marathi_sentiment_indic_bert_bilstm_DL.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "marathi sentiment_indic_bert_bilstm_DL.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "T4-RKh__QMPQ",
        "gaDc0mN-QJA_",
        "6cPfl8pHQreT",
        "MiCJrhH8RObG",
        "1NkzS2D-f9XA",
        "xcYEBaOHgh1c",
        "S-Q6fKgFyYIV"
      ],
      "machine_shape": "hm",
      "authorship_tag": "ABX9TyMa4vi47AHokWJ7ojyUxkxA",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU",
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "205da120f68b40a9b6ef8c9a73910569": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_ff2320313d024d01a561e8ffccb9b616",
              "IPY_MODEL_650ddd441ea64aaa82f56892fb125b1b",
              "IPY_MODEL_edcd578d3ce4484284e1e5ecf3f09c17"
            ],
            "layout": "IPY_MODEL_1b5a4c041e2b494d98bbefa4597da650"
          }
        },
        "ff2320313d024d01a561e8ffccb9b616": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_b3d348e9c0c64dd996cfec2f998d1275",
            "placeholder": "​",
            "style": "IPY_MODEL_b759f5e72e124f30bea95236514dad19",
            "value": "Downloading: 100%"
          }
        },
        "650ddd441ea64aaa82f56892fb125b1b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_92af723e66324254b18eeacd8ce3a305",
            "max": 507,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_85df9309721c4b8ca644c7fe6e2e8eba",
            "value": 507
          }
        },
        "edcd578d3ce4484284e1e5ecf3f09c17": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_0aad3c7e591a4908975822111a23b857",
            "placeholder": "​",
            "style": "IPY_MODEL_a4fc3ed1925346aeaaa235090163ae73",
            "value": " 507/507 [00:00&lt;00:00, 16.1kB/s]"
          }
        },
        "1b5a4c041e2b494d98bbefa4597da650": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "b3d348e9c0c64dd996cfec2f998d1275": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "b759f5e72e124f30bea95236514dad19": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "92af723e66324254b18eeacd8ce3a305": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "85df9309721c4b8ca644c7fe6e2e8eba": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "0aad3c7e591a4908975822111a23b857": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "a4fc3ed1925346aeaaa235090163ae73": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "196bb0c2b73c428285d41dafd686b027": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_4141001d54b64235b10f3182d6fa81f0",
              "IPY_MODEL_7598b04f85684651bdb5c119695220c3",
              "IPY_MODEL_0fb72d9bb09b4a5791a18d8bb4437db2"
            ],
            "layout": "IPY_MODEL_0857c152da454d22a92c990510bc9ba5"
          }
        },
        "4141001d54b64235b10f3182d6fa81f0": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_e2c90d00749b4ed4a3b820edd77896d5",
            "placeholder": "​",
            "style": "IPY_MODEL_41a9a7d4d2af416da34d190d5efb3ff7",
            "value": "Downloading: 100%"
          }
        },
        "7598b04f85684651bdb5c119695220c3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_1254013e7f8344b4bc60e8c3c570c0e2",
            "max": 134982446,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_97b556cca93f41b880eb65de7ac5f6ac",
            "value": 134982446
          }
        },
        "0fb72d9bb09b4a5791a18d8bb4437db2": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_d7203eff7f0349858adb4791bd037671",
            "placeholder": "​",
            "style": "IPY_MODEL_7c4fe79867d545fabb599582f9dc15e2",
            "value": " 129M/129M [00:03&lt;00:00, 43.5MB/s]"
          }
        },
        "0857c152da454d22a92c990510bc9ba5": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "e2c90d00749b4ed4a3b820edd77896d5": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "41a9a7d4d2af416da34d190d5efb3ff7": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "1254013e7f8344b4bc60e8c3c570c0e2": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "97b556cca93f41b880eb65de7ac5f6ac": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "d7203eff7f0349858adb4791bd037671": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "7c4fe79867d545fabb599582f9dc15e2": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "17b58c8377fb4692946110f094c01f4e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_0bc7ac8fa2e84935be5dae2dbc50e04a",
              "IPY_MODEL_1f21a2cb72324c1daed3dee4834346c3",
              "IPY_MODEL_7f83447b215b410c84edf9c5c84c0daf"
            ],
            "layout": "IPY_MODEL_1fe9370d7bfe43c495e3c29b97f43618"
          }
        },
        "0bc7ac8fa2e84935be5dae2dbc50e04a": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_1eef7d8fd36f4aa4946528c86b9fc4d6",
            "placeholder": "​",
            "style": "IPY_MODEL_35947046566947bcaea56b10f4b99338",
            "value": "Downloading: 100%"
          }
        },
        "1f21a2cb72324c1daed3dee4834346c3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_288df8c356a44461a50de71a9fdbd155",
            "max": 5646064,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_990781bd6e4643789f7126f10d87a785",
            "value": 5646064
          }
        },
        "7f83447b215b410c84edf9c5c84c0daf": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_cd75f3286d004c5ca90c079e86adf0ef",
            "placeholder": "​",
            "style": "IPY_MODEL_2150b555e79a480cb746691fc6113345",
            "value": " 5.38M/5.38M [00:00&lt;00:00, 10.4MB/s]"
          }
        },
        "1fe9370d7bfe43c495e3c29b97f43618": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "1eef7d8fd36f4aa4946528c86b9fc4d6": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "35947046566947bcaea56b10f4b99338": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "288df8c356a44461a50de71a9fdbd155": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "990781bd6e4643789f7126f10d87a785": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "cd75f3286d004c5ca90c079e86adf0ef": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "2150b555e79a480cb746691fc6113345": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Daishinkan002/Sentiment-Analysis-of-Marathi-Tweets/blob/main/marathi_sentiment_indic_bert_bilstm_DL.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Import Libraries"
      ],
      "metadata": {
        "id": "T4-RKh__QMPQ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import os\n",
        "import re\n",
        "import string\n",
        "from tqdm import tqdm\n",
        "import pandas as pd\n",
        "from multiprocessing import Pool\n",
        "import numpy as np\n",
        "import pickle\n",
        "os.system('pip install -qq transformers')\n",
        "os.system('pip install -q -U watermark')\n",
        "import transformers\n",
        "from transformers import BertModel, BertTokenizer, AdamW, get_linear_schedule_with_warmup\n",
        "import torch\n",
        "from transformers import AutoModel, AutoTokenizer \n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "from pylab import rcParams\n",
        "import matplotlib.pyplot as plt\n",
        "from matplotlib import rc\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import confusion_matrix, classification_report\n",
        "from collections import defaultdict\n",
        "from textwrap import wrap\n",
        "\n",
        "from torch import nn, optim\n",
        "from torch.utils.data import Dataset, DataLoader\n",
        "import torch.nn.functional as F\n",
        "\n",
        "%matplotlib inline\n",
        "%config InlineBackend.figure_format='retina'\n",
        "\n",
        "sns.set(style='whitegrid', palette='muted', font_scale=1.2)\n",
        "\n",
        "HAPPY_COLORS_PALETTE = [\"#01BEFE\", \"#FFDD00\", \"#FF7D00\", \"#FF006D\", \"#ADFF02\", \"#8F00FF\"]\n",
        "\n",
        "sns.set_palette(sns.color_palette(HAPPY_COLORS_PALETTE))\n",
        "\n",
        "rcParams['figure.figsize'] = 12, 8\n",
        "\n",
        "RANDOM_SEED = 42\n",
        "np.random.seed(RANDOM_SEED)\n",
        "torch.manual_seed(RANDOM_SEED)\n",
        "\n",
        "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
        "device"
      ],
      "metadata": {
        "id": "aV6NTQHLP9A7"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "%reload_ext watermark\n",
        "%watermark -v -p numpy,pandas,torch,transformers"
      ],
      "metadata": {
        "id": "UlnjNs4UQGHI"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Dataset Loading"
      ],
      "metadata": {
        "id": "gaDc0mN-QJA_"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/gdrive')"
      ],
      "metadata": {
        "id": "nzLM9i3cQYqk"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_train = pd.read_pickle('/content/gdrive/MyDrive/Major_Project/Marathi-Sentiment/Refined_Dataset/marathi_train.pkl')\n",
        "df_test = pd.read_pickle('/content/gdrive/MyDrive/Major_Project/Marathi-Sentiment/Refined_Dataset/marathi_test.pkl')\n",
        "df_valid = pd.read_pickle('/content/gdrive/MyDrive/Major_Project/Marathi-Sentiment/Refined_Dataset/marathi_valid.pkl')"
      ],
      "metadata": {
        "id": "zUSPe5MAQb6N"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Data analysis"
      ],
      "metadata": {
        "id": "6cPfl8pHQreT"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "sns.countplot(df_train.label)\n",
        "plt.xlabel('Sentiment')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qxDwk1pgQ_j1",
        "outputId": "081064a9-e1c6-473c-d67d-3493b529a036"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
            "  FutureWarning\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 0, 'Sentiment')"
            ]
          },
          "metadata": {},
          "execution_count": 5
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "image/png": {
              "width": 410,
              "height": 272
            },
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "sns.countplot(df_test.label)\n",
        "plt.xlabel('Sentiment')"
      ],
      "metadata": {
        "id": "WNfwyyGiRCxK"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "sns.countplot(df_valid.label)\n",
        "plt.xlabel('Sentiment')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0wIRz7GKRDKf",
        "outputId": "a441b2c9-d0f1-43ed-b9ec-df9772320e50"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
            "  FutureWarning\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 0, 'Sentiment')"
            ]
          },
          "metadata": {},
          "execution_count": 7
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "image/png": {
              "width": 401,
              "height": 272
            },
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "pd.set_option(\"display.max_colwidth\", -1)\n",
        "df_train.head(10)"
      ],
      "metadata": {
        "id": "aSmHI1k8Z7Ji"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Preprocessing with transformers"
      ],
      "metadata": {
        "id": "MiCJrhH8RObG"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install sentencepiece"
      ],
      "metadata": {
        "id": "IK87JavGuMQd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# PRE_TRAINED_MODEL_NAME = 'google/muril-base-cased'\n",
        "PRE_TRAINED_MODEL_NAME = 'ai4bharat/indic-bert'"
      ],
      "metadata": {
        "id": "WGCtLbViSGC3"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "indic_bert_model = AutoModel.from_pretrained(PRE_TRAINED_MODEL_NAME)\n",
        "tokenizer = AutoTokenizer.from_pretrained(PRE_TRAINED_MODEL_NAME)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "referenced_widgets": [
            "205da120f68b40a9b6ef8c9a73910569",
            "ff2320313d024d01a561e8ffccb9b616",
            "650ddd441ea64aaa82f56892fb125b1b",
            "edcd578d3ce4484284e1e5ecf3f09c17",
            "1b5a4c041e2b494d98bbefa4597da650",
            "b3d348e9c0c64dd996cfec2f998d1275",
            "b759f5e72e124f30bea95236514dad19",
            "92af723e66324254b18eeacd8ce3a305",
            "85df9309721c4b8ca644c7fe6e2e8eba",
            "0aad3c7e591a4908975822111a23b857",
            "a4fc3ed1925346aeaaa235090163ae73",
            "196bb0c2b73c428285d41dafd686b027",
            "4141001d54b64235b10f3182d6fa81f0",
            "7598b04f85684651bdb5c119695220c3",
            "0fb72d9bb09b4a5791a18d8bb4437db2",
            "0857c152da454d22a92c990510bc9ba5",
            "e2c90d00749b4ed4a3b820edd77896d5",
            "41a9a7d4d2af416da34d190d5efb3ff7",
            "1254013e7f8344b4bc60e8c3c570c0e2",
            "97b556cca93f41b880eb65de7ac5f6ac",
            "d7203eff7f0349858adb4791bd037671",
            "7c4fe79867d545fabb599582f9dc15e2",
            "17b58c8377fb4692946110f094c01f4e",
            "0bc7ac8fa2e84935be5dae2dbc50e04a",
            "1f21a2cb72324c1daed3dee4834346c3",
            "7f83447b215b410c84edf9c5c84c0daf",
            "1fe9370d7bfe43c495e3c29b97f43618",
            "1eef7d8fd36f4aa4946528c86b9fc4d6",
            "35947046566947bcaea56b10f4b99338",
            "288df8c356a44461a50de71a9fdbd155",
            "990781bd6e4643789f7126f10d87a785",
            "cd75f3286d004c5ca90c079e86adf0ef",
            "2150b555e79a480cb746691fc6113345"
          ]
        },
        "id": "-1-0FchCSoAh",
        "outputId": "b3ad0084-a1f9-4525-89de-5bf692902ed4"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "Downloading:   0%|          | 0.00/507 [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "205da120f68b40a9b6ef8c9a73910569"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "Downloading:   0%|          | 0.00/129M [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "196bb0c2b73c428285d41dafd686b027"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Some weights of the model checkpoint at ai4bharat/indic-bert were not used when initializing AlbertModel: ['sop_classifier.classifier.bias', 'sop_classifier.classifier.weight', 'predictions.LayerNorm.bias', 'predictions.decoder.weight', 'predictions.dense.bias', 'predictions.LayerNorm.weight', 'predictions.dense.weight', 'predictions.bias', 'predictions.decoder.bias']\n",
            "- This IS expected if you are initializing AlbertModel 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 AlbertModel from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "Downloading:   0%|          | 0.00/5.38M [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "17b58c8377fb4692946110f094c01f4e"
            }
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "sample_txt = 'मेरी आणि सामंथा बस स्थानकावर लवकर पोहोचल्या पण दुपारपर्यंत बसची वाट पाहत बसले'"
      ],
      "metadata": {
        "id": "DSG3e5IOSqQS"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "\n",
        "tokens = tokenizer.tokenize(sample_txt)\n",
        "token_ids = tokenizer.convert_tokens_to_ids(tokens)\n",
        "\n",
        "print(f' Sentence: {sample_txt}')\n",
        "print(f'   Tokens: {tokens}')\n",
        "print(f'Token IDs: {token_ids}')\n",
        "\n"
      ],
      "metadata": {
        "id": "v-XPV5KJSrud"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(tokenizer.sep_token, tokenizer.sep_token_id)\n",
        "print(tokenizer.cls_token, tokenizer.cls_token_id)\n",
        "print(tokenizer.pad_token, tokenizer.pad_token_id)\n",
        "print(tokenizer.unk_token, tokenizer.unk_token_id)"
      ],
      "metadata": {
        "id": "bz7lIAYmStFW"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "\n",
        "encoding = tokenizer.encode_plus(\n",
        "  sample_txt,\n",
        "  max_length=32, # Max no. of tokens \n",
        "  add_special_tokens=True, # Add '[CLS]' and '[SEP]'\n",
        "  return_token_type_ids=True,\n",
        "  pad_to_max_length=True,\n",
        "  return_attention_mask=True,\n",
        "  return_tensors='pt',  # Return PyTorch tensors\n",
        ")\n",
        "\n",
        "encoding.keys()\n",
        "\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UZbOPMb8S0pe",
        "outputId": "de742603-4681-4b4d-cd0a-d57f100e9f16"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Truncation was not explicitly activated but `max_length` is provided a specific value, please use `truncation=True` to explicitly truncate examples to max length. Defaulting to 'longest_first' truncation strategy. If you encode pairs of sequences (GLUE-style) with the tokenizer you can select this strategy more precisely by providing a specific strategy to `truncation`.\n",
            "/usr/local/lib/python3.7/dist-packages/transformers/tokenization_utils_base.py:2277: FutureWarning: The `pad_to_max_length` argument is deprecated and will be removed in a future version, use `padding=True` or `padding='longest'` to pad to the longest sequence in the batch, or use `padding='max_length'` to pad to a max length. In this case, you can give a specific length with `max_length` (e.g. `max_length=45`) or leave max_length to None to pad to the maximal input size of the model (e.g. 512 for Bert).\n",
            "  FutureWarning,\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys(['input_ids', 'token_type_ids', 'attention_mask'])"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(len(encoding['input_ids'][0]))\n",
        "encoding['input_ids']"
      ],
      "metadata": {
        "id": "Qg7ggFc1S2NK"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(len(encoding['token_type_ids'][0]))\n",
        "encoding['token_type_ids']"
      ],
      "metadata": {
        "id": "ERedciK3S8a0"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(len(encoding['attention_mask'][0]))\n",
        "encoding['attention_mask']"
      ],
      "metadata": {
        "id": "71aO4e5uS-Um"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "tokenizer.convert_ids_to_tokens(encoding['input_ids'][0])"
      ],
      "metadata": {
        "id": "LIH_xNWhS_Jg"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "token_lens = []\n",
        "for txt in tqdm(df_train.tuned_tweet):\n",
        "  tokens = tokenizer.encode(txt, max_length=512)\n",
        "  token_lens.append(len(tokens))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "y6ezNEkWTA9W",
        "outputId": "02c5e021-e47f-49d3-e611-f8ea0f05027d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 12109/12109 [00:04<00:00, 2999.19it/s]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "sns.distplot(token_lens)\n",
        "plt.xlim([0, 256]);\n",
        "plt.xlabel('Token count');\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cb0h1WmTTmtZ",
        "outputId": "9e0c8e2d-d944-4b3d-cbc8-312f2e154369"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/seaborn/distributions.py:2619: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
            "  warnings.warn(msg, FutureWarning)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "image/png": {
              "width": 419,
              "height": 272
            },
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "MAX_LEN = 160"
      ],
      "metadata": {
        "id": "arGFeZvZT7ao"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "class GPReviewDataset(Dataset):\n",
        "\n",
        "  def __init__(self, text, targets, tokenizer, max_len):\n",
        "    self.text = text\n",
        "    self.targets = targets\n",
        "    self.tokenizer = tokenizer\n",
        "    self.max_len = max_len\n",
        "  \n",
        "  def __len__(self):\n",
        "    return len(self.text)\n",
        "  \n",
        "  def __getitem__(self, item):\n",
        "    text = str(self.text[item])\n",
        "    target = self.targets[item]\n",
        "\n",
        "    encoding = self.tokenizer.encode_plus(\n",
        "      text,\n",
        "      add_special_tokens=True,\n",
        "      max_length=self.max_len,\n",
        "      return_token_type_ids=True,\n",
        "      return_attention_mask=True,\n",
        "      padding= 'max_length',\n",
        "      # pad_to_max_length = True,\n",
        "      return_tensors='pt',\n",
        "    )\n",
        "\n",
        "    return {\n",
        "      'tweet_text': text,\n",
        "      'input_ids': encoding['input_ids'].flatten(),\n",
        "      'attention_mask': encoding['attention_mask'].flatten(),\n",
        "      'type_ids': encoding['token_type_ids'].flatten(),\n",
        "      'targets': torch.tensor(target, dtype=torch.long)\n",
        "    }"
      ],
      "metadata": {
        "id": "xy5-ExfZUDyj"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_train.shape, df_valid.shape, df_test.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_Ryi4gp5UFF2",
        "outputId": "514e29d2-f737-41a8-e81d-a8df44881adc"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((12109, 3), (1499, 3), (2250, 3))"
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def create_data_loader(df, tokenizer, max_len, batch_size):\n",
        "  ds = GPReviewDataset(\n",
        "    text=df.tuned_tweet.to_numpy(),\n",
        "    targets=df.label.to_numpy(),\n",
        "    tokenizer=tokenizer,\n",
        "    max_len=max_len\n",
        "  )\n",
        "\n",
        "  return DataLoader(\n",
        "    ds,\n",
        "    batch_size=batch_size,\n",
        "    num_workers=2\n",
        "  )"
      ],
      "metadata": {
        "id": "CfQBBL3xUIpx"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "BATCH_SIZE = 16\n",
        "\n",
        "train_data_loader = create_data_loader(df_train, tokenizer, MAX_LEN, BATCH_SIZE)\n",
        "val_data_loader = create_data_loader(df_valid, tokenizer, MAX_LEN, BATCH_SIZE)\n",
        "test_data_loader = create_data_loader(df_test, tokenizer, MAX_LEN, BATCH_SIZE)\n"
      ],
      "metadata": {
        "id": "4D70nKptUKto"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "data = next(iter(train_data_loader))\n",
        "data.keys()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nkt-MX8HUPzE",
        "outputId": "1e139c76-7ad1-42cd-a84a-17d16daa7854"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys(['tweet_text', 'input_ids', 'attention_mask', 'type_ids', 'targets'])"
            ]
          },
          "metadata": {},
          "execution_count": 27
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(data['input_ids'].shape)\n",
        "print(data['attention_mask'].shape)\n",
        "print(data['targets'].shape)"
      ],
      "metadata": {
        "id": "OodIlW5VUaJF"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Classification"
      ],
      "metadata": {
        "id": "FHUsLaOaUdko"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(encoding['attention_mask'] , encoding['token_type_ids'])"
      ],
      "metadata": {
        "id": "HAMUHya8Ugk0"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "x = indic_bert_model(\n",
        "  input_ids=encoding['input_ids'], \n",
        "  attention_mask=encoding['attention_mask'],\n",
        "  token_type_ids=encoding['token_type_ids']\n",
        ")"
      ],
      "metadata": {
        "id": "aQ9B9m3PUqaw"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(x['last_hidden_state'].shape, x['pooler_output'].shape)"
      ],
      "metadata": {
        "id": "smRnivzjUtlq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "indic_bert_model.config.hidden_size"
      ],
      "metadata": {
        "id": "ZGepiN8AUvP2"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "class SentimentClassifier(nn.Module):\n",
        "\n",
        "  def __init__(self, n_classes, base):\n",
        "    super(SentimentClassifier, self).__init__()\n",
        "    self.indic = base\n",
        "    self.drop1 = nn.Dropout(p=0.3)\n",
        "    self.lstm = nn.LSTM(base.config.hidden_size, 150,bidirectional=True,dropout=0.2)\n",
        "    self.out1 = nn.Linear(300,64)\n",
        "    self.relu = nn.ReLU()\n",
        "    self.drop2 = nn.Dropout(p=0.1) \n",
        "    self.out = nn.Linear(64, n_classes)\n",
        "  \n",
        "  def forward(self, input_ids, attention_mask, token_type_ids):\n",
        "    x = self.indic(\n",
        "      input_ids=input_ids,\n",
        "      attention_mask=attention_mask,\n",
        "      token_type_ids=token_type_ids\n",
        "    )\n",
        "    pooled_output = x['pooler_output']\n",
        "    output = self.drop1(pooled_output)\n",
        "    output = output.view(len(output), 1, -1)\n",
        "    output, _ = self.lstm(output)\n",
        "    output = output.view(len(output), output.shape[-1])\n",
        "    output =self.relu(self.out1(output))\n",
        "    output = self.drop2(output)\n",
        "    return self.out(output)"
      ],
      "metadata": {
        "id": "mtaUQJiKUwn7"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "class_names = ['-1','0', '1']\n",
        "model = SentimentClassifier(len(class_names), indic_bert_model)\n",
        "model = model.to(device)"
      ],
      "metadata": {
        "id": "r79z4AhLU2YC"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "data.keys()"
      ],
      "metadata": {
        "id": "DqHaT_ZHU6x1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "input_ids = data['input_ids'].to(device)\n",
        "attention_mask = data['attention_mask'].to(device)\n",
        "token_type_ids = data['type_ids'].to(device)\n",
        "\n",
        "print(input_ids.shape) # batch size x seq length\n",
        "print(attention_mask.shape) # batch size x seq length\n",
        "print(token_type_ids.shape) # batch size x seq_length"
      ],
      "metadata": {
        "id": "vwKNj4T5U8M2"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "data"
      ],
      "metadata": {
        "id": "ExubRkjtU9g4"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "input_ids"
      ],
      "metadata": {
        "id": "_lrayt93VASj"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model(input_ids, attention_mask, token_type_ids)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KEmieWKDVG7c",
        "outputId": "bd21cf05-8a2d-4f67-9fe0-74f4a2fee34f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[-0.0040, -0.1024, -0.1204],\n",
              "        [ 0.0009, -0.1082, -0.1124],\n",
              "        [-0.0142, -0.1053, -0.1194],\n",
              "        [-0.0206, -0.0996, -0.1291],\n",
              "        [-0.0030, -0.1070, -0.1201],\n",
              "        [-0.0142, -0.1014, -0.1211],\n",
              "        [-0.0062, -0.1035, -0.1159],\n",
              "        [-0.0067, -0.1026, -0.1130],\n",
              "        [-0.0087, -0.1041, -0.1220],\n",
              "        [-0.0138, -0.0945, -0.1211],\n",
              "        [-0.0026, -0.1104, -0.1166],\n",
              "        [-0.0135, -0.0989, -0.1203],\n",
              "        [-0.0151, -0.0988, -0.1185],\n",
              "        [-0.0096, -0.1003, -0.1040],\n",
              "        [-0.0193, -0.1013, -0.1219],\n",
              "        [-0.0121, -0.1132, -0.1156]], device='cuda:0',\n",
              "       grad_fn=<AddmmBackward0>)"
            ]
          },
          "metadata": {},
          "execution_count": 39
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "F.softmax(model(input_ids, attention_mask, token_type_ids), dim=1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "W5Fn4IAqVIP-",
        "outputId": "4905fd70-5180-4878-8f31-57e8ad0c5ba2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[0.3572, 0.3213, 0.3215],\n",
              "        [0.3540, 0.3254, 0.3206],\n",
              "        [0.3542, 0.3280, 0.3178],\n",
              "        [0.3559, 0.3242, 0.3199],\n",
              "        [0.3551, 0.3273, 0.3175],\n",
              "        [0.3546, 0.3283, 0.3171],\n",
              "        [0.3545, 0.3276, 0.3179],\n",
              "        [0.3563, 0.3260, 0.3177],\n",
              "        [0.3559, 0.3275, 0.3166],\n",
              "        [0.3587, 0.3233, 0.3179],\n",
              "        [0.3551, 0.3267, 0.3183],\n",
              "        [0.3558, 0.3248, 0.3195],\n",
              "        [0.3571, 0.3262, 0.3167],\n",
              "        [0.3569, 0.3266, 0.3165],\n",
              "        [0.3564, 0.3260, 0.3176],\n",
              "        [0.3527, 0.3288, 0.3185]], device='cuda:0', grad_fn=<SoftmaxBackward0>)"
            ]
          },
          "metadata": {},
          "execution_count": 40
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "EPOCHS = 10\n",
        "optimizer = AdamW(model.parameters(), lr=1e-5, correct_bias=False)\n",
        "total_steps = len(train_data_loader) * EPOCHS\n",
        "\n",
        "scheduler = get_linear_schedule_with_warmup(\n",
        "  optimizer,\n",
        "  num_warmup_steps=0,\n",
        "  num_training_steps=total_steps\n",
        ")\n",
        "\n",
        "loss_fn = nn.CrossEntropyLoss().to(device)"
      ],
      "metadata": {
        "id": "Svns7YsrVKCY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def train_epoch(model, data_loader, loss_fn, optimizer, device, scheduler, n_examples):\n",
        "  model = model.train()\n",
        "\n",
        "  losses = []\n",
        "  correct_predictions = 0\n",
        "  \n",
        "  for d in tqdm(data_loader):\n",
        "    input_ids = d[\"input_ids\"].to(device)\n",
        "    attention_mask = d[\"attention_mask\"].to(device)\n",
        "    token_type_ids = d['type_ids'].to(device)\n",
        "    targets = d[\"targets\"].to(device)\n",
        "    targets+=1\n",
        "\n",
        "    outputs = model(\n",
        "      input_ids=input_ids,\n",
        "      attention_mask=attention_mask,\n",
        "      token_type_ids=token_type_ids\n",
        "    )\n",
        "\n",
        "    _, preds = torch.max(outputs, dim=1)\n",
        "  \n",
        "    loss = loss_fn(outputs, targets)\n",
        "\n",
        "    correct_predictions += torch.sum(preds == targets)\n",
        "    losses.append(loss.item())\n",
        "\n",
        "    loss.backward()\n",
        "    nn.utils.clip_grad_norm_(model.parameters(), max_norm=1.0)\n",
        "    optimizer.step()\n",
        "    scheduler.step()\n",
        "    optimizer.zero_grad()\n",
        "\n",
        "  return correct_predictions.double() / n_examples, np.mean(losses)\n"
      ],
      "metadata": {
        "id": "laG5vgmIVMam"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def eval_model(model, data_loader, loss_fn, device, n_examples):\n",
        "  model = model.eval()\n",
        "\n",
        "  losses = []\n",
        "  correct_predictions = 0\n",
        "\n",
        "  with torch.no_grad():\n",
        "    for d in data_loader:\n",
        "      input_ids = d[\"input_ids\"].to(device)\n",
        "      attention_mask = d[\"attention_mask\"].to(device)\n",
        "      token_type_ids = d['type_ids'].to(device)\n",
        "      targets = d[\"targets\"].to(device)\n",
        "      targets+=1\n",
        "\n",
        "      outputs = model(\n",
        "        input_ids=input_ids,\n",
        "        attention_mask=attention_mask,\n",
        "        token_type_ids=token_type_ids\n",
        "\n",
        "      )\n",
        "      _, preds = torch.max(outputs, dim=1)\n",
        "\n",
        "      loss = loss_fn(outputs, targets)\n",
        "\n",
        "      correct_predictions += torch.sum(preds == targets)\n",
        "      losses.append(loss.item())\n",
        "\n",
        "  return correct_predictions.double() / n_examples, np.mean(losses)\n"
      ],
      "metadata": {
        "id": "TNBvhja9VO-H"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "history = defaultdict(list)\n",
        "best_accuracy = 0\n",
        "\n",
        "for epoch in range(EPOCHS):\n",
        "\n",
        "  print(f'Epoch {epoch + 1}/{EPOCHS}')\n",
        "  print('-' * 10)\n",
        "\n",
        "  train_acc, train_loss = train_epoch(\n",
        "    model,\n",
        "    train_data_loader,    \n",
        "    loss_fn, \n",
        "    optimizer, \n",
        "    device, \n",
        "    scheduler, \n",
        "    len(df_train)\n",
        "  )\n",
        "\n",
        "  print(f'\\n--> Train loss {train_loss} accuracy {train_acc}')\n",
        "\n",
        "  val_acc, val_loss = eval_model(\n",
        "    model,\n",
        "    val_data_loader,\n",
        "    loss_fn, \n",
        "    device, \n",
        "    len(df_valid)\n",
        "  )\n",
        "\n",
        "  print(f'--> Val   loss {val_loss} accuracy {val_acc}')\n",
        "  print()\n",
        "\n",
        "  history['train_acc'].append(train_acc)\n",
        "  history['train_loss'].append(train_loss)\n",
        "  history['val_acc'].append(val_acc)\n",
        "  history['val_loss'].append(val_loss)\n",
        "\n",
        "  if val_acc > best_accuracy:\n",
        "    torch.save(model.state_dict(), 'indic_bert_bilstm_model_state.bin')\n",
        "    best_accuracy = val_acc\n"
      ],
      "metadata": {
        "id": "VPjImpMiVRd8"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Results"
      ],
      "metadata": {
        "id": "1NkzS2D-f9XA"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "for i in range(len(history[\"train_acc\"])):\n",
        "  history[\"train_acc\"][i]=history[\"train_acc\"][i].cpu() "
      ],
      "metadata": {
        "id": "Up4xxKpHf_hq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for i in range(len(history[\"val_acc\"])):\n",
        "  history[\"val_acc\"][i]=history[\"val_acc\"][i].cpu()  "
      ],
      "metadata": {
        "id": "CNzDUH_XgDGe"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(history['train_acc'], label='train accuracy')\n",
        "plt.plot(history['val_acc'], label='validation accuracy')\n",
        "\n",
        "plt.title('Training history')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend()\n",
        "plt.ylim([.8, 1]);"
      ],
      "metadata": {
        "id": "7hPPTemggFnu"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Evaluation"
      ],
      "metadata": {
        "id": "xcYEBaOHgh1c"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model2 = SentimentClassifier(len(class_names), indic_bert_model)\n",
        "model2.load_state_dict(torch.load('indic_bert_bilstm_model_state.bin'))\n",
        "model2.eval()"
      ],
      "metadata": {
        "id": "kEUGlUHvZEfK"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model2 = model2.to(device)"
      ],
      "metadata": {
        "id": "KMn_DI0dZrYV"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "test_acc, _ = eval_model(\n",
        "  model2,\n",
        "  test_data_loader,\n",
        "  loss_fn,\n",
        "  device,\n",
        "  len(df_test)\n",
        ")\n",
        "\n",
        "test_acc.item()"
      ],
      "metadata": {
        "id": "1DbgLmPHgkjg",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8417ac98-d293-4e37-cc3d-52d3665bdd85"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.7471111111111112"
            ]
          },
          "metadata": {},
          "execution_count": 50
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def get_predictions(model, data_loader):\n",
        "  model = model.eval()\n",
        "  \n",
        "  tweet_texts = []\n",
        "  predictions = []\n",
        "  prediction_probs = []\n",
        "  real_values = []\n",
        "\n",
        "  with torch.no_grad():\n",
        "    for d in data_loader:\n",
        "\n",
        "      texts = d[\"tweet_text\"]\n",
        "      input_ids = d[\"input_ids\"].to(device)\n",
        "      attention_mask = d[\"attention_mask\"].to(device)\n",
        "      token_type_ids = d['type_ids'].to(device)\n",
        "      targets = d[\"targets\"].to(device)\n",
        "\n",
        "      outputs = model(\n",
        "        input_ids=input_ids,\n",
        "        attention_mask=attention_mask,\n",
        "        token_type_ids = token_type_ids\n",
        "      )\n",
        "      _, preds = torch.max(outputs, dim=1)\n",
        "\n",
        "      probs = F.softmax(outputs, dim=1)\n",
        "\n",
        "      tweet_texts.extend(texts)\n",
        "      predictions.extend(preds)\n",
        "      prediction_probs.extend(probs)\n",
        "      real_values.extend(targets)\n",
        "\n",
        "  predictions = torch.stack(predictions).cpu()\n",
        "  prediction_probs = torch.stack(prediction_probs).cpu()\n",
        "  real_values = torch.stack(real_values).cpu()\n",
        "  return tweet_texts, predictions, prediction_probs, real_values"
      ],
      "metadata": {
        "id": "qG9tcBizgoTx"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "y_tweet_texts, y_pred, y_pred_probs, y_test = get_predictions(\n",
        "  model,\n",
        "  test_data_loader\n",
        ")"
      ],
      "metadata": {
        "id": "Jy1hyBzdgp4w"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "y_pred-=1"
      ],
      "metadata": {
        "id": "WAg0axWf2sQN"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(classification_report(y_test, y_pred, target_names=class_names))"
      ],
      "metadata": {
        "id": "gSGiEUjVgsd-",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "01baf5a0-efe7-42aa-80c2-fb6ced91aa61"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "          -1       0.81      0.82      0.81       750\n",
            "           0       0.66      0.80      0.73       750\n",
            "           1       0.80      0.63      0.70       750\n",
            "\n",
            "    accuracy                           0.75      2250\n",
            "   macro avg       0.76      0.75      0.75      2250\n",
            "weighted avg       0.76      0.75      0.75      2250\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def show_confusion_matrix(confusion_matrix):\n",
        "  hmap = sns.heatmap(confusion_matrix, annot=True, fmt=\"d\", cmap=\"Blues\")\n",
        "  hmap.yaxis.set_ticklabels(hmap.yaxis.get_ticklabels(), rotation=0, ha='right')\n",
        "  hmap.xaxis.set_ticklabels(hmap.xaxis.get_ticklabels(), rotation=0, ha='right')\n",
        "  plt.ylabel('True sentiment')\n",
        "  plt.xlabel('Predicted sentiment');\n",
        "\n",
        "cm = confusion_matrix(y_test, y_pred)\n",
        "df_cm = pd.DataFrame(cm, index=class_names, columns=class_names)\n",
        "show_confusion_matrix(df_cm)"
      ],
      "metadata": {
        "id": "X_a_HrrDgvfh",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "fa7884a2-907c-46be-d024-6841222b2218"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "image/png": {
              "width": 385,
              "height": 272
            },
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "y_test[50:60]"
      ],
      "metadata": {
        "id": "7nq0Dy-fg0C1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "idx = 50\n",
        "\n",
        "tweet_text = y_tweet_texts[idx]\n",
        "true_sentiment = y_test[idx]\n",
        "pred_df = pd.DataFrame({\n",
        "  'class_names': class_names,\n",
        "  'values': y_pred_probs[idx]\n",
        "})"
      ],
      "metadata": {
        "id": "OvUwvAxLg3J8"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print('\\n\\n',tweet_text, '\\n\\n')\n",
        "\n",
        "sns.barplot(x='values', y='class_names', data=pred_df, orient='h')\n",
        "plt.ylabel('sentiment')\n",
        "plt.xlabel('probability')\n",
        "plt.xlim([0, 1]);"
      ],
      "metadata": {
        "id": "Q1lsnXkag4z8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ed184bfe-eabb-43f9-8ba3-14a17de54436"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "\n",
            " आमचे मार्गदर्शक, अहमदनगर शहराचे माजी महापौर व माजी नगराध्यक्ष दीप चव्हाण साहेबांना वाढदिवसाच्या मनःपूर्वक शुभेच्छा \n",
            "\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "image/png": {
              "width": 400,
              "height": 272
            },
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "y_test[14], y_tweet_texts[14]"
      ],
      "metadata": {
        "id": "Fut-BCNug6SP",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6796dca9-a9e2-461c-c79e-2a1187ca60f3"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(tensor(-1),\n",
              " 'भीमा-पाटस साखर कारखान्याच्या कामगारांना २२ महिन्यांचा पगार मिळाला नाही,मग दौंडच्या आमदारांनी तालुक्यासाठी काय केलं याउलट सूतगिरणी अडचणीत आली,म्हणून आपण टेक्स्टाईल पार्क उभारलं ज्यातून ६००० पेक्षा जास्त महिलांना रोजगार मिळाला या भाजपा-शिवसेनावाल्यांनी बारामतीत कोणता उद्योग आणला')"
            ]
          },
          "metadata": {},
          "execution_count": 59
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Custom Prediction"
      ],
      "metadata": {
        "id": "S-Q6fKgFyYIV"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# strip -> punctuations  -> @ mentions delete -> url delete ->  remove # from #-tokens -> stop words delete (no delete )  -> rare words no delete)\n",
        "def preprocess(line):\n",
        "\n",
        "  line = line.strip()\n",
        "  line = re.sub(r'\\S*\\.com\\S*',\" \",line)\n",
        "  line = re.sub(r'\\S*\\.in\\S*',\" \",line)\n",
        "  line = re.sub(r'\\S*\\.net\\S*',\" \",line)\n",
        "  line = re.sub(r'\\S*\\.me\\S*',\" \",line)\n",
        "  line = re.sub(r'\\s+:\\s+',\" \",line)\n",
        "  line = re.sub(r'\\s+-\\s+',\" \",line)\n",
        "  \n",
        "  \n",
        "  line = re.sub(r'/https?\\/\\/(.*\\.(in))\\//',\" \",line)\n",
        "  line = re.split(r'\\s+|[!$%&()*+./;<=>?\\^`{|}~\\'\\\"]\\s*', line)\n",
        "  line = [token.strip() for token in line if token != '' and token != ' ' and token != '\\n']\n",
        "  \n",
        "  if len(line) > 1:\n",
        "    tokens = []\n",
        "    for i in range(len(line)):\n",
        "      if line[i][0] == '#':\n",
        "        tokens.append(line[i][1:])\n",
        "      elif line[i][0] == '@' or line[i][0] in string.punctuation :\n",
        "        continue\n",
        "      else:\n",
        "        tokens.append(line[i])\n",
        "    return tokens\n"
      ],
      "metadata": {
        "id": "MSw8o7WT0vD-"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "tweet_text = \"नमो सरकारला साधी अडचने पार पाडण्याची बुद्धी नाही आहे. एकदम मूर्खपणाचे लक्षण आहे\"\n",
        "tweet_text = \"प्रौढ प्रताप पुरंदर, क्षत्रिय कुलावतंस, सिंहासनाधीश्वर, महाराजाधिराज महाराज श्रीमंत श्री छत्रपती शिवाजी महाराज की जय\"\n",
        "tweet_text = \"मनाचे वेधना किती सोसावे सामान्य माणसाने\"\n",
        "tweet_text = \"माननीय महोदय संजय विचारे यांच्या अपघाताची बातमी आज कळाली. खूप वाईट वाटले.\"\n",
        "tweet_text = \"नमो सरकारला साधी अडचने पार पाडण्याची बुद्धी नाही आहे. एकदम मूर्खपणाचे लक्षण आहे\"\n",
        "tweet_text = \"गढ आला पण माझा सिंह गेला\"\n",
        "tweet_text = \"मैदानात हजारो कंडोम पाकीटं कुणी फेकली?; खुलासा होताच अधिकाऱ्यांनी डोक्यावर हात मारला\"\n",
        "tweet_text = \"महाशिवरात्रीच्या सर्वांना हार्दिक हार्दिक शुभेच्छा\"\n",
        "tweet_text= \"नमो सरकारला साधी अडचने पार पाडण्याची बुद्धी नाही आहे. एकदम मूर्खपणाचे लक्षण आहे\"\n",
        "tweet_text = \"हिंदू धर्मात मुलांना लहानपणापासून आपला धर्मग्रंथ गीता शिकवला जात नाही हे सत्य आहे.\"\n",
        "tweet_text = \"प्रा. यारकुंडकरांसारख्या थोर वयकतीमत्वाचा परिचय देताना, मला आज प्रचंड अभिमान वाटतोय.\"\n",
        "tweet_text = preprocess(tweet_text)\n",
        "tweet_text = \" \".join(tweet_text)"
      ],
      "metadata": {
        "id": "qNzE5THryqEX"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "encoded_tweet = tokenizer.encode_plus(\n",
        "  tweet_text,\n",
        "  max_length=MAX_LEN,\n",
        "  add_special_tokens=True,\n",
        "  return_token_type_ids=True,\n",
        "  pad_to_max_length=True,\n",
        "  return_attention_mask=True,\n",
        "  return_tensors='pt',\n",
        ")\n"
      ],
      "metadata": {
        "id": "mcj_N_NPyv3D"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "input_ids = encoded_tweet['input_ids'].to(device)\n",
        "attention_mask = encoded_tweet['attention_mask'].to(device)\n",
        "token_type_ids = encoded_tweet['token_type_ids'].to(device)\n",
        "\n",
        "output = model(input_ids, attention_mask, token_type_ids)\n",
        "_, prediction = torch.max(output, dim=1)\n",
        "\n",
        "print(f'Tweet text: {tweet_text}')\n",
        "print(f'Sarcasm  : {class_names[prediction]}')"
      ],
      "metadata": {
        "id": "lQ86vuoRyyrh",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5dbc029b-1d20-4efa-ff7f-57c307b4c235"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Tweet text: प्रा यारकुंडकरांसारख्या थोर वयकतीमत्वाचा परिचय देताना, मला आज प्रचंड अभिमान वाटतोय\n",
            "Sarcasm  : 1\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "Ib0RmKH-y0hT"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}