DL_assign_2.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "DL_Assign_2.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyMWJArO8BkIp6VD01lqcT6o"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "332f06dea9ae4154a2a66b457f3f7dfc": {
          "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_df9bd5d9bb3a47f8880157fc5f9da3e9",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_bd46a75de7124140911ac0f2f2d03af5",
              "IPY_MODEL_1423c5774d0f49d68ca20c3261b20d4e"
            ]
          }
        },
        "df9bd5d9bb3a47f8880157fc5f9da3e9": {
          "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
          }
        },
        "bd46a75de7124140911ac0f2f2d03af5": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_bc244f45e2d548588a37fa5a94784e11",
            "_dom_classes": [],
            "description": "",
            "_model_name": "FloatProgressModel",
            "bar_style": "info",
            "max": 1,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 1,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_db44fa2176454f948e567b3104aca393"
          }
        },
        "1423c5774d0f49d68ca20c3261b20d4e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_2ed3b68fedb94c6c9fc656ce13d9f42c",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 9920512/? [00:19&lt;00:00, 9679968.15it/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_a637b741f5a54404b42dc8758cadaab8"
          }
        },
        "bc244f45e2d548588a37fa5a94784e11": {
          "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"
          }
        },
        "db44fa2176454f948e567b3104aca393": {
          "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
          }
        },
        "2ed3b68fedb94c6c9fc656ce13d9f42c": {
          "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"
          }
        },
        "a637b741f5a54404b42dc8758cadaab8": {
          "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
          }
        },
        "637eaffa33234c67bb741ea7ad1b310d": {
          "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_18eed2f80f4145128b62c0933f61e2a3",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_db019ea3c06b4139aa78e7a11f9ea62a",
              "IPY_MODEL_5b4381e5ca3d4db68dbeb1ffe8673915"
            ]
          }
        },
        "18eed2f80f4145128b62c0933f61e2a3": {
          "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
          }
        },
        "db019ea3c06b4139aa78e7a11f9ea62a": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_b1128e8fb0f9436d9cddabb4cd180d51",
            "_dom_classes": [],
            "description": "",
            "_model_name": "FloatProgressModel",
            "bar_style": "info",
            "max": 1,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 1,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_7993fb0c878c43b19bbba158e791c65b"
          }
        },
        "5b4381e5ca3d4db68dbeb1ffe8673915": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_640119456f1e400d840c46b2d505a7c2",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 32768/? [00:18&lt;00:00, 156965.46it/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_bf2af578de484c078ee420fb87cf18cd"
          }
        },
        "b1128e8fb0f9436d9cddabb4cd180d51": {
          "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"
          }
        },
        "7993fb0c878c43b19bbba158e791c65b": {
          "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
          }
        },
        "640119456f1e400d840c46b2d505a7c2": {
          "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"
          }
        },
        "bf2af578de484c078ee420fb87cf18cd": {
          "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
          }
        },
        "4fa4f92b46004fd1bd926018d2843ed7": {
          "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_4dca01a2abf744a1b3bdcb3a5c8685f8",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_9f5a919e621d4d7995466e7d279c87c9",
              "IPY_MODEL_398a954624944a43bde530c6c6bb3159"
            ]
          }
        },
        "4dca01a2abf744a1b3bdcb3a5c8685f8": {
          "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
          }
        },
        "9f5a919e621d4d7995466e7d279c87c9": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_b8a61d06e5254d48a04282c2d9e07e86",
            "_dom_classes": [],
            "description": "",
            "_model_name": "FloatProgressModel",
            "bar_style": "info",
            "max": 1,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 1,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_a4181cf9604b495eaf624d6a680ed6c2"
          }
        },
        "398a954624944a43bde530c6c6bb3159": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_8857c1040ccb489c964a168b8356404f",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 1654784/? [00:18&lt;00:00, 10083347.47it/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_4b313604d97544f590783a9b7b6bf332"
          }
        },
        "b8a61d06e5254d48a04282c2d9e07e86": {
          "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"
          }
        },
        "a4181cf9604b495eaf624d6a680ed6c2": {
          "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
          }
        },
        "8857c1040ccb489c964a168b8356404f": {
          "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"
          }
        },
        "4b313604d97544f590783a9b7b6bf332": {
          "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
          }
        },
        "59eccc1321f24994bfed5af565317e73": {
          "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_8b8703dc9c914988bc42b274cd4ae9e5",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_d43b9c723c0a4333a1873122b1680f63",
              "IPY_MODEL_6c8b4d42ace240db9a7f4a4e2966a414"
            ]
          }
        },
        "8b8703dc9c914988bc42b274cd4ae9e5": {
          "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
          }
        },
        "d43b9c723c0a4333a1873122b1680f63": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_a924245365bc4efba379e57ce0cef92d",
            "_dom_classes": [],
            "description": "  0%",
            "_model_name": "FloatProgressModel",
            "bar_style": "info",
            "max": 1,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 0,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_fc555b0577fc4be2ad1a8e83fe57d873"
          }
        },
        "6c8b4d42ace240db9a7f4a4e2966a414": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_83e30d7c86584ae9abce9dad677afa48",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 0/4542 [00:00&lt;?, ?it/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_f9f033cec4a4478c98c0bfda0dcd1398"
          }
        },
        "a924245365bc4efba379e57ce0cef92d": {
          "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"
          }
        },
        "fc555b0577fc4be2ad1a8e83fe57d873": {
          "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
          }
        },
        "83e30d7c86584ae9abce9dad677afa48": {
          "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"
          }
        },
        "f9f033cec4a4478c98c0bfda0dcd1398": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "_6CDabfi_gJk"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "import torch.utils.data\n",
        "import torch.optim as optim \n",
        "import torchvision.datasets as dataset\n",
        "import torchvision.transforms as transforms"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "k_vOz1CJBE71"
      },
      "source": [
        "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dSBhmhKIBoLj"
      },
      "source": [
        "####Downloading the data and normalize"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 313,
          "referenced_widgets": [
            "332f06dea9ae4154a2a66b457f3f7dfc",
            "df9bd5d9bb3a47f8880157fc5f9da3e9",
            "bd46a75de7124140911ac0f2f2d03af5",
            "1423c5774d0f49d68ca20c3261b20d4e",
            "bc244f45e2d548588a37fa5a94784e11",
            "db44fa2176454f948e567b3104aca393",
            "2ed3b68fedb94c6c9fc656ce13d9f42c",
            "a637b741f5a54404b42dc8758cadaab8",
            "637eaffa33234c67bb741ea7ad1b310d",
            "18eed2f80f4145128b62c0933f61e2a3",
            "db019ea3c06b4139aa78e7a11f9ea62a",
            "5b4381e5ca3d4db68dbeb1ffe8673915",
            "b1128e8fb0f9436d9cddabb4cd180d51",
            "7993fb0c878c43b19bbba158e791c65b",
            "640119456f1e400d840c46b2d505a7c2",
            "bf2af578de484c078ee420fb87cf18cd",
            "4fa4f92b46004fd1bd926018d2843ed7",
            "4dca01a2abf744a1b3bdcb3a5c8685f8",
            "9f5a919e621d4d7995466e7d279c87c9",
            "398a954624944a43bde530c6c6bb3159",
            "b8a61d06e5254d48a04282c2d9e07e86",
            "a4181cf9604b495eaf624d6a680ed6c2",
            "8857c1040ccb489c964a168b8356404f",
            "4b313604d97544f590783a9b7b6bf332",
            "59eccc1321f24994bfed5af565317e73",
            "8b8703dc9c914988bc42b274cd4ae9e5",
            "d43b9c723c0a4333a1873122b1680f63",
            "6c8b4d42ace240db9a7f4a4e2966a414",
            "a924245365bc4efba379e57ce0cef92d",
            "fc555b0577fc4be2ad1a8e83fe57d873",
            "83e30d7c86584ae9abce9dad677afa48",
            "f9f033cec4a4478c98c0bfda0dcd1398"
          ]
        },
        "id": "Y5IBkR15zezO",
        "outputId": "854610b3-cb6a-4557-9d3f-c1ccca3cddd8"
      },
      "source": [
        "train_set = dataset.MNIST(\n",
        "    root = './data',\n",
        "    train = True,\n",
        "    download = True,\n",
        "    transform = transforms.Compose([\n",
        "        transforms.ToTensor(),\n",
        "        transforms.Normalize((0.1307,), (0.3081))                                 \n",
        "    ])\n",
        ")\n",
        "\n",
        "test_set = dataset.MNIST(\n",
        "    root = './data',\n",
        "    train = False,\n",
        "    download = True,\n",
        "    transform = transforms.Compose([\n",
        "        transforms.ToTensor(),\n",
        "        transforms.Normalize((0.1307,), (0.3081))                                 \n",
        "    ])\n",
        ")"
      ],
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz to ./data/MNIST/raw/train-images-idx3-ubyte.gz\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "332f06dea9ae4154a2a66b457f3f7dfc",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=1.0, bar_style='info', max=1.0), HTML(value='')))"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Extracting ./data/MNIST/raw/train-images-idx3-ubyte.gz to ./data/MNIST/raw\n",
            "Downloading http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz to ./data/MNIST/raw/train-labels-idx1-ubyte.gz\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "637eaffa33234c67bb741ea7ad1b310d",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=1.0, bar_style='info', max=1.0), HTML(value='')))"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Extracting ./data/MNIST/raw/train-labels-idx1-ubyte.gz to ./data/MNIST/raw\n",
            "Downloading http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz to ./data/MNIST/raw/t10k-images-idx3-ubyte.gz\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "4fa4f92b46004fd1bd926018d2843ed7",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=1.0, bar_style='info', max=1.0), HTML(value='')))"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Extracting ./data/MNIST/raw/t10k-images-idx3-ubyte.gz to ./data/MNIST/raw\n",
            "Downloading http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz to ./data/MNIST/raw/t10k-labels-idx1-ubyte.gz\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "59eccc1321f24994bfed5af565317e73",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=1.0, bar_style='info', max=1.0), HTML(value='')))"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Extracting ./data/MNIST/raw/t10k-labels-idx1-ubyte.gz to ./data/MNIST/raw\n",
            "Processing...\n",
            "Done!\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6o8cCdiiKyv7"
      },
      "source": [
        "#batchsize for training and testing\n",
        "batch_size = 1000\n",
        "\n",
        "train_loader = torch.utils.data.DataLoader(train_set, batch_size = batch_size, shuffle=True)\n",
        "test_loader = torch.utils.data.DataLoader(test_set, batch_size = batch_size, shuffle=True)\n"
      ],
      "execution_count": 45,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bahYK7hRMckC"
      },
      "source": [
        "data = enumerate(train_loader)\n",
        "batch_idx, (images_data, images_labels) = next(data)"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0Jw9VsDDhOi8",
        "outputId": "ca7a8302-d25f-4fee-c1f2-766997c80e53"
      },
      "source": [
        "images_data.shape"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([1000, 1, 28, 28])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 551
        },
        "id": "3QCP-C8VhRlK",
        "outputId": "47fc890c-5728-4e8b-97c4-dbd1bdb7e231"
      },
      "source": [
        "fig = plt.figure()\n",
        "for i in range(25):\n",
        "  plt.subplot(5,5,i+1)\n",
        "  plt.tight_layout()\n",
        "  plt.imshow(images_data[i][0], interpolation='none')\n",
        "  plt.title(\"Label: {}\".format(images_labels[i]))\n",
        "  plt.xticks([])\n",
        "  plt.yticks([])\n",
        "fig "
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 25 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 25 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NVljCLeVm-SR"
      },
      "source": [
        "To calculate output size, we will use the formula (W - K + 2P/S) + 1\n",
        "where W : width = 28, K : filter size(kernel_size) = 5, S : number of strides = 1, P : padding = 2\n",
        "image dimensions are 28 * 28 * 1\n",
        "\n",
        "The output from the 1st layer is 14 * 14 * 32\n",
        "\n",
        "The output from the 2nd layer is 7 * 7 * 64\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aZbahz_rpZ0r"
      },
      "source": [
        "#### Defining the Neural Network structure"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UgYmhEZMjL0a"
      },
      "source": [
        "class ConvNet(nn.Module):\n",
        "    def __init__(self):\n",
        "        super(ConvNet,self).__init__()\n",
        "        #Convolutional Layers\n",
        "        self.conv1 = nn.Conv2d(in_channels=1, out_channels=32, kernel_size=5, stride = 1, padding=2)\n",
        "        self.conv2 = nn.Conv2d(in_channels=32, out_channels=64, kernel_size=5, stride = 1, padding=2)\n",
        "        #Max Pooling Layer\n",
        "        self.max_pool = nn.MaxPool2d(kernel_size=2, stride=2)\n",
        "        #Fully Connected Layer\n",
        "        self.fc1 = nn.Linear(7*7*64, 1000)\n",
        "        self.fc2 = nn.Linear(1000,10)\n",
        "    \n",
        "    def forward(self,x):\n",
        "        x = F.relu(self.conv1(x))\n",
        "        x = self.max_pool(x)\n",
        "        x = F.relu(self.conv2(x))\n",
        "        x = self.max_pool(x)\n",
        "        # flatten the data dimension from 7*7*64 into 3136*1 (i.e., reshaping the tensor)\n",
        "        x = x.view(-1, 7*7*64)\n",
        "        x = self.fc1(x)\n",
        "        x = self.fc2(x)\n",
        "# No need to use softmax function as it is already included in CrossEntropyLoss()\n",
        "        return x"
      ],
      "execution_count": 29,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zEVYaXLOm8z-"
      },
      "source": [
        "epochs = 100\n",
        "log_interval = 10 \n",
        "learning_rate = 0.001\n",
        "train_counter = []\n",
        "train_loss = []\n",
        "test_losses = []\n",
        "test_counter = [i*len(train_loader.dataset) for i in range(n_epochs + 1)]"
      ],
      "execution_count": 38,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iAD8PVeOp3nD"
      },
      "source": [
        "#loss functions and optimization algo\n",
        "model = ConvNet()\n",
        "model.to(device) #transfer to GPU if available\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr = learning_rate)\n",
        "loss_func = nn.CrossEntropyLoss()"
      ],
      "execution_count": 31,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YCNmWew82uC1"
      },
      "source": [
        "####Training the model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LDoC1l6vqVbY",
        "outputId": "2e9cedc0-67f0-45c5-8e9a-a79a6d1f226c"
      },
      "source": [
        "gradient_norm = []\n",
        "for epoch in range(epochs):\n",
        "    print('Epoch {}/{}'.format(epoch, epochs - 1))\n",
        "    print('-' * 55)\n",
        "# Train the model\n",
        "    model.train() \n",
        "    for batch_idx, data_set in enumerate(train_loader):\n",
        "        \n",
        "        data , target = data_set\n",
        "        data = data.to(device)\n",
        "        target = target.to(device)\n",
        "        # Forward propagation\n",
        "        optimizer.zero_grad()  # set gradients to zero\n",
        "        output = model(data)   \n",
        "        output.to(device)\n",
        "        loss = loss_func(output,target) # compute loss\n",
        "\n",
        "        # Back propagation and perform adam optimization\n",
        "        loss.backward() # calculate gradients\n",
        "        optimizer.step()  # update weights\n",
        "\n",
        "        if batch_idx % log_interval == 0:\n",
        "         print('Train Epoch: {} [{}/{} ({:.0f}%)]\\t Loss: {:.6f}'.format(\n",
        "           epoch, batch_idx * len(data), len(train_loader.dataset),\n",
        "           100. * batch_idx / len(train_loader), loss.item()))\n",
        "         \n",
        "        train_loss.append(loss.item())\n",
        "        train_counter.append((batch_idx*100) + ((epoch-1)*len(train_loader.dataset)))\n",
        "    \n",
        "    total_tr_loss = sum(train_loss)/(batch_idx + 1)\n",
        "    print(f\"Total train loss : {total_tr_loss}\")\n",
        "    # Calculating gradient norm\n",
        "    gradients = []\n",
        "    for name, param in model.named_parameters():\n",
        "        if param.requires_grad and \"bias\" not in name:\n",
        "            # Uncomment any of the below line if you want to see the gradient values\n",
        "            #print (name, param.grad.data)\n",
        "            #print(f'{name}.grad', param.grad, epoch)\n",
        "            gradients.append(torch.flatten(param.grad.data))\n",
        "    print(f'Grad_norm : {torch.norm(torch.cat([gradients[0],gradients[1],gradients[2],gradients[3]],dim=0))}')\n",
        "    gradient_norm.append(torch.norm(torch.cat([gradients[0],gradients[1],gradients[2],gradients[3]],dim=0)).item())"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 0/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 0 [0/60000 (0%)]\t Loss: 0.010447\n",
            "Train Epoch: 0 [10000/60000 (17%)]\t Loss: 0.004814\n",
            "Train Epoch: 0 [20000/60000 (33%)]\t Loss: 0.002930\n",
            "Train Epoch: 0 [30000/60000 (50%)]\t Loss: 0.010728\n",
            "Train Epoch: 0 [40000/60000 (67%)]\t Loss: 0.002728\n",
            "Train Epoch: 0 [50000/60000 (83%)]\t Loss: 0.014765\n",
            "Total train loss : 0.007159978830410789\n",
            "Grad_norm : 0.2707873582839966\n",
            "Epoch 1/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 1 [0/60000 (0%)]\t Loss: 0.004237\n",
            "Train Epoch: 1 [10000/60000 (17%)]\t Loss: 0.002817\n",
            "Train Epoch: 1 [20000/60000 (33%)]\t Loss: 0.004179\n",
            "Train Epoch: 1 [30000/60000 (50%)]\t Loss: 0.006608\n",
            "Train Epoch: 1 [40000/60000 (67%)]\t Loss: 0.003489\n",
            "Train Epoch: 1 [50000/60000 (83%)]\t Loss: 0.012051\n",
            "Total train loss : 0.01392101778377158\n",
            "Grad_norm : 0.1276119500398636\n",
            "Epoch 2/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 2 [0/60000 (0%)]\t Loss: 0.004340\n",
            "Train Epoch: 2 [10000/60000 (17%)]\t Loss: 0.007861\n",
            "Train Epoch: 2 [20000/60000 (33%)]\t Loss: 0.009216\n",
            "Train Epoch: 2 [30000/60000 (50%)]\t Loss: 0.003708\n",
            "Train Epoch: 2 [40000/60000 (67%)]\t Loss: 0.004720\n",
            "Train Epoch: 2 [50000/60000 (83%)]\t Loss: 0.007240\n",
            "Total train loss : 0.019708929045979554\n",
            "Grad_norm : 0.0879933312535286\n",
            "Epoch 3/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 3 [0/60000 (0%)]\t Loss: 0.002675\n",
            "Train Epoch: 3 [10000/60000 (17%)]\t Loss: 0.004522\n",
            "Train Epoch: 3 [20000/60000 (33%)]\t Loss: 0.003136\n",
            "Train Epoch: 3 [30000/60000 (50%)]\t Loss: 0.002558\n",
            "Train Epoch: 3 [40000/60000 (67%)]\t Loss: 0.003543\n",
            "Train Epoch: 3 [50000/60000 (83%)]\t Loss: 0.005429\n",
            "Total train loss : 0.025263497505026558\n",
            "Grad_norm : 0.30089297890663147\n",
            "Epoch 4/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 4 [0/60000 (0%)]\t Loss: 0.001813\n",
            "Train Epoch: 4 [10000/60000 (17%)]\t Loss: 0.001557\n",
            "Train Epoch: 4 [20000/60000 (33%)]\t Loss: 0.004349\n",
            "Train Epoch: 4 [30000/60000 (50%)]\t Loss: 0.001727\n",
            "Train Epoch: 4 [40000/60000 (67%)]\t Loss: 0.003274\n",
            "Train Epoch: 4 [50000/60000 (83%)]\t Loss: 0.002169\n",
            "Total train loss : 0.030062260036356748\n",
            "Grad_norm : 0.18500757217407227\n",
            "Epoch 5/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 5 [0/60000 (0%)]\t Loss: 0.001581\n",
            "Train Epoch: 5 [10000/60000 (17%)]\t Loss: 0.001283\n",
            "Train Epoch: 5 [20000/60000 (33%)]\t Loss: 0.002294\n",
            "Train Epoch: 5 [30000/60000 (50%)]\t Loss: 0.008966\n",
            "Train Epoch: 5 [40000/60000 (67%)]\t Loss: 0.005079\n",
            "Train Epoch: 5 [50000/60000 (83%)]\t Loss: 0.003737\n",
            "Total train loss : 0.034305543909431434\n",
            "Grad_norm : 0.09194664657115936\n",
            "Epoch 6/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 6 [0/60000 (0%)]\t Loss: 0.003273\n",
            "Train Epoch: 6 [10000/60000 (17%)]\t Loss: 0.003256\n",
            "Train Epoch: 6 [20000/60000 (33%)]\t Loss: 0.005094\n",
            "Train Epoch: 6 [30000/60000 (50%)]\t Loss: 0.006264\n",
            "Train Epoch: 6 [40000/60000 (67%)]\t Loss: 0.002137\n",
            "Train Epoch: 6 [50000/60000 (83%)]\t Loss: 0.003328\n",
            "Total train loss : 0.038639558205613864\n",
            "Grad_norm : 0.3215278387069702\n",
            "Epoch 7/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 7 [0/60000 (0%)]\t Loss: 0.001029\n",
            "Train Epoch: 7 [10000/60000 (17%)]\t Loss: 0.008952\n",
            "Train Epoch: 7 [20000/60000 (33%)]\t Loss: 0.004933\n",
            "Train Epoch: 7 [30000/60000 (50%)]\t Loss: 0.004924\n",
            "Train Epoch: 7 [40000/60000 (67%)]\t Loss: 0.002620\n",
            "Train Epoch: 7 [50000/60000 (83%)]\t Loss: 0.000981\n",
            "Total train loss : 0.043976879555460376\n",
            "Grad_norm : 0.14471127092838287\n",
            "Epoch 8/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 8 [0/60000 (0%)]\t Loss: 0.001788\n",
            "Train Epoch: 8 [10000/60000 (17%)]\t Loss: 0.002828\n",
            "Train Epoch: 8 [20000/60000 (33%)]\t Loss: 0.002627\n",
            "Train Epoch: 8 [30000/60000 (50%)]\t Loss: 0.001168\n",
            "Train Epoch: 8 [40000/60000 (67%)]\t Loss: 0.004568\n",
            "Train Epoch: 8 [50000/60000 (83%)]\t Loss: 0.002093\n",
            "Total train loss : 0.04752422679254475\n",
            "Grad_norm : 0.14858460426330566\n",
            "Epoch 9/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 9 [0/60000 (0%)]\t Loss: 0.002444\n",
            "Train Epoch: 9 [10000/60000 (17%)]\t Loss: 0.004473\n",
            "Train Epoch: 9 [20000/60000 (33%)]\t Loss: 0.001103\n",
            "Train Epoch: 9 [30000/60000 (50%)]\t Loss: 0.011726\n",
            "Train Epoch: 9 [40000/60000 (67%)]\t Loss: 0.007546\n",
            "Train Epoch: 9 [50000/60000 (83%)]\t Loss: 0.001710\n",
            "Total train loss : 0.0522087161099383\n",
            "Grad_norm : 0.2857171595096588\n",
            "Epoch 10/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 10 [0/60000 (0%)]\t Loss: 0.002310\n",
            "Train Epoch: 10 [10000/60000 (17%)]\t Loss: 0.002871\n",
            "Train Epoch: 10 [20000/60000 (33%)]\t Loss: 0.005847\n",
            "Train Epoch: 10 [30000/60000 (50%)]\t Loss: 0.006560\n",
            "Train Epoch: 10 [40000/60000 (67%)]\t Loss: 0.002084\n",
            "Train Epoch: 10 [50000/60000 (83%)]\t Loss: 0.006413\n",
            "Total train loss : 0.05644216859849015\n",
            "Grad_norm : 0.12509453296661377\n",
            "Epoch 11/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 11 [0/60000 (0%)]\t Loss: 0.000805\n",
            "Train Epoch: 11 [10000/60000 (17%)]\t Loss: 0.001842\n",
            "Train Epoch: 11 [20000/60000 (33%)]\t Loss: 0.000721\n",
            "Train Epoch: 11 [30000/60000 (50%)]\t Loss: 0.002007\n",
            "Train Epoch: 11 [40000/60000 (67%)]\t Loss: 0.001775\n",
            "Train Epoch: 11 [50000/60000 (83%)]\t Loss: 0.001773\n",
            "Total train loss : 0.059288082611116504\n",
            "Grad_norm : 0.2332208752632141\n",
            "Epoch 12/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 12 [0/60000 (0%)]\t Loss: 0.000593\n",
            "Train Epoch: 12 [10000/60000 (17%)]\t Loss: 0.000771\n",
            "Train Epoch: 12 [20000/60000 (33%)]\t Loss: 0.001854\n",
            "Train Epoch: 12 [30000/60000 (50%)]\t Loss: 0.000704\n",
            "Train Epoch: 12 [40000/60000 (67%)]\t Loss: 0.000572\n",
            "Train Epoch: 12 [50000/60000 (83%)]\t Loss: 0.000454\n",
            "Total train loss : 0.060468337216298094\n",
            "Grad_norm : 0.14655953645706177\n",
            "Epoch 13/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 13 [0/60000 (0%)]\t Loss: 0.003977\n",
            "Train Epoch: 13 [10000/60000 (17%)]\t Loss: 0.001143\n",
            "Train Epoch: 13 [20000/60000 (33%)]\t Loss: 0.001482\n",
            "Train Epoch: 13 [30000/60000 (50%)]\t Loss: 0.001323\n",
            "Train Epoch: 13 [40000/60000 (67%)]\t Loss: 0.002442\n",
            "Train Epoch: 13 [50000/60000 (83%)]\t Loss: 0.006445\n",
            "Total train loss : 0.061721646925919534\n",
            "Grad_norm : 0.09745344519615173\n",
            "Epoch 14/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 14 [0/60000 (0%)]\t Loss: 0.000558\n",
            "Train Epoch: 14 [10000/60000 (17%)]\t Loss: 0.000443\n",
            "Train Epoch: 14 [20000/60000 (33%)]\t Loss: 0.001372\n",
            "Train Epoch: 14 [30000/60000 (50%)]\t Loss: 0.000292\n",
            "Train Epoch: 14 [40000/60000 (67%)]\t Loss: 0.000467\n",
            "Train Epoch: 14 [50000/60000 (83%)]\t Loss: 0.001372\n",
            "Total train loss : 0.06282087530507852\n",
            "Grad_norm : 0.18200236558914185\n",
            "Epoch 15/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 15 [0/60000 (0%)]\t Loss: 0.000463\n",
            "Train Epoch: 15 [10000/60000 (17%)]\t Loss: 0.000538\n",
            "Train Epoch: 15 [20000/60000 (33%)]\t Loss: 0.000300\n",
            "Train Epoch: 15 [30000/60000 (50%)]\t Loss: 0.000177\n",
            "Train Epoch: 15 [40000/60000 (67%)]\t Loss: 0.000638\n",
            "Train Epoch: 15 [50000/60000 (83%)]\t Loss: 0.000157\n",
            "Total train loss : 0.06359025311515627\n",
            "Grad_norm : 0.01051246840506792\n",
            "Epoch 16/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 16 [0/60000 (0%)]\t Loss: 0.000164\n",
            "Train Epoch: 16 [10000/60000 (17%)]\t Loss: 0.000129\n",
            "Train Epoch: 16 [20000/60000 (33%)]\t Loss: 0.000118\n",
            "Train Epoch: 16 [30000/60000 (50%)]\t Loss: 0.000087\n",
            "Train Epoch: 16 [40000/60000 (67%)]\t Loss: 0.000061\n",
            "Train Epoch: 16 [50000/60000 (83%)]\t Loss: 0.001108\n",
            "Total train loss : 0.06521723405785451\n",
            "Grad_norm : 0.05849435552954674\n",
            "Epoch 17/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 17 [0/60000 (0%)]\t Loss: 0.007295\n",
            "Train Epoch: 17 [10000/60000 (17%)]\t Loss: 0.000446\n",
            "Train Epoch: 17 [20000/60000 (33%)]\t Loss: 0.006915\n",
            "Train Epoch: 17 [30000/60000 (50%)]\t Loss: 0.017335\n",
            "Train Epoch: 17 [40000/60000 (67%)]\t Loss: 0.003736\n",
            "Train Epoch: 17 [50000/60000 (83%)]\t Loss: 0.003384\n",
            "Total train loss : 0.07279823344955123\n",
            "Grad_norm : 0.30399301648139954\n",
            "Epoch 18/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 18 [0/60000 (0%)]\t Loss: 0.005024\n",
            "Train Epoch: 18 [10000/60000 (17%)]\t Loss: 0.003849\n",
            "Train Epoch: 18 [20000/60000 (33%)]\t Loss: 0.011329\n",
            "Train Epoch: 18 [30000/60000 (50%)]\t Loss: 0.000502\n",
            "Train Epoch: 18 [40000/60000 (67%)]\t Loss: 0.006097\n",
            "Train Epoch: 18 [50000/60000 (83%)]\t Loss: 0.003210\n",
            "Total train loss : 0.07859961480365503\n",
            "Grad_norm : 0.2200431078672409\n",
            "Epoch 19/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 19 [0/60000 (0%)]\t Loss: 0.002754\n",
            "Train Epoch: 19 [10000/60000 (17%)]\t Loss: 0.002498\n",
            "Train Epoch: 19 [20000/60000 (33%)]\t Loss: 0.004559\n",
            "Train Epoch: 19 [30000/60000 (50%)]\t Loss: 0.006432\n",
            "Train Epoch: 19 [40000/60000 (67%)]\t Loss: 0.004065\n",
            "Train Epoch: 19 [50000/60000 (83%)]\t Loss: 0.003829\n",
            "Total train loss : 0.08489612798087667\n",
            "Grad_norm : 0.10878072679042816\n",
            "Epoch 20/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 20 [0/60000 (0%)]\t Loss: 0.002443\n",
            "Train Epoch: 20 [10000/60000 (17%)]\t Loss: 0.008285\n",
            "Train Epoch: 20 [20000/60000 (33%)]\t Loss: 0.008291\n",
            "Train Epoch: 20 [30000/60000 (50%)]\t Loss: 0.014471\n",
            "Train Epoch: 20 [40000/60000 (67%)]\t Loss: 0.015141\n",
            "Train Epoch: 20 [50000/60000 (83%)]\t Loss: 0.003824\n",
            "Total train loss : 0.09280496433044998\n",
            "Grad_norm : 0.25449100136756897\n",
            "Epoch 21/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 21 [0/60000 (0%)]\t Loss: 0.003601\n",
            "Train Epoch: 21 [10000/60000 (17%)]\t Loss: 0.004630\n",
            "Train Epoch: 21 [20000/60000 (33%)]\t Loss: 0.004226\n",
            "Train Epoch: 21 [30000/60000 (50%)]\t Loss: 0.001684\n",
            "Train Epoch: 21 [40000/60000 (67%)]\t Loss: 0.001698\n",
            "Train Epoch: 21 [50000/60000 (83%)]\t Loss: 0.010292\n",
            "Total train loss : 0.09792079332243399\n",
            "Grad_norm : 0.2603347897529602\n",
            "Epoch 22/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 22 [0/60000 (0%)]\t Loss: 0.003333\n",
            "Train Epoch: 22 [10000/60000 (17%)]\t Loss: 0.006274\n",
            "Train Epoch: 22 [20000/60000 (33%)]\t Loss: 0.000585\n",
            "Train Epoch: 22 [30000/60000 (50%)]\t Loss: 0.004555\n",
            "Train Epoch: 22 [40000/60000 (67%)]\t Loss: 0.003793\n",
            "Train Epoch: 22 [50000/60000 (83%)]\t Loss: 0.003589\n",
            "Total train loss : 0.10232146346230972\n",
            "Grad_norm : 0.15701566636562347\n",
            "Epoch 23/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 23 [0/60000 (0%)]\t Loss: 0.001824\n",
            "Train Epoch: 23 [10000/60000 (17%)]\t Loss: 0.003342\n",
            "Train Epoch: 23 [20000/60000 (33%)]\t Loss: 0.003782\n",
            "Train Epoch: 23 [30000/60000 (50%)]\t Loss: 0.001609\n",
            "Train Epoch: 23 [40000/60000 (67%)]\t Loss: 0.006418\n",
            "Train Epoch: 23 [50000/60000 (83%)]\t Loss: 0.001553\n",
            "Total train loss : 0.10613780458164304\n",
            "Grad_norm : 0.010080594569444656\n",
            "Epoch 24/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 24 [0/60000 (0%)]\t Loss: 0.000663\n",
            "Train Epoch: 24 [10000/60000 (17%)]\t Loss: 0.002865\n",
            "Train Epoch: 24 [20000/60000 (33%)]\t Loss: 0.002967\n",
            "Train Epoch: 24 [30000/60000 (50%)]\t Loss: 0.000626\n",
            "Train Epoch: 24 [40000/60000 (67%)]\t Loss: 0.000588\n",
            "Train Epoch: 24 [50000/60000 (83%)]\t Loss: 0.001524\n",
            "Total train loss : 0.1106713712267568\n",
            "Grad_norm : 0.07167573273181915\n",
            "Epoch 25/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 25 [0/60000 (0%)]\t Loss: 0.001318\n",
            "Train Epoch: 25 [10000/60000 (17%)]\t Loss: 0.000768\n",
            "Train Epoch: 25 [20000/60000 (33%)]\t Loss: 0.004294\n",
            "Train Epoch: 25 [30000/60000 (50%)]\t Loss: 0.001395\n",
            "Train Epoch: 25 [40000/60000 (67%)]\t Loss: 0.001320\n",
            "Train Epoch: 25 [50000/60000 (83%)]\t Loss: 0.003637\n",
            "Total train loss : 0.1126033249557319\n",
            "Grad_norm : 0.04945910722017288\n",
            "Epoch 26/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 26 [0/60000 (0%)]\t Loss: 0.003059\n",
            "Train Epoch: 26 [10000/60000 (17%)]\t Loss: 0.000111\n",
            "Train Epoch: 26 [20000/60000 (33%)]\t Loss: 0.000824\n",
            "Train Epoch: 26 [30000/60000 (50%)]\t Loss: 0.003483\n",
            "Train Epoch: 26 [40000/60000 (67%)]\t Loss: 0.000322\n",
            "Train Epoch: 26 [50000/60000 (83%)]\t Loss: 0.000534\n",
            "Total train loss : 0.11328285424148514\n",
            "Grad_norm : 0.008260210044682026\n",
            "Epoch 27/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 27 [0/60000 (0%)]\t Loss: 0.000008\n",
            "Train Epoch: 27 [10000/60000 (17%)]\t Loss: 0.000164\n",
            "Train Epoch: 27 [20000/60000 (33%)]\t Loss: 0.002263\n",
            "Train Epoch: 27 [30000/60000 (50%)]\t Loss: 0.000067\n",
            "Train Epoch: 27 [40000/60000 (67%)]\t Loss: 0.000121\n",
            "Train Epoch: 27 [50000/60000 (83%)]\t Loss: 0.002459\n",
            "Total train loss : 0.11384193811074207\n",
            "Grad_norm : 0.01938587985932827\n",
            "Epoch 28/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 28 [0/60000 (0%)]\t Loss: 0.000203\n",
            "Train Epoch: 28 [10000/60000 (17%)]\t Loss: 0.005163\n",
            "Train Epoch: 28 [20000/60000 (33%)]\t Loss: 0.000178\n",
            "Train Epoch: 28 [30000/60000 (50%)]\t Loss: 0.000462\n",
            "Train Epoch: 28 [40000/60000 (67%)]\t Loss: 0.000123\n",
            "Train Epoch: 28 [50000/60000 (83%)]\t Loss: 0.000738\n",
            "Total train loss : 0.11652666018267155\n",
            "Grad_norm : 0.11531450599431992\n",
            "Epoch 29/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 29 [0/60000 (0%)]\t Loss: 0.000064\n",
            "Train Epoch: 29 [10000/60000 (17%)]\t Loss: 0.000272\n",
            "Train Epoch: 29 [20000/60000 (33%)]\t Loss: 0.000113\n",
            "Train Epoch: 29 [30000/60000 (50%)]\t Loss: 0.002901\n",
            "Train Epoch: 29 [40000/60000 (67%)]\t Loss: 0.001516\n",
            "Train Epoch: 29 [50000/60000 (83%)]\t Loss: 0.000549\n",
            "Total train loss : 0.11794532788955318\n",
            "Grad_norm : 0.11001938581466675\n",
            "Epoch 30/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 30 [0/60000 (0%)]\t Loss: 0.000751\n",
            "Train Epoch: 30 [10000/60000 (17%)]\t Loss: 0.000231\n",
            "Train Epoch: 30 [20000/60000 (33%)]\t Loss: 0.000229\n",
            "Train Epoch: 30 [30000/60000 (50%)]\t Loss: 0.000452\n",
            "Train Epoch: 30 [40000/60000 (67%)]\t Loss: 0.000920\n",
            "Train Epoch: 30 [50000/60000 (83%)]\t Loss: 0.000992\n",
            "Total train loss : 0.12054516131570381\n",
            "Grad_norm : 0.10923459380865097\n",
            "Epoch 31/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 31 [0/60000 (0%)]\t Loss: 0.001736\n",
            "Train Epoch: 31 [10000/60000 (17%)]\t Loss: 0.001363\n",
            "Train Epoch: 31 [20000/60000 (33%)]\t Loss: 0.000386\n",
            "Train Epoch: 31 [30000/60000 (50%)]\t Loss: 0.004614\n",
            "Train Epoch: 31 [40000/60000 (67%)]\t Loss: 0.005082\n",
            "Train Epoch: 31 [50000/60000 (83%)]\t Loss: 0.000282\n",
            "Total train loss : 0.12265284424734091\n",
            "Grad_norm : 0.01158500649034977\n",
            "Epoch 32/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 32 [0/60000 (0%)]\t Loss: 0.001908\n",
            "Train Epoch: 32 [10000/60000 (17%)]\t Loss: 0.000124\n",
            "Train Epoch: 32 [20000/60000 (33%)]\t Loss: 0.000072\n",
            "Train Epoch: 32 [30000/60000 (50%)]\t Loss: 0.000651\n",
            "Train Epoch: 32 [40000/60000 (67%)]\t Loss: 0.000722\n",
            "Train Epoch: 32 [50000/60000 (83%)]\t Loss: 0.001434\n",
            "Total train loss : 0.12403942420583007\n",
            "Grad_norm : 0.2317529022693634\n",
            "Epoch 33/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 33 [0/60000 (0%)]\t Loss: 0.017266\n",
            "Train Epoch: 33 [10000/60000 (17%)]\t Loss: 0.000488\n",
            "Train Epoch: 33 [20000/60000 (33%)]\t Loss: 0.003145\n",
            "Train Epoch: 33 [30000/60000 (50%)]\t Loss: 0.000316\n",
            "Train Epoch: 33 [40000/60000 (67%)]\t Loss: 0.017131\n",
            "Train Epoch: 33 [50000/60000 (83%)]\t Loss: 0.006131\n",
            "Total train loss : 0.12898316606007634\n",
            "Grad_norm : 0.2429129034280777\n",
            "Epoch 34/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 34 [0/60000 (0%)]\t Loss: 0.002632\n",
            "Train Epoch: 34 [10000/60000 (17%)]\t Loss: 0.003665\n",
            "Train Epoch: 34 [20000/60000 (33%)]\t Loss: 0.001303\n",
            "Train Epoch: 34 [30000/60000 (50%)]\t Loss: 0.006654\n",
            "Train Epoch: 34 [40000/60000 (67%)]\t Loss: 0.011424\n",
            "Train Epoch: 34 [50000/60000 (83%)]\t Loss: 0.002758\n",
            "Total train loss : 0.1356212914191474\n",
            "Grad_norm : 0.2789785861968994\n",
            "Epoch 35/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 35 [0/60000 (0%)]\t Loss: 0.003372\n",
            "Train Epoch: 35 [10000/60000 (17%)]\t Loss: 0.002702\n",
            "Train Epoch: 35 [20000/60000 (33%)]\t Loss: 0.005356\n",
            "Train Epoch: 35 [30000/60000 (50%)]\t Loss: 0.001742\n",
            "Train Epoch: 35 [40000/60000 (67%)]\t Loss: 0.004335\n",
            "Train Epoch: 35 [50000/60000 (83%)]\t Loss: 0.001191\n",
            "Total train loss : 0.13874398311213554\n",
            "Grad_norm : 0.14112526178359985\n",
            "Epoch 36/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 36 [0/60000 (0%)]\t Loss: 0.003858\n",
            "Train Epoch: 36 [10000/60000 (17%)]\t Loss: 0.000602\n",
            "Train Epoch: 36 [20000/60000 (33%)]\t Loss: 0.000349\n",
            "Train Epoch: 36 [30000/60000 (50%)]\t Loss: 0.000137\n",
            "Train Epoch: 36 [40000/60000 (67%)]\t Loss: 0.007700\n",
            "Train Epoch: 36 [50000/60000 (83%)]\t Loss: 0.003999\n",
            "Total train loss : 0.14152463727224737\n",
            "Grad_norm : 0.209483340382576\n",
            "Epoch 37/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 37 [0/60000 (0%)]\t Loss: 0.023936\n",
            "Train Epoch: 37 [10000/60000 (17%)]\t Loss: 0.011947\n",
            "Train Epoch: 37 [20000/60000 (33%)]\t Loss: 0.005660\n",
            "Train Epoch: 37 [30000/60000 (50%)]\t Loss: 0.003617\n",
            "Train Epoch: 37 [40000/60000 (67%)]\t Loss: 0.000063\n",
            "Train Epoch: 37 [50000/60000 (83%)]\t Loss: 0.000491\n",
            "Total train loss : 0.1450002541798464\n",
            "Grad_norm : 0.016209393739700317\n",
            "Epoch 38/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 38 [0/60000 (0%)]\t Loss: 0.004155\n",
            "Train Epoch: 38 [10000/60000 (17%)]\t Loss: 0.000641\n",
            "Train Epoch: 38 [20000/60000 (33%)]\t Loss: 0.014645\n",
            "Train Epoch: 38 [30000/60000 (50%)]\t Loss: 0.002997\n",
            "Train Epoch: 38 [40000/60000 (67%)]\t Loss: 0.003331\n",
            "Train Epoch: 38 [50000/60000 (83%)]\t Loss: 0.003273\n",
            "Total train loss : 0.15056578150495928\n",
            "Grad_norm : 0.7279402613639832\n",
            "Epoch 39/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 39 [0/60000 (0%)]\t Loss: 0.007580\n",
            "Train Epoch: 39 [10000/60000 (17%)]\t Loss: 0.006053\n",
            "Train Epoch: 39 [20000/60000 (33%)]\t Loss: 0.002100\n",
            "Train Epoch: 39 [30000/60000 (50%)]\t Loss: 0.010477\n",
            "Train Epoch: 39 [40000/60000 (67%)]\t Loss: 0.006718\n",
            "Train Epoch: 39 [50000/60000 (83%)]\t Loss: 0.003191\n",
            "Total train loss : 0.1575844949635363\n",
            "Grad_norm : 0.21069638431072235\n",
            "Epoch 40/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 40 [0/60000 (0%)]\t Loss: 0.000101\n",
            "Train Epoch: 40 [10000/60000 (17%)]\t Loss: 0.011098\n",
            "Train Epoch: 40 [20000/60000 (33%)]\t Loss: 0.003432\n",
            "Train Epoch: 40 [30000/60000 (50%)]\t Loss: 0.003241\n",
            "Train Epoch: 40 [40000/60000 (67%)]\t Loss: 0.007597\n",
            "Train Epoch: 40 [50000/60000 (83%)]\t Loss: 0.001876\n",
            "Total train loss : 0.16103957671537802\n",
            "Grad_norm : 0.04361412674188614\n",
            "Epoch 41/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 41 [0/60000 (0%)]\t Loss: 0.001043\n",
            "Train Epoch: 41 [10000/60000 (17%)]\t Loss: 0.009023\n",
            "Train Epoch: 41 [20000/60000 (33%)]\t Loss: 0.001601\n",
            "Train Epoch: 41 [30000/60000 (50%)]\t Loss: 0.000049\n",
            "Train Epoch: 41 [40000/60000 (67%)]\t Loss: 0.001593\n",
            "Train Epoch: 41 [50000/60000 (83%)]\t Loss: 0.000203\n",
            "Total train loss : 0.16253292169417363\n",
            "Grad_norm : 0.0016763255698606372\n",
            "Epoch 42/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 42 [0/60000 (0%)]\t Loss: 0.000367\n",
            "Train Epoch: 42 [10000/60000 (17%)]\t Loss: 0.000293\n",
            "Train Epoch: 42 [20000/60000 (33%)]\t Loss: 0.000045\n",
            "Train Epoch: 42 [30000/60000 (50%)]\t Loss: 0.004236\n",
            "Train Epoch: 42 [40000/60000 (67%)]\t Loss: 0.000013\n",
            "Train Epoch: 42 [50000/60000 (83%)]\t Loss: 0.000058\n",
            "Total train loss : 0.16417570286525157\n",
            "Grad_norm : 0.15856300294399261\n",
            "Epoch 43/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 43 [0/60000 (0%)]\t Loss: 0.001288\n",
            "Train Epoch: 43 [10000/60000 (17%)]\t Loss: 0.004243\n",
            "Train Epoch: 43 [20000/60000 (33%)]\t Loss: 0.003431\n",
            "Train Epoch: 43 [30000/60000 (50%)]\t Loss: 0.001100\n",
            "Train Epoch: 43 [40000/60000 (67%)]\t Loss: 0.000222\n",
            "Train Epoch: 43 [50000/60000 (83%)]\t Loss: 0.000414\n",
            "Total train loss : 0.16546148088073853\n",
            "Grad_norm : 0.03635428845882416\n",
            "Epoch 44/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 44 [0/60000 (0%)]\t Loss: 0.000187\n",
            "Train Epoch: 44 [10000/60000 (17%)]\t Loss: 0.000448\n",
            "Train Epoch: 44 [20000/60000 (33%)]\t Loss: 0.000044\n",
            "Train Epoch: 44 [30000/60000 (50%)]\t Loss: 0.000832\n",
            "Train Epoch: 44 [40000/60000 (67%)]\t Loss: 0.000122\n",
            "Train Epoch: 44 [50000/60000 (83%)]\t Loss: 0.000062\n",
            "Total train loss : 0.16640137468856817\n",
            "Grad_norm : 0.010048238560557365\n",
            "Epoch 45/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 45 [0/60000 (0%)]\t Loss: 0.009552\n",
            "Train Epoch: 45 [10000/60000 (17%)]\t Loss: 0.000222\n",
            "Train Epoch: 45 [20000/60000 (33%)]\t Loss: 0.000599\n",
            "Train Epoch: 45 [30000/60000 (50%)]\t Loss: 0.000004\n",
            "Train Epoch: 45 [40000/60000 (67%)]\t Loss: 0.000057\n",
            "Train Epoch: 45 [50000/60000 (83%)]\t Loss: 0.000376\n",
            "Total train loss : 0.167211231913916\n",
            "Grad_norm : 0.05363277718424797\n",
            "Epoch 46/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 46 [0/60000 (0%)]\t Loss: 0.000020\n",
            "Train Epoch: 46 [10000/60000 (17%)]\t Loss: 0.003307\n",
            "Train Epoch: 46 [20000/60000 (33%)]\t Loss: 0.003606\n",
            "Train Epoch: 46 [30000/60000 (50%)]\t Loss: 0.005060\n",
            "Train Epoch: 46 [40000/60000 (67%)]\t Loss: 0.005579\n",
            "Train Epoch: 46 [50000/60000 (83%)]\t Loss: 0.000306\n",
            "Total train loss : 0.1709023631236126\n",
            "Grad_norm : 0.19910122454166412\n",
            "Epoch 47/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 47 [0/60000 (0%)]\t Loss: 0.000717\n",
            "Train Epoch: 47 [10000/60000 (17%)]\t Loss: 0.022024\n",
            "Train Epoch: 47 [20000/60000 (33%)]\t Loss: 0.001329\n",
            "Train Epoch: 47 [30000/60000 (50%)]\t Loss: 0.002104\n",
            "Train Epoch: 47 [40000/60000 (67%)]\t Loss: 0.000585\n",
            "Train Epoch: 47 [50000/60000 (83%)]\t Loss: 0.000056\n",
            "Total train loss : 0.174129044934557\n",
            "Grad_norm : 0.022655902430415154\n",
            "Epoch 48/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 48 [0/60000 (0%)]\t Loss: 0.000394\n",
            "Train Epoch: 48 [10000/60000 (17%)]\t Loss: 0.006395\n",
            "Train Epoch: 48 [20000/60000 (33%)]\t Loss: 0.000221\n",
            "Train Epoch: 48 [30000/60000 (50%)]\t Loss: 0.000061\n",
            "Train Epoch: 48 [40000/60000 (67%)]\t Loss: 0.001551\n",
            "Train Epoch: 48 [50000/60000 (83%)]\t Loss: 0.000948\n",
            "Total train loss : 0.17542975115650278\n",
            "Grad_norm : 0.015337440185248852\n",
            "Epoch 49/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 49 [0/60000 (0%)]\t Loss: 0.000032\n",
            "Train Epoch: 49 [10000/60000 (17%)]\t Loss: 0.001000\n",
            "Train Epoch: 49 [20000/60000 (33%)]\t Loss: 0.000536\n",
            "Train Epoch: 49 [30000/60000 (50%)]\t Loss: 0.000112\n",
            "Train Epoch: 49 [40000/60000 (67%)]\t Loss: 0.002588\n",
            "Train Epoch: 49 [50000/60000 (83%)]\t Loss: 0.000798\n",
            "Total train loss : 0.17703468262085\n",
            "Grad_norm : 0.2635476887226105\n",
            "Epoch 50/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 50 [0/60000 (0%)]\t Loss: 0.003840\n",
            "Train Epoch: 50 [10000/60000 (17%)]\t Loss: 0.000357\n",
            "Train Epoch: 50 [20000/60000 (33%)]\t Loss: 0.000024\n",
            "Train Epoch: 50 [30000/60000 (50%)]\t Loss: 0.006495\n",
            "Train Epoch: 50 [40000/60000 (67%)]\t Loss: 0.009717\n",
            "Train Epoch: 50 [50000/60000 (83%)]\t Loss: 0.001779\n",
            "Total train loss : 0.1799864085538123\n",
            "Grad_norm : 0.29044240713119507\n",
            "Epoch 51/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 51 [0/60000 (0%)]\t Loss: 0.000564\n",
            "Train Epoch: 51 [10000/60000 (17%)]\t Loss: 0.002240\n",
            "Train Epoch: 51 [20000/60000 (33%)]\t Loss: 0.000396\n",
            "Train Epoch: 51 [30000/60000 (50%)]\t Loss: 0.002036\n",
            "Train Epoch: 51 [40000/60000 (67%)]\t Loss: 0.005576\n",
            "Train Epoch: 51 [50000/60000 (83%)]\t Loss: 0.000597\n",
            "Total train loss : 0.18266317970273274\n",
            "Grad_norm : 0.03804609552025795\n",
            "Epoch 52/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 52 [0/60000 (0%)]\t Loss: 0.000380\n",
            "Train Epoch: 52 [10000/60000 (17%)]\t Loss: 0.000176\n",
            "Train Epoch: 52 [20000/60000 (33%)]\t Loss: 0.000011\n",
            "Train Epoch: 52 [30000/60000 (50%)]\t Loss: 0.000354\n",
            "Train Epoch: 52 [40000/60000 (67%)]\t Loss: 0.000132\n",
            "Train Epoch: 52 [50000/60000 (83%)]\t Loss: 0.000111\n",
            "Total train loss : 0.1840924571172536\n",
            "Grad_norm : 0.11544378846883774\n",
            "Epoch 53/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 53 [0/60000 (0%)]\t Loss: 0.000131\n",
            "Train Epoch: 53 [10000/60000 (17%)]\t Loss: 0.000980\n",
            "Train Epoch: 53 [20000/60000 (33%)]\t Loss: 0.000130\n",
            "Train Epoch: 53 [30000/60000 (50%)]\t Loss: 0.001202\n",
            "Train Epoch: 53 [40000/60000 (67%)]\t Loss: 0.000566\n",
            "Train Epoch: 53 [50000/60000 (83%)]\t Loss: 0.000045\n",
            "Total train loss : 0.18541554167840482\n",
            "Grad_norm : 0.2120893895626068\n",
            "Epoch 54/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 54 [0/60000 (0%)]\t Loss: 0.001266\n",
            "Train Epoch: 54 [10000/60000 (17%)]\t Loss: 0.000201\n",
            "Train Epoch: 54 [20000/60000 (33%)]\t Loss: 0.004063\n",
            "Train Epoch: 54 [30000/60000 (50%)]\t Loss: 0.003675\n",
            "Train Epoch: 54 [40000/60000 (67%)]\t Loss: 0.001626\n",
            "Train Epoch: 54 [50000/60000 (83%)]\t Loss: 0.002397\n",
            "Total train loss : 0.1879176158061379\n",
            "Grad_norm : 0.054813336580991745\n",
            "Epoch 55/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 55 [0/60000 (0%)]\t Loss: 0.012594\n",
            "Train Epoch: 55 [10000/60000 (17%)]\t Loss: 0.001837\n",
            "Train Epoch: 55 [20000/60000 (33%)]\t Loss: 0.001489\n",
            "Train Epoch: 55 [30000/60000 (50%)]\t Loss: 0.000151\n",
            "Train Epoch: 55 [40000/60000 (67%)]\t Loss: 0.006288\n",
            "Train Epoch: 55 [50000/60000 (83%)]\t Loss: 0.020281\n",
            "Total train loss : 0.1940904680285333\n",
            "Grad_norm : 0.0551433339715004\n",
            "Epoch 56/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 56 [0/60000 (0%)]\t Loss: 0.002838\n",
            "Train Epoch: 56 [10000/60000 (17%)]\t Loss: 0.000153\n",
            "Train Epoch: 56 [20000/60000 (33%)]\t Loss: 0.016042\n",
            "Train Epoch: 56 [30000/60000 (50%)]\t Loss: 0.002690\n",
            "Train Epoch: 56 [40000/60000 (67%)]\t Loss: 0.018409\n",
            "Train Epoch: 56 [50000/60000 (83%)]\t Loss: 0.002401\n",
            "Total train loss : 0.1986823528865519\n",
            "Grad_norm : 0.3267475962638855\n",
            "Epoch 57/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 57 [0/60000 (0%)]\t Loss: 0.005199\n",
            "Train Epoch: 57 [10000/60000 (17%)]\t Loss: 0.001264\n",
            "Train Epoch: 57 [20000/60000 (33%)]\t Loss: 0.002367\n",
            "Train Epoch: 57 [30000/60000 (50%)]\t Loss: 0.002604\n",
            "Train Epoch: 57 [40000/60000 (67%)]\t Loss: 0.001342\n",
            "Train Epoch: 57 [50000/60000 (83%)]\t Loss: 0.000385\n",
            "Total train loss : 0.20388694722265274\n",
            "Grad_norm : 0.37504711747169495\n",
            "Epoch 58/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 58 [0/60000 (0%)]\t Loss: 0.003993\n",
            "Train Epoch: 58 [10000/60000 (17%)]\t Loss: 0.001716\n",
            "Train Epoch: 58 [20000/60000 (33%)]\t Loss: 0.006408\n",
            "Train Epoch: 58 [30000/60000 (50%)]\t Loss: 0.005648\n",
            "Train Epoch: 58 [40000/60000 (67%)]\t Loss: 0.009336\n",
            "Train Epoch: 58 [50000/60000 (83%)]\t Loss: 0.001785\n",
            "Total train loss : 0.21074048762541225\n",
            "Grad_norm : 0.26636695861816406\n",
            "Epoch 59/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 59 [0/60000 (0%)]\t Loss: 0.004152\n",
            "Train Epoch: 59 [10000/60000 (17%)]\t Loss: 0.001606\n",
            "Train Epoch: 59 [20000/60000 (33%)]\t Loss: 0.014188\n",
            "Train Epoch: 59 [30000/60000 (50%)]\t Loss: 0.001144\n",
            "Train Epoch: 59 [40000/60000 (67%)]\t Loss: 0.003940\n",
            "Train Epoch: 59 [50000/60000 (83%)]\t Loss: 0.022233\n",
            "Total train loss : 0.2173970759049477\n",
            "Grad_norm : 0.24370932579040527\n",
            "Epoch 60/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 60 [0/60000 (0%)]\t Loss: 0.001907\n",
            "Train Epoch: 60 [10000/60000 (17%)]\t Loss: 0.001565\n",
            "Train Epoch: 60 [20000/60000 (33%)]\t Loss: 0.001806\n",
            "Train Epoch: 60 [30000/60000 (50%)]\t Loss: 0.001124\n",
            "Train Epoch: 60 [40000/60000 (67%)]\t Loss: 0.001254\n",
            "Train Epoch: 60 [50000/60000 (83%)]\t Loss: 0.000118\n",
            "Total train loss : 0.22068589427528879\n",
            "Grad_norm : 0.07319959253072739\n",
            "Epoch 61/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 61 [0/60000 (0%)]\t Loss: 0.000060\n",
            "Train Epoch: 61 [10000/60000 (17%)]\t Loss: 0.003958\n",
            "Train Epoch: 61 [20000/60000 (33%)]\t Loss: 0.000274\n",
            "Train Epoch: 61 [30000/60000 (50%)]\t Loss: 0.003098\n",
            "Train Epoch: 61 [40000/60000 (67%)]\t Loss: 0.002433\n",
            "Train Epoch: 61 [50000/60000 (83%)]\t Loss: 0.003819\n",
            "Total train loss : 0.22220197777795267\n",
            "Grad_norm : 0.0014390387805178761\n",
            "Epoch 62/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 62 [0/60000 (0%)]\t Loss: 0.000713\n",
            "Train Epoch: 62 [10000/60000 (17%)]\t Loss: 0.000006\n",
            "Train Epoch: 62 [20000/60000 (33%)]\t Loss: 0.000106\n",
            "Train Epoch: 62 [30000/60000 (50%)]\t Loss: 0.000946\n",
            "Train Epoch: 62 [40000/60000 (67%)]\t Loss: 0.003612\n",
            "Train Epoch: 62 [50000/60000 (83%)]\t Loss: 0.001020\n",
            "Total train loss : 0.2230830350876128\n",
            "Grad_norm : 0.18676835298538208\n",
            "Epoch 63/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 63 [0/60000 (0%)]\t Loss: 0.000667\n",
            "Train Epoch: 63 [10000/60000 (17%)]\t Loss: 0.000059\n",
            "Train Epoch: 63 [20000/60000 (33%)]\t Loss: 0.000007\n",
            "Train Epoch: 63 [30000/60000 (50%)]\t Loss: 0.000045\n",
            "Train Epoch: 63 [40000/60000 (67%)]\t Loss: 0.000014\n",
            "Train Epoch: 63 [50000/60000 (83%)]\t Loss: 0.000007\n",
            "Total train loss : 0.22321109903493078\n",
            "Grad_norm : 0.14335551857948303\n",
            "Epoch 64/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 64 [0/60000 (0%)]\t Loss: 0.000007\n",
            "Train Epoch: 64 [10000/60000 (17%)]\t Loss: 0.000069\n",
            "Train Epoch: 64 [20000/60000 (33%)]\t Loss: 0.000007\n",
            "Train Epoch: 64 [30000/60000 (50%)]\t Loss: 0.000006\n",
            "Train Epoch: 64 [40000/60000 (67%)]\t Loss: 0.000563\n",
            "Train Epoch: 64 [50000/60000 (83%)]\t Loss: 0.000001\n",
            "Total train loss : 0.2232523949337389\n",
            "Grad_norm : 0.0002689876710064709\n",
            "Epoch 65/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 65 [0/60000 (0%)]\t Loss: 0.000008\n",
            "Train Epoch: 65 [10000/60000 (17%)]\t Loss: 0.000019\n",
            "Train Epoch: 65 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 65 [30000/60000 (50%)]\t Loss: 0.000002\n",
            "Train Epoch: 65 [40000/60000 (67%)]\t Loss: 0.000002\n",
            "Train Epoch: 65 [50000/60000 (83%)]\t Loss: 0.000002\n",
            "Total train loss : 0.22326915803329944\n",
            "Grad_norm : 0.0013195296050980687\n",
            "Epoch 66/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 66 [0/60000 (0%)]\t Loss: 0.000003\n",
            "Train Epoch: 66 [10000/60000 (17%)]\t Loss: 0.000003\n",
            "Train Epoch: 66 [20000/60000 (33%)]\t Loss: 0.000006\n",
            "Train Epoch: 66 [30000/60000 (50%)]\t Loss: 0.000004\n",
            "Train Epoch: 66 [40000/60000 (67%)]\t Loss: 0.000003\n",
            "Train Epoch: 66 [50000/60000 (83%)]\t Loss: 0.000006\n",
            "Total train loss : 0.2232751248762187\n",
            "Grad_norm : 5.357662303140387e-05\n",
            "Epoch 67/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 67 [0/60000 (0%)]\t Loss: 0.000007\n",
            "Train Epoch: 67 [10000/60000 (17%)]\t Loss: 0.000003\n",
            "Train Epoch: 67 [20000/60000 (33%)]\t Loss: 0.000003\n",
            "Train Epoch: 67 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 67 [40000/60000 (67%)]\t Loss: 0.000006\n",
            "Train Epoch: 67 [50000/60000 (83%)]\t Loss: 0.000005\n",
            "Total train loss : 0.2232780166741764\n",
            "Grad_norm : 0.0008377316407859325\n",
            "Epoch 68/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 68 [0/60000 (0%)]\t Loss: 0.000003\n",
            "Train Epoch: 68 [10000/60000 (17%)]\t Loss: 0.000004\n",
            "Train Epoch: 68 [20000/60000 (33%)]\t Loss: 0.000000\n",
            "Train Epoch: 68 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 68 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 68 [50000/60000 (83%)]\t Loss: 0.000003\n",
            "Total train loss : 0.22328053499520664\n",
            "Grad_norm : 0.000458411785075441\n",
            "Epoch 69/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 69 [0/60000 (0%)]\t Loss: 0.000002\n",
            "Train Epoch: 69 [10000/60000 (17%)]\t Loss: 0.000003\n",
            "Train Epoch: 69 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 69 [30000/60000 (50%)]\t Loss: 0.000002\n",
            "Train Epoch: 69 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 69 [50000/60000 (83%)]\t Loss: 0.000000\n",
            "Total train loss : 0.22328280231325456\n",
            "Grad_norm : 0.00017118206596933305\n",
            "Epoch 70/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 70 [0/60000 (0%)]\t Loss: 0.000006\n",
            "Train Epoch: 70 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 70 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 70 [30000/60000 (50%)]\t Loss: 0.000002\n",
            "Train Epoch: 70 [40000/60000 (67%)]\t Loss: 0.000002\n",
            "Train Epoch: 70 [50000/60000 (83%)]\t Loss: 0.000002\n",
            "Total train loss : 0.22328487369631442\n",
            "Grad_norm : 0.00034211401361972094\n",
            "Epoch 71/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 71 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 71 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 71 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 71 [30000/60000 (50%)]\t Loss: 0.000004\n",
            "Train Epoch: 71 [40000/60000 (67%)]\t Loss: 0.000003\n",
            "Train Epoch: 71 [50000/60000 (83%)]\t Loss: 0.000002\n",
            "Total train loss : 0.22328679263377324\n",
            "Grad_norm : 5.0448361434973776e-05\n",
            "Epoch 72/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 72 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 72 [10000/60000 (17%)]\t Loss: 0.000005\n",
            "Train Epoch: 72 [20000/60000 (33%)]\t Loss: 0.000002\n",
            "Train Epoch: 72 [30000/60000 (50%)]\t Loss: 0.000004\n",
            "Train Epoch: 72 [40000/60000 (67%)]\t Loss: 0.000004\n",
            "Train Epoch: 72 [50000/60000 (83%)]\t Loss: 0.000000\n",
            "Total train loss : 0.22328858380863562\n",
            "Grad_norm : 0.00033103598980233073\n",
            "Epoch 73/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 73 [0/60000 (0%)]\t Loss: 0.000002\n",
            "Train Epoch: 73 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 73 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 73 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 73 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 73 [50000/60000 (83%)]\t Loss: 0.000001\n",
            "Total train loss : 0.22329026893242107\n",
            "Grad_norm : 8.893268386600539e-05\n",
            "Epoch 74/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 74 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 74 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 74 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 74 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 74 [40000/60000 (67%)]\t Loss: 0.000000\n",
            "Train Epoch: 74 [50000/60000 (83%)]\t Loss: 0.000002\n",
            "Total train loss : 0.22329186121049388\n",
            "Grad_norm : 0.00021832215134054422\n",
            "Epoch 75/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 75 [0/60000 (0%)]\t Loss: 0.000002\n",
            "Train Epoch: 75 [10000/60000 (17%)]\t Loss: 0.000002\n",
            "Train Epoch: 75 [20000/60000 (33%)]\t Loss: 0.000002\n",
            "Train Epoch: 75 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 75 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 75 [50000/60000 (83%)]\t Loss: 0.000004\n",
            "Total train loss : 0.22329336883718062\n",
            "Grad_norm : 9.172625141218305e-05\n",
            "Epoch 76/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 76 [0/60000 (0%)]\t Loss: 0.000002\n",
            "Train Epoch: 76 [10000/60000 (17%)]\t Loss: 0.000003\n",
            "Train Epoch: 76 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 76 [30000/60000 (50%)]\t Loss: 0.000002\n",
            "Train Epoch: 76 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 76 [50000/60000 (83%)]\t Loss: 0.000000\n",
            "Total train loss : 0.22329480275579977\n",
            "Grad_norm : 0.0001323405740549788\n",
            "Epoch 77/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 77 [0/60000 (0%)]\t Loss: 0.000000\n",
            "Train Epoch: 77 [10000/60000 (17%)]\t Loss: 0.000003\n",
            "Train Epoch: 77 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 77 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 77 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 77 [50000/60000 (83%)]\t Loss: 0.000002\n",
            "Total train loss : 0.2232961705113766\n",
            "Grad_norm : 0.00011577058467082679\n",
            "Epoch 78/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 78 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 78 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 78 [20000/60000 (33%)]\t Loss: 0.000003\n",
            "Train Epoch: 78 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 78 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 78 [50000/60000 (83%)]\t Loss: 0.000000\n",
            "Total train loss : 0.2232974795630772\n",
            "Grad_norm : 0.00010949397255899385\n",
            "Epoch 79/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 79 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 79 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 79 [20000/60000 (33%)]\t Loss: 0.000002\n",
            "Train Epoch: 79 [30000/60000 (50%)]\t Loss: 0.000002\n",
            "Train Epoch: 79 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 79 [50000/60000 (83%)]\t Loss: 0.000001\n",
            "Total train loss : 0.22329873422669805\n",
            "Grad_norm : 0.00011359799827914685\n",
            "Epoch 80/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 80 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 80 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 80 [20000/60000 (33%)]\t Loss: 0.000000\n",
            "Train Epoch: 80 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 80 [40000/60000 (67%)]\t Loss: 0.000002\n",
            "Train Epoch: 80 [50000/60000 (83%)]\t Loss: 0.000002\n",
            "Total train loss : 0.22329993834133788\n",
            "Grad_norm : 0.0001227987522725016\n",
            "Epoch 81/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 81 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 81 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 81 [20000/60000 (33%)]\t Loss: 0.000000\n",
            "Train Epoch: 81 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 81 [40000/60000 (67%)]\t Loss: 0.000000\n",
            "Train Epoch: 81 [50000/60000 (83%)]\t Loss: 0.000001\n",
            "Total train loss : 0.22330109456476396\n",
            "Grad_norm : 0.00013934436719864607\n",
            "Epoch 82/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 82 [0/60000 (0%)]\t Loss: 0.000002\n",
            "Train Epoch: 82 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 82 [20000/60000 (33%)]\t Loss: 0.000000\n",
            "Train Epoch: 82 [30000/60000 (50%)]\t Loss: 0.000002\n",
            "Train Epoch: 82 [40000/60000 (67%)]\t Loss: 0.000002\n",
            "Train Epoch: 82 [50000/60000 (83%)]\t Loss: 0.000000\n",
            "Total train loss : 0.2233022072707333\n",
            "Grad_norm : 6.795304216211662e-05\n",
            "Epoch 83/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 83 [0/60000 (0%)]\t Loss: 0.000002\n",
            "Train Epoch: 83 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 83 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 83 [30000/60000 (50%)]\t Loss: 0.000002\n",
            "Train Epoch: 83 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 83 [50000/60000 (83%)]\t Loss: 0.000001\n",
            "Total train loss : 0.22330327926887225\n",
            "Grad_norm : 0.0001903880329336971\n",
            "Epoch 84/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 84 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 84 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 84 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 84 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 84 [40000/60000 (67%)]\t Loss: 0.000002\n",
            "Train Epoch: 84 [50000/60000 (83%)]\t Loss: 0.000001\n",
            "Total train loss : 0.22330431307443158\n",
            "Grad_norm : 0.00014165080210659653\n",
            "Epoch 85/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 85 [0/60000 (0%)]\t Loss: 0.000000\n",
            "Train Epoch: 85 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 85 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 85 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 85 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 85 [50000/60000 (83%)]\t Loss: 0.000001\n",
            "Total train loss : 0.22330531101246365\n",
            "Grad_norm : 2.9771123081445694e-05\n",
            "Epoch 86/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 86 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 86 [10000/60000 (17%)]\t Loss: 0.000002\n",
            "Train Epoch: 86 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 86 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 86 [40000/60000 (67%)]\t Loss: 0.000000\n",
            "Train Epoch: 86 [50000/60000 (83%)]\t Loss: 0.000000\n",
            "Total train loss : 0.2233062747033287\n",
            "Grad_norm : 6.137921445770189e-05\n",
            "Epoch 87/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 87 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 87 [10000/60000 (17%)]\t Loss: 0.000000\n",
            "Train Epoch: 87 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 87 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 87 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 87 [50000/60000 (83%)]\t Loss: 0.000001\n",
            "Total train loss : 0.22330720687061004\n",
            "Grad_norm : 0.00012739127851091325\n",
            "Epoch 88/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 88 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 88 [10000/60000 (17%)]\t Loss: 0.000000\n",
            "Train Epoch: 88 [20000/60000 (33%)]\t Loss: 0.000000\n",
            "Train Epoch: 88 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 88 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 88 [50000/60000 (83%)]\t Loss: 0.000000\n",
            "Total train loss : 0.2233081080045764\n",
            "Grad_norm : 0.00010915282473433763\n",
            "Epoch 89/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 89 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 89 [10000/60000 (17%)]\t Loss: 0.000002\n",
            "Train Epoch: 89 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 89 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 89 [40000/60000 (67%)]\t Loss: 0.000000\n",
            "Train Epoch: 89 [50000/60000 (83%)]\t Loss: 0.000001\n",
            "Total train loss : 0.22330897899609306\n",
            "Grad_norm : 5.6150769523810595e-05\n",
            "Epoch 90/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 90 [0/60000 (0%)]\t Loss: 0.000000\n",
            "Train Epoch: 90 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 90 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 90 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 90 [40000/60000 (67%)]\t Loss: 0.000000\n",
            "Train Epoch: 90 [50000/60000 (83%)]\t Loss: 0.000003\n",
            "Total train loss : 0.22330982307392375\n",
            "Grad_norm : 0.00011353783338563517\n",
            "Epoch 91/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 91 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 91 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 91 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 91 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 91 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 91 [50000/60000 (83%)]\t Loss: 0.000001\n",
            "Total train loss : 0.22331064100647352\n",
            "Grad_norm : 6.031629891367629e-05\n",
            "Epoch 92/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 92 [0/60000 (0%)]\t Loss: 0.000000\n",
            "Train Epoch: 92 [10000/60000 (17%)]\t Loss: 0.000000\n",
            "Train Epoch: 92 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 92 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 92 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 92 [50000/60000 (83%)]\t Loss: 0.000002\n",
            "Total train loss : 0.22331143338189322\n",
            "Grad_norm : 5.065678124083206e-05\n",
            "Epoch 93/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 93 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 93 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 93 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 93 [30000/60000 (50%)]\t Loss: 0.000000\n",
            "Train Epoch: 93 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 93 [50000/60000 (83%)]\t Loss: 0.000001\n",
            "Total train loss : 0.2233122023923194\n",
            "Grad_norm : 0.00010040574125014246\n",
            "Epoch 94/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 94 [0/60000 (0%)]\t Loss: 0.000000\n",
            "Train Epoch: 94 [10000/60000 (17%)]\t Loss: 0.000000\n",
            "Train Epoch: 94 [20000/60000 (33%)]\t Loss: 0.000000\n",
            "Train Epoch: 94 [30000/60000 (50%)]\t Loss: 0.000000\n",
            "Train Epoch: 94 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 94 [50000/60000 (83%)]\t Loss: 0.000000\n",
            "Total train loss : 0.22331294920700248\n",
            "Grad_norm : 9.770508768269792e-05\n",
            "Epoch 95/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 95 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 95 [10000/60000 (17%)]\t Loss: 0.000000\n",
            "Train Epoch: 95 [20000/60000 (33%)]\t Loss: 0.000000\n",
            "Train Epoch: 95 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 95 [40000/60000 (67%)]\t Loss: 0.000000\n",
            "Train Epoch: 95 [50000/60000 (83%)]\t Loss: 0.000000\n",
            "Total train loss : 0.22331367346144285\n",
            "Grad_norm : 6.27876870566979e-05\n",
            "Epoch 96/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 96 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 96 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 96 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 96 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 96 [40000/60000 (67%)]\t Loss: 0.000000\n",
            "Train Epoch: 96 [50000/60000 (83%)]\t Loss: 0.000001\n",
            "Total train loss : 0.223314376531266\n",
            "Grad_norm : 3.350187398609705e-05\n",
            "Epoch 97/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 97 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 97 [10000/60000 (17%)]\t Loss: 0.000000\n",
            "Train Epoch: 97 [20000/60000 (33%)]\t Loss: 0.000000\n",
            "Train Epoch: 97 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 97 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 97 [50000/60000 (83%)]\t Loss: 0.000001\n",
            "Total train loss : 0.22331505961413925\n",
            "Grad_norm : 3.959398600272834e-05\n",
            "Epoch 98/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 98 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 98 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 98 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 98 [30000/60000 (50%)]\t Loss: 0.000000\n",
            "Train Epoch: 98 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 98 [50000/60000 (83%)]\t Loss: 0.000001\n",
            "Total train loss : 0.2233157224515073\n",
            "Grad_norm : 2.8423246476450004e-05\n",
            "Epoch 99/99\n",
            "-------------------------------------------------------\n",
            "Train Epoch: 99 [0/60000 (0%)]\t Loss: 0.000001\n",
            "Train Epoch: 99 [10000/60000 (17%)]\t Loss: 0.000001\n",
            "Train Epoch: 99 [20000/60000 (33%)]\t Loss: 0.000001\n",
            "Train Epoch: 99 [30000/60000 (50%)]\t Loss: 0.000001\n",
            "Train Epoch: 99 [40000/60000 (67%)]\t Loss: 0.000001\n",
            "Train Epoch: 99 [50000/60000 (83%)]\t Loss: 0.000001\n",
            "Total train loss : 0.22331636707937955\n",
            "Grad_norm : 2.2257234377320856e-05\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mfV_c1j8yazc",
        "outputId": "bb9dfc02-0819-4fcf-cf5f-4bf4f43da5f9"
      },
      "source": [
        "print(len(gradient_norm))"
      ],
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "100\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yTY84Ka82bT2"
      },
      "source": [
        "ep = [e for e in range(epochs)]\n",
        "#ep"
      ],
      "execution_count": 41,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VA06f9j22p17"
      },
      "source": [
        "#### Evaluating and testing the model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-62LwAG02ldT",
        "outputId": "25b79322-a807-465a-dc34-e05417a09fa3"
      },
      "source": [
        "correct = 0\n",
        "total = 0\n",
        "pred = []\n",
        "# Test the model\n",
        "model.eval()\n",
        "test_loss = 0\n",
        "for data, target in test_loader:\n",
        "  data, target = data.to(device), target.to(device)\n",
        "  output = model(data)\n",
        "  test_loss += loss_func(output,target)\n",
        "  _, predicted = torch.max(output.data,1)\n",
        "  pred.append(output)\n",
        "  total += target.size(0)\n",
        "  correct += (predicted == target).sum().item()\n",
        "\n",
        "test_loss = test_loss/len(test_loader.dataset)\n",
        "print('Average loss for the test set is {} \\nThe test accuracy of the 2-layered cnn model on the {} test images: {} %'.format(test_loss, len(test_set),(correct / (total)) * 100))"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Average loss for the test set is 5.196604615775868e-05 \n",
            "The test accuracy of the 2-layered cnn model on the 10000 test images: 99.28 %\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Q1unSdYg3H0K"
      },
      "source": [
        "####Plot of gradient norm vs epoch"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 513
        },
        "id": "tp2-DucO3HHI",
        "outputId": "9d4a0a6e-5012-4b83-a797-e971de92d5af"
      },
      "source": [
        "#line graph\n",
        "fig = plt.figure()\n",
        "fig.set_size_inches(16, 8)\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Norm of the gradient\")\n",
        "plt.scatter(ep,gradient_norm)\n",
        "plt.plot(ep,gradient_norm)\n",
        "plt.title(\"Gradient norm vs Epoch\")\n",
        "plt.savefig('line plot of gradient norm vs epoch (with 100 epoch).pdf',dpi = 300)"
      ],
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 513
        },
        "id": "ywLmHeii3Tm3",
        "outputId": "c1fd1381-66c3-4eb9-a0da-4cabe1cb61f3"
      },
      "source": [
        "#scatter plot\n",
        "fig = plt.figure()\n",
        "fig.set_size_inches(16, 8)\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Norm of the gradient\")\n",
        "plt.scatter(ep,gradient_norm)\n",
        "plt.title(\"Gradient norm vs Epoch\")\n",
        "plt.savefig('scatter plot of gradient norm vs epoch (with 100 epochs).pdf',dpi = 300)"
      ],
      "execution_count": 57,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6E49Qdxf1XoR"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}