples with predicted output and true output\n",
        "images, labels = iter(test_loader).next()\n",
        "# images = images.numpy()\n",
        "if is_cuda:\n",
        "  images = images.cuda()\n",
        "  labels = labels.cuda()\n",
        "\n",
        "out = model(images)\n",
        "_, preds = torch.max(out, dim=1)\n",
        "\n",
        "images = images.to('cpu').numpy()\n",
        "\n",
        "fig = plt.figure(figsize=(15,15))\n",
        "for i in np.arange(1, 13):\n",
        "    ax = fig.add_subplot(4, 3, i)\n",
        "    plt.imshow(images[i][0])\n",
        "    ax.set_title(\"Predicted: {}/ Actual: {}\".format(image_dict[preds[i].item()], image_dict[labels[i].item()]), \n",
        "                color=('green' if preds[i] == labels[i] else 'red'))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x1080 with 12 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}