textClassification_ncv_Fine_Tunning_Bert_beta1.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "textClassification_ncv_Fine_Tunning_Bert-beta1.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.8.5"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "80bb686f18294b3089e7a1a6895b4f2a": {
          "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_25300b1ce6dd42a5938be2c70282d0f6",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_e6099193d78c4fa9bd837d5d73c3e990",
              "IPY_MODEL_317caba37f83404c85121e6a3370ef2e"
            ]
          }
        },
        "25300b1ce6dd42a5938be2c70282d0f6": {
          "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
          }
        },
        "e6099193d78c4fa9bd837d5d73c3e990": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_17bccd0f294d4a3d9743a6858441c93b",
            "_dom_classes": [],
            "description": "Downloading: 100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 231508,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 231508,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_453ea79107af4a6695bd26a010080ff6"
          }
        },
        "317caba37f83404c85121e6a3370ef2e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_f450869ea47a4213b4ca33b6a8f09a6e",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 232k/232k [00:01&lt;00:00, 124kB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_1d237d946165474dba7554c0ca3ae604"
          }
        },
        "17bccd0f294d4a3d9743a6858441c93b": {
          "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"
          }
        },
        "453ea79107af4a6695bd26a010080ff6": {
          "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
          }
        },
        "f450869ea47a4213b4ca33b6a8f09a6e": {
          "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"
          }
        },
        "1d237d946165474dba7554c0ca3ae604": {
          "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
          }
        },
        "184a311dbbfc462a9215688bf3ac74bb": {
          "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_5c4f5a1e77294fecada31be2833a40eb",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_60c7459612804b69a5863ce680fbdc13",
              "IPY_MODEL_caec3ff5c84d45f8b3e2d4144cee2249"
            ]
          }
        },
        "5c4f5a1e77294fecada31be2833a40eb": {
          "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
          }
        },
        "60c7459612804b69a5863ce680fbdc13": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_f84b3d21f2b24bbc8604e06f442c37df",
            "_dom_classes": [],
            "description": "Downloading: 100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 466062,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 466062,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_419b31bd03ea4dbdaca65ffd9fb8741c"
          }
        },
        "caec3ff5c84d45f8b3e2d4144cee2249": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_3b2b63efcd2646c8a884353ad40041d6",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 466k/466k [00:16&lt;00:00, 29.0kB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_e0476fffa97547e0baaa6bbd44accf19"
          }
        },
        "f84b3d21f2b24bbc8604e06f442c37df": {
          "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"
          }
        },
        "419b31bd03ea4dbdaca65ffd9fb8741c": {
          "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
          }
        },
        "3b2b63efcd2646c8a884353ad40041d6": {
          "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"
          }
        },
        "e0476fffa97547e0baaa6bbd44accf19": {
          "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
          }
        },
        "deb2a286c0b949f49afca2a44add412c": {
          "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_0f2059ac53c54ccfaa659f0cd6a97472",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_ef9e8a34ac1c46719fe2892d008591fa",
              "IPY_MODEL_1caa12aeb3064d659b42353ce575c0e2"
            ]
          }
        },
        "0f2059ac53c54ccfaa659f0cd6a97472": {
          "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
          }
        },
        "ef9e8a34ac1c46719fe2892d008591fa": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_9137e74a56654a3b92bbd418d897daee",
            "_dom_classes": [],
            "description": "Downloading: 100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 536063208,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 536063208,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_4a99031f626141428097c6217b129d96"
          }
        },
        "1caa12aeb3064d659b42353ce575c0e2": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_d9c1b25317d24557acc74a75ead8d9e6",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 536M/536M [00:14&lt;00:00, 37.9MB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_293a246ac4894c25a0ec293665c38838"
          }
        },
        "9137e74a56654a3b92bbd418d897daee": {
          "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"
          }
        },
        "4a99031f626141428097c6217b129d96": {
          "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
          }
        },
        "d9c1b25317d24557acc74a75ead8d9e6": {
          "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"
          }
        },
        "293a246ac4894c25a0ec293665c38838": {
          "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": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Lu6CCLW8OWEV",
        "outputId": "dc210bdf-788a-42f3-f041-cb7aa51175d7"
      },
      "source": [
        "%tensorflow_version 2.x\n",
        "import tensorflow as tf\n",
        "device_name = tf.test.gpu_device_name()\n",
        "if device_name != '/device:GPU:0':\n",
        "  raise SystemError('GPU device not found')\n",
        "print('Found GPU at: {}'.format(device_name))"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found GPU at: /device:GPU:0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oxedPL5lOhDv",
        "outputId": "2f69bb07-b76c-49f5-f6e3-76b6a7528e9c"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/gdrive')"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/gdrive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6MkXbU4HOrUl",
        "outputId": "6bc56949-b4f5-44cf-887d-ec9b13a3c2c4"
      },
      "source": [
        "!pip install silence_tensorflow\n",
        "!pip install pycm\n",
        "!pip install lime\n",
        "!pip install transformers\n",
        "import nltk\n",
        "nltk.download('stopwords')\n",
        "nltk.download('wordnet')"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting silence_tensorflow\n",
            "  Downloading https://files.pythonhosted.org/packages/96/d7/076b21d0e79cfc8a085f623e6577b754c50a42cfbcce51d77d0d2206988c/silence_tensorflow-1.1.1.tar.gz\n",
            "Building wheels for collected packages: silence-tensorflow\n",
            "  Building wheel for silence-tensorflow (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for silence-tensorflow: filename=silence_tensorflow-1.1.1-cp37-none-any.whl size=3743 sha256=597355cd1a1d720e3da28d4724764ea4449271722f84ba930d07115f65d4ac90\n",
            "  Stored in directory: /root/.cache/pip/wheels/51/0b/35/cf3020764bee61daa81fa249df3a448e3806344a087fc12292\n",
            "Successfully built silence-tensorflow\n",
            "Installing collected packages: silence-tensorflow\n",
            "Successfully installed silence-tensorflow-1.1.1\n",
            "Collecting pycm\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/07/e0/de92301a06bf5e2d24e04edc178ee457b49dbad2606c3252949222110f4d/pycm-3.0-py2.py3-none-any.whl (61kB)\n",
            "\u001b[K     |████████████████████████████████| 71kB 3.6MB/s \n",
            "\u001b[?25hRequirement already satisfied: numpy>=1.9.0 in /usr/local/lib/python3.7/dist-packages (from pycm) (1.19.5)\n",
            "Collecting art>=1.8\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/83/37/c28d69ba1005889e8ff9633b8e7eccaaddc03549058a51ef5f51853cf183/art-5.1-py2.py3-none-any.whl (567kB)\n",
            "\u001b[K     |████████████████████████████████| 573kB 7.4MB/s \n",
            "\u001b[?25hInstalling collected packages: art, pycm\n",
            "Successfully installed art-5.1 pycm-3.0\n",
            "Collecting lime\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/f5/86/91a13127d83d793ecb50eb75e716f76e6eda809b6803c5a4ff462339789e/lime-0.2.0.1.tar.gz (275kB)\n",
            "\u001b[K     |████████████████████████████████| 276kB 6.0MB/s \n",
            "\u001b[?25hRequirement already satisfied: matplotlib in /usr/local/lib/python3.7/dist-packages (from lime) (3.2.2)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.7/dist-packages (from lime) (1.19.5)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.7/dist-packages (from lime) (1.4.1)\n",
            "Requirement already satisfied: tqdm in /usr/local/lib/python3.7/dist-packages (from lime) (4.41.1)\n",
            "Requirement already satisfied: scikit-learn>=0.18 in /usr/local/lib/python3.7/dist-packages (from lime) (0.22.2.post1)\n",
            "Requirement already satisfied: scikit-image>=0.12 in /usr/local/lib/python3.7/dist-packages (from lime) (0.16.2)\n",
            "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->lime) (2.4.7)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->lime) (1.3.1)\n",
            "Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->lime) (2.8.1)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.7/dist-packages (from matplotlib->lime) (0.10.0)\n",
            "Requirement already satisfied: joblib>=0.11 in /usr/local/lib/python3.7/dist-packages (from scikit-learn>=0.18->lime) (1.0.1)\n",
            "Requirement already satisfied: PyWavelets>=0.4.0 in /usr/local/lib/python3.7/dist-packages (from scikit-image>=0.12->lime) (1.1.1)\n",
            "Requirement already satisfied: networkx>=2.0 in /usr/local/lib/python3.7/dist-packages (from scikit-image>=0.12->lime) (2.5)\n",
            "Requirement already satisfied: pillow>=4.3.0 in /usr/local/lib/python3.7/dist-packages (from scikit-image>=0.12->lime) (7.0.0)\n",
            "Requirement already satisfied: imageio>=2.3.0 in /usr/local/lib/python3.7/dist-packages (from scikit-image>=0.12->lime) (2.4.1)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/dist-packages (from python-dateutil>=2.1->matplotlib->lime) (1.15.0)\n",
            "Requirement already satisfied: decorator>=4.3.0 in /usr/local/lib/python3.7/dist-packages (from networkx>=2.0->scikit-image>=0.12->lime) (4.4.2)\n",
            "Building wheels for collected packages: lime\n",
            "  Building wheel for lime (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for lime: filename=lime-0.2.0.1-cp37-none-any.whl size=283846 sha256=ab6aa1031a962d5558b45ed788cb626b6b4be4953ff73cde78631de6a7f13065\n",
            "  Stored in directory: /root/.cache/pip/wheels/4c/4f/a5/0bc765457bd41378bf3ce8d17d7495369d6e7ca3b712c60c89\n",
            "Successfully built lime\n",
            "Installing collected packages: lime\n",
            "Successfully installed lime-0.2.0.1\n",
            "Collecting transformers\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/f9/54/5ca07ec9569d2f232f3166de5457b63943882f7950ddfcc887732fc7fb23/transformers-4.3.3-py3-none-any.whl (1.9MB)\n",
            "\u001b[K     |████████████████████████████████| 1.9MB 4.2MB/s \n",
            "\u001b[?25hRequirement already satisfied: tqdm>=4.27 in /usr/local/lib/python3.7/dist-packages (from transformers) (4.41.1)\n",
            "Collecting tokenizers<0.11,>=0.10.1\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/71/23/2ddc317b2121117bf34dd00f5b0de194158f2a44ee2bf5e47c7166878a97/tokenizers-0.10.1-cp37-cp37m-manylinux2010_x86_64.whl (3.2MB)\n",
            "\u001b[K     |████████████████████████████████| 3.2MB 41.7MB/s \n",
            "\u001b[?25hRequirement already satisfied: numpy>=1.17 in /usr/local/lib/python3.7/dist-packages (from transformers) (1.19.5)\n",
            "Requirement already satisfied: importlib-metadata; python_version < \"3.8\" in /usr/local/lib/python3.7/dist-packages (from transformers) (3.7.0)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.7/dist-packages (from transformers) (3.0.12)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.7/dist-packages (from transformers) (2.23.0)\n",
            "Requirement already satisfied: packaging in /usr/local/lib/python3.7/dist-packages (from transformers) (20.9)\n",
            "Collecting sacremoses\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/7d/34/09d19aff26edcc8eb2a01bed8e98f13a1537005d31e95233fd48216eed10/sacremoses-0.0.43.tar.gz (883kB)\n",
            "\u001b[K     |████████████████████████████████| 890kB 26.0MB/s \n",
            "\u001b[?25hRequirement already satisfied: regex!=2019.12.17 in /usr/local/lib/python3.7/dist-packages (from transformers) (2019.12.20)\n",
            "Requirement already satisfied: typing-extensions>=3.6.4; python_version < \"3.8\" in /usr/local/lib/python3.7/dist-packages (from importlib-metadata; python_version < \"3.8\"->transformers) (3.7.4.3)\n",
            "Requirement already satisfied: zipp>=0.5 in /usr/local/lib/python3.7/dist-packages (from importlib-metadata; python_version < \"3.8\"->transformers) (3.4.0)\n",
            "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.7/dist-packages (from requests->transformers) (1.24.3)\n",
            "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.7/dist-packages (from requests->transformers) (3.0.4)\n",
            "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.7/dist-packages (from requests->transformers) (2.10)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.7/dist-packages (from requests->transformers) (2020.12.5)\n",
            "Requirement already satisfied: pyparsing>=2.0.2 in /usr/local/lib/python3.7/dist-packages (from packaging->transformers) (2.4.7)\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.7/dist-packages (from sacremoses->transformers) (1.15.0)\n",
            "Requirement already satisfied: click in /usr/local/lib/python3.7/dist-packages (from sacremoses->transformers) (7.1.2)\n",
            "Requirement already satisfied: joblib in /usr/local/lib/python3.7/dist-packages (from sacremoses->transformers) (1.0.1)\n",
            "Building wheels for collected packages: sacremoses\n",
            "  Building wheel for sacremoses (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for sacremoses: filename=sacremoses-0.0.43-cp37-none-any.whl size=893262 sha256=291a6d22a6b911f9001e794389fdd5886901520bd33b1f6606f6a623e0c2a32f\n",
            "  Stored in directory: /root/.cache/pip/wheels/29/3c/fd/7ce5c3f0666dab31a50123635e6fb5e19ceb42ce38d4e58f45\n",
            "Successfully built sacremoses\n",
            "Installing collected packages: tokenizers, sacremoses, transformers\n",
            "Successfully installed sacremoses-0.0.43 tokenizers-0.10.1 transformers-4.3.3\n",
            "[nltk_data] Downloading package stopwords to /root/nltk_data...\n",
            "[nltk_data]   Unzipping corpora/stopwords.zip.\n",
            "[nltk_data] Downloading package wordnet to /root/nltk_data...\n",
            "[nltk_data]   Unzipping corpora/wordnet.zip.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qqaSuDi0PMuq"
      },
      "source": [
        "import os\n",
        "default_path = \"/content/gdrive/MyDrive/Colab Notebooks/Automatic Software Artifacts\"\n",
        "os.chdir(default_path)"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rlEQzlfvNvmX"
      },
      "source": [
        "import warnings\n",
        "warnings.filterwarnings(\"ignore\")\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "## for deep learning\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras.layers import Embedding\n",
        "from tensorflow.keras import layers, preprocessing as kprocessing\n",
        "from tensorflow.keras import backend as K\n",
        "## for bert language model\n",
        "import transformers\n",
        "from transformers import BertTokenizer, TFBertModel, BertConfig, TFDistilBertModel\n",
        "from tensorflow.keras.models import Sequential, Model\n",
        "from tensorflow.keras.layers import Dense, Dropout, Activation, Conv1D, GlobalMaxPooling1D, LSTM, SpatialDropout1D, Input, Flatten, MaxPooling1D\n",
        "from sklearn.utils import shuffle\n",
        "from tensorflow.keras.preprocessing import text, sequence\n",
        "from sklearn.model_selection import train_test_split\n",
        "from tensorflow.keras.optimizers import Adam\n",
        "from tensorflow.keras import utils\n",
        "\n",
        "## for plotting\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "## for processing\n",
        "import nltk\n",
        "import re\n",
        "## for bag-of-words\n",
        "from sklearn.feature_extraction.text import TfidfVectorizer\n",
        "from sklearn import feature_extraction, model_selection, naive_bayes, pipeline, manifold, preprocessing\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "from sklearn import feature_selection\n",
        "# Classifiers\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "from sklearn.ensemble import RandomForestClassifier, AdaBoostClassifier, VotingClassifier\n",
        "from sklearn import svm\n",
        "from sklearn.svm import SVC, LinearSVC\n",
        "from sklearn.tree import DecisionTreeClassifier\n",
        "from sklearn.naive_bayes import MultinomialNB, GaussianNB, ComplementNB\n",
        "from sklearn.model_selection import StratifiedShuffleSplit\n",
        "# Metrics\n",
        "from sklearn import metrics\n",
        "from sklearn.metrics import confusion_matrix\n",
        "from pycm import *\n",
        "## for explainer\n",
        "from lime import lime_text\n",
        "## for word embedding\n",
        "import gensim\n",
        "import gensim.downloader as gensim_api\n",
        "\n",
        "import utils_functions as uf\n",
        "\n",
        "import matplotlib as mpl\n",
        "# Set the default color cycle\n",
        "mpl.rcParams['axes.prop_cycle'] = mpl.cycler(color=[\"b\", \"k\", \"c\"]) \n",
        "\n",
        "np.random.seed(7)"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sJIZ4lsONvmf"
      },
      "source": [
        "### Some Utils Functions"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EnJ7ms3y1atk"
      },
      "source": [
        "# encode y_train\n",
        "def encode_y(y_in):\n",
        "    dic_y_mapping = {n:label for n,label in enumerate(np.unique(y_in))}\n",
        "    inverse_dic = {v:k for k,v in dic_y_mapping.items()}\n",
        "    y_in = np.array([inverse_dic[y] for y in y_in])\n",
        "    return y_in, dic_y_mapping"
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8YZHVr33Nvmg"
      },
      "source": [
        "# '''\n",
        "# Compute different text length metrics.\n",
        "# :parameter\n",
        "#     :param dtf: dataframe - dtf with a text column\n",
        "#     :param column: string - name of column containing text\n",
        "# :return\n",
        "#     dtf: input dataframe with 2 new columns\n",
        "# '''\n",
        "def add_text_length(dtf, column):\n",
        "    dtf['word_count'] = dtf[column].apply(lambda x: len(str(x).split(\" \")))\n",
        "    dtf['char_count'] = dtf[column].apply(lambda x: sum(len(word) for word in str(x).split(\" \")))\n",
        "    dtf['sentence_count'] = dtf[column].apply(lambda x: len(str(x).split(\".\")))\n",
        "    dtf['avg_word_length'] = dtf['char_count'] / dtf['word_count']\n",
        "    dtf['avg_sentence_lenght'] = dtf['word_count'] / dtf['sentence_count']\n",
        "    print(dtf[['word_count','char_count','sentence_count','avg_word_length','avg_sentence_lenght']].describe().T[[\"min\",\"mean\",\"max\"]])\n",
        "    return dtf"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "x95STmQiNvmh"
      },
      "source": [
        "from sklearn.decomposition import IncrementalPCA    # inital reduction\n",
        "from sklearn.manifold import TSNE                   # final reduction\n",
        "import numpy as np                                  # array handling\n",
        "\n",
        "def display_closestwords_tsnescatterplot(word_vect, dim, words):\n",
        "    \n",
        "    arr = np.empty((0,dim), dtype='f')\n",
        "    word_labels = words\n",
        "\n",
        "    # get close words\n",
        "    #close_words = [model.similar_by_word(word) for word in words]\n",
        "    \n",
        "    # add the vector for each of the closest words to the array\n",
        "    close_words=[]\n",
        "    for word in words:\n",
        "        arr = np.append(arr, np.array([word_vect[word]]), axis=0)\n",
        "        close_words +=model.similar_by_word(word)\n",
        "        \n",
        "    for wrd_score in close_words:\n",
        "        wrd_vector = word_vect[wrd_score[0]]\n",
        "        word_labels.append(wrd_score[0])\n",
        "        arr = np.append(arr, np.array([wrd_vector]), axis=0)\n",
        "        \n",
        "    # find tsne coords for 2 dimensions\n",
        "    tsne = TSNE(n_components=2, random_state=0)\n",
        "    #np.set_printoptions(suppress=True)\n",
        "    Y = tsne.fit_transform(arr)\n",
        "\n",
        "    x_coords = Y[:, 0]\n",
        "    y_coords = Y[:, 1]\n",
        "    # display scatter plot\n",
        "    plt.scatter(x_coords, y_coords)\n",
        "\n",
        "    for label, x, y in zip(word_labels, x_coords, y_coords):\n",
        "        plt.annotate(label, xy=(x, y), xytext=(0, 0), textcoords='offset points')\n",
        "    plt.xlim(x_coords.min()+0.00005, x_coords.max()+0.00005)\n",
        "    plt.ylim(y_coords.min()+0.00005, y_coords.max()+0.00005)\n",
        "    plt.show()\n",
        "    \n",
        "def tsne_plot(word_vect, words):\n",
        "    \"Creates and TSNE model and plots it\"\n",
        "    labels = []\n",
        "    tokens = []\n",
        "\n",
        "    #for word in model.wv.vocab:\n",
        "    for word in words:\n",
        "        tokens.append(word_vect[word])\n",
        "        labels.append(word)\n",
        "    \n",
        "    tsne_model = TSNE(perplexity=40, n_components=2, init='pca', n_iter=2500, random_state=23)\n",
        "    new_values = tsne_model.fit_transform(tokens)\n",
        "\n",
        "    x = []\n",
        "    y = []\n",
        "    for value in new_values:\n",
        "        x.append(value[0])\n",
        "        y.append(value[1])\n",
        "        \n",
        "    plt.figure(figsize=(14, 10)) \n",
        "    for i in range(len(x)):\n",
        "        plt.scatter(x[i],y[i])\n",
        "        plt.annotate(labels[i],\n",
        "                     xy=(x[i], y[i]),\n",
        "                     xytext=(5, 2),\n",
        "                     textcoords='offset points',\n",
        "                     ha='right',\n",
        "                     va='bottom')\n",
        "    plt.show()"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NTBTCXLuNvmk"
      },
      "source": [
        "### Read and Show a Data Sample"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WYC6OB5GNvml",
        "tags": [
          "outputPrepend"
        ]
      },
      "source": [
        "file = 'dataset.csv'\n",
        "df = pd.read_csv(file,encoding='ISO-8859-1')\n",
        "\n",
        "labels = df.NumLabel\n",
        "labels = labels.to_numpy()"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 195
        },
        "id": "zKe-o96UNvml",
        "outputId": "c974fa19-8c21-41ad-dedb-b88a799718ee"
      },
      "source": [
        "df.sample(5)"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Label</th>\n",
              "      <th>NumLabel</th>\n",
              "      <th>Text</th>\n",
              "      <th>FileName</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>Design Document</td>\n",
              "      <td>0</td>\n",
              "      <td>place your ssl tls server keys and certificate...</td>\n",
              "      <td>../MSRDataset/Oracle/Design Document/2521-dock...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>128</th>\n",
              "      <td>License</td>\n",
              "      <td>4</td>\n",
              "      <td>mit licenced copyright Ã¢ wisembly permission ...</td>\n",
              "      <td>../MSRDataset/Oracle/License/21787-elephant.io...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>158</th>\n",
              "      <td>Contributors' Guide</td>\n",
              "      <td>5</td>\n",
              "      <td>running the tests you ll need npm https npmjs ...</td>\n",
              "      <td>../MSRDataset/Oracle/Contributors' Guide/15685...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>98</th>\n",
              "      <td>License</td>\n",
              "      <td>4</td>\n",
              "      <td>the mit license copyright c automattic dev clo...</td>\n",
              "      <td>../MSRDataset/Oracle/License/21443-socket.io-n...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>131</th>\n",
              "      <td>Contributors' Guide</td>\n",
              "      <td>5</td>\n",
              "      <td>angular js enterprise edition boilerplate this...</td>\n",
              "      <td>../MSRDataset/Oracle/Contributors' Guide/5371-...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                   Label  ...                                           FileName\n",
              "11       Design Document  ...  ../MSRDataset/Oracle/Design Document/2521-dock...\n",
              "128              License  ...  ../MSRDataset/Oracle/License/21787-elephant.io...\n",
              "158  Contributors' Guide  ...  ../MSRDataset/Oracle/Contributors' Guide/15685...\n",
              "98               License  ...  ../MSRDataset/Oracle/License/21443-socket.io-n...\n",
              "131  Contributors' Guide  ...  ../MSRDataset/Oracle/Contributors' Guide/5371-...\n",
              "\n",
              "[5 rows x 4 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 294
        },
        "id": "x5BpAMmQNvmn",
        "outputId": "99bdca0b-a519-4da5-8b11-877c9becfc57"
      },
      "source": [
        "fig, ax = plt.subplots()\n",
        "fig.suptitle(\"Univariate Distribution\", fontsize=12)\n",
        "df[\"Label\"].reset_index().groupby(\"Label\").count().sort_values(by= \n",
        "       \"index\").plot(kind=\"barh\", legend=False, \n",
        "        ax=ax).grid(axis='x')\n",
        "plt.show()"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nSQhYuSINvmo"
      },
      "source": [
        "### Run Preprocessing Function to Clean Text"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 195
        },
        "id": "lgu-abzONvmo",
        "outputId": "75e021c4-4ebb-4fed-f9cb-5d20b73c4cea"
      },
      "source": [
        "# Load stopwords from NLTK\n",
        "lst_stopwords = nltk.corpus.stopwords.words(\"english\")\n",
        "df[\"text_clean\"] = df[\"Text\"].apply(lambda x: \n",
        "                      uf.utils_preprocess_text(x, flg_stemm=True, flg_lemm=True, \n",
        "                      lst_stopwords=lst_stopwords))\n",
        "df.sample(5)"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Label</th>\n",
              "      <th>NumLabel</th>\n",
              "      <th>Text</th>\n",
              "      <th>FileName</th>\n",
              "      <th>text_clean</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>31</th>\n",
              "      <td>List of Contributors</td>\n",
              "      <td>1</td>\n",
              "      <td>copyrightbegin copyright ericsson ab all right...</td>\n",
              "      <td>../MSRDataset/Oracle/List of Contributors/1544...</td>\n",
              "      <td>copyrightbegin copyright ericsson ab right res...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>205</th>\n",
              "      <td>Release Notes</td>\n",
              "      <td>6</td>\n",
              "      <td>fixes fixed status code error on supervisord s...</td>\n",
              "      <td>../MSRDataset/Oracle/Release Notes/7215-dextr-...</td>\n",
              "      <td>fix fix statu code error supervisord superviso...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>60</th>\n",
              "      <td>Setup Files</td>\n",
              "      <td>2</td>\n",
              "      <td>name device js version homepage https github c...</td>\n",
              "      <td>../MSRDataset/Oracle/Setup Files/23829-device....</td>\n",
              "      <td>name devic j version homepag http github com m...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>30</th>\n",
              "      <td>List of Contributors</td>\n",
              "      <td>1</td>\n",
              "      <td>original authors niclas eklund contributors</td>\n",
              "      <td>../MSRDataset/Oracle/List of Contributors/1544...</td>\n",
              "      <td>origin author nicla eklund contributor</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>Design Document</td>\n",
              "      <td>0</td>\n",
              "      <td>mechanize forms span class docwarning this doc...</td>\n",
              "      <td>../MSRDataset/Oracle/Design Document/1641-mech...</td>\n",
              "      <td>mechan form span class docwarn document need r...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                    Label  ...                                         text_clean\n",
              "31   List of Contributors  ...  copyrightbegin copyright ericsson ab right res...\n",
              "205         Release Notes  ...  fix fix statu code error supervisord superviso...\n",
              "60            Setup Files  ...  name devic j version homepag http github com m...\n",
              "30   List of Contributors  ...             origin author nicla eklund contributor\n",
              "6         Design Document  ...  mechan form span class docwarn document need r...\n",
              "\n",
              "[5 rows x 5 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AgJ__rpMNvmp"
      },
      "source": [
        "### Text Length Analysis"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 402
        },
        "id": "zRaqpeKINvmp",
        "outputId": "2bb1c887-9aef-49ea-f2ed-9fdcef39ec05"
      },
      "source": [
        "dtf = add_text_length(df, \"text_clean\")\n",
        "dtf.head()"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "                      min         mean            max\n",
            "word_count            3.0  1186.144231   40837.000000\n",
            "char_count           16.0  5896.274038  206636.000000\n",
            "sentence_count        1.0     1.000000       1.000000\n",
            "avg_word_length       3.5     5.384789       7.153846\n",
            "avg_sentence_lenght   3.0  1186.144231   40837.000000\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Label</th>\n",
              "      <th>NumLabel</th>\n",
              "      <th>Text</th>\n",
              "      <th>FileName</th>\n",
              "      <th>text_clean</th>\n",
              "      <th>word_count</th>\n",
              "      <th>char_count</th>\n",
              "      <th>sentence_count</th>\n",
              "      <th>avg_word_length</th>\n",
              "      <th>avg_sentence_lenght</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Design Document</td>\n",
              "      <td>0</td>\n",
              "      <td>design operation and maintenance manual for me...</td>\n",
              "      <td>../MSRDataset/Oracle/Design Document/13663-cmg...</td>\n",
              "      <td>design oper mainten manual messagelogg message...</td>\n",
              "      <td>2485</td>\n",
              "      <td>14533</td>\n",
              "      <td>1</td>\n",
              "      <td>5.848290</td>\n",
              "      <td>2485.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Design Document</td>\n",
              "      <td>0</td>\n",
              "      <td>gspread documentation master file created by s...</td>\n",
              "      <td>../MSRDataset/Oracle/Design Document/30200-gsp...</td>\n",
              "      <td>gspread document master file creat sphinx quic...</td>\n",
              "      <td>133</td>\n",
              "      <td>899</td>\n",
              "      <td>1</td>\n",
              "      <td>6.759398</td>\n",
              "      <td>133.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Design Document</td>\n",
              "      <td>0</td>\n",
              "      <td>non blocking trace setting introduction before...</td>\n",
              "      <td>../MSRDataset/Oracle/Design Document/15549-otp...</td>\n",
              "      <td>non block trace set introduct otp trace set ch...</td>\n",
              "      <td>772</td>\n",
              "      <td>4302</td>\n",
              "      <td>1</td>\n",
              "      <td>5.572539</td>\n",
              "      <td>772.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Design Document</td>\n",
              "      <td>0</td>\n",
              "      <td>layout doc title delite popup delite popup del...</td>\n",
              "      <td>../MSRDataset/Oracle/Design Document/12383-del...</td>\n",
              "      <td>layout doc titl delit popup delit popup delit ...</td>\n",
              "      <td>983</td>\n",
              "      <td>5086</td>\n",
              "      <td>1</td>\n",
              "      <td>5.173957</td>\n",
              "      <td>983.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Design Document</td>\n",
              "      <td>0</td>\n",
              "      <td>design patterns observer</td>\n",
              "      <td>../MSRDataset/Oracle/Design Document/55544-des...</td>\n",
              "      <td>design pattern observ</td>\n",
              "      <td>3</td>\n",
              "      <td>19</td>\n",
              "      <td>1</td>\n",
              "      <td>6.333333</td>\n",
              "      <td>3.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "             Label  NumLabel  ... avg_word_length avg_sentence_lenght\n",
              "0  Design Document         0  ...        5.848290              2485.0\n",
              "1  Design Document         0  ...        6.759398               133.0\n",
              "2  Design Document         0  ...        5.572539               772.0\n",
              "3  Design Document         0  ...        5.173957               983.0\n",
              "4  Design Document         0  ...        6.333333                 3.0\n",
              "\n",
              "[5 rows x 10 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nxz4gLekNvms"
      },
      "source": [
        "# encode y_train\n",
        "def encode_y_train(y_train):\n",
        "    dic_y_mapping = {n:label for n,label in enumerate(np.unique(y_train))}\n",
        "    inverse_dic = {v:k for k,v in dic_y_mapping.items()}\n",
        "    y_train = np.array([inverse_dic[y] for y in y_train])\n",
        "    return y_train, dic_y_mapping"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 116,
          "referenced_widgets": [
            "80bb686f18294b3089e7a1a6895b4f2a",
            "25300b1ce6dd42a5938be2c70282d0f6",
            "e6099193d78c4fa9bd837d5d73c3e990",
            "317caba37f83404c85121e6a3370ef2e",
            "17bccd0f294d4a3d9743a6858441c93b",
            "453ea79107af4a6695bd26a010080ff6",
            "f450869ea47a4213b4ca33b6a8f09a6e",
            "1d237d946165474dba7554c0ca3ae604",
            "184a311dbbfc462a9215688bf3ac74bb",
            "5c4f5a1e77294fecada31be2833a40eb",
            "60c7459612804b69a5863ce680fbdc13",
            "caec3ff5c84d45f8b3e2d4144cee2249",
            "f84b3d21f2b24bbc8604e06f442c37df",
            "419b31bd03ea4dbdaca65ffd9fb8741c",
            "3b2b63efcd2646c8a884353ad40041d6",
            "e0476fffa97547e0baaa6bbd44accf19"
          ]
        },
        "id": "4BMsMwpZcWZo",
        "outputId": "8a4e4086-a7b1-4873-905f-c3f054f77fdd"
      },
      "source": [
        "from transformers import AutoModel, BertTokenizerFast\n",
        "\n",
        "# import BERT-base pretrained model\n",
        "# base_bert = AutoModel.from_pretrained('bert-base-uncased')\n",
        "\n",
        "# Load the BERT tokenizer\n",
        "tokenizer = BertTokenizerFast.from_pretrained('bert-base-uncased')"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "80bb686f18294b3089e7a1a6895b4f2a",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=231508.0, style=ProgressStyle(descripti…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "184a311dbbfc462a9215688bf3ac74bb",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=466062.0, style=ProgressStyle(descripti…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uFtqfq-CNvms"
      },
      "source": [
        "## BERT\n",
        "Transfer learning from a pre-trained of BERT:\n",
        "24-layer, 1024-hidden, 16-heads, 336M parameters.\n",
        "Trained on lower-cased English text."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5cTaZ3LNNvms"
      },
      "source": [
        "### Tokenizer Definition"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u1XYDhyJNvmt"
      },
      "source": [
        "base_bert = 'bert-base-uncased' # Pick any desired pre-trained model\n",
        "\n",
        "# Defining DistilBERT tokonizer\n",
        "# tokenizer = BertTokenizer.from_pretrained(base_bert, do_lower_case=True, add_special_tokens=True,\n",
        "#                                                 max_length=256, pad_to_max_length=True)"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "efRAoc4aNvmt"
      },
      "source": [
        "### Load Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 174,
          "referenced_widgets": [
            "deb2a286c0b949f49afca2a44add412c",
            "0f2059ac53c54ccfaa659f0cd6a97472",
            "ef9e8a34ac1c46719fe2892d008591fa",
            "1caa12aeb3064d659b42353ce575c0e2",
            "9137e74a56654a3b92bbd418d897daee",
            "4a99031f626141428097c6217b129d96",
            "d9c1b25317d24557acc74a75ead8d9e6",
            "293a246ac4894c25a0ec293665c38838"
          ]
        },
        "id": "95o78P2pNvmt",
        "outputId": "b4af8d37-a203-4bf0-8e36-3cc91fa349f0"
      },
      "source": [
        "config = BertConfig(dropout=0.2, attention_dropout=0.2)\n",
        "config.output_hidden_states = False\n",
        "transformer_model = TFBertModel.from_pretrained(base_bert, config = config)"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "deb2a286c0b949f49afca2a44add412c",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=536063208.0, style=ProgressStyle(descri…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "Some layers from the model checkpoint at bert-base-uncased were not used when initializing TFBertModel: ['nsp___cls', 'mlm___cls']\n",
            "- This IS expected if you are initializing TFBertModel from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n",
            "- This IS NOT expected if you are initializing TFBertModel from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n",
            "All the layers of TFBertModel were initialized from the model checkpoint at bert-base-uncased.\n",
            "If your task is similar to the task the model of the checkpoint was trained on, you can already use TFBertModel for predictions without further training.\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-ciu4AQUNvmu"
      },
      "source": [
        "### Tokenization of Documents"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UEILN-ChNvmu"
      },
      "source": [
        "def bert_tokenize(sentences, tokenizer):\n",
        "    corpus = []\n",
        "    sentences_list = []\n",
        "    sentences_list = sentences.tolist()\n",
        "#     for i in range(len(sentences_list)):\n",
        "#         split_sentences = get_split(sentences_list[i])\n",
        "#         corpus = corpus + split_sentences\n",
        "#     print(\"Sentenças no Corpus:\", len(corpus))\n",
        "    input_ids, input_masks, input_segments = [],[],[]\n",
        "#     for sentence in tqdm(sentences):\n",
        "    for sentence in sentences_list:\n",
        "#         print(sentence)\n",
        "#         print(len(sentence.split()))\n",
        "        inputs = tokenizer.encode_plus(sentence, add_special_tokens=True, max_length=128, pad_to_max_length=True, \n",
        "                                             return_attention_mask=True, return_token_type_ids=True)\n",
        "        input_ids.append(inputs['input_ids'])\n",
        "        input_masks.append(inputs['attention_mask'])\n",
        "#         print(input_ids)\n",
        "#         print(len(input_ids[0]))\n",
        "#     print(len(input_ids))\n",
        "#         input_segments.append(inputs['token_type_ids'])        \n",
        "        \n",
        "#     return np.asarray(input_ids, dtype='int32'), np.asarray(input_masks, dtype='int32'), np.asarray(input_segments, dtype='int32')\n",
        "    return np.asarray(input_ids, dtype='int32'), np.asarray(input_masks, dtype='int32')\n",
        "\n",
        "from tensorflow.keras.constraints import max_norm\n",
        "def fit_bert_classif(X_train, y_train, X_test, encode_y=False, dic_y_mapping=None, model=None, epochs=1, batch_size=8):\n",
        "    ## encode y\n",
        "    if encode_y is True:\n",
        "        dic_y_mapping = {n:label for n,label in enumerate(np.unique(y_train))}\n",
        "        inverse_dic = {v:k for k,v in dic_y_mapping.items()}\n",
        "        y_train = np.array( [inverse_dic[y] for y in y_train] )\n",
        "    # print(dic_y_mapping)\n",
        "    \n",
        "    ## model\n",
        "    if model is None:\n",
        "        ### inputs\n",
        "        input_ids_in = tf.keras.layers.Input(shape=(128,), name='input_token', dtype='int32')\n",
        "        input_masks_in = tf.keras.layers.Input(shape=(128,), name='masked_token', dtype='int32')\n",
        "#         input_segments_in = tf.keras.layers.Input(shape=(128,), name='segment_token', dtype='int32') \n",
        "\n",
        "        \n",
        "        ### pre-trained bert\n",
        "        X = transformer_model(input_ids_in, attention_mask=input_masks_in)[0]\n",
        "    \n",
        "        ### fine-tuning\n",
        "        # 2 layers of bidirectional lstm\n",
        "        X = layers.Bidirectional(layers.LSTM(180, dropout=0.0,\n",
        "                              return_sequences=True))(X)\n",
        "        X = layers.Bidirectional(layers.LSTM(90, dropout=0.0))(X)\n",
        "\n",
        "        # ## final dense layers\n",
        "        X = layers.Dense(32, activation='tanh', kernel_constraint=max_norm(11))(X)\n",
        "        # X = layers.Dropout(0.1)(X)\n",
        "\n",
        "        y_out = layers.Dense(len(np.unique(y_train)), activation='softmax')(X)\n",
        "\n",
        "\n",
        "        ### compile\n",
        "        model = Model(inputs=[input_ids_in, input_masks_in], outputs = y_out)\n",
        "\n",
        "        for layer in model.layers[:3]:\n",
        "            layer.trainable = False\n",
        "\n",
        "        model.compile(loss='sparse_categorical_crossentropy', optimizer='Nadam', metrics=['accuracy'])\n",
        "        # print(model.summary())\n",
        "        \n",
        "    ## train\n",
        "    verbose = 1 if epochs > 1 else 2\n",
        "    training = model.fit(x=X_train, y=y_train, batch_size=batch_size, epochs=epochs, shuffle=True, verbose=verbose, validation_split=0.2)\n",
        "    if epochs > 1:\n",
        "        utils_plot_keras_training(training)\n",
        "    \n",
        "    ## test\n",
        "    predicted_prob = model.predict(X_test)\n",
        "    predicted = [dic_y_mapping[np.argmax(pred)] for pred in predicted_prob] if encode_y is True else [np.argmax(pred)]\n",
        "    return training.model, predicted_prob, predicted\n",
        "\n",
        "\n",
        "'''\n",
        "Plot loss and metrics of keras training.\n",
        "'''\n",
        "def utils_plot_keras_training(training):\n",
        "    metrics = [k for k in training.history.keys() if (\"loss\" not in k) and (\"val\" not in k)]\n",
        "    fig, ax = plt.subplots(nrows=1, ncols=2, sharey=True, figsize=(15,3))\n",
        "    \n",
        "    ## training\n",
        "    ax[0].set(title=\"Training\")\n",
        "    ax11 = ax[0].twinx()\n",
        "    ax[0].plot(training.history['loss'], color='black')\n",
        "    ax[0].set_xlabel('Epochs')\n",
        "    ax[0].set_ylabel('Loss', color='black')\n",
        "    for metric in metrics:\n",
        "        ax11.plot(training.history[metric], label=metric)\n",
        "    ax11.set_ylabel(\"Score\", color='steelblue')\n",
        "    ax11.legend()\n",
        "    \n",
        "    ## validation\n",
        "    ax[1].set(title=\"Validation\")\n",
        "    ax22 = ax[1].twinx()\n",
        "    ax[1].plot(training.history['val_loss'], color='black')\n",
        "    ax[1].set_xlabel('Epochs')\n",
        "    ax[1].set_ylabel('Loss', color='black')\n",
        "    for metric in metrics:\n",
        "        ax22.plot(training.history['val_'+metric], label=metric)\n",
        "    ax22.set_ylabel(\"Score\", color=\"steelblue\")\n",
        "    plt.show()"
      ],
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ky4M2_woNvmr"
      },
      "source": [
        "### Define Target"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "G-k9KC9cNvmv"
      },
      "source": [
        "X = df[\"Text\"]\n",
        "y = df[\"Label\"].values"
      ],
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4JA5Nohe_lg2"
      },
      "source": [
        "# Split original sentence in small sentences.\n",
        "def get_split(text1):\n",
        "    l_total   = []\n",
        "    l_partial = []\n",
        "    len_final = 0\n",
        "    if len(text1.split())//150 > 0:\n",
        "        n = len(text1.split())//150\n",
        "    else:\n",
        "        n = 1\n",
        "    for w in range(n):\n",
        "        if w == 0:\n",
        "            l_partial = text1.split()[:200]\n",
        "            l_total.append(\" \".join(l_partial))\n",
        "            len_final = len_final+ (len(l_total[w].split()))\n",
        "        else:\n",
        "            l_partial = text1.split()[w*150:w*150+200]\n",
        "            l_total.append(\" \".join(l_partial))\n",
        "            len_final = len_final+ (len(l_total[w].split()))\n",
        "        if (w == n-1 and n != 1) and (len_final-50 < len(text1.split())):\n",
        "            l_partial = text1.split()[n*200-50:len(text1.split())]\n",
        "            l_total.append(\" \".join(l_partial))\n",
        "    return l_total"
      ],
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Et0gV6Jm_lg3",
        "outputId": "1b18f687-8bb9-4b01-fae6-d8b28bbbe909"
      },
      "source": [
        "# Function to replicate the same number of targets (classes) for each feature that is added.\n",
        "novas_features = []\n",
        "idx = 0\n",
        "total_geral = 0\n",
        "for i, val in X.iteritems():\n",
        "    total = get_split(X[i]) # divides long sentences into smaller sentences\n",
        "    novas_features = novas_features + total\n",
        "    if len(X[i].split()) > 200:\n",
        "        total_geral = total_geral + len(total)\n",
        "        for j in range(len(total)-1):\n",
        "            y = np.insert(y, idx, y[idx])\n",
        "            idx += 1\n",
        "    idx += 1\n",
        "print(f'Summary - Initial Samples: {len (df [\"Label\"].values)} | Total New Samples: {len (y)}')\n",
        "\n",
        "X = pd.Series(novas_features)"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Summary - Initial Samples: 208 | Total New Samples: 2315\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FFLtLQhx_lg3"
      },
      "source": [
        "### Plot data\r\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "en9gFTT__lg3"
      },
      "source": [
        "df_target = pd.DataFrame(y, columns=['Label'])\n",
        "df_data = pd.concat([X,df_target], axis=1)"
      ],
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 294
        },
        "id": "KjJA9nm2_lg4",
        "outputId": "04e37108-e9ec-4968-d09f-6c4bff4d991c"
      },
      "source": [
        "fig, ax = plt.subplots()\n",
        "fig.suptitle(\"Univariate Distribution\", fontsize=12)\n",
        "df_data[\"Label\"].reset_index().groupby(\"Label\").count().sort_values(by= \n",
        "       \"index\").plot(kind=\"barh\", legend=False, \n",
        "        ax=ax).grid(axis='x')\n",
        "plt.show()"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1IAvwG5yNvmv"
      },
      "source": [
        "### Training and Fine-tuning"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "VkYSVgkyNvmv",
        "outputId": "8cb86a61-293a-401b-96b2-9a3c8280bfaf"
      },
      "source": [
        "final_mean = []\n",
        "conf_matrix_list_of_arrays = []\n",
        "for i in range(1):\n",
        "\n",
        "    accrs = []\n",
        "    fold=0\n",
        "#     conf_matrix_list_of_arrays = []\n",
        "\n",
        "    k=2\n",
        "    sss = StratifiedShuffleSplit(n_splits=k, test_size=0.2, random_state=42)\n",
        "\n",
        "    for train_index, test_index in sss.split(X, y):\n",
        "        X_train, X_test = X[train_index], X[test_index]\n",
        "        y_train, y_test = y[train_index], y[test_index]\n",
        "\n",
        "        fold +=1\n",
        "\n",
        "        #create tensors\n",
        "        X_train = bert_tokenize(sentences=X_train, tokenizer=tokenizer)\n",
        "        X_test = bert_tokenize(sentences=X_test, tokenizer=tokenizer)\n",
        "\n",
        "        model, predicted_prob, predicted = fit_bert_classif(X_train, y_train, X_test, encode_y=True, \n",
        "                                                            model=None, epochs=2, batch_size=8)\n",
        "        # Validade Test Set\n",
        "        accr = metrics.accuracy_score(y_test, predicted)\n",
        "        #\n",
        "        conf_matrix = confusion_matrix(y_test, predicted)\n",
        "        conf_matrix_list_of_arrays.append(conf_matrix)\n",
        "        #\n",
        "        accrs.append(accr)\n",
        "\n",
        "        print('\\n--------------------------------------------------------------------')\n",
        "        print(f'Iteração: {i+1} Fold: {fold} Accuracy: {accr}, Mean Accuracy: {sum(accrs)/fold}')\n",
        "        print('\\n--------------------------------------------------------------------')\n",
        "    final_mean.append(sum(accrs)/k)\n",
        "\n",
        "# means for Confusion Matrix\n",
        "mean_of_conf_matrix_arrays = np.mean(conf_matrix_list_of_arrays, axis=0)"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:The parameters `output_attentions`, `output_hidden_states` and `use_cache` cannot be updated when calling a model.They have to be set to True/False in the config object (i.e.: `config=XConfig.from_pretrained('name', output_attentions=True)`).\n",
            "WARNING:tensorflow:The parameter `return_dict` cannot be set in graph mode and will always be set to `True`.\n",
            "Epoch 1/2\n",
            "WARNING:tensorflow:The parameters `output_attentions`, `output_hidden_states` and `use_cache` cannot be updated when calling a model.They have to be set to True/False in the config object (i.e.: `config=XConfig.from_pretrained('name', output_attentions=True)`).\n",
            "WARNING:tensorflow:The parameter `return_dict` cannot be set in graph mode and will always be set to `True`.\n",
            "WARNING:tensorflow:The parameters `output_attentions`, `output_hidden_states` and `use_cache` cannot be updated when calling a model.They have to be set to True/False in the config object (i.e.: `config=XConfig.from_pretrained('name', output_attentions=True)`).\n",
            "WARNING:tensorflow:The parameter `return_dict` cannot be set in graph mode and will always be set to `True`.\n",
            "186/186 [==============================] - ETA: 0s - loss: 0.9484 - accuracy: 0.6685WARNING:tensorflow:The parameters `output_attentions`, `output_hidden_states` and `use_cache` cannot be updated when calling a model.They have to be set to True/False in the config object (i.e.: `config=XConfig.from_pretrained('name', output_attentions=True)`).\n",
            "WARNING:tensorflow:The parameter `return_dict` cannot be set in graph mode and will always be set to `True`.\n",
            "186/186 [==============================] - 75s 315ms/step - loss: 0.9471 - accuracy: 0.6690 - val_loss: 0.6000 - val_accuracy: 0.7951\n",
            "Epoch 2/2\n",
            "186/186 [==============================] - 56s 301ms/step - loss: 0.4597 - accuracy: 0.8280 - val_loss: 1.3750 - val_accuracy: 0.5660\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA6MAAADgCAYAAAD2QrBDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdd5iU5fn28e+9Sy9SlL4gaOwoEFBjQwgqRECwIHVn7FGMpP5sCYmvqDFFjYImIYrO0LGgSAfFYBdskaKIgLD0siCdLdf7xzMLw7Jsn5mdZ8/PcezBzswzM/esyLXXPOdctzMzREREREREROIpJdELEBERERERkcpHzaiIiIiIiIjEnZpRERERERERiTs1oyIiIiIiIhJ3akZFREREREQk7tSMioiIiIiISNypGRWpgJxzs5xzwfI+VkREpCJyzplz7keR7//lnBtenGNL8TyDnXNzS7tOESlfTvuMipQP59yeqIu1gINATuTyz81sfPxXJSIiEh/OudnAJ2b2x3zX9wH+DaSZWfZx7mvAaWa2shjPU6xjnXOtgdVA1eM9r4gkls6MipQTM6uT9wWsBXpHXXe4EXXOVUncKkVERGImBAxxzrl816cD49UQikh+akZFYsw518U5l+Gcu885twl40TnXwDk33Tm31TmXGfk+Leo+7zjnbot8f5Nz7j3n3N8jx652zv2slMe2cc4tdM7tds7Nd84965wbF8cfh4iI+NfrwInAZXlXOOcaAL2Aac65D51zO51zG51zo5xz1Qp6EOfcS865R6Iu/1/kPhucc7fkO7anc+5z59wPzrl1zrmHom5eGPlzp3Nuj3Puorw6GXX/i51zi5xzuyJ/Xhx12zvOuRHOufcjdXOuc+6kMvx8RCQfNaMi8dEUaAicDNyB9//ei5HLrYD9wKhC7n8h8A1wEvBX4IUC3nkuzrETgE/wfll4CO/dahERkTIzs/3AFCAQdfWNwNfAHuDXeLXpIqAbMLSox3TO9QB+B1wJnAZcke+QvZHnqw/0BO5yzvWN3NY58mf9SErpw3yP3RCYATyDVxefBGY4506MOmwQcDPQGKgWWYuIlBM1oyLxkQv8ycwOmtl+M9tuZq+a2T4z2w08ClxeyP2/N7P/mFkOXgyqGdCkJMc651oB5wN/NLNDZvYeMK28XqCIiAhe3bnBOVcjcjkAhMzsUzP7yMyyzWwN3mdIC6t7eW4EXjSzJWa2F++N1MPM7B0z+8rMcs3sf8DEYj4ueM3rt2Y2NrKuiXiNc++oY140sxVRjXb7Yj62iBSDmlGR+NhqZgfyLjjnajnn/u2c+9459wNelKi+cy71OPfflPeNme2LfFunhMc2B3ZEXQewroSvQ0RE5Lgib3RuA/o6504FLgAmOOdOj3wkZVOk7j2Gd5a0KM05ulZ9H32jc+5C59yCyMdedgF3FvNx8x77+3zXfQ+0iLq8Ker7fRy/9opIKagZFYmP/GOrfwucAVxoZidwJEp0vOhtedgINHTO1Yq6rmUMn09ERCqnMN4Z0SHAHDPbDPwT76zjaZG69yDFq3kbObpWtcp3+wS8lE9LM6sH/CvqcYvaMmID3sdlorUC1hdjXSJSDtSMiiRGXbzPie6MfGblT7F+QjP7HlgMPOScq+acu4ijo0giIiLlIYz32c7b8WK74NW9H4A9zrkzgbuK+VhTgJucc2dH3kzNXy/r4qV+DjjnLsD7jGeerXgfkznlOI89EzjdOTfIOVfFOdcfOBuYXsy1iUgZqRkVSYx/ADXxokwfAbPj9LyD8QZHbAceASbj7YcqIiJSLiKfCf0AqM2R2QS/w2sUdwP/was/xXmsWXg1821gZeTPaEOBh51zu4E/4jWveffdhzeT4f3IFN+f5Hvs7XiTfn+LVxfvBXqZ2bbivlYRKRtnVlSCQUT8yjk3GfjazGJ+ZlZEREREJJrOjIpUIs65851zpzrnUiLj8vvg7QsnIiIiIhJXVRK9ABGJq6bAa3j7qWUAd5nZ54ldkoiIiIhURorpioiIiIiISNwppisiIiIiIiJxF7Nm1Dk3xjm3xTm3pIjjznfOZTvnbojVWkRERERERKRiiVlM1znXGdgDhM2s7XGOSQXmAQeAMWb2SlGPm5KSYjVr1izXtYqISMW0b98+MzOleIpJNVJEpHIoTn2MDKt8GkgFnjezx/Pd3gpvL+D6kWPuN7OZkdseAG4FcoBhZjan/F9FDAcYmdlC51zrIg67B3gVOL+4j1uzZk327t1bhpWJiEiycM7tT/QakolqpIhI5VBUfYyc9HsWuBJvaOUi59w0M1sWddgfgClm9k/n3NnATKB15PsBwDlAc2C+c+50M8sp79eRsHebnXMtgGuBfyZqDSIiIiIiIj50AbDSzFaZ2SFgEt6WftEMOCHyfT1gQ+T7PsAkMztoZquBlZHHK3eJjD79A7jPzHKLOtA5d4dzbrFzbnF2dnYcliYiIpIcVCNFRKQALYB1UZczItdFewgY4pzLwDsrek8J7lsuEtmMdgImOefWADcAzznn+hZ0oJmNNrNOZtapShVtjSoiIpJHNVJEpFKqkvdGZOTrjlI8xkDgJTNLA64Gxjrn4tofJqxqmVmbvO+dcy8B083s9dI8VlZWFhkZGRw4cKC8llep1KhRg7S0NKpWrZropYiIiIiISNGyzaxTIbevB1pGXU6LXBftVqAHgJl96JyrAZxUzPuWi5g1o865iUAX4KTIqd8/AVUBzOxf5flcGRkZ1K1bl9atW+OcK8+H9j0zY/v27WRkZNCmTZui7yAiUojsbHjzTZg4EcaPB73HlVyefhp++AHuuAOaNEn0akRE/OWbb74hHA5z8sknc8cdpTmRWSKLgNOcc23wGskBwKB8x6wFugEvOefOAmoAW4FpwATn3JN4A4xOAz6JxSJjOU13YAmOvaksz3XgwAE1oqXknOPEE09k69atiV6KiCSxHTvg+efhuefg+++hVStYvRpOPz3RK5OS+PRTGDsWHnkEBgyAYcOgY8dEr0pEJHnt2LGDSZMmEQqF+OSTT0hJSeHOO++M+fOaWbZz7hfAHLxtW8aY2VLn3MPAYjObBvwW+I9z7td4w4xuMm/fz6XOuSnAMiAbuDsWk3QhhvuMxkrt2rUt/9j65cuXc9ZZZyVoRf6gn6GIlMZXX8HIkTBuHOzfD126eA1M795QHh9fdM7tM7PaZX+kyqGgGllS33wDo0bBSy/Bnj1w0UXef9Prr9eZbhGR4jh06BCzZs0iHA7z5ptvkpWVRdu2bQkGgwwePJhmzZqV+Tn8Uh816UBEREokL4r7zDPwzjtQowakp8MvfgHnnZfo1UlZnXGG9wbDI494DemoUTBwIDRvDnfeqQiviEhBzIzPPvuMUCjExIkT2bZtG40aNeLuu+8mEAjQvn17pTgLoDOjSSY7O5tYTEusTD9DESmdgqK4d98Nt94KJ54Ym+f0yzu/8VIeZ0bzy82F2bO9Nx/mzIFq1RThFRHJs379esaPH084HGbp0qVUq1aNa665hmAwSPfu3WM2INQv9TGRW7v4Tt++fenYsSPnnHMOo0ePBmD27Nn8+Mc/pl27dnTr1g2APXv2cPPNN3Puuedy3nnn8eqrrwJQp06dw4/1yiuvcNNNNwFw0003ceedd3LhhRdy77338sknn3DRRRfRoUMHLr74Yr755hsAcnJy+N3vfkfbtm0577zzGDlyJG+//TZ9+x7ZMWfevHlce+218fhxiIhPfPWVdzYsLQ3uuw/atIHXXoPvvoN7741dIyoVQ0oKXH2115B+/bX3d+G116BTJ7jkEpg0CbKyEr1KEZH42bdvH+PHj6d79+60atWK++67jxNOOIF//vOfbNq0iZdffplevXppp4pi8F1M91e/gi++KN/HbN8e/vGPoo8bM2YMDRs2ZP/+/Zx//vn06dOH22+/nYULF9KmTRt27NgBwIgRI6hXrx5fffUVAJmZmUU+dkZGBh988AGpqan88MMPvPvuu1SpUoX58+fz4IMP8uqrrzJ69GjWrFnDF198QZUqVdixYwcNGjRg6NChbN26lUaNGvHiiy9yyy23lOnnISL+pyiuFKSwCO9dd3mNauPGiV6liEj5y83N5d133yUcDvPyyy+ze/duWrVqxYMPPkh6ejqna2JfqfiuGU2kZ555hqlTpwKwbt06Ro8eTefOnQ9vmdKwYUMA5s+fz6RJkw7fr0GDBkU+dr9+/UhNTQVg165dBINBvv32W5xzZEXekp4/fz533nnn4Rhv3vOlp6czbtw4br75Zj788EPC4XA5vWIR8ZuCorh/+Utso7iSfOrVg1/+Eu6550iEd/hwGDFCEV4R8ZeVK1cSDocZO3Ysa9asoU6dOvTr149AIEDnzp1JSVHQtCx814wW5wxmLLzzzjvMnz+fDz/8kFq1atGlSxfat2/P119/XezHiP5Q84EDB466rXbtI5Hw4cOH07VrV6ZOncqaNWvo0qVLoY97880307t3b2rUqEG/fv1i8plTEUluBU3FffJJuOaa8pmKK/6UF+G9+movwjtqFIRCEA7DxRd7zaqm8IpIstm5cyeTJ08mHA7zwQcf4Jzjiiuu4JFHHqFv375H/V4uZaNWvpzs2rWLBg0aUKtWLb7++ms++ugjDhw4wMKFC1m9ejXA4ZjulVdeybPPPnv4vnkx3SZNmrB8+XJyc3MPn2E93nO1aNECgJdeeunw9VdeeSX//ve/yc7OPur5mjdvTvPmzXnkkUe4+eaby+9Fi0hSy8mBqVOha1cvejt2LAweDF9+CQsWwHXXqRGV4jvzTK8Zzcjw3hjevNmL8LZu7cV6t2xJ9ApFRI4vKyuLGTNmcOONN9K0aVPuvPNOMjMzefzxx1m7di1z585l8ODBakTLmZrRctKjRw+ys7M566yzuP/++/nJT35Co0aNGD16NNdddx3t2rWjf//+APzhD38gMzOTtm3b0q5dOxYsWADA448/Tq9evbj44osL3X/o3nvv5YEHHqBDhw6HG0+A2267jVatWnHeeefRrl07JkyYcPi2wYMH07JlS03MFRF27IC//Q1OPdVrOFet8qK4GRnwn//oM6FSNnkR3hUrYMYMOPdcL8LbsiUEg/Dpp4leoYjIEV988QW/+c1vSEtLo1evXrz99tvccccdLFq0iKVLl3LfffeRlpaW6GX6lrZ2qSR+8Ytf0KFDB2699dYCb9fPUMT/Cori3nNPxY7iFjW63jnXA3gaSAWeN7PH893eCggB9SPH3G9mM51zVwKPA9WAQ8D/mdnbMXoZcROLrV3KQ3SEd88eRXhFJLE2bdp0eDuW//3vf1StWpVevXoRDAb52c9+RrVq1RK9xCL5ZWsXNaOVQMeOHalduzbz5s2jevXqBR6jn6GIP+XkwLRpR0/FHTLEawSS4QxoYcXWOZcKrACuBDKARcBAM1sWdcxo4HMz+6dz7mxgppm1ds51ADab2QbnXFtgjpm1iPkLirGK2ozm2bXLm8I7cqS3NZCm8IpIvOzfv59p06YRCoWYM2cOubm5XHDBBQQCAQYMGMCJSTalzy/NaAV9L1zK06fKRIlUOjt2wAsvwLPP+noq7gXASjNbBeCcmwT0AZZFHWPACZHv6wEbAMzs86hjlgI1nXPVzexgzFddiUVP4Z01y2tKNYVXRGLFzHj//fcJh8NMmTKFXbt2kZaWxn333Ud6erpOxFQAakZFRHzEh1NxqzjnFkddHm1moyPftwDWRd2WAVyY7/4PAXOdc/cAtYErCniO64HP1IjGT0oK9OzpfRU0hXfYMO/zzIrwikhprF69mnA4TDgcZtWqVdSqVYvrr7+eYDBIly5dDm+XKImXnL+aFMDMjtoaRYov2aLaInK0vCjuyJHeFNxki+IWIdvMOpXh/gOBl8zsCefcRcBY51xbM8sFcM6dA/wFuKoc1iqlkDeF99FHj0R4BwxQhFdESmbXrl288sorhEIh3n33XZxzdO3alT/+8Y9cf/311KlTJ9FLlAL4YppujRo12L59u5qqUjAztm/fTo0aNRK9FBEpofxTcb/7rtJNxV0PtIy6nBa5LtqtwBQAM/sQqAGcBOCcSwOmAgEz+y7mq5VCRU/hnT5dU3hFpGjZ2dnMnj2bQYMG0bRpU2677TY2b97Mo48+ypo1a3jrrbcIBoNqRCswXwwwysrKIiMjgwMHDiRoVcmtRo0apKWlUVV5KJGkkIxTcUuriAFGVfAGGHXDa0IXAYPMbGnUMbOAyWb2knPuLOAtvHhvPeC/wP8zs9di/DLipqIPMCqpgqbwKsIrIkuWLCEUCjF+/Hg2btxIgwYNGDBgAMFgkAsuuKBSpCX9MsDIF82oiIjf5eTAm296U3F9GMU9rmJs7XI18A+8bVvGmNmjzrmHgcVmNi0yQfc/QB28YUb3mtlc59wfgAeAb6Me7ioz2xKzFxMHfq2RmsIrIlu2bGHixImEQiE+//xzqlSpwtVXX00gEKBXr17H3THCr9SMJohfC62ISEEKmop7992+m4p7XH4ptvHi9xqZm3tkCu+cOVCtGgwc6L0poym8Iv5z8OBB3nzzTcLhMLNmzSI7O5uOHTse3o6lcSV+N8ov9VHNqIhIBZQ/inv55V480Y9R3ML4pdjGS2WqkXkR3pdegr17FeEV8Qsz4+OPPyYUCjF58mQyMzNp1qwZQ4YMIRAI0LZt20QvsULwS31UMyoiUkFU1ihuYfxSbOOlMtZIRXhF/OH7779n3LhxhMNhVqxYQc2aNbn22msJBAJcccUV2o4lH7/URzWjIiIJVtmjuIXxS7GNl8pcIxXhFUk+u3fv5tVXXyUcDrNgwQIAOnfuTDAY5IYbbuCEE05I8AorLr/URzWjIiIJoihu0fxSbONFNdKjCK9IxZWTk8OCBQsIhUK89tpr7Nu3j1NPPZVAIEB6ejpt2rRJ9BKTgl/qo5pREZE4UhS3ZPxSbONFNfJou3bBiy96jakivCKJtXz5csLhMOPGjSMjI4N69erRv39/AoEAF198caXYjqU8+aU+qhkVEYmDgqK4Q4fCbbcpilsYvxTbeFGNLFhehPeZZ2DuXEV4ReJl27ZtTJo0iXA4zKJFi0hNTaVHjx4EAgF69+5NzZo1E73EpOWX+hizZtQ5NwboBWwxs2PGXjnnBgP3AQ7YDdxlZl8W9bgqtCKSTJYs8aK4Y8cqilsafim28aIaWTRFeEVi69ChQ8yYMYNwOMyMGTPIysqiXbt2BINBBg4cSNOmTRO9RF/wS32MZTPaGdgDhI/TjF4MLDezTOfcz4CHzOzCoh5XhVZEKjpFccuPX4ptvKhGFl9BEd6hQ70Ib6NGiV6dSHIxMxYvXkwoFGLSpEls376dJk2aMHjwYAKBAO3atUv0En2nOPXROdcDeBpIBZ43s8fz3f4U0DVysRbQ2MzqR27LAb6K3LbWzK4pz/UfXkMsY7rOudbA9IKa0XzHNQCWmFmLoh5ThVZEKqq8KO5zz8GaNdCypTcVV1Hc0lMzWjKqkSWnCK9I6WVkZBzejmX58uVUr16dvn37EggEuOqqq6iiCFDMFFUfnXOpwArgSiADWAQMNLNlxzn+HqCDmd0SubzHzOqU/8qPVlH+htwKzEr0IkRESqOgKO4TTyiKK5IMUlKgZ0/vKzrCGwopwitSkL179/Laa68RDod56623MDMuueQSRo8eTb9+/ahfv36ilyieC4CVZrYKwDk3CegDFNiMAgOBP8VpbYelxPsJ83POdcVrRu8r5Jg7nHOLnXOLs7Oz47c4EZHjyMmB11+Hn/4Uzj0XwmEYPBi+/BLeecf75VWNqMSDamT5OfNMrxldvx6eego2b4YBA6B1a3j0Udi6NdErFEmM3NxcFixYwE033USTJk0IBAKsXLmS4cOH8+233/Lee+9x++23qxGNryp5//ZHvu7Id3sLYF3U5YzIdcdwzp0MtAHejrq6RuRxP3LO9S3XlUc/dyJjus6584CpwM/MbEVxHlMRJBFJpMzMI1NxFcWNPcV0S0Y1snzlj/BWr+41p8OGwY9/nOjVicTeihUrCIfDjB07lrVr11K3bl1uvPFGAoEAl156KSkpCT+vVWkVI6Z7A9DDzG6LXE4HLjSzXxRw7H1AmpndE3VdCzNb75w7Ba9J7WZm35X360jY+/bOuVbAa0B6cRtREZFEURRXpPIpLMJ7ySXe50oV4RW/2bFjB5MnTyYcDvPRRx+RkpLClVdeyeOPP06fPn2oVatWopcoxbMeaBl1OS1yXUEGAHdHX2Fm6yN/rnLOvQN0AMq9GY3lNN2JQBfgJGAzXga5KoCZ/cs59zxwPfB95C7ZZtapqMfVu74iEi8FTcUdPNj7BVSDAeNDZ0ZLRjUy9vKm8I4cCatWaQqv+ENWVhazZ88mFArx5ptvcujQIc455xyCwSCDBw+mefPmiV6i5FOMM6NV8AYYdcNrQhcBg8xsab7jzgRmA20s0hhGhsvuM7ODzrmTgA+BPscbflSm1xHLmG4sqNCKSKwpiltxqBktGdXI+FGEV5KdmfH5558TDoeZMGECW7dupVGjRgwaNIhAIECHDh1wziV6mXIcxdza5WrgH3hbu4wxs0edcw8Di81sWuSYh4AaZnZ/1P0uBv4N5OLNGPqHmb0Qk9ehZlRExFNQFHfYMEVxE0nNaMmoRiZGdIR3714vwjtsGFx7rSK8UvFs2LCB8ePHEw6HWbJkCdWqVaN3794Eg0F69OhBVf2lTQp+qY9qRkWkUsuL4o4cCW+/rShuReOXYhsvqpGJlT/C26IF3HWXIrySePv27eONN94gFAoxb948cnNz+clPfkIgEKB///40bNgw0UuUEvJLfVQzKiKVkqK4ycEvxTZeVCMrhpwcL8I7cqQivJI4ubm5vPfee4TDYaZMmcLu3btp1aoV6enppKenc8YZZyR6iVIGfqmPakZFpFJRFDe5+KXYxotqZMWzfLkX4Q2FFOGV+Pjuu+8Ob8eyevVqateuTb9+/QgEAlx++eXajsUn/FIf1YyKiO8pipu8/FJs40U1suJShFdiaefOnbz88suEQiHef/99nHN069aNQCDAddddR+3a+mfUb/xSH9WMiohvKYqb/PxSbONFNbLiU4RXykt2djZz584lHA7z+uuvc/DgQc4888zD27G0bNmy6AeRpOWX+qhmVER8p6Ao7j33QJ8+iuImG78U23hRjUwuivBKafzvf/8jFAoxfvx4Nm/eTMOGDRk4cCDBYJBOnTppO5ZKwi/1Uc2oiPiCorj+5JdiGy+qkclp505vWxhFeOV4Nm/ezIQJEwiFQnz55ZdUrVqVnj17EggE6NmzJ9WqVUv0EiXO/FIf1YyKSFJTFNff/FJs40U1MrkVFOEdONB7U00R3srnwIEDTJs2jXA4zOzZs8nJyeH8888nEAgwYMAATjrppEQvURLIL/VRzaiIJKW8KO64cbBvH3Tu7MXbFMX1F78U23hRjfQPRXgrJzPjww8/JBQKMXnyZHbt2kWLFi0Ob8dy9tlnJ3qJUkH4pT6qGRWRpKEobuXjl2IbL6qR/qMIb+WwZs0axo4dSzgcZuXKldSqVYvrrruOYDBI165dSU1NTfQSpYLxS31UMyoiFZ6iuJWXX4ptvKhG+ldehPeZZ2DePEV4/eCHH37glVdeIRQKsXDhQgC6du1KIBDg+uuvp27dugleoVRkfqmPakZFpMJauvTIVFxFcSsnvxTbeFGNrBwU4U1eOTk5vPXWW4RCIaZOncr+/fs57bTTCAaDDBkyhJNPPjnRS5Qk4Zf6qGZURCqUnByYPt17919RXPFLsY0X1cjKpaAI79ChcPvtivBWNEuXLiUcDjNu3Dg2bNhA/fr1GTBgAMFgkAsvvFDbsUiJ+aU+qhkVkQqhoCju0KFeFFcDAysvvxTbeFGNrJwU4a2Ytm7dysSJEwmHw3z66aekpqZy9dVXEwgE6NWrFzVq1Ej0EiWJ+aU+qhkVkYRSFFcK45diGy+qkaIIb2IdPHiQ6dOnEw6HmTlzJtnZ2XTo0IFgMMjAgQNp3LhxopcoPuGX+qhmVETiTlFcKa6iiq1zrgfwNJAKPG9mj+e7vRUQAupHjrnfzGZGbnsAuBXIAYaZ2ZzYvIr4UY2UPDt3wosveo2pIryxZWZ88sknhEIhJk2aRGZmJk2bNmXIkCEEAgHOPffcRC9RfEjNaIKo0IokL0VxpaQKK7bOuVRgBXAlkAEsAgaa2bKoY0YDn5vZP51zZwMzzax15PuJwAVAc2A+cLqZ5cT2FcWWaqTkd7wI77Bh0KFDoleX3NauXcu4ceMIh8N888031KhRg2uvvZZAIMAVV1xBFcV7JIb80ozq/xIRibmCorh//7uiuFJmFwArzWwVgHNuEtAHWBZ1jAEnRL6vB2yIfN8HmGRmB4HVzrmVkcf7MB4LF4mX1FTo1cv7io7wvvQSXHqpl0hRhLf49uzZw6uvvko4HGbBggWYGZdddhn/93//xw033EC9evUSvUSRpJKS6AWIiD/l5MAbb0C3btC2rffLz8CB8MUX8N//wvXXqxGVMmsBrIu6nBG5LtpDwBDnXAYwE7inBPcV8ZWzzvKSKRkZ8OSTsGED9O8PbdrAY4/B1q2JXmHFlLcdSyAQoEmTJtx0002sWbOGP/3pT3z33XcsXLiQW2+9VY2oSCmoGRWRcpWZ6Z31/NGPoG9f+PZb+POfYd06eP55fSZUSqyKc25x1NcdJbz/QOAlM0sDrgbGOudU+6RSq18ffv1rWLEC3nwTzj4bfv9776MTN98Mn3+e6BVWDF9//TUPPvggrVu35oorruCNN95g8ODBvPvuu6xcuZI//elPnHLKKYlepkhS03kJESkXiuJKjGSbWafj3LYeaBl1OS1yXbRbgR4AZvahc64GcFIx7yvia4rwHmv79u1MmjSJcDjMJ598QkpKCt27d+fvf/8711xzDTVr1kz0EkV8RQOMRKTUCpqKO2iQ9wtM+/aJXp34QREDjKrgDTDqhtdILgIGmdnSqGNmAZPN7CXn3FnAW3hx3LOBCRwZYPQWcJoGGEllVxmn8B46dIhZs2YRCoWYPn06WVlZnHfeeQSDQQYNGkTTpk0TvRsbpl4AACAASURBVESRY/hlgJGaUREpsfxTcdPS4O67NRVXyl8xtna5GvgH3rYtY8zsUefcw8BiM5sWmZr7H6AO3jCje81sbuS+vwduAbKBX5nZrBi/nJhTjZTy4vcpvGbGp59+SjgcZuLEiWzbto3GjRszePBggsEg7fSZEqng1IwW9cDOjQF6AVvMrG0Btzu8veGuBvYBN5nZZ0U9rgqtSOIUFMUdNkxRXIkdvxTbeFGNlFhYtsw7UxoOw969yR3hXb9+/eHtWJYtW0b16tXp06cPgUCAq666iqrJ9oKk0ipOfSzGXtxPAV0jF2sBjc2sfuS2IPCHyG2PmFmoPNd/eA0xbEY7A3uA8HGa0avxphpeDVwIPG1mFxb1uCq0IvGVF8UdORLeektRXIkvNaMloxopsZSsEd69e/fy+uuvEwqFmD9/PmbGxRdfTDAYpF+/fjRo0CDRSxQpsWIkh4rcizvf8fcAHczsFudcQ2Ax0AkvVfQp0NHMMsv5ZcRumq6ZLQR2FHJIH7xG1czsI6C+c65ZrNYjIiWTmQlPPHFkKu433xyZivvCC2pERUQqm2Sawpubm8s777zDLbfcQtOmTRkyZAgrVqzgD3/4AytWrOD999/njjvuUCMqfnZ4L24zOwTk7cV9PAOBiZHvuwPzzGxHpAGdR2QYYHlLZLDueHu8bUzMckQECo7i/u1vXkOqKK6IiERP4Y2O8FaEKbzffvst4XCYsWPH8v3331O3bl1uvPFGAoEAl112GSkp2tlJKo2Ceq0CU6jOuZOBNsDbhdw3JntxJ8WvlpF95e4AqFatWoJXI+I/iuKKJC/VSEmks8+G556Dxx47EuHt3z++Ed7MzEymTJlCKBTiww8/JCUlhSuuuILHHnuMvn37UqtWrdguQCQxqjjnFkddHm1mo0v5WAOAVxIxUT6Rbw8Ve483MxttZp3MrFMVnZoRKTeK4oqUTPcRM2p2HzHjjESvI5pqpFQE0RHeadOOjvDeckv5R3izsrKYPn06N954I82aNePOO+9k165d/OUvf2Ht2rXMmTOHQYMGqREVP8vO+7c/8pW/ES3JftoDOBLRLel9gdLXx0Q2o9OAgPP8BNhlZoroisTB0qVw553eliy/+x20agUvvwyrV8P992t7FpGCdB8xozfwBTA7crl99xEzpiV2VSIVS2oq9O4Nc+d6teaWW2DKFPjxj+Gyy7xak5VVusc2M7744gt+/etfk5aWRu/evVmwYAE///nPWbx4MUuWLOHee++lRYuYpAlFks0i4DTnXBvnXDW8hvOYmuWcOxNoAHwYdfUc4CrnXAPnXAPgqsh1BSpLfYxZM+qcm4j3os5wzmU45251zt3pnLszcshMYBWwEm8PuKGxWouIeFHcN96AK66Atm0hFIIBA7x3q//7X7jhBn0mVKQID+ENhNgJMGd4zy/wPmMjIgXIi/BmZMCTT8KGDXDjjXDKKV6sd+vW4j3Oxo0beeKJJ2jXrh0dOnTg2Wef5dJLL+WNN95g/fr1PP3003Ts2BFv10ARATCzbOAXeE3kcmCKmS11zj3snLsm6tABwCSL2mLFzHYAI/Aa2kXAw5HrjuchSlkfi/Wrp3OuNrDfzHKdc6cDZwKzzOy4722Z2cDCHjPygu8uzvOLSOllZsKYMd7neNas8c6G/vnPcNttOgMqUkJZc4b33NV9xIzo62KzP1pEaeqvSEWTF+EdNgxmzoRnnvEivA8/fGQ+QYcOR99n//79vPHGG4RCIebOnUtubi4XXnghzz77LP379+fEE09MzIsRSSJmNhPvBGD0dX/Md/mh49x3DDCmmE9V6vpYrH1GnXOfApfhncJ9H69DPmRmg4u5wHKjPdREiqegqbj33KOpuJJcKtI+o91HzHgBeAu4H7geGAZUnTO8552F3rEMSlp/VSMlWeRN4Q2FvBrlTeE1Gjf+gPHjX2LKlCn88MMPtGzZkvT0dNLT0znzzDMTvWyRCsMv9bG4v5I6M9vnnLsVeM7M/uqc+6LUKxaRmMjJgRkzvHed33oLqleHwYM1FVeknNwD/B44CEzAiz49EuPnVP0VX4qewvv3v2/nuecc/fs3BFpRtWoLrr02nZ///Dq6dOmi7VhEKr5S18diN6POuYuAwcCtketSS7hIEYmRvCjus896Q4gUxRUpX91HzEgFZswZ3rMrXsGNF9Vf8aVdu3bx8ssvEw6Heffdd4FUzjvvfnJyhrJ06UO88QbUrg0NGhwb4RWRiqOs9bG4bzX9CngAmBr54OspwIKSPpmIlK/8U3FbttRUXJFYmDO8Zw6Q233EjHpxfmrVX/GN7OxsZs+ezcCBA2natCm33347W7Zs4bHHHuP771fx5ZePsGRJ88NTeCdPLp8pvCISO2Wtj8X6zOhRd3AuBahjZj+U5gnLSp+HkcpOUVypTCrYZ2LeADoA84DDhWjO8J7D4vH8xam/qpFSEX311VeEw2HGjRvHpk2baNCgAQMHDiQYDHL++ecfdwruzp1HUj+rVnlvvN51F9x+OzRqFOcXIVLB+KU+Fnea7gTgTiAHb3jCCc65p83sb6VasYiUWEFR3Mce84qyzoCKxMVrka+4Uf2VZLVlyxYmTJhAKBTiiy++oEqVKvTs2ZNAIEDPnj2pXr16kY9Rvz785jfwy18WfwqviCREqetjcafpfmFm7Z1zg4Ef401K+tTMzivNk5aF3vWVymbZMm8qbjjsTRy87DJvPL6m4kplUJHe+QXoPmJGNeD0yMVv5gzvGdPgYEnrr2qkJNKBAweYPn06oVCIWbNmkZOTQ6dOnQgEAgwYMIBG5XA6s6ApvHk1sWrVcngRIknCL/WxuM3oUqA93nSkUWb2X+fcl2bWrrQLLi0VWqkMFMUV8VSkYtt9xIwuQAhYAzigJRCcM7znwlg9Z0nrr2qkxJuZ8dFHHxEKhZg8eTI7d+6kefPmh7djOeecc2LyvHkR3lGjjqSFFOGVysQv9bG451X+HXnwL4GFzrmTgYR8ZlTEzxTFFanQngCumjO85zcA3UfMOB2YCHSM4XOq/kqF9P333zN27FjC4TDffvstNWvW5LrrriMYDPLTn/6U1NTYDn1WhFekQil1fSzWNF0ze8bMWpjZ1eb5HuhalhWLyBHLlnnv6OZNxU1LOzIV94EH1IiKVBBV8wotwJzhPVcAMQ0Gqv5KRbJ7925efPFFunbtSuvWrRk+fDjNmzdnzJgxbN68mXHjxnHllVfGvBGNlpoKvXvDvHkcM4W3c2evlmZnx205IpVVqetjcQcY1QP+BHSOXPVf4GFgV8nWKSJ5FMUVSTqLu4+Y8TwwLnJ5MLA4lk+o+iuJlpOTw9tvv00oFOK1115j//79/OhHP2LEiBEMGTKE1q1bJ3qJh519Njz3nJcoyovw3nij9wbv0KFKGYnEUKnrY3H3GR0D7AZujHz9ALxYwkWKCF4U94kn4LTToE8f+OYbr3BmZMALL6gRFanA7gKWAcMiX8si18WS6q8kxLJly7j//vs5+eSTueqqq5gxYwaBQIAPPviAFStW8Ic//KFCNaLR8iK8334L06bBmWfCgw96Tektt8Dnnyd6hSK+U+r6WNzPjJ5qZtdHXf5/zrkvSrREkUquoKm4f/2rpuKKJJEqwNNzhvd8EqD7iBmpQNH7U5SN6q/EzbZt25g4cSLhcJjFixeTmprKz372M5566il69+5NjRo1Er3EEsmL8PbuffQU3hdf9GrwPffAtdeqBouUg1LXx+KeGd3vnLs074Jz7hJgf0lXKVLZ5OR478pecQWcc45XAAcM8N6VXbgQbrhBRVAkibwF1Iy6XBOYH+PnVP2VmDp06BBTp06lb9++NGvWjGHDhpGdnc1TTz3F+vXrefPNN+nXr1/SNaL55UV416/30kkZGV6Et00b+POfYdu2RK9QJKmVuj4Wtxm9E3jWObfGObcGGAX8vCQrFKlMFMUV8aUac4b33JN3IfJ9rRg/p+qvlDsz45NPPuEXv/gFzZo147rrruPjjz/ml7/8JV9++SWff/45v/rVr2jSpEmil1ruFOEViYlS18fiTtPN29PsPOA8M+sA/LQ0KxXxM03FFfG1vd1HzPhx3oXuI2Z0IsZnKVV/pTytW7eOP//5z5x99tlceOGFvPDCC1x11VXMnDmTdevW8fe//53zzjsv0cuMi/xTeG++WVN4Rcqg1PXRmVmpntE5t9bMWpXqzmWgDb2lotFUXJHYqWCbep8PTAI2RK5qBvSfM7znp/FcR2H1VzVS8tuzZw9Tp04lFArx9ttvY2ZceumlBINB+vXrR7169RK9xAojM9P7OM2oUUf2+tYUXqmo/FIfixvTLYgrw31Fkt7OnfDkk4riivhd9xEzzu8+YkbTOcN7LgLOBCYDWcBsYHUClqT6K4XKzc3l7bff5qabbqJp06YEAgFWrVrFH//4R1auXMm7777LbbfdpkY0nwYNjkR433jj2AjvFxodJnKU8qiPZWlGS3dKVSTJ5UVxW7SA3/5WUVyRSuDfwKHI9xcBDwLPApnA6ASsR/VXCvTNN9/w+9//ntatW9OtWzemTp3KwIEDWbhwId999x0PPfQQp556aqKXWeGlpsI11xwb4e3QwYvwvvKKIrwiEWWuj4XO8XTO7abgouc4emKSiK/lRXFHjoT58xXFFalkUucM77kj8n1/YPSc4T1fBV7tPmJGTM6VqP5Kce3YsYPJkycTCoX4+OOPSUlJ4aqrruKvf/0rffr0oWZN/XUpi7PPhn/+00s+5UV4+/VThFckosz1sdAzo2ZW18xOKOCrrplpQwrxvfxR3K+/VhRXpBJK7T5iRl7N6wa8HXVbTGqh6q8UJisri2nTpnHDDTfQrFkzhg4dyt69e/nb3/5GRkYGs2bNYsCAAWpEy5EivCIFKnN9VEETKcCyZd5Z0HAY9u3zNsf+y1+gb1+oWjXRqxOROJsI/Lf7iBnb8KYDvgvQfcSMHwG7ErkwqTzMjM8//5xQKMSECRPYtm0bjRo1YujQoQQCAdq3b49z+jhxrOVFeK+5xovwjhrl/a7w4ove7wrDhnm/K2gPcakkylwfy/KZURFfycmBN9+EK6+Ec87xCkv//vDZZ7BwoRfLUSMqUvnMGd7zUeC3wEvApXOG98yLz6YA9yRqXVI5bNiwgb/97W+ce+65dOzYkX/961906dKFN998k/Xr1/PUU0/RoUMHNaIJcM45XoQ3I8PbWzwjw/tdoU0b+POfYdu2RK9QJLbKoz6WemuXRNHYeilvO3fCmDHHjnK/7TZo1CjRqxOp3CrS6PpkoBrpD/v27eP1118nHA4zb948cnNzueiiiwgEAvTv358GDRokeolSAM2XkHjyS32MaTPqnOsBPA2kAs+b2eP5bm8FhID6kWPuN7OZhT2mCq2Ul4KiuPfcoyiuSEVSVLEtRp15CugauVgLaGxm9SO3/RXoifcO7jzgl5Zs79DmoxqZvHJzc3nvvfcIhUK8/PLL7N69m1atWhEIBEhPT+f0009P9BKlBKIjvHm/YyjCK+WpOM1oUTUycsyNwEN4Q/O+NLNBketzgK8ih601s2vKcflHnj9Wddc5lwqsAK4EMoBFwEAzWxZ1zGjgczP7p3PubGCmmbUu7HFVaKUscnJg5kx45pkj71oOGuQ1oR06JHp1IpJfYcW2OHUm3/H3AB3M7Bbn3MXA34DOkZvfAx4ws3fK+SXElWpk8lm5ciVjx44lHA6zZs0a6tSpQ79+/QgEAnTu3JmUFH2iKpllZh6ZwhudvtIUXimrYrxZW5xe7DRgCvBTM8t0zjU2sy2R2/aYWZ2Yvghi+5nRC4CVZrbKzA4Bk4A++Y4x4ITI9/WADTFcj1Ri0VNxr7kGli/3puKuW+dFdNWIiiSl4tSZaAPxhi2AV39qANWA6kBVYHMM1ypy2M6dOxk9ejSXXHIJp512GiNGjOC0005j3LhxbNq0iTFjxtClSxc1oj5Q2BTeW2/VFF6JqeLUyNuBZ80sEyCvEY2nWP4r1wJYF3U5I3JdtIeAIc65DGAmGgQh5WzZMrjrLmjRAn77W+8f/ylTvHcnH3hAnwkVSXLFqTMAOOdOBtoQGTtvZh8CC4CNka85ZrY8pquVSi07O5uZM2fSv39/mjZtys9//nMyMzN5/PHHWbt2LXPnzmXw4MHUrp30HwGTAuRN4Z03D5YsgZtvhkmTvDfDO3eGV16B7OxEr1KSTBXn3OKorzvy3V6cGnk6cLpz7n3n3EeRWG+eGpHH/cg51zcG6wcSv7XLQOAlM3vCOXcRMNY519bMcqMPivxw7wCoVq1aApYpyURRXBFfqeKcWxx1ebSZjS7F4wwAXjGzHADn3I+As4C0yO3znHOXmdm7ZVtu/KlGVmxffvnl4e1YNm/ezIknnsgdd9xBIBCgY8eOmoJbCeVN4X3sMS+d9eyz3hReRXilhLLNrFMZH6MKcBrQBa8eLnTOnWtmO4GTzWy9c+4U4G3n3Fdm9l0Zn+8YsTwzuh5oGXU5LXJdtFvxcsp571LXAI7538/MRptZJzPrVEWf+pbjUBRXxJey8/79j3xFN6LFqTN5BnAkogtwLfCRme0xsz3ALOCi8lx4vKhGVjybNm3iySefpH379rRv355Ro0Zx8cUX8/rrr7NhwwaeeeYZOnXqpEa0kmvQwEtt5UV4zzhDEV4pV8WpkRnANDPLMrPVeJ8xPQ3AzNZH/lwFvAPE5DfpWDaji4DTnHNtnHPV8H4RmJbvmLVANwDn3Fl4zejWGK5JfGjZMu+dxLwobosWiuKKVBLFqTM4584EGgAfRl29FrjcOVfFOVcVuBxQTFdK7cCBA0yZMoWePXuSlpbGb3/7W6pXr86oUaPYuHEjr732Gn369NHZazlGXoR3/nxFeKVcFadGvo53VhTn3El4sd1VzrkGzrnqUddfAhQ4HLCsYr21y9XAP/DGCY8xs0edcw8Di81sWmSC7n+AOnjDJO41s7mFPaYmBQooiitSWRRjWmChdSZyzENADTO7P+p+qcBzeNN0DZhtZr+J3SuJD9XI+DIzPvjgA8LhMJMnT2bXrl2kpaWRnp5Oeno6Z511VqKXKEkqM/NIhHf1amjZ0puBoQiv5Cnm1i5F9WIOeALoAeQAj5rZpMjE+X8DuXgnL/9hZi/E5HUk25ZqKrSV286d3j/OeSPSW7SAu++G227TGVARP/LLpt7xohoZH6tXrz68Hct3331HrVq1uP766wkGg3Tp0oXU1NREL1F8IicHZszw3nx/6y3vzffBg70339u3T/TqJJH8Uh/VjEpSWLbMa0BDIW/z6EsvPbJ5dNWqiV6diMSKX4ptvKhGxs4PP/zAyy+/TDgcZuHChTjn6Nq1K4FAgOuvv546dWK+HZ9UckuXer8LhcPe70KXXXbkdyF9XLzy8Ut9VDMqFZaiuCLil2IbL6qR5SsnJ4f58+cTCoWYOnUqBw4c4PTTTycYDDJkyBBatWqV6CVKJVRQhHfoUC8lpghv5eGX+qhmVCqcnTvhxRe9d/9WrfKiuHmjzhXFFalc/FJs40U1snwsWbKEcDjMuHHj2LhxIw0aNGDAgAEEg0EuuOACTcGVCiF/hLdGjSNv2ivC639+qY9qRqXCyIvihsOwd6+iuCLin2IbL6qRpbd161YmTJhAOBzms88+o0qVKlx99dUEAgF69epF9erVE71EkePKH+Ht3NlrShXh9S+/1Ec1o5JQiuKKSGH8UmzjRTWyZA4ePMj06dMJhULMmjWL7OxsOnbsSCAQYMCAATRu3DjRSxQpkbwI76hRsGaNIrx+5pf6qGZUEkJRXBEpDr8U23hRjSyamfHxxx8TDoeZNGkSmZmZNGvWjCFDhhAIBGjbtm2ilyhSZorw+p9f6qOaUYmr5cth5EhFcUWkePxSbONFNfL41q5de3g7lhUrVlCzZk2uvfZaAoEAV1xxhbZjEd8qKMI7bBj06aMIbzLzS31UMyoxlxfFHTkS5s1TFFdEis8vxTZeVCOPtmfPHl599VVCoRDvvPMOZkbnzp0JBoPccMMNnHDCCYleokjcKMLrL36pj5WuGV2xYgUPP/wwrVq1omXLlkd9NWjQQBPyypGiuCJSVn4ptvGiZtTbjmXBggWEw2FeffVV9u3bx6mnnkogECA9PZ02bdokeokiCZWTA9OneycJFOFNXn6pj5WuGX3vvfdIT08nIyOD7Ozso26rVavWMQ1q/q+6deuW9SX4nqK4IlJe/FJs46UyN6Nff/01oVCIcePGkZGRQb169ejfvz+BQICLL75YbzaLFGDJEu+kwdixivAmG7/Ux0rXjObJyclh8+bNrFu37rhfGzduJP/Pp379+oU2q2lpadSoUaPM60s2BUVxBw703mX78Y8TvToRSVZ+KbbxUtma0e3btzNp0iRCoRCLFi0iNTWVHj16EAgE6N27NzVr1kz0EkWSgiK8yccv9bHSNqPFkZWVxYYNG47brK5du5Zt27Ydc79GjRodt1lt1aoVzZs3p4pP3m5SFFdEYskvxTZeKkMzeujQIWbOnEk4HGb69OlkZWXRrl07gsEgAwcOpGnTpoleokjSUoQ3efilPqoZLaP9+/eTkZFR6BnWXbt2HXWflJQUmjVrVugZ1iZNmpCSkpKgV1U0RXFFJB78UmzjpaLVyPJiZnz66aeEQiEmTpzI9u3badKkCYMHDyYQCNCuXbtEL1HEd/IivOEw7N+vCG9F45f6qGY0Dnbv3n3MGdX8Dev+/fuPuk/VqlVp0aJFgWdW875v2LBhXD8DoyiuiMSbX4ptvCRjjSxMRkYG48ePJxQKsXz5cqpXr07fvn0JBAJcddVVvkkZiVRkivBWTH6pj2pGKwAzY8eOHYWeXc3IyCArK+uo+9WsWbPIgUvlMbZeUVwRSRS/FNt48UON3Lt3L1OnTiUcDjN//nzMjEsuuYRgMEi/fv2oX79+opcoUikVFOEdPNg7KaFwQvz5pT6qGU0Subm5bNmypcCzqtEDl3Jzc4+63wknnHDMGdX8A5eON+BBUVwRSTS/FNt4SdYamZuby3//+1/C4TCvvPIKe/bsoXXr1oe3Y/nRj36U6CWKSBRFeBPPL/VRzaiPZGVlsXHjxkLPsG7ZsuWY+5100kmHm9MWLVqxd+/lfPbZJSxZ0oxq1YwBA4xf/jJFUVwRiTu/FNt4SbYauWLFCsLhMGPHjmXt2rXUrVuXG2+8kUAgwKWXXlqhZyeIiCK8ieSX+qhmtJI5cOBAgQOXVq3awZIlndi8uR+5uW2ADOA54D84t73IgUtNmzbVLw0iUu78UmzjJRlqZGZmJpMnTyYcDvPhhx+SkpLClVdeSTAYpE+fPtSqVSvRSxSRElKEN/78Uh/VjFZyy5d772aFQkeiuLfffoD27VezadPxhy7t27fvqMepUqVKgQOXogcvnXjiidp0XERKxC/FNl4qao3Myspi9uzZhMNhpk2bxqFDhzjnnHMIBoMMHjyY5s2bJ3qJIlJOFOGND7/URzWjlVBODsyaBc88U7qpuGZGZmZmkQOXDh06dNT9atSoUeTApXr16sXoVYtIMvJLsY2XilQjzYwvvviCcDjMhAkT2LJlC40aNWLQoEEEAgE6dOigNyhFfEwR3tjyS31UM1qJxHMqbm5uLlu3bi104NKGDRuOGbhUt27dQrezSUtLU4RLpBLxS7GNl4pQIzdu3Hh4O5YlS5ZQrVo1evfuTTAYpEePHlTVBDyRSiUvwvvMM/D224rwlhe/1Ec1o5VA/ijuJZd4cYlrr03sVNzs7OwiBy5t3rz5mPudeOKJhZ5dbdGiBdWqVUvAKxKR8uaXYhsviaqR+/fv5/XXXyccDjN37lxyc3P5yU9+QiAQoH///jRs2DDuaxKRikcR3vLjl/qoZtSncnNh5szSR3ErioMHDxY4cCn6KzMz86j7OOdo0qRJgWdWowcupaamJuhViUhx+aXYxks8a6SZ8d577xEKhXj55Zf54YcfaNWqFenp6aSnp3PGGWfEZR0iknx27PAivM8+eyTCe/fdXoT3xBMTvbrk4Jf6qGbUZ/KiuM8+C999F9sobkWxZ8+eYxrW/PHg/H9nqlSpQvPmzQs9w9qoUSN9nkkkwfxSbOMlHjXyu+++Y+zYsYTDYVavXk3t2rXp168fgUCAyy+/XJPVRaTYFOEtveLUR+dcD+BpIBV43sweL+CYG4GHAAO+NLNBkeuDwB8ihz1iZqFyXP6R549lM1qWH8DxqBktWEWN4lYEZsbOnTsLPbu6bt26YwYuVa9evVgDl9SwisSOmtGSiVWN3LVrF1OmTCEcDvPee+/hnKNbt24EAgGuu+46atfWfyIRKZv8Ed7LL/eaUkV4C1ZUfXTOpQIrgCvx9mxcBAw0s2VRx5wGTAF+amaZzrnGZrbFOdcQWAx0wuvRPgU6mllm/ucp8+uIVTNalh9AYY+rZvSI/FHcatVg0KDki+JWBGZWrIFLOTk5R92vTp06hW5n07JlSw1cEikDNaMlU541Mjs7m3nz5hEKhXjjjTc4cOAAZ5555uHtWFq2bFkuzyMiEi0vwjtqFHz/vSK8x1OMZvQi4CEz6x65/ACAmf056pi/AivM7Pl89x0IdDGzn0cu/xt4x8wmlvvriGEzWuofQGHUjBYcxb3rLi+K27hxolfnX9nZ2WzatOm4zeratWsLHLjUsGHDIgcuVa9ePQGvSKTiUzNaMuVRI//3v/8RDocZP348mzZtomHDhgwcOJBgMEinTp2UBhGRuFCEt3DOuUPAV1FXjTaz0VG33wD0MLPbIpfTgQvN7BdRx7yOd/LwErwk60NmNts59zughpk9EjluOLDfzP5e3q8jlie9WwDroi5nABfmO+Z0AOfc+0T9AGK4pqRWUBT3sccUxY2XKlWqkJaWRlpaGhdddFGBxxw8eJD169cft2H94IMP2LFjxzH3ix64VNDg/0EGzQAADw5JREFUpWbNmmngkojE3LXXXsvrr79O1apV6dmzJ4FAgJ49e2pCuYjEXWqqF9Ht0+foCO8LLyjCG5FtZp3K+BhVgNOALkAasNA5d25ZF1bSBSRSgT8AM9sZfZBz7g7gDqDSFURFcZNL9erVOeWUUzjllFOOe8zevXsLnBC8du1avv76a+bNm8eePXuOuk9qamqRA5caN26sMxYilVB51shevXrRrVs3BgwYwEnalV5EKoi2beFf//JOwuRFeG+4QRHeIqwHoj9PkRa5LloG8LGZZQGrnXMr8Hqz9Xj9WfR934nFIhMd0/0X3g/gxcjlt4D7zWzR8R63ssR0FcWtvMyMXbt2FTlw6eDBg0fdr1q1aqSlpRW6pU39+vXVsEpSUUy3ZCpLjRSRyk0R3mJ9ZrQKXgS3G15zuQgYZGZLo47pgTfTJ+icOwn4HGjPkaFFeae+PsMbYHRsvK+sryOGzWipfwBmtv14j+v3QqupuFIcZsa2bdsK3MYm72v9+vXHDFyqXbt2kROC69Spk6BXJXIsNaMl4/caKSKS35IlMHIkjB1buabwFnNrl6uBf+B9HHKMmT3qnHsYWGxm05x3huIJoAeQAzxqZpMi970FeDDyUI/mnTws99cR461dSv0DOB4/Ftq8KO7IkTB3rqK4Uj5ycnKKNXAp/78BDRo0KLRZTUtL08AliRs1oyXjxxopIlIclW0Kr1/qY0yb0VjwU6HdtcuL4o4apSiuJMahQ4cKHbi0bt06tm8/NqjQuHHjQre0adasGVX8/HakxE0xYkiF7mftnHsK6Bq5WAtobGb1I7e1Ap7H+0yNAVeb2ZpyfxFx5KcaKSJSGpUlwqtmNEH8UGgVxZVksm/fvuMOXMr7fvfu3UfdJyUlpVgDl1JSUhL0qiRZFFZsi7Ofdb7j7wE6mNktkcvv4CVy5jnn6gC5ZrYvBi8jbvxQI0VEyoufI7xqRhMkWQttQVHcgQO9/yE6dkz06kTKpjgDlw4cOHDUfapVq0aLFi0K3dKmQYMGGrhUyRXRjBY5KC/f8R8Af4o0n2fj7cl2aYyWnhDJWiNFRGLJjxFeNaMJkmyFNn8Ut3lzGDpUUVypXMyM7du3FzlwKTs7+6j71apVq8iBS3Xr1k3Qq5J4KGxT7+Js6B31OCcDHwFpZpbjnOsL3AYcAtoA8/Gmuefkv28ySbYaKSISTwVFeIcM8U4OnXdeoldXMmpGEyRZCq2iuCIlk5OTw+b/397dxshVlmEcv66+aLehXaGLpZTWxVhQBAXSmOoHUUHT1KQkYAQiKoRorFiMGkITP0jUL2g0ptIEIeXFV1ASSYlANVDFCEVQoFCCpmJDt620tS9L05a+7O2HcyrTPbM70+7uc86e+f+STc+cnTl9zt3ZvXrPeeaZ114b9urq1q1bCwsudXd3D/lRNkcXXJoyZUpJZ4WRanFl9Hia0ZuUNaJLGx67UtIFkl6VdJ+khyJi5dicSRrjJSMBoGzNpvDecIO0ePH4mMJLM1qSKgctU3GBsXXw4EFt2bJl2IZ1x44dhcedeuqpw15dPf300zWZV4kqabSm6dp+VtL1EfFEfnuBpFsi4qL89mclLYiI68fmTNKockYCQBWN1ym8NKMlqWLQMhUXqI79+/cXFlwaPDW4v7//mMdMmDBBs2bNGrZhnTlzJgsulaBFM9ry86zz+71b0iOSzow89PLFj/4u6ZKI2G77LmUfO7Zi7M5m7FUxIwFgPDhyRHrwwWwK75o11Z/CSzNakioF7csvZ1dBG6fiLl0qXXYZU3GBKuvv72+54NL+/fuPeczkyZObLrjUuPDSKaecwoJLo6yNj3YZ9vOs8/vcLGlKRCwb9NiPK/usa0v6m6QvRsTBsTmTNKqUkQAwXr3wQnaRqcpTeGlGS1J20DIVF6i/iNDOnTtbLrh06NChYx7X1dXVcsGl6dOnl3RW41NdwjaVsjMSAOpk8BTeuXOz2Y9VmMJbl3ykGW0TU3EBNBoYGGi54NKWLVsKCy5Nnz59yI+yObrgUldXV0lnVT11CdtUaEYBYPRVcQpvXfKRZrQFpuICOFGHDh1queDS9u3bC4/r6ekZ9urq7NmzO2bBpbqEbSo0owAwtqoyhbcu+Ugz2sTAgPTww9mrH0zFBTCWDhw4UFhwafD04D179hzzGNs67bTThv1Im5kzZ2rixIklndXoqUvYpkIzCgBp7NwprVwprVhRzhTeuuQjzWgDpuICqKLXX3+95YJL+/btO+YxkyZNGnbBpTlz5qinp6fyCy7VJWxToRkFgLTKmsJbl3ykGVU2FffWW6W772YqLoDxJyK0a9euIT/KZtOmTerr6yssuDRlypSWCy51d3eXdFaZuoRtKjSjAFCelFN465KPHduMMhUXQCcZGBjQtm3bWi64NDAwcMzjpk2bNuRH2RxdcGnq1KljNu66hG0qNKMAUL4UU3jrko8d14wyFRcAmjt8+HDLBZe2bdtWeNyMGTOaNqyLFy/WtGnTRjSmuoRtKjSjAFAdYzmFty752HHN6OrV0sKFTMUFgBNx4MABbd68edgFl3bv3i1J6uvr0+zZs0f099UlbFOhGQWAamqcwrtokXT//SM7Xl3yseOa0YEBad066fzzR3FQAID/27t3rzZt2qSzzz5bEyZMGNGx6hK2qdCMAkC17dwp9fdLvb0jO05d8rHjmlEAwPhRl7BNhYwEgM5Ql3wc2UvWAAAAAACcAJpRAAAAAEByNKMAAAAAgORoRgEAAAAAydGMAgAAAACSG3er6doekLR/hIeZJOnwKAynbqhLETUpoiZF1KRotGrSFRG8cNomMnLMUJMiatIcdSmiJkWjUZNa5OO4a0ZHg+1nImJ+2eOoGupSRE2KqEkRNSmiJuMX/3ZF1KSImjRHXYqoSRE1edO476YBAAAAAOMPzSgAAAAAILlObUZvL3sAFUVdiqhJETUpoiZF1GT84t+uiJoUUZPmqEsRNSmiJrmOfM8oAAAAAKBcnXplFAAAAABQolo3o7YX2v6H7Q22lzX5/ltt35d//ynbvelHmVYbNfm67Zdsr7P9qO13lDHO1FrVpeF+l9sO27VfAa2dmtj+dP58WW/7l6nHmFobPz9zba+x/Wz+M7SojHGmZPtO29tsvzjE9217eV6zdbYvTD1GNEdGFpGRReRjEflYRD4WkY9tiohafkmaKOlfkt4p6S2Snpd0zqD7fFnSbfn2lZLuK3vcFajJRyVNzbeX1L0m7dYlv980SY9LWitpftnjLrsmkuZJelbSyfntt5c97grU5HZJS/LtcyRtLHvcCeryYUkXSnpxiO8vkvSwJEtaIOmpssfMFxk5gpp0VEaSjyf8PCEfyUfysc2vOl8Z/YCkDRHxSkQclHSvpEsH3edSSffk2/dLuti2E44xtZY1iYg1EbEvv7lW0hmJx1iGdp4rkvQdSbdIOpBycCVppyZfkLQiInZJUkRsSzzG1NqpSUianm93S9qScHyliIjHJe0c5i6XSvppZNZKepvtWWlGh2GQkUVkZBH5WEQ+FpGPTZCP7alzMzpb0qaG2335vqb3iYjDkvZImpFkdOVopyaNrlP2ik3dtaxLPnViTkT8LuXAStTOc+UsSWfZ/ovttbYXJhtdOdqpyc2SrrbdJ+khSUvTDK3Sjvf3DtIgI4vIyCLysYh8LCIfTwz5KGlS2QNANdm+WtJ8SReVPZay2Z4g6YeSril5KFUzSdlUpI8ouzrwuO3zImJ3qaMq11WS7o6IH9j+oKSf2T43IgbKHhiA0UNGZsjHIZGPReQjmqrzldHNkuY03D4j39f0PrYnKZs28N8koytHOzWR7UskfVPS4oh4I9HYytSqLtMknSvpj7Y3KpvXv6rmizS081zpk7QqIg5FxL8l/VNZ+NZVOzW5TtKvJSkinpQ0RVJPktFVV1u/d5AcGVlERhaRj0XkYxH5eGLIR9W7GX1a0jzbZ9p+i7LFF1YNus8qSZ/Ptz8l6bHI31FcUy1rYvsCST9RFrJ1f4/DUcPWJSL2RERPRPRGRK+y9wktjohnyhluEu38/Dyg7FVf2e5RNi3plZSDTKydmrwq6WJJsv0eZWG7Pekoq2eVpM/lqwYukLQnIraWPSiQkU2QkUXkYxH5WEQ+nhjyUTWephsRh21/RdJqZat83RkR621/W9IzEbFK0kpl0wQ2KHuD8ZXljXjstVmT70s6SdJv8nUqXo2IxaUNOoE269JR2qzJakmfsP2SpCOSboyI2l41abMm35B0h+2vKVus4Zqa/+ddtn+l7D9dPfl7gb4labIkRcRtyt4btEjSBkn7JF1bzkjRiIwsIiOLyMci8rGIfGyOfGyPa/48AAAAAABUUJ2n6QIAAAAAKopmFAAAAACQHM0oAAAAACA5mlEAAAAAQHI0owAAAACA5GhGgQRsH7H9XMPXslE8dq/tF0freAAApEI+Ap2ttp8zClTM/og4v+xBAABQMeQj0MG4MgqUyPZG29+z/YLtv9p+V76/1/ZjttfZftT23Hz/TNu/tf18/vWh/FATbd9he73t39vuyu9/g+2X8uPcW9JpAgBwXMhHoDPQjAJpdA2ahnRFw/f2RMR5km6V9KN8348l3RMR75P0C0nL8/3LJf0pIt4v6UJJ6/P98yStiIj3Stot6fJ8/zJJF+TH+dJYnRwAACeIfAQ6mCOi7DEAtWd7b0Sc1GT/Rkkfi4hXbE+W9J+ImGF7h6RZEXEo3781Inpsb5d0RkS80XCMXkl/iIh5+e2bJE2OiO/afkTSXkkPSHogIvaO8akCANA28hHobFwZBcoXQ2wfjzcato/ozfeDf1LSCmWvEj9tm/eJAwDGC/IRqDmaUaB8VzT8+WS+/YSkK/Ptz0j6c779qKQlkmR7ou3uoQ5qe4KkORGxRtJNkrolFV59BgCgoshHoOZ4FQhIo8v2cw23H4mIo8vXn2x7nbJXb6/K9y2VdJftGyVtl3Rtvv+rkm63fZ2yV3iXSNo6xN85UdLP80C2pOURsXvUzggAgJEjH4EOxntGgRLl74mZHxE7yh4LAABVQT4CnYFpugAAAACA5LgyCgAAAABIjiujAAAAAIDkaEYBAAAAAMnRjAIAAAAAkqMZBQAAAAAkRzMKAAAAAEiOZhQAAAAAkNz/AGnCYPLMWZCrAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 1080x216 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:The parameters `output_attentions`, `output_hidden_states` and `use_cache` cannot be updated when calling a model.They have to be set to True/False in the config object (i.e.: `config=XConfig.from_pretrained('name', output_attentions=True)`).\n",
            "WARNING:tensorflow:The parameter `return_dict` cannot be set in graph mode and will always be set to `True`.\n",
            "\n",
            "--------------------------------------------------------------------\n",
            "Iteração: 1 Fold: 1 Accuracy: 0.5356371490280778, Mean Accuracy: 0.5356371490280778\n",
            "\n",
            "--------------------------------------------------------------------\n",
            "WARNING:tensorflow:The parameters `output_attentions`, `output_hidden_states` and `use_cache` cannot be updated when calling a model.They have to be set to True/False in the config object (i.e.: `config=XConfig.from_pretrained('name', output_attentions=True)`).\n",
            "WARNING:tensorflow:The parameter `return_dict` cannot be set in graph mode and will always be set to `True`.\n",
            "Epoch 1/2\n",
            "WARNING:tensorflow:The parameters `output_attentions`, `output_hidden_states` and `use_cache` cannot be updated when calling a model.They have to be set to True/False in the config object (i.e.: `config=XConfig.from_pretrained('name', output_attentions=True)`).\n",
            "WARNING:tensorflow:The parameter `return_dict` cannot be set in graph mode and will always be set to `True`.\n",
            "WARNING:tensorflow:The parameters `output_attentions`, `output_hidden_states` and `use_cache` cannot be updated when calling a model.They have to be set to True/False in the config object (i.e.: `config=XConfig.from_pretrained('name', output_attentions=True)`).\n",
            "WARNING:tensorflow:The parameter `return_dict` cannot be set in graph mode and will always be set to `True`.\n",
            "186/186 [==============================] - ETA: 0s - loss: 0.9606 - accuracy: 0.6688WARNING:tensorflow:The parameters `output_attentions`, `output_hidden_states` and `use_cache` cannot be updated when calling a model.They have to be set to True/False in the config object (i.e.: `config=XConfig.from_pretrained('name', output_attentions=True)`).\n",
            "WARNING:tensorflow:The parameter `return_dict` cannot be set in graph mode and will always be set to `True`.\n",
            "186/186 [==============================] - 74s 315ms/step - loss: 0.9594 - accuracy: 0.6692 - val_loss: 0.5882 - val_accuracy: 0.7790\n",
            "Epoch 2/2\n",
            "186/186 [==============================] - 56s 300ms/step - loss: 0.4438 - accuracy: 0.8368 - val_loss: 0.3655 - val_accuracy: 0.8679\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x216 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:The parameters `output_attentions`, `output_hidden_states` and `use_cache` cannot be updated when calling a model.They have to be set to True/False in the config object (i.e.: `config=XConfig.from_pretrained('name', output_attentions=True)`).\n",
            "WARNING:tensorflow:The parameter `return_dict` cannot be set in graph mode and will always be set to `True`.\n",
            "\n",
            "--------------------------------------------------------------------\n",
            "Iteração: 1 Fold: 2 Accuracy: 0.8790496760259179, Mean Accuracy: 0.7073434125269978\n",
            "\n",
            "--------------------------------------------------------------------\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 778
        },
        "id": "0MjS-OGPNvmw",
        "scrolled": false,
        "outputId": "c5fcaf56-286d-4586-97d5-876035bbbfbe"
      },
      "source": [
        "uf.custom_metrics(mean_of_conf_matrix_arrays, np.unique(y_test))"
      ],
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "                                          Metrics Resume                                          \n",
            "==================================================================================================\n",
            "Class Name          \t Precision\t    Recall\t  F1-Score\t       MCC\t       AUC\n",
            "Contributors' Guide \t      0.62\t      0.35\t      0.44\t      0.42\t      0.66\n",
            "Design Document     \t      0.75\t      0.43\t      0.55\t      0.55\t      0.71\n",
            "License             \t      0.91\t       1.0\t      0.95\t      0.95\t      0.99\n",
            "List of Contributors\t      0.98\t       0.9\t      0.94\t      0.92\t      0.95\n",
            "Release Notes       \t      0.96\t      0.47\t      0.63\t      0.62\t      0.73\n",
            "Requirement Document\t      0.36\t      0.81\t       0.5\t      0.41\t      0.77\n",
            "Setup Files         \t      0.79\t      0.79\t      0.79\t      0.78\t      0.89\n",
            "\n",
            "Macro Average       \t      0.77\t      0.68\t      0.68\t      0.66\t      0.81\n",
            "Micro Average       \t      0.71\t      0.71\t      0.71\t      0.78\t      0.83\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sQuD0pJjOdEe"
      },
      "source": [
        "Plot ROC Curve"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 485
        },
        "id": "c49U2D5_OP_r",
        "outputId": "19447950-cb7d-4763-edb8-ecb518a47130"
      },
      "source": [
        "classes = np.unique(y_test)\r\n",
        "y_test_array = pd.get_dummies(y_test, drop_first=False).values\r\n",
        "\r\n",
        "import matplotlib as mpl\r\n",
        "mpl.style.use('seaborn')\r\n",
        "\r\n",
        "fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(15,5))\r\n",
        "## Plot roc\r\n",
        "for i in range(len(classes)):\r\n",
        "    fpr, tpr, thresholds = metrics.roc_curve(y_test_array[:,i],  \r\n",
        "                           predicted_prob[:,i])\r\n",
        "    ax[0].plot(fpr, tpr, lw=3, \r\n",
        "              label='{0} (area={1:0.2f})'.format(classes[i], \r\n",
        "                              metrics.auc(fpr, tpr))\r\n",
        "               )\r\n",
        "ax[0].plot([0,1], [0,1], color='navy', lw=3, linestyle='--')\r\n",
        "ax[0].set(xlim=[-0.05,1.0], ylim=[0.0,1.05], \r\n",
        "          xlabel='False Positive Rate', \r\n",
        "          ylabel=\"True Positive Rate (Recall)\", \r\n",
        "          title=\"Receiver operating characteristic\")\r\n",
        "ax[0].legend(loc=\"lower right\")\r\n",
        "ax[0].grid(True)\r\n",
        "    \r\n",
        "## Plot precision-recall curve\r\n",
        "for i in range(len(classes)):\r\n",
        "    precision, recall, thresholds = metrics.precision_recall_curve(\r\n",
        "                 y_test_array[:,i], predicted_prob[:,i])\r\n",
        "    ax[1].plot(recall, precision, lw=3, \r\n",
        "               label='{0} (area={1:0.2f})'.format(classes[i], \r\n",
        "                                  metrics.auc(recall, precision))\r\n",
        "              )\r\n",
        "ax[1].set(xlim=[0.0,1.05], ylim=[0.0,1.05], xlabel='Recall', \r\n",
        "          ylabel=\"Precision\", title=\"Precision-Recall curve\")\r\n",
        "ax[1].legend(loc=\"best\")\r\n",
        "ax[1].grid(True)\r\n",
        "plt.show()"
      ],
      "execution_count": 78,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1500x500 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Pyo3QnyqNvmw"
      },
      "source": [
        "### BERT with Conventional ML"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Q_pRpjXb_lg5"
      },
      "source": [
        "def utils_bert_embedding(txt, tokenizer, nlp, log=False):\n",
        "    idx = tokenizer.encode(txt, add_special_tokens=True, max_length=256, pad_to_max_length=True, \n",
        "                                             return_attention_mask=True, return_token_type_ids=True)\n",
        "    print(idx)\n",
        "    if log is True:\n",
        "        print(\"tokens:\", tokenizer.convert_ids_to_tokens(idx))\n",
        "        print(\"ids   :\", tokenizer.encode(txt))\n",
        "    idx = np.array(idx)[None,:]\n",
        "    embedding = nlp(idx)\n",
        "    X = np.array(embedding[0][0][1:-1])\n",
        "    return X\n",
        "'''\n",
        "Creates a feature matrix (num_docs x vector_size)\n",
        ":parameter\n",
        "    :param x: string or list\n",
        "    :param tokenizer: transformers tokenizer\n",
        "    :param nlp: transformers bert\n",
        "    :param log: bool - print tokens\n",
        ":return\n",
        "    vector or matrix \n",
        "'''\n",
        "def embedding_bert(x, tokenizer=None, nlp=None, log=False):\n",
        "    tokenizer = transformers.BertTokenizer.from_pretrained('bert-base-uncased') if tokenizer is None else tokenizer\n",
        "    nlp = transformers.TFBertModel.from_pretrained('bert-base-uncased') if nlp is None else nlp\n",
        "    \n",
        "    ## single word --> vec (size,)\n",
        "    if (type(x) is str) and (len(x.split()) == 1):\n",
        "        X = utils_bert_embedding(x, tokenizer, nlp, log).reshape(-1)\n",
        "    \n",
        "    ## list of words --> matrix (n, size)\n",
        "    elif (type(x) is list) and (type(x[0]) is str) and (len(x[0].split()) == 1):\n",
        "        X = utils_bert_embedding(x, tokenizer, nlp, log)\n",
        "    \n",
        "    ## list of lists of words --> matrix (n mean vectors, size)\n",
        "    elif (type(x) is list) and (type(x[0]) is list):\n",
        "        lst_mean_vecs = [utils_bert_embedding(lst, tokenizer, nlp, log).mean(0) for lst in x]\n",
        "        X = np.array(lst_mean_vecs)\n",
        "    \n",
        "    ## single text --> matrix (n words, size)\n",
        "    elif (type(x) is str) and (len(x.split()) > 1):\n",
        "        X = utils_bert_embedding(x, tokenizer, nlp, log)\n",
        "        \n",
        "    ## list of texts --> matrix (n mean vectors, size)\n",
        "    else:\n",
        "        lst_mean_vecs = [utils_bert_embedding(txt, tokenizer, nlp, log).mean(0) for txt in x]\n",
        "        X = np.array(lst_mean_vecs)\n",
        "    return X"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qd4VJ6fs_lg5"
      },
      "source": [
        "# Use for mNB\n",
        "from sklearn.preprocessing import MinMaxScaler, normalize\n",
        "def normalizer(X):\n",
        "    scaler = MinMaxScaler()\n",
        "    scaled_X = scaler.fit_transform(X)\n",
        "    normalized_X = normalize(scaled_X, norm='max', axis=0, copy=True)\n",
        "    return(normalized_X)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sACmVeF1_lg5"
      },
      "source": [
        "#### Classifier"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5T_bDmGL_lg5"
      },
      "source": [
        "# Model Design\n",
        "clf1 = svm.SVC(gamma='scale')\n",
        "clf2 = RandomForestClassifier(n_estimators=500, random_state=1)\n",
        "clf3 = MultinomialNB(alpha=0.01)\n",
        "clf4 = DecisionTreeClassifier(random_state=0, max_depth=500)\n",
        "\n",
        "svc = SVC(probability=True, kernel='linear')\n",
        "ada = AdaBoostClassifier(n_estimators=50, base_estimator=svc,learning_rate=1)\n",
        "ada2 = AdaBoostClassifier(\n",
        "    DecisionTreeClassifier(max_depth=2),\n",
        "    n_estimators=600,\n",
        "    learning_rate=1.5,\n",
        "    algorithm=\"SAMME\")\n",
        "ada3 = AdaBoostClassifier()\n",
        "rf = RandomForestClassifier(max_depth=100, n_estimators=100, max_features=10)\n",
        "# nb = naive_bayes.MultinomialNB(alpha=0.01)\n",
        "nb = naive_bayes.MultinomialNB(alpha=0.0)\n",
        "lr = LogisticRegression(random_state=0, solver='lbfgs', multi_class='multinomial', class_weight='balanced', C=1500)\n",
        "dt = DecisionTreeClassifier(max_depth=500)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zvVOTUmp_lg6"
      },
      "source": [
        "# Disable bert warnings\n",
        "import logging\n",
        "logging.getLogger(\"transformers.tokenization_utils\").setLevel(logging.ERROR)\n",
        "from tqdm import tqdm\n",
        "accrs = []\n",
        "preds = []\n",
        "conf_matrix_list_of_arrays = []\n",
        "fold = 0\n",
        "k=10\n",
        "\n",
        "sss = StratifiedShuffleSplit(n_splits=k, test_size=0.2, random_state=42)\n",
        "\n",
        "for train_index, test_index in sss.split(X, y):\n",
        "    X_train, X_test = X[train_index], X[test_index]\n",
        "    y_train, y_test = y[train_index], y[test_index]\n",
        "\n",
        "    fold +=1\n",
        "\n",
        "    X_train = embedding_bert(X_train,tokenizer, transformer_model, log=False)\n",
        "    X_train = normalizer(X_train) # Use for mNB\n",
        "    X_test = embedding_bert(X_test,tokenizer, transformer_model, log=False)\n",
        "    X_test = normalizer(X_test) # Use for mNB\n",
        "    \n",
        "    dic_y_mapping = {n:label for n,label in enumerate(np.unique(y_train))}\n",
        "\n",
        "    # train classifier\n",
        "    model = nb.fit(X_train, y_train)\n",
        "    \n",
        "    # test\n",
        "    predicted_prob = model.predict_proba(X_test)\n",
        "    predicted = [dic_y_mapping[np.argmax(pred)] for pred in predicted_prob]\n",
        "    \n",
        "    # Validade Test Set\n",
        "    accr = metrics.accuracy_score(y_test, predicted)\n",
        "\n",
        "    conf_matrix = confusion_matrix(y_test, predicted)\n",
        "    conf_matrix_list_of_arrays.append(conf_matrix)\n",
        "\n",
        "    accrs.append(accr)\n",
        "    print('\\n--------------------------------------------------------------------')\n",
        "    print(f'Fold: {fold} Accuracy: {accr}, Mean Accuracy: {sum(accrs)/fold}')\n",
        "    print('--------------------------------------------------------------------')\n",
        "    \n",
        "\n",
        "# means for Confusion Matrix\n",
        "mean_of_conf_matrix_arrays = np.mean(conf_matrix_list_of_arrays, axis=0)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Q_82BkAz_lg6"
      },
      "source": [
        "### Plot Metrics Resume and Confusion Matrix"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DkW-CNBA_lg6"
      },
      "source": [
        "uf.custom_metrics(mean_of_conf_matrix_arrays, np.unique(y_test))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pocKvbyC_lg6"
      },
      "source": [
        "### Voting Classifier"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hQRgBz3t_lg6"
      },
      "source": [
        "accrs = []\n",
        "conf_matrix_list_of_arrays = []\n",
        "kfolds=10\n",
        "fold = 0\n",
        "sss = StratifiedShuffleSplit(n_splits=kfolds, test_size=0.2, random_state=42)\n",
        "\n",
        "for train_index, test_index in sss.split(X, y):\n",
        "    X_train, X_test = X[train_index], X[test_index]\n",
        "    y_train, y_test = y[train_index], y[test_index]\n",
        "    \n",
        "    fold +=1\n",
        "\n",
        "    X_train = embedding_bert(X_train,tokenizer, transformer_model, log=False)\n",
        "    X_train = normalizer(X_train) # Use for mNB\n",
        "    X_test = embedding_bert(X_test,tokenizer, transformer_model, log=False)\n",
        "    X_test = normalizer(X_test) # Use for mNB\n",
        "    \n",
        "    dic_y_mapping = {n:label for n,label in enumerate(np.unique(y_train))}\n",
        "    \n",
        "    model = VotingClassifier(estimators=[\n",
        "        ('svm', clf1),('rf', clf2), ('mnb', clf3), ('tree', clf4)], voting='hard',weights=[1,1,2,1])\n",
        "    model = model.fit(X_train, y_train)\n",
        "    \n",
        "    predicted = model.predict(X_test)\n",
        "    accr = metrics.accuracy_score(y_test, predicted)\n",
        "    #\n",
        "    conf_matrix = confusion_matrix(y_test, predicted)\n",
        "    conf_matrix_list_of_arrays .append(conf_matrix)\n",
        "    #    \n",
        "    accrs.append(accr)\n",
        "#     print(metrics.classification_report(y_test, predicted))\n",
        "\n",
        "    print(f'Fold: {fold} Accuracy: {accr}, Mean Accuracy: {sum(accrs)/fold}')\n",
        "\n",
        "# means for Confusion Matrix\n",
        "mean_of_conf_matrix_arrays = np.mean(conf_matrix_list_of_arrays, axis=0)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4gQeI_ap_lg8"
      },
      "source": [
        "### Plot Metrics Resume and Confusion Matrix"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MzInU-xm_lg8"
      },
      "source": [
        "uf.custom_metrics(mean_of_conf_matrix_arrays, np.unique(y_test))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "x8-LXDZY_lg9"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}