T5_COT_TextVQA_Final.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU",
    "gpuClass": "standard",
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "948bb33688ff4396a2c4cd8e572cc22c": {
          "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_230b3fa41a524f8592aee73c586652a9",
              "IPY_MODEL_ff4e4f5f416c4d31a95611d1e78b115e",
              "IPY_MODEL_ff0a96b733ff4a66b5e0a5a1ce0ab93e"
            ],
            "layout": "IPY_MODEL_092836f77ad943b1ba7029af26a0e26c"
          }
        },
        "230b3fa41a524f8592aee73c586652a9": {
          "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_6f5a2390a4ca4cff8dc63d784b01831b",
            "placeholder": "​",
            "style": "IPY_MODEL_13643eefb62845fbb24ac2c0ecfd662a",
            "value": "Map: 100%"
          }
        },
        "ff4e4f5f416c4d31a95611d1e78b115e": {
          "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": "",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_004d88d8a59645b7ba7c5d11af9d8873",
            "max": 34602,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_3c9b8f5d4b474f188e992d5687f9a1ce",
            "value": 34602
          }
        },
        "ff0a96b733ff4a66b5e0a5a1ce0ab93e": {
          "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_ff5de3793f4045748d705f8746136d89",
            "placeholder": "​",
            "style": "IPY_MODEL_8a79f7459c934f04a78bc73aeb1a7b6f",
            "value": " 34602/34602 [3:10:43&lt;00:00,  3.28 examples/s]"
          }
        },
        "092836f77ad943b1ba7029af26a0e26c": {
          "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": "hidden",
            "width": null
          }
        },
        "6f5a2390a4ca4cff8dc63d784b01831b": {
          "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
          }
        },
        "13643eefb62845fbb24ac2c0ecfd662a": {
          "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": ""
          }
        },
        "004d88d8a59645b7ba7c5d11af9d8873": {
          "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
          }
        },
        "3c9b8f5d4b474f188e992d5687f9a1ce": {
          "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": ""
          }
        },
        "ff5de3793f4045748d705f8746136d89": {
          "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
          }
        },
        "8a79f7459c934f04a78bc73aeb1a7b6f": {
          "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": ""
          }
        },
        "5fca702b8b6740dba11f0331a24699da": {
          "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_93b65bbc085c4b2db3af239a8f2ba6d4",
              "IPY_MODEL_c7c4150aa35d4640845d016e4a9cce2d",
              "IPY_MODEL_41e0e28c03b541d1bee623bba6ef4e74"
            ],
            "layout": "IPY_MODEL_34691d196c724a88a9cd469a523c4406"
          }
        },
        "93b65bbc085c4b2db3af239a8f2ba6d4": {
          "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_411baf3961704ed8a353d7f5652a5f75",
            "placeholder": "​",
            "style": "IPY_MODEL_323d9315ac4840429b266972fabb8cbf",
            "value": "Map: 100%"
          }
        },
        "c7c4150aa35d4640845d016e4a9cce2d": {
          "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": "",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_3c69be4548814a50a3db7263390967ad",
            "max": 5000,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_f3fba36b19444640a780b86be0ce760e",
            "value": 5000
          }
        },
        "41e0e28c03b541d1bee623bba6ef4e74": {
          "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_f99065e7e9ff4fc68b1d5876894f98ed",
            "placeholder": "​",
            "style": "IPY_MODEL_f7763b72390b4f44a5bceb41e1ce3e4a",
            "value": " 5000/5000 [28:09&lt;00:00,  2.88 examples/s]"
          }
        },
        "34691d196c724a88a9cd469a523c4406": {
          "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": "hidden",
            "width": null
          }
        },
        "411baf3961704ed8a353d7f5652a5f75": {
          "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
          }
        },
        "323d9315ac4840429b266972fabb8cbf": {
          "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": ""
          }
        },
        "3c69be4548814a50a3db7263390967ad": {
          "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
          }
        },
        "f3fba36b19444640a780b86be0ce760e": {
          "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": ""
          }
        },
        "f99065e7e9ff4fc68b1d5876894f98ed": {
          "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
          }
        },
        "f7763b72390b4f44a5bceb41e1ce3e4a": {
          "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": ""
          }
        },
        "f94458f1a7b14371b89bc579fbcff2dc": {
          "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_83325b142d4b4061afbb4dd33bacd1a5",
              "IPY_MODEL_4e8860edcc654b88bfdda219a56c4555",
              "IPY_MODEL_785902b5e8744d898310ab169c161f98"
            ],
            "layout": "IPY_MODEL_127eb977723b4ff2a4829b34ca0e53e5"
          }
        },
        "83325b142d4b4061afbb4dd33bacd1a5": {
          "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_a282978a379c4dd78c6e6dc0d7784ede",
            "placeholder": "​",
            "style": "IPY_MODEL_e3942b7d347b411093c4e05dd17bbd55",
            "value": "Map: 100%"
          }
        },
        "4e8860edcc654b88bfdda219a56c4555": {
          "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": "",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_2a403930db2948ceb5e21c6d3ae5b47c",
            "max": 24221,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_17186c399bd04e5d906cd0c1cd930fed",
            "value": 24221
          }
        },
        "785902b5e8744d898310ab169c161f98": {
          "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_cc12f31c82664da48514dfbb6c006cdd",
            "placeholder": "​",
            "style": "IPY_MODEL_374e03e68cb546748dde8b241cb84395",
            "value": " 24221/24221 [00:06&lt;00:00, 3838.19 examples/s]"
          }
        },
        "127eb977723b4ff2a4829b34ca0e53e5": {
          "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": "hidden",
            "width": null
          }
        },
        "a282978a379c4dd78c6e6dc0d7784ede": {
          "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
          }
        },
        "e3942b7d347b411093c4e05dd17bbd55": {
          "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": ""
          }
        },
        "2a403930db2948ceb5e21c6d3ae5b47c": {
          "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
          }
        },
        "17186c399bd04e5d906cd0c1cd930fed": {
          "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": ""
          }
        },
        "cc12f31c82664da48514dfbb6c006cdd": {
          "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
          }
        },
        "374e03e68cb546748dde8b241cb84395": {
          "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": ""
          }
        },
        "4269ad7cf59d40febeccd99fd481224d": {
          "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_7c128ae652904df19fa0e64c07bd973b",
              "IPY_MODEL_a542ac81470845f4afde6ad7057e5c1f",
              "IPY_MODEL_b75da31dd18443a8a856d93b2be3334a"
            ],
            "layout": "IPY_MODEL_ffa730f4a187493290393877894cf3d6"
          }
        },
        "7c128ae652904df19fa0e64c07bd973b": {
          "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_77596335f63744ebab6fbf34c2410d71",
            "placeholder": "​",
            "style": "IPY_MODEL_3d933fbbe80e4d6996a4986a6f05a5bd",
            "value": "Map:  99%"
          }
        },
        "a542ac81470845f4afde6ad7057e5c1f": {
          "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": "",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_185c39f1534a4f83828f02ac555eb70c",
            "max": 24221,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_8aa01ebd101a40ff99b28ed8911531a8",
            "value": 24221
          }
        },
        "b75da31dd18443a8a856d93b2be3334a": {
          "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_06585e141d3a453ea24920332def171f",
            "placeholder": "​",
            "style": "IPY_MODEL_eb955ae085a04edab814dfa600457666",
            "value": " 24000/24221 [00:05&lt;00:00, 4567.61 examples/s]"
          }
        },
        "ffa730f4a187493290393877894cf3d6": {
          "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": "hidden",
            "width": null
          }
        },
        "77596335f63744ebab6fbf34c2410d71": {
          "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
          }
        },
        "3d933fbbe80e4d6996a4986a6f05a5bd": {
          "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": ""
          }
        },
        "185c39f1534a4f83828f02ac555eb70c": {
          "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
          }
        },
        "8aa01ebd101a40ff99b28ed8911531a8": {
          "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": ""
          }
        },
        "06585e141d3a453ea24920332def171f": {
          "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
          }
        },
        "eb955ae085a04edab814dfa600457666": {
          "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": ""
          }
        },
        "7472735e3d6248d289d9a5e288b62901": {
          "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_d99ae61970874f50b68594c478d8a4ec",
              "IPY_MODEL_2614bbc433bf4dc098a59fff6937e6ef",
              "IPY_MODEL_bb9940320ef44dec943984e1046c516c"
            ],
            "layout": "IPY_MODEL_3d7addab8f5246ee90cb6ac2b5c3618c"
          }
        },
        "d99ae61970874f50b68594c478d8a4ec": {
          "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_77dcef9b9c0e45bfa4e671124f964dc3",
            "placeholder": "​",
            "style": "IPY_MODEL_71107d219d6c4e06be4beb8c21c9be78",
            "value": "Map:  96%"
          }
        },
        "2614bbc433bf4dc098a59fff6937e6ef": {
          "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": "",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_27179b00ba804cdfa1e2de0c7b6e1cda",
            "max": 5000,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_bc8af15c64e6432cb357e7fa7574f240",
            "value": 5000
          }
        },
        "bb9940320ef44dec943984e1046c516c": {
          "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_c3b251cfdda248b7bb4ac16f9c2f459a",
            "placeholder": "​",
            "style": "IPY_MODEL_3296f169295c468aa219b8f5f0a9c6fb",
            "value": " 4783/5000 [00:01&lt;00:00, 4353.00 examples/s]"
          }
        },
        "3d7addab8f5246ee90cb6ac2b5c3618c": {
          "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": "hidden",
            "width": null
          }
        },
        "77dcef9b9c0e45bfa4e671124f964dc3": {
          "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
          }
        },
        "71107d219d6c4e06be4beb8c21c9be78": {
          "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": ""
          }
        },
        "27179b00ba804cdfa1e2de0c7b6e1cda": {
          "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
          }
        },
        "bc8af15c64e6432cb357e7fa7574f240": {
          "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": ""
          }
        },
        "c3b251cfdda248b7bb4ac16f9c2f459a": {
          "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
          }
        },
        "3296f169295c468aa219b8f5f0a9c6fb": {
          "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": ""
          }
        },
        "3e948967685545afb171cee25de25232": {
          "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_085b0387d66b4bf4997a69d08621fd48",
              "IPY_MODEL_e68c9f97febf410696387eb5e75a1956",
              "IPY_MODEL_871cdb26fb0447628f01dde867c872b8"
            ],
            "layout": "IPY_MODEL_d0da573b42074bf5908a8f15e83df554"
          }
        },
        "085b0387d66b4bf4997a69d08621fd48": {
          "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_850980052d1040a99373fb037fb2bfc6",
            "placeholder": "​",
            "style": "IPY_MODEL_a7e13fdb02c44e9fb5401fb20f95c16c",
            "value": "Map: 100%"
          }
        },
        "e68c9f97febf410696387eb5e75a1956": {
          "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": "",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_a45aa7a3030f4720b42235e0e8ee4324",
            "max": 5000,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_a7ebba7a3e5c4e73af154cea8a8f20ba",
            "value": 5000
          }
        },
        "871cdb26fb0447628f01dde867c872b8": {
          "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_9727d979449445b89f696720d0a03924",
            "placeholder": "​",
            "style": "IPY_MODEL_39cd4073af7544fbb3c17031a3705c3c",
            "value": " 5000/5000 [00:01&lt;00:00, 4348.42 examples/s]"
          }
        },
        "d0da573b42074bf5908a8f15e83df554": {
          "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": "hidden",
            "width": null
          }
        },
        "850980052d1040a99373fb037fb2bfc6": {
          "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
          }
        },
        "a7e13fdb02c44e9fb5401fb20f95c16c": {
          "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": ""
          }
        },
        "a45aa7a3030f4720b42235e0e8ee4324": {
          "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
          }
        },
        "a7ebba7a3e5c4e73af154cea8a8f20ba": {
          "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": ""
          }
        },
        "9727d979449445b89f696720d0a03924": {
          "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
          }
        },
        "39cd4073af7544fbb3c17031a3705c3c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "LlJj1YWtxrbV"
      },
      "outputs": [],
      "source": [
        "!pip install datasets\n",
        "!pip install -q git+https://github.com/huggingface/transformers.git"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import transformers\n",
        "from transformers import T5Tokenizer, T5ForConditionalGeneration, AutoTokenizer\n",
        "from torch.utils.data import DataLoader, Dataset\n",
        "import torch\n",
        "import json\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "from datasets import load_dataset, load_metric\n",
        "\n",
        "torch.backends.cuda.max_split_size_mb = 1024\n",
        "# Check if a GPU is available\n",
        "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
        "\n",
        "# Define the model and tokenizer\n",
        "tokenizer = AutoTokenizer.from_pretrained(\"t5-small\")\n",
        "model = T5ForConditionalGeneration.from_pretrained(\"t5-small\").to(device)"
      ],
      "metadata": {
        "id": "tQmd_Zc_ylyD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "device"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "t70ZrlseE2z6",
        "outputId": "f9148a50-b0c5-4a0d-94d0-7e7f3163806d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "device(type='cuda')"
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# import datasets"
      ],
      "metadata": {
        "id": "IMiclgcoBlc3"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# dataset = load_dataset(\"textvqa\")\n",
        "# train_data, val_data = dataset[\"train\"], dataset[\"validation\"]"
      ],
      "metadata": {
        "id": "iJ2DqyNlyoXD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_data_dict = train_data.to_dict()"
      ],
      "metadata": {
        "id": "vsqNVNFj0DKn"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# train_data_dict.keys()"
      ],
      "metadata": {
        "id": "QUyhxDyx1jut"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "M5wB0P2D1lTU"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# %mkdir textvqa"
      ],
      "metadata": {
        "id": "jJ6-1Ku3z41U"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from transformers import VisionEncoderDecoderModel, ViTImageProcessor, AutoTokenizer\n",
        "import torch\n",
        "from PIL import Image\n",
        "\n",
        "vit_model = VisionEncoderDecoderModel.from_pretrained(\"nlpconnect/vit-gpt2-image-captioning\")\n",
        "feature_extractor = ViTImageProcessor.from_pretrained(\"nlpconnect/vit-gpt2-image-captioning\")\n",
        "tokenizer = AutoTokenizer.from_pretrained(\"nlpconnect/vit-gpt2-image-captioning\")\n",
        "\n",
        "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
        "vit_model.to(device)\n",
        "\n",
        "\n",
        "\n",
        "max_length = 64\n",
        "num_beams = 4\n",
        "gen_kwargs = {\"max_length\": max_length, \"num_beams\": num_beams}\n",
        "def predict_step(example):\n",
        "  images = []\n",
        "  # for image_path in image_paths:\n",
        "  #   i_image = Image.open(image_path)\n",
        "  #   if i_image.mode != \"RGB\":\n",
        "  #     i_image = i_image.convert(mode=\"RGB\")\n",
        "  image = example[\"image\"]\n",
        "  \n",
        "  if image:\n",
        "    if image.mode != \"RGB\":\n",
        "      image = image.convert(mode=\"RGB\")\n",
        "    images.append(image)\n",
        "\n",
        "    pixel_values = feature_extractor(images=images, return_tensors=\"pt\").pixel_values\n",
        "    pixel_values = pixel_values.to(device)\n",
        "\n",
        "    output_ids = vit_model.generate(pixel_values, **gen_kwargs)\n",
        "\n",
        "    preds = tokenizer.batch_decode(output_ids, skip_special_tokens=True)\n",
        "    preds = [pred.strip() for pred in preds]\n",
        "\n",
        "    example[\"captions\"] = preds[0]\n",
        "  else:\n",
        "    example[\"captions\"] = \"\"\n",
        "  return example"
      ],
      "metadata": {
        "id": "tX5awALZz_dQ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#Mapping the predict_step function to generate image captions for training set\n",
        "train_data_captions = train_data.map(predict_step)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 17,
          "referenced_widgets": [
            "948bb33688ff4396a2c4cd8e572cc22c",
            "230b3fa41a524f8592aee73c586652a9",
            "ff4e4f5f416c4d31a95611d1e78b115e",
            "ff0a96b733ff4a66b5e0a5a1ce0ab93e",
            "092836f77ad943b1ba7029af26a0e26c",
            "6f5a2390a4ca4cff8dc63d784b01831b",
            "13643eefb62845fbb24ac2c0ecfd662a",
            "004d88d8a59645b7ba7c5d11af9d8873",
            "3c9b8f5d4b474f188e992d5687f9a1ce",
            "ff5de3793f4045748d705f8746136d89",
            "8a79f7459c934f04a78bc73aeb1a7b6f"
          ]
        },
        "id": "wvFs5CKs1uq8",
        "outputId": "e3cb7da3-d7d6-4e83-a2fc-6c2a5b86ce14"
      },
      "execution_count": null,
      "outputs": [
        {
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "948bb33688ff4396a2c4cd8e572cc22c",
              "version_major": 2,
              "version_minor": 0
            },
            "text/plain": [
              "Map:   0%|          | 0/34602 [00:00<?, ? examples/s]"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_data_captions[0]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vwPj8ZxPiIzP",
        "outputId": "3040c2e8-0d78-43f1-b2e3-ff3365e4c2de"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'image_id': '0054c91397f2fe05',\n",
              " 'question_id': 0,\n",
              " 'question': 'what is the brand of phone?',\n",
              " 'question_tokens': ['what', 'is', 'the', 'brand', 'of', 'phone'],\n",
              " 'image': <PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=1024x730 at 0x7F7683033370>,\n",
              " 'image_width': 1024,\n",
              " 'image_height': 730,\n",
              " 'flickr_original_url': 'https://farm6.staticflickr.com/2891/9134076951_f65b421097_o.jpg',\n",
              " 'flickr_300k_url': 'https://c4.staticflickr.com/3/2891/9134076951_9db89d3e0f_z.jpg',\n",
              " 'answers': ['nokia',\n",
              "  'nokia',\n",
              "  'nokia',\n",
              "  'nokia',\n",
              "  'toshiba',\n",
              "  'nokia',\n",
              "  'nokia',\n",
              "  'nokia',\n",
              "  'nokia',\n",
              "  'nokia'],\n",
              " 'image_classes': ['Belt',\n",
              "  'Headphones',\n",
              "  'Goggles',\n",
              "  'Scale',\n",
              "  'Bottle opener',\n",
              "  'Mobile phone',\n",
              "  'Mirror',\n",
              "  'Digital clock',\n",
              "  'Television',\n",
              "  'Telephone',\n",
              "  'Tool',\n",
              "  'Wheel',\n",
              "  'Camera',\n",
              "  'Watch',\n",
              "  'Glasses',\n",
              "  'Aircraft'],\n",
              " 'set_name': 'train',\n",
              " 'captions': 'a black and white photo of a cell phone'}"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "len(val_data)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HAh0HUdGit7e",
        "outputId": "aa5f502b-d7cf-4b38-94ef-e00f6ce8fbd9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "5000"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#Mapping the predict_step function to generate image captions for training set\n",
        "val_data_captions = val_data.map(predict_step)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 17,
          "referenced_widgets": [
            "5fca702b8b6740dba11f0331a24699da",
            "93b65bbc085c4b2db3af239a8f2ba6d4",
            "c7c4150aa35d4640845d016e4a9cce2d",
            "41e0e28c03b541d1bee623bba6ef4e74",
            "34691d196c724a88a9cd469a523c4406",
            "411baf3961704ed8a353d7f5652a5f75",
            "323d9315ac4840429b266972fabb8cbf",
            "3c69be4548814a50a3db7263390967ad",
            "f3fba36b19444640a780b86be0ce760e",
            "f99065e7e9ff4fc68b1d5876894f98ed",
            "f7763b72390b4f44a5bceb41e1ce3e4a"
          ]
        },
        "id": "UAtGCcuf1xK6",
        "outputId": "f8602b5e-b3ba-40f4-dc92-a5c6f9a73639"
      },
      "execution_count": null,
      "outputs": [
        {
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "5fca702b8b6740dba11f0331a24699da",
              "version_major": 2,
              "version_minor": 0
            },
            "text/plain": [
              "Map:   0%|          | 0/5000 [00:00<?, ? examples/s]"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "val_data_captions[0]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "c8-L1JCWspDs",
        "outputId": "4d653e93-4abd-4ba5-fe38-eda094d5bd23"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'image_id': '003a8ae2ef43b901',\n",
              " 'question_id': 34602,\n",
              " 'question': 'what is the brand of this camera?',\n",
              " 'question_tokens': ['what', 'is', 'the', 'brand', 'of', 'this', 'camera'],\n",
              " 'image': <PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=1024x664 at 0x7F7683627730>,\n",
              " 'image_width': 1024,\n",
              " 'image_height': 664,\n",
              " 'flickr_original_url': 'https://farm2.staticflickr.com/4/5566811_bc00d504a6_o.jpg',\n",
              " 'flickr_300k_url': 'https://farm2.staticflickr.com/4/5566811_bc00d504a6_o.jpg',\n",
              " 'answers': ['nous les gosses',\n",
              "  'dakota',\n",
              "  'clos culombu',\n",
              "  'dakota digital',\n",
              "  'dakota',\n",
              "  'dakota',\n",
              "  'dakota digital',\n",
              "  'dakota digital',\n",
              "  'dakota',\n",
              "  'dakota'],\n",
              " 'image_classes': ['Cassette deck',\n",
              "  'Printer',\n",
              "  'Medical equipment',\n",
              "  'Computer mouse',\n",
              "  'Scale',\n",
              "  'Telephone',\n",
              "  'Camera',\n",
              "  'Ipod',\n",
              "  'Remote control'],\n",
              " 'set_name': 'val',\n",
              " 'captions': 'a close up picture of a cell phone'}"
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "try:\n",
        "    import cPickle as pickle\n",
        "except ImportError:  # Python 3.x\n",
        "    import pickle"
      ],
      "metadata": {
        "id": "2lE7SwLI11Fo"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_data_captions_dict = train_data_captions.to_dict()"
      ],
      "metadata": {
        "id": "KJ4qPkJB16_w"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_data_captions_dict.keys()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xLkBoMhDiMzx",
        "outputId": "a0f6159a-9eec-4104-a35d-9e0ac0f1043e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys(['image_id', 'question_id', 'question', 'question_tokens', 'image', 'image_width', 'image_height', 'flickr_original_url', 'flickr_300k_url', 'answers', 'image_classes', 'set_name', 'captions'])"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# train_data_captions_dict[\"answers\"]"
      ],
      "metadata": {
        "id": "vtr8UxsEi4Gz"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "with open('./textvqa/train_captionsdata.pickle', 'wb') as handle:\n",
        "    pickle.dump(train_data_captions_dict, handle, protocol=pickle.HIGHEST_PROTOCOL)"
      ],
      "metadata": {
        "id": "8fb4pAbw11iS"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "with open('./textvqa/train_captionsdata.pickle', 'rb') as handle:\n",
        "    train_data_captions_dict = pickle.load(handle)"
      ],
      "metadata": {
        "id": "KaM_q0MREGkI"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "val_data_captions_dict = val_data_captions.to_dict()"
      ],
      "metadata": {
        "id": "9N56X0Lb14wQ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "with open('./textvqa/valid_captionsdata.pickle', 'wb') as handle:\n",
        "    pickle.dump(val_data_captions_dict, handle, protocol=pickle.HIGHEST_PROTOCOL)"
      ],
      "metadata": {
        "id": "5DP4N8PG1-1t"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "with open('./textvqa/valid_captionsdata.pickle', 'rb') as handle:\n",
        "    val_data_captions_dict = pickle.load(handle)"
      ],
      "metadata": {
        "id": "8G9k994JEJQe"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#Functions to create unique list of choices and the ground truth answer\n",
        "def create_choices(answers):\n",
        "    choices_list = list(set(answers))\n",
        "    return choices_list\n",
        "def label_answer(answers):\n",
        "    answer = max(answers,key=answers.count)\n",
        "    return answer"
      ],
      "metadata": {
        "id": "0W3_j9bY_idd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_data_captions_dict[\"choices\"] = list(map(create_choices,train_data_captions_dict[\"answers\"]))"
      ],
      "metadata": {
        "id": "m61rojXTA3y1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_data_captions_dict[\"answer\"] = list(map(label_answer,train_data_captions_dict[\"answers\"]))"
      ],
      "metadata": {
        "id": "EmA-CP0fBWi2"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_data_captions_new = datasets.Dataset.from_pandas(pd.DataFrame(data=train_data_captions_dict))"
      ],
      "metadata": {
        "id": "rkkx6IVkBZiD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_data_captions_new[0]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "T1WWt771BiZp",
        "outputId": "875d9852-0fe4-400e-92dc-84a05a03eb76"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'image_id': '0054c91397f2fe05',\n",
              " 'question_id': 0,\n",
              " 'question': 'what is the brand of phone?',\n",
              " 'question_tokens': ['what', 'is', 'the', 'brand', 'of', 'phone'],\n",
              " 'image': {'bytes': None,\n",
              "  'path': '/root/.cache/huggingface/datasets/downloads/extracted/f8b3358ef194bf4536b05cbeff7819f8ddc108d2929f31f1662029638691522e/train_images/0054c91397f2fe05.jpg'},\n",
              " 'image_width': 1024,\n",
              " 'image_height': 730,\n",
              " 'flickr_original_url': 'https://farm6.staticflickr.com/2891/9134076951_f65b421097_o.jpg',\n",
              " 'flickr_300k_url': 'https://c4.staticflickr.com/3/2891/9134076951_9db89d3e0f_z.jpg',\n",
              " 'answers': ['nokia',\n",
              "  'nokia',\n",
              "  'nokia',\n",
              "  'nokia',\n",
              "  'toshiba',\n",
              "  'nokia',\n",
              "  'nokia',\n",
              "  'nokia',\n",
              "  'nokia',\n",
              "  'nokia'],\n",
              " 'image_classes': ['Belt',\n",
              "  'Headphones',\n",
              "  'Goggles',\n",
              "  'Scale',\n",
              "  'Bottle opener',\n",
              "  'Mobile phone',\n",
              "  'Mirror',\n",
              "  'Digital clock',\n",
              "  'Television',\n",
              "  'Telephone',\n",
              "  'Tool',\n",
              "  'Wheel',\n",
              "  'Camera',\n",
              "  'Watch',\n",
              "  'Glasses',\n",
              "  'Aircraft'],\n",
              " 'set_name': 'train',\n",
              " 'captions': 'a black and white photo of a cell phone',\n",
              " 'choices': ['nokia', 'toshiba'],\n",
              " 'answer': 'nokia'}"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "len([i for i in train_data_captions_new if i[\"image_id\"]])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bxNkPSwOZntj",
        "outputId": "19ffb352-82df-4124-c16c-953d3a2cba13"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "34602"
            ]
          },
          "metadata": {},
          "execution_count": 103
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "len(train_data_captions_new)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rMjbv0cKZ-pw",
        "outputId": "c70b7930-5932-461f-d70b-da58fc43a813"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "34602"
            ]
          },
          "metadata": {},
          "execution_count": 104
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "val_data_captions_dict[\"choices\"] = list(map(create_choices,val_data_captions_dict[\"answers\"]))\n",
        "val_data_captions_dict[\"answer\"] = list(map(label_answer,val_data_captions_dict[\"answers\"]))"
      ],
      "metadata": {
        "id": "TkO329usBokD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "val_data_captions_new = datasets.Dataset.from_pandas(pd.DataFrame(data=val_data_captions_dict))"
      ],
      "metadata": {
        "id": "GBT7hrbVBzjg"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "val_data_captions_new[0]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "A4bPZmh_B3ry",
        "outputId": "08072f13-6180-42cc-e7aa-0e0453e8b987"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'image_id': '003a8ae2ef43b901',\n",
              " 'question_id': 34602,\n",
              " 'question': 'what is the brand of this camera?',\n",
              " 'question_tokens': ['what', 'is', 'the', 'brand', 'of', 'this', 'camera'],\n",
              " 'image': {'bytes': None,\n",
              "  'path': '/root/.cache/huggingface/datasets/downloads/extracted/f8b3358ef194bf4536b05cbeff7819f8ddc108d2929f31f1662029638691522e/train_images/003a8ae2ef43b901.jpg'},\n",
              " 'image_width': 1024,\n",
              " 'image_height': 664,\n",
              " 'flickr_original_url': 'https://farm2.staticflickr.com/4/5566811_bc00d504a6_o.jpg',\n",
              " 'flickr_300k_url': 'https://farm2.staticflickr.com/4/5566811_bc00d504a6_o.jpg',\n",
              " 'answers': ['nous les gosses',\n",
              "  'dakota',\n",
              "  'clos culombu',\n",
              "  'dakota digital',\n",
              "  'dakota',\n",
              "  'dakota',\n",
              "  'dakota digital',\n",
              "  'dakota digital',\n",
              "  'dakota',\n",
              "  'dakota'],\n",
              " 'image_classes': ['Cassette deck',\n",
              "  'Printer',\n",
              "  'Medical equipment',\n",
              "  'Computer mouse',\n",
              "  'Scale',\n",
              "  'Telephone',\n",
              "  'Camera',\n",
              "  'Ipod',\n",
              "  'Remote control'],\n",
              " 'set_name': 'val',\n",
              " 'captions': 'a close up picture of a cell phone',\n",
              " 'choices': ['nous les gosses', 'dakota digital', 'dakota', 'clos culombu'],\n",
              " 'answer': 'dakota'}"
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "sample_train_dataset = train_data_captions_new.train_test_split(train_size=0.7, seed = 42)[\"train\"]"
      ],
      "metadata": {
        "id": "kae1i4L7GiO-"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "len(sample_train_dataset)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Qew3OFpwGocr",
        "outputId": "105778da-9e55-4036-fcf4-cb84ebe574e6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "24221"
            ]
          },
          "metadata": {},
          "execution_count": 106
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Define the model and tokenizer\n",
        "tokenizer = AutoTokenizer.from_pretrained(\"t5-small\")\n",
        "model = T5ForConditionalGeneration.from_pretrained(\"t5-small\").to(device) \n",
        "#Change this to the new model from checkpoints."
      ],
      "metadata": {
        "id": "gjvVVH4RB5hp"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# process the examples in input and target text format and the eos token at the end \n",
        "\n",
        "#For answer ONLY (not solution and explanation)\n",
        "def add_eos_to_examples(example):\n",
        "    example['input_text'] = 'question: %s  choices: %s  context: %s </s>' % (example['question'].lower(), example['choices'], example['captions'].lower())\n",
        "    example['target_text'] = 'answer: %s </s>' % ([example['answer']])\n",
        "    return example\n",
        "\n",
        "# tokenize the examples\n",
        "def convert_to_features(example_batch):\n",
        "    input_encodings = tokenizer.batch_encode_plus(example_batch['input_text'], pad_to_max_length=True, max_length=512, truncation=True)\n",
        "    target_encodings = tokenizer.batch_encode_plus(example_batch['target_text'], pad_to_max_length=True, max_length=128, truncation=True)\n",
        "    encodings = {\n",
        "        'input_ids': input_encodings['input_ids'], \n",
        "        'attention_mask': input_encodings['attention_mask'],\n",
        "        'target_ids': target_encodings['input_ids'],\n",
        "        'target_attention_mask': target_encodings['attention_mask']\n",
        "    }\n",
        "\n",
        "    return encodings"
      ],
      "metadata": {
        "id": "kifaoY0rB-gv"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#For generated examples as input:\n",
        "\n",
        "# map add_eos_to_examples function to the dataset example wise \n",
        "train_dataset = sample_train_dataset.map(add_eos_to_examples, load_from_cache_file=False)\n",
        "# map convert_to_features batch wise\n",
        "train_dataset = train_dataset.map(convert_to_features, batched=True)\n",
        "\n",
        "valid_dataset = val_data_captions_new.map(add_eos_to_examples, load_from_cache_file=False)\n",
        "valid_dataset = valid_dataset.map(convert_to_features, batched=True, load_from_cache_file=False)\n",
        "\n",
        "\n",
        "# set the tensor type and the columns which the dataset should return\n",
        "columns = ['input_ids', 'target_ids', 'attention_mask', 'target_attention_mask']\n",
        "train_dataset.set_format(type='torch', columns=columns)\n",
        "valid_dataset.set_format(type='torch', columns=columns)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 73,
          "referenced_widgets": [
            "f94458f1a7b14371b89bc579fbcff2dc",
            "83325b142d4b4061afbb4dd33bacd1a5",
            "4e8860edcc654b88bfdda219a56c4555",
            "785902b5e8744d898310ab169c161f98",
            "127eb977723b4ff2a4829b34ca0e53e5",
            "a282978a379c4dd78c6e6dc0d7784ede",
            "e3942b7d347b411093c4e05dd17bbd55",
            "2a403930db2948ceb5e21c6d3ae5b47c",
            "17186c399bd04e5d906cd0c1cd930fed",
            "cc12f31c82664da48514dfbb6c006cdd",
            "374e03e68cb546748dde8b241cb84395",
            "4269ad7cf59d40febeccd99fd481224d",
            "7c128ae652904df19fa0e64c07bd973b",
            "a542ac81470845f4afde6ad7057e5c1f",
            "b75da31dd18443a8a856d93b2be3334a",
            "ffa730f4a187493290393877894cf3d6",
            "77596335f63744ebab6fbf34c2410d71",
            "3d933fbbe80e4d6996a4986a6f05a5bd",
            "185c39f1534a4f83828f02ac555eb70c",
            "8aa01ebd101a40ff99b28ed8911531a8",
            "06585e141d3a453ea24920332def171f",
            "eb955ae085a04edab814dfa600457666",
            "7472735e3d6248d289d9a5e288b62901",
            "d99ae61970874f50b68594c478d8a4ec",
            "2614bbc433bf4dc098a59fff6937e6ef",
            "bb9940320ef44dec943984e1046c516c",
            "3d7addab8f5246ee90cb6ac2b5c3618c",
            "77dcef9b9c0e45bfa4e671124f964dc3",
            "71107d219d6c4e06be4beb8c21c9be78",
            "27179b00ba804cdfa1e2de0c7b6e1cda",
            "bc8af15c64e6432cb357e7fa7574f240",
            "c3b251cfdda248b7bb4ac16f9c2f459a",
            "3296f169295c468aa219b8f5f0a9c6fb",
            "3e948967685545afb171cee25de25232",
            "085b0387d66b4bf4997a69d08621fd48",
            "e68c9f97febf410696387eb5e75a1956",
            "871cdb26fb0447628f01dde867c872b8",
            "d0da573b42074bf5908a8f15e83df554",
            "850980052d1040a99373fb037fb2bfc6",
            "a7e13fdb02c44e9fb5401fb20f95c16c",
            "a45aa7a3030f4720b42235e0e8ee4324",
            "a7ebba7a3e5c4e73af154cea8a8f20ba",
            "9727d979449445b89f696720d0a03924",
            "39cd4073af7544fbb3c17031a3705c3c"
          ]
        },
        "id": "yDxWmEuTCSrp",
        "outputId": "9dd73ba3-3482-49d1-ff2f-ae8ecfe244d1"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "Map:   0%|          | 0/24221 [00:00<?, ? examples/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "f94458f1a7b14371b89bc579fbcff2dc"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "Map:   0%|          | 0/24221 [00:00<?, ? examples/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "4269ad7cf59d40febeccd99fd481224d"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.9/dist-packages/transformers/tokenization_utils_base.py:2364: 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",
            "  warnings.warn(\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "Map:   0%|          | 0/5000 [00:00<?, ? examples/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "7472735e3d6248d289d9a5e288b62901"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "Map:   0%|          | 0/5000 [00:00<?, ? examples/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "3e948967685545afb171cee25de25232"
            }
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "len(train_dataset), len(valid_dataset)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HTiHJlSOCZD6",
        "outputId": "f2418e48-b7ab-4d9f-9c5e-6aa64ef52348"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(24221, 5000)"
            ]
          },
          "metadata": {},
          "execution_count": 110
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# train_dataset[0]"
      ],
      "metadata": {
        "id": "bKp8BSXpGDlu"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_dataloader = DataLoader(train_dataset, shuffle=True, batch_size=8)\n",
        "valid_dataloader = DataLoader(valid_dataset, batch_size=8)"
      ],
      "metadata": {
        "id": "ZjznGQjMCd4l"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "batch = next(iter(train_dataloader))\n",
        "print(batch.keys())"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yFmpXiq1CfL7",
        "outputId": "5986541d-6a34-482a-c13f-aa1d6101d29d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "dict_keys(['input_ids', 'attention_mask', 'target_ids', 'target_attention_mask'])\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "tokenizer.decode(batch['input_ids'][7])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 178
        },
        "id": "W4hvjMHvCgbX",
        "outputId": "91312978-04f3-4a77-d984-91539896925a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "\"question: what does the sign say on the building on the right? choices: ['plole', 'pioloe', 'piole'] context: a city street at night with a large clock tower</s></s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>\""
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 113
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "tokenizer.decode(batch['target_ids'][7])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 107
        },
        "id": "fGNOHbsfCh6P",
        "outputId": "926c1a26-9003-4033-9f58-01648aa346f8"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "\"answer: ['piole']</s></s><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad><pad>\""
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 114
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from transformers import DataCollator, AdamW, get_linear_schedule_with_warmup\n",
        "import numpy as np"
      ],
      "metadata": {
        "id": "jRKQq3uLCj94"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def data_collator(batch):\n",
        "    input_ids = torch.stack([example['input_ids'] for example in batch])\n",
        "    attention_mask = torch.stack([example['attention_mask'] for example in batch])\n",
        "\n",
        "    target_ids = torch.stack([example['target_ids'] for example in batch])\n",
        "    target_attention_mask = torch.stack([example['target_attention_mask'] for example in batch])\n",
        "    \n",
        "    return {\n",
        "        'input_ids': input_ids,\n",
        "        'attention_mask': attention_mask,\n",
        "        'target_ids': target_ids,\n",
        "        'target_attention_mask': target_attention_mask\n",
        "    }"
      ],
      "metadata": {
        "id": "aeT5TRoUClQC"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "type(train_dataset)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yWrEDwy6GJRW",
        "outputId": "ce7d3d67-31b0-4ca8-b1bb-76eb872b83f1"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "datasets.arrow_dataset.Dataset"
            ]
          },
          "metadata": {},
          "execution_count": 117
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from tqdm import tqdm\n",
        "optimizer = AdamW(model.parameters(), lr=1e-5)\n",
        "num_training_steps = 10000\n",
        "num_warmup_steps = 500\n",
        "lr_scheduler = get_linear_schedule_with_warmup(optimizer, num_warmup_steps, num_training_steps)\n",
        "\n",
        "# Define the training loop\n",
        "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
        "model.to(device)\n",
        "train_loss = []\n",
        "val_loss = []\n",
        "for epoch in range(5):\n",
        "    print(\"Training: \")\n",
        "    model.train(True)\n",
        "    train_dataloader = DataLoader(train_dataset, batch_size=4, shuffle=True, collate_fn=data_collator)\n",
        "    total_loss = 0\n",
        "    torch.cuda.empty_cache()\n",
        "    for batch in tqdm(train_dataloader):\n",
        "        input_ids = batch['input_ids'].to(device)\n",
        "        attention_mask = batch['attention_mask'].to(device)\n",
        "        target_ids = batch['target_ids'].to(device)\n",
        "        target_attention_mask = batch['target_attention_mask'].to(device)\n",
        "        \n",
        "        outputs = model(\n",
        "            input_ids=input_ids,\n",
        "            attention_mask=attention_mask,\n",
        "            labels=target_ids,\n",
        "            decoder_attention_mask=target_attention_mask,\n",
        "            use_cache=False\n",
        "        )\n",
        "        \n",
        "        loss = outputs.loss\n",
        "        loss.backward()\n",
        "        total_loss += loss.item()\n",
        "        torch.nn.utils.clip_grad_norm_(model.parameters(), 1.0)\n",
        "        optimizer.step()\n",
        "        lr_scheduler.step()\n",
        "        optimizer.zero_grad()\n",
        "    train_epoch_loss = total_loss / len(train_dataloader) \n",
        "    train_loss.append(train_epoch_loss)\n",
        "    print(f\"Training Loss at epoch {epoch}: {train_epoch_loss}\")\n",
        "\n",
        "    print(\"Validation:\")\n",
        "    valid_dataloader = DataLoader(valid_dataset, batch_size=4, shuffle=False, collate_fn=data_collator)\n",
        "    model.eval()\n",
        "    total_loss = 0\n",
        "    with torch.no_grad():\n",
        "        for batch in tqdm(valid_dataloader):\n",
        "            input_ids = batch['input_ids'].to(device)\n",
        "            attention_mask = batch['attention_mask'].to(device)\n",
        "            target_ids = batch['target_ids'].to(device)\n",
        "            target_attention_mask = batch['target_attention_mask'].to(device)\n",
        "\n",
        "            outputs = model(\n",
        "                input_ids=input_ids,\n",
        "                attention_mask=attention_mask,\n",
        "                labels=target_ids,\n",
        "                decoder_attention_mask=target_attention_mask,\n",
        "                use_cache=False\n",
        "            )\n",
        "\n",
        "            loss = outputs.loss\n",
        "            total_loss += loss.item()\n",
        "    val_epoch_loss = total_loss / len(valid_dataloader) \n",
        "    val_loss.append(val_epoch_loss)\n",
        "    print(f\"Validation Loss at epoch {epoch}: {val_epoch_loss}\")\n",
        "    # val_loss.append(total_loss / len(valid_dataloader))\n",
        "    # print(val_loss)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 820
        },
        "id": "V7UqVZOmCmqA",
        "outputId": "e1dc2f33-7c64-42a4-ddde-eb594d5d3450"
      },
      "execution_count": null,
      "outputs": [
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.9/dist-packages/transformers/optimization.py:407: FutureWarning: This implementation of AdamW is deprecated and will be removed in a future version. Use the PyTorch implementation torch.optim.AdamW instead, or set `no_deprecation_warning=True` to disable this warning\n",
            "  warnings.warn(\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Training: \n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 6056/6056 [17:57<00:00,  5.62it/s]\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Training Loss at epoch 0: 0.48205909228804505\n",
            "Validation:\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 1250/1250 [01:11<00:00, 17.37it/s]\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Validation Loss at epoch 0: 0.008576889139041305\n",
            "Training: \n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 6056/6056 [17:58<00:00,  5.61it/s]\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Training Loss at epoch 1: 0.011513533774923129\n",
            "Validation:\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 1250/1250 [01:12<00:00, 17.22it/s]\n"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Validation Loss at epoch 1: 0.008320784086035565\n",
            "Training: \n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 6056/6056 [18:00<00:00,  5.60it/s]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training Loss at epoch 2: 0.01130285238939273\n",
            "Validation:\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 1250/1250 [01:12<00:00, 17.22it/s]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Validation Loss at epoch 2: 0.008320784086035565\n",
            "Training: \n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "  0%|          | 4/6056 [00:00<21:49,  4.62it/s]\n"
          ]
        },
        {
          "output_type": "error",
          "ename": "KeyboardInterrupt",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-118-af4630acd797>\u001b[0m in \u001b[0;36m<cell line: 12>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     31\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     32\u001b[0m         \u001b[0mloss\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0moutputs\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mloss\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 33\u001b[0;31m         \u001b[0mloss\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbackward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     34\u001b[0m         \u001b[0mtotal_loss\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0mloss\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mitem\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     35\u001b[0m         \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mutils\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mclip_grad_norm_\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mparameters\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1.0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.9/dist-packages/torch/_tensor.py\u001b[0m in \u001b[0;36mbackward\u001b[0;34m(self, gradient, retain_graph, create_graph, inputs)\u001b[0m\n\u001b[1;32m    485\u001b[0m                 \u001b[0minputs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    486\u001b[0m             )\n\u001b[0;32m--> 487\u001b[0;31m         torch.autograd.backward(\n\u001b[0m\u001b[1;32m    488\u001b[0m             \u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mgradient\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mretain_graph\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcreate_graph\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minputs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    489\u001b[0m         )\n",
            "\u001b[0;32m/usr/local/lib/python3.9/dist-packages/torch/autograd/__init__.py\u001b[0m in \u001b[0;36mbackward\u001b[0;34m(tensors, grad_tensors, retain_graph, create_graph, grad_variables, inputs)\u001b[0m\n\u001b[1;32m    198\u001b[0m     \u001b[0;31m# some Python versions print out the first line of a multi-line function\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    199\u001b[0m     \u001b[0;31m# calls in the traceback and some print out the last line\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 200\u001b[0;31m     Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass\n\u001b[0m\u001b[1;32m    201\u001b[0m         \u001b[0mtensors\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mgrad_tensors_\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mretain_graph\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcreate_graph\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    202\u001b[0m         allow_unreachable=True, accumulate_grad=True)  # Calls into the C++ engine to run the backward pass\n",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "# Plot the training and validation losses\n",
        "plt.plot(train_loss, label='Train Loss')\n",
        "plt.plot(val_loss, label='Valid Loss')\n",
        "plt.xlabel('Epoch')\n",
        "plt.ylabel('Loss')\n",
        "plt.title(\"Training and Validation Loss for Answer - TextVQA (5 Epochs)\")\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "id": "1e8ECioDCoSr",
        "outputId": "b6cc8641-0c1f-4fd5-aa90-fa8d6f353655"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "# Plot the training and validation losses\n",
        "plt.plot(train_loss[1:5], label='Train Loss')\n",
        "plt.plot(val_loss[1:5], label='Valid Loss')\n",
        "plt.xlabel('Epoch')\n",
        "plt.ylabel('Loss')\n",
        "plt.title(\"Training and Validation Loss for Answer with Generated Explanation as Input\")\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "id": "bHeRTlVmCr6l",
        "outputId": "d8fad5ca-ed90-4e99-c27e-3367324ec2c8"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# %mkdir models_saved"
      ],
      "metadata": {
        "id": "tuYXISMhC3v3"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#Parameters for answer only model\n",
        "input_max_len = 512\n",
        "target_max_len = 128\n",
        "num_epochs = 10\n",
        "# learning_rate = 1e-5\n",
        "train_batch_size = 8\n",
        "val_batch_size = 8\n",
        "# no_of_training_steps = 10000\n",
        "# no_of_warmup_steps = 500\n",
        "model_name = f\"t5small_answer_model2_{input_max_len}_{target_max_len}_{num_epochs}_{train_batch_size}_{val_batch_size}\""
      ],
      "metadata": {
        "id": "3Z2NzF5ACsjx"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#Answer only model\n",
        "model.save_pretrained(f\"./textvqa/models_saved/{model_name}\", from_pt=True)"
      ],
      "metadata": {
        "id": "0PI2jgxgCvyQ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#evaluating the text generation on training set\n",
        "model.eval()\n",
        "train_dataloader = DataLoader(train_dataset, batch_size=8, collate_fn=data_collator)\n",
        "\n",
        "model_answers_train = []\n",
        "target_answers_train = []\n",
        "\n",
        "with torch.no_grad():\n",
        "    for batch in tqdm(train_dataloader):\n",
        "        input_ids = batch['input_ids'].to(device)\n",
        "        attention_mask = batch['attention_mask'].to(device)\n",
        "        target_ids = batch['target_ids'].to(device)\n",
        "        target_attention_mask = batch['target_attention_mask'].to(device)\n",
        "        \n",
        "        outputs = model.generate(\n",
        "            input_ids=input_ids,\n",
        "            attention_mask=attention_mask,\n",
        "            #decoder_attention_mask=target_attention_mask,\n",
        "            max_length=512,\n",
        "            num_beams=10,\n",
        "            early_stopping=True\n",
        "        )\n",
        "        \n",
        "        # Convert the generated output to text\n",
        "        output_text = tokenizer.batch_decode(outputs, skip_special_tokens=True)\n",
        "        target_text = tokenizer.batch_decode(target_ids, skip_special_tokens=True)\n",
        "        # print(output_text)\n",
        "        # print(\"--\")\n",
        "        # print(target_text)\n",
        "        model_answers_train.append(output_text)\n",
        "        target_answers_train.append(target_text)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "je-l8uw2DCVt",
        "outputId": "faa4403b-93a6-4efb-aab8-5e902500ca86"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 3028/3028 [25:55<00:00,  1.95it/s]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "with open('./textvqa/model2_answers_train', 'wb') as handle:\n",
        "    pickle.dump(model_answers_train, handle, protocol=pickle.HIGHEST_PROTOCOL)\n",
        "with open('./textvqa/model2_target_answers_train', 'wb') as handle:\n",
        "    pickle.dump(target_answers_train, handle, protocol=pickle.HIGHEST_PROTOCOL)"
      ],
      "metadata": {
        "id": "_-eRdkBgDKxQ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model_answers_train_overall,target_answers_train_overall=sum(model_answers_train,[]),sum(target_answers_train,[])\n",
        "accuracy_boolean_train=[a.lower().split(\"explanation:\")[0].strip()==b.lower().split(\"explanation:\")[0].strip() for a,b in zip((model_answers_train_overall),(target_answers_train_overall))]\n",
        "acc_train=(sum(accuracy_boolean_train)/len(accuracy_boolean_train))*100\n",
        "print(f\"The training accuracy of the model is {acc_train:.2f} %\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sC-l9sbfDcZ6",
        "outputId": "1c3aba56-6262-4b78-d308-30f6d68409da"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The training accuracy of the model is 65.92 %\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "sum(model_answers_train,[])[0].split(\"explanation:\")[0]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "id": "T8IqeJvpUC26",
        "outputId": "7338e462-5128-4e31-e37a-bf92e77d19fd"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "\"answer: ['douglas adams']\""
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 92
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model_answers_train_overall,target_answers_train_overall=sum(model_answers_train,[]),sum(target_answers_train,[])\n",
        "accuracy_boolean_train=[a.lower().split(\"answer:\")[1].strip()==b.lower().split(\"answer:\")[1].strip() for a,b in zip((model_answers_train_overall),(target_answers_train_overall))]\n",
        "acc_train=(sum(accuracy_boolean_train)/len(accuracy_boolean_train))*100\n",
        "print(f\"The training accuracy of the model is {acc_train:.2f} %\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "H-GspdYUX3Ow",
        "outputId": "8ff8cbca-cdca-4ddb-c7ef-14ce8a7bbd92"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The training accuracy of the model is 64.17 %\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.eval()\n",
        "valid_dataloader = DataLoader(valid_dataset, batch_size=8, collate_fn=data_collator)\n",
        "\n",
        "model_answers = []\n",
        "target_answers = []\n",
        "\n",
        "with torch.no_grad():\n",
        "    for batch in tqdm(valid_dataloader):\n",
        "        input_ids = batch['input_ids'].to(device)\n",
        "        attention_mask = batch['attention_mask'].to(device)\n",
        "        target_ids = batch['target_ids'].to(device)\n",
        "        target_attention_mask = batch['target_attention_mask'].to(device)\n",
        "        \n",
        "        outputs = model.generate(\n",
        "            input_ids=input_ids,\n",
        "            attention_mask=attention_mask,\n",
        "            # decoder_attention_mask=target_attention_mask,\n",
        "            max_length=512,\n",
        "            num_beams=10,\n",
        "            early_stopping=True\n",
        "        )\n",
        "        \n",
        "        # Convert the generated output to text\n",
        "        output_text = tokenizer.batch_decode(outputs, skip_special_tokens=True)\n",
        "        target_text = tokenizer.batch_decode(target_ids, skip_special_tokens=True)\n",
        "        model_answers.append(output_text)\n",
        "        target_answers.append(target_text)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LQwLEU8qDd2B",
        "outputId": "ec1248e9-dca5-40f4-b453-18bf3aebbfd8"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 625/625 [05:21<00:00,  1.95it/s]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "with open('./textvqa/model1_model_answers', 'wb') as handle:\n",
        "    pickle.dump(model_answers, handle, protocol=pickle.HIGHEST_PROTOCOL)\n",
        "with open('./textvqa/model1_target_answers', 'wb') as handle:\n",
        "    pickle.dump(target_answers, handle, protocol=pickle.HIGHEST_PROTOCOL)"
      ],
      "metadata": {
        "id": "JjTLPIWrDlJy"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model_answers_overall,target_answers_overall=sum(model_answers,[]),sum(target_answers,[])\n",
        "\n",
        "accuracy_boolean_valid = [a.lower().split(\"answer:\")[1].strip()==b.lower().split(\"answer:\")[1].strip() for a,b in zip((model_answers_overall),(target_answers_overall))]\n",
        "acc_train=(sum(accuracy_boolean_valid)/len(accuracy_boolean_valid))*100\n",
        "print(f\"The validation accuracy of the model is {acc_train:.2f} %\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iLJa_SgCDosS",
        "outputId": "f181842d-1e74-472f-8dd9-1d97ab1c2d01"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The validation accuracy of the model is 62.66 %\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "u-zpBoqzXRQs"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}