multimodal_analysis.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "display_name": "PyCharm (MultiModal-Movie-Genre-Analysis)",
      "language": "python",
      "name": "pycharm-90cbde0b"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.7"
    },
    "colab": {
      "name": "multimodal_analysis.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true
    },
    "accelerator": "GPU",
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "61edecabeab9432ab6ce256da04f9069": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_770f821ce5494fa9b22d800ddc15d7a6",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_470e13d7b3954108bdad07a6b498e5ba",
              "IPY_MODEL_b47e45e0bccc47888339d364bb3d1415"
            ]
          }
        },
        "770f821ce5494fa9b22d800ddc15d7a6": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "470e13d7b3954108bdad07a6b498e5ba": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_f981c2e4073b444faab95985a4a1961d",
            "_dom_classes": [],
            "description": "Downloading: 100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 433,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 433,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_a844d4bc5bbe4b91864b55e7b5178a2c"
          }
        },
        "b47e45e0bccc47888339d364bb3d1415": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_a8b373092fe54af88cd5bdbe362f8ba0",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 433/433 [00:28&lt;00:00, 15.4B/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_37590395d4804e0b8fea9248ba73f8bd"
          }
        },
        "f981c2e4073b444faab95985a4a1961d": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "a844d4bc5bbe4b91864b55e7b5178a2c": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "a8b373092fe54af88cd5bdbe362f8ba0": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "37590395d4804e0b8fea9248ba73f8bd": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "64be918799fe43f49881c30d9726602b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_fee8f9ed429644eabffd9b9724b27fc2",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_49f864b6c1a147769f9ef75b97600c41",
              "IPY_MODEL_1343b8904bf24584b0536024df9c9fcb"
            ]
          }
        },
        "fee8f9ed429644eabffd9b9724b27fc2": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "49f864b6c1a147769f9ef75b97600c41": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_52a8b8a2e9a0465da6f9f2b784f9957e",
            "_dom_classes": [],
            "description": "Downloading: 100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 440473133,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 440473133,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_8ecbcec1c16c422580bc266252ed3323"
          }
        },
        "1343b8904bf24584b0536024df9c9fcb": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_4613c6dd537f46d382fe9631e544d2a0",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 440M/440M [00:25&lt;00:00, 17.6MB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_b213394b031549f99c3a4d6b2cbc08bf"
          }
        },
        "52a8b8a2e9a0465da6f9f2b784f9957e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "8ecbcec1c16c422580bc266252ed3323": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "4613c6dd537f46d382fe9631e544d2a0": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "b213394b031549f99c3a4d6b2cbc08bf": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "76479bb53ab2494aa62ba6ed72daa029": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_c20db9b38bd445c5bef09c6ce7047feb",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_eb90697c571648568a82e608077bb1d3",
              "IPY_MODEL_87ff9f5188904c3bbf1e65877867e4df"
            ]
          }
        },
        "c20db9b38bd445c5bef09c6ce7047feb": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "eb90697c571648568a82e608077bb1d3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_d8271404acf3416ea2c667becbadc79f",
            "_dom_classes": [],
            "description": "100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 46827520,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 46827520,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_605fc8339a1149779748ee1f17ad1253"
          }
        },
        "87ff9f5188904c3bbf1e65877867e4df": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_283f895a54be4e1aa404c43f2a0d5b6d",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 44.7M/44.7M [00:00&lt;00:00, 61.1MB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_871bf420589442f3bac2c1e3aa8d3327"
          }
        },
        "d8271404acf3416ea2c667becbadc79f": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "605fc8339a1149779748ee1f17ad1253": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "283f895a54be4e1aa404c43f2a0d5b6d": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "871bf420589442f3bac2c1e3aa8d3327": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "6705e37e708b4c4fae7920646dadffc7": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_81dab32e9c184e25bf6f1fb59dcf9b9f",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_aea9df890ea040f8bc910cc0da04df55",
              "IPY_MODEL_6ca2dbc448b447758511a32b85b0bfc0"
            ]
          }
        },
        "81dab32e9c184e25bf6f1fb59dcf9b9f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "aea9df890ea040f8bc910cc0da04df55": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_b1f6f39e8d664da39ff4de94566f53bd",
            "_dom_classes": [],
            "description": "Downloading: 100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 231508,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 231508,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_b7ea70a545ca42ab8786c354eb9b0550"
          }
        },
        "6ca2dbc448b447758511a32b85b0bfc0": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_a727c7d783e1423b9ac89e9beab0fcab",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 232k/232k [00:00&lt;00:00, 789kB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_c8dc266c76304c368de4e5ee9cf94172"
          }
        },
        "b1f6f39e8d664da39ff4de94566f53bd": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "b7ea70a545ca42ab8786c354eb9b0550": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "a727c7d783e1423b9ac89e9beab0fcab": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "c8dc266c76304c368de4e5ee9cf94172": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PmzejPg62ikF"
      },
      "source": [
        "### Training Gated MultiModal Unit"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pO3BWcty2ikF"
      },
      "source": [
        "###### load pretrained BERT model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7kzYtj9DLKlK",
        "outputId": "701900ff-8c6a-46fa-b97a-4c032b6ba206"
      },
      "source": [
        "!wget -nc http://www.cs.virginia.edu/~vicente/vislang/mmimdb-256.tar.gz\n",
        "!tar xf mmimdb-256.tar.gz -C ./data/mmimdb-256/"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2020-12-04 00:52:09--  http://www.cs.virginia.edu/~vicente/vislang/mmimdb-256.tar.gz\n",
            "Resolving www.cs.virginia.edu (www.cs.virginia.edu)... 128.143.67.11\n",
            "Connecting to www.cs.virginia.edu (www.cs.virginia.edu)|128.143.67.11|:80... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 643479473 (614M) [application/x-gzip]\n",
            "Saving to: ‘mmimdb-256.tar.gz’\n",
            "\n",
            "mmimdb-256.tar.gz   100%[===================>] 613.67M  37.8MB/s    in 17s     \n",
            "\n",
            "2020-12-04 00:52:26 (36.5 MB/s) - ‘mmimdb-256.tar.gz’ saved [643479473/643479473]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UG8ZLHsZ2ikF",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2476f0d0-b0dc-40f0-8a16-c47a829e985e"
      },
      "source": [
        "!pip install livelossplot --quiet\n",
        "!pip -q install transformers"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[K     |████████████████████████████████| 1.4MB 7.8MB/s \n",
            "\u001b[K     |████████████████████████████████| 890kB 16.5MB/s \n",
            "\u001b[K     |████████████████████████████████| 2.9MB 43.0MB/s \n",
            "\u001b[?25h  Building wheel for sacremoses (setup.py) ... \u001b[?25l\u001b[?25hdone\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VkbOF-ZV2ikF"
      },
      "source": [
        "import torch, os, json\n",
        "import matplotlib.pyplot as plt\n",
        "from transformers import BertTokenizer\n",
        "from transformers import BertForSequenceClassification\n",
        "from PIL import Image\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZE7wkB5CxUfW"
      },
      "source": [
        "### Load Pretrained Bert Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 220,
          "referenced_widgets": [
            "61edecabeab9432ab6ce256da04f9069",
            "770f821ce5494fa9b22d800ddc15d7a6",
            "470e13d7b3954108bdad07a6b498e5ba",
            "b47e45e0bccc47888339d364bb3d1415",
            "f981c2e4073b444faab95985a4a1961d",
            "a844d4bc5bbe4b91864b55e7b5178a2c",
            "a8b373092fe54af88cd5bdbe362f8ba0",
            "37590395d4804e0b8fea9248ba73f8bd",
            "64be918799fe43f49881c30d9726602b",
            "fee8f9ed429644eabffd9b9724b27fc2",
            "49f864b6c1a147769f9ef75b97600c41",
            "1343b8904bf24584b0536024df9c9fcb",
            "52a8b8a2e9a0465da6f9f2b784f9957e",
            "8ecbcec1c16c422580bc266252ed3323",
            "4613c6dd537f46d382fe9631e544d2a0",
            "b213394b031549f99c3a4d6b2cbc08bf"
          ]
        },
        "id": "qRweBu7M2ikF",
        "outputId": "d9348b82-381f-47da-baac-0241566a5942"
      },
      "source": [
        "text_model = BertForSequenceClassification.from_pretrained('bert-base-uncased', \n",
        "    num_labels = 27,  output_attentions = False, \n",
        "    output_hidden_states = False);"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "61edecabeab9432ab6ce256da04f9069",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=433.0, style=ProgressStyle(description_…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "64be918799fe43f49881c30d9726602b",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=440473133.0, style=ProgressStyle(descri…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "Some weights of the model checkpoint at bert-base-uncased were not used when initializing BertForSequenceClassification: ['cls.predictions.bias', 'cls.predictions.transform.dense.weight', 'cls.predictions.transform.dense.bias', 'cls.predictions.decoder.weight', 'cls.seq_relationship.weight', 'cls.seq_relationship.bias', 'cls.predictions.transform.LayerNorm.weight', 'cls.predictions.transform.LayerNorm.bias']\n",
            "- This IS expected if you are initializing BertForSequenceClassification from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n",
            "- This IS NOT expected if you are initializing BertForSequenceClassification from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n",
            "Some weights of BertForSequenceClassification were not initialized from the model checkpoint at bert-base-uncased and are newly initialized: ['classifier.weight', 'classifier.bias']\n",
            "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "goWyCZ5m2ors"
      },
      "source": [
        "text_model.cuda();"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "llvYp8tg2ikG",
        "outputId": "edaf1998-e45b-41de-d84c-237edca75dea"
      },
      "source": [
        "text_model.load_state_dict(torch.load('best_model_bert.pth'))"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<All keys matched successfully>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HGowlcQtxdGs"
      },
      "source": [
        "### Load Pretrained Resnet "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JpU5BjA1x_va",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 103,
          "referenced_widgets": [
            "76479bb53ab2494aa62ba6ed72daa029",
            "c20db9b38bd445c5bef09c6ce7047feb",
            "eb90697c571648568a82e608077bb1d3",
            "87ff9f5188904c3bbf1e65877867e4df",
            "d8271404acf3416ea2c667becbadc79f",
            "605fc8339a1149779748ee1f17ad1253",
            "283f895a54be4e1aa404c43f2a0d5b6d",
            "871bf420589442f3bac2c1e3aa8d3327"
          ]
        },
        "outputId": "1422f255-6f51-4042-bbec-b144cabb2691"
      },
      "source": [
        "\n",
        "from torchvision import models\n",
        "\n",
        "image_model = models.resnet18(pretrained=True) \n",
        "image_model.cuda()\n",
        "num_ftrs = image_model.fc.in_features\n",
        "image_model.fc = nn.Linear(num_ftrs, 27)"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading: \"https://download.pytorch.org/models/resnet18-5c106cde.pth\" to /root/.cache/torch/hub/checkpoints/resnet18-5c106cde.pth\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "76479bb53ab2494aa62ba6ed72daa029",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, max=46827520.0), HTML(value='')))"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EfjdqgBgIwgg",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d624589a-9f80-4b46-ef10-cab06574818d"
      },
      "source": [
        "image_model.load_state_dict(torch.load('best_img_model.pth'))"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<All keys matched successfully>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oBsfg7dAd9A5"
      },
      "source": [
        "Equations governing GMU are as follows: <br>\n",
        "$h_{v} = tanh(W_{v} \\cdot x_{})$ <br>\n",
        "$h_{t} = tanh(W_{t} \\cdot x_{t})$ <br>\n",
        "$z = \\sigma(W_{z} \\cdot[x_{v} \\cdot x_{t}])$ <br>\n",
        "$h = z \\ast h_{v} + (1 - z) \\ast h_{t}$ <br>\n",
        "$\\Theta = \\{W_{v}, W_{t}, W_{z}\\}$ <br>\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "d58fFE-6GrmD"
      },
      "source": [
        "\n",
        "\n",
        "# used to find h_v and h_t\n",
        "class LinearClassifier(nn.Module):  \n",
        "  def __init__(self, encoding_size):\n",
        "    super(LinearClassifier, self).__init__()\n",
        "    self.linear = nn.Linear(encoding_size, encoding_size)\n",
        "\n",
        "  def forward(self, x):\n",
        "    x = self.linear(x)\n",
        "    x = torch.tanh(x)\n",
        "    return x\n"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZnodpjBXdkLm"
      },
      "source": [
        "class LinearCombine(nn.Module):\n",
        "  def __init__(self, encoding_size):\n",
        "    super().__init__()\n",
        "    self.linear = nn.Linear(encoding_size, 27) # 27 is the number of genres\n",
        "\n",
        "    self.sigmoid = nn.Sigmoid()\n",
        "\n",
        "  def forward(self, x, y):\n",
        "    x = self.linear(torch.cat((x, y), dim=1))\n",
        "    x = self.sigmoid(x)\n",
        "    return x\n"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-TPGQPOMg-q7"
      },
      "source": [
        "class Gated_MultiModal_Unit():\n",
        "\n",
        "  def __init__(self, img_model, text_model):\n",
        "    super().__init__()\n",
        "    self.bert_model = text_model\n",
        "    self.resnet_model = img_model\n",
        "    self.hv_gate = LinearClassifier(27) # num of categories\n",
        "    self.ht_gate = LinearClassifier(27) # num of categories\n",
        "    self.z_gate = LinearCombine(54) # 2* num of categories, for data fusion\n",
        "\n",
        "\n",
        "\n",
        "  def forward(self, mode,imgs, texts, text_masks, labels):\n",
        "    if (mode == 'train'):\n",
        "      self.hv_gate.train()\n",
        "      self.ht_gate.train()\n",
        "      self.z_gate.train()\n",
        "\n",
        "    else:\n",
        "      self.hv_gate.eval()\n",
        "      self.ht_gate.eval()\n",
        "      self.z_gate.eval()\n",
        "\n",
        "    self.resnet_model.eval()\n",
        "    self.bert_model.eval()\n",
        "    self.resnet_model.cuda()\n",
        "    self.bert_model.cuda()\n",
        "    img_pred = self.resnet_model(imgs)\n",
        "    text_pred = self.bert_model(texts, text_masks).logits\n",
        "\n",
        "    \n",
        "    self.hv_gate.cuda()\n",
        "    self.ht_gate.cuda()\n",
        "    self.z_gate.cuda()\n",
        "\n",
        "\n",
        "    h_v = self.hv_gate(img_pred)\n",
        "    h_t = self.ht_gate(text_pred)\n",
        "\n",
        "    z = self.z_gate(img_pred, text_pred)\n",
        "\n",
        "    h = (torch.mul(z, h_v)) + (torch.mul((1-z), h_t))\n",
        "\n",
        "    predictions = h\n",
        "    return predictions\n",
        "\n",
        "\n"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "t1pAD8y9LXCb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 270,
          "referenced_widgets": [
            "6705e37e708b4c4fae7920646dadffc7",
            "81dab32e9c184e25bf6f1fb59dcf9b9f",
            "aea9df890ea040f8bc910cc0da04df55",
            "6ca2dbc448b447758511a32b85b0bfc0",
            "b1f6f39e8d664da39ff4de94566f53bd",
            "b7ea70a545ca42ab8786c354eb9b0550",
            "a727c7d783e1423b9ac89e9beab0fcab",
            "c8dc266c76304c368de4e5ee9cf94172"
          ]
        },
        "outputId": "bb6f1be2-300a-404c-f3db-2a2638d80a65"
      },
      "source": [
        "class MovieDataset(torch.utils.data.Dataset):\n",
        "    def __init__(self, folder = 'data/mmimdb-256/dataset-resized-256max', split = 'dev',\n",
        "                 image_transform = None):\n",
        "        self.json_dir = os.path.join(folder, split, 'metadata')\n",
        "        self.image_dir = os.path.join(folder, split, 'images')\n",
        "        self.image_transform = image_transform\n",
        "        self.tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')\n",
        "\n",
        "        # Category definitions of movies.\n",
        "        self.categories = ['Action', 'Adult', 'Adventure', 'Animation', 'Biography', \n",
        "                           'Comedy', 'Crime', 'Documentary', 'Drama', \n",
        "                           'Family', 'Fantasy', 'Film-Noir', 'History', \n",
        "                           'Horror', 'Music', 'Musical', 'Mystery', 'News', \n",
        "                           'Reality-TV', 'Romance', 'Sci-Fi', 'Short', \n",
        "                           'Sport', 'Talk-Show', 'Thriller', 'War', 'Western']\n",
        "        self.categories2ids = {category: id for (id, category) \n",
        "                               in enumerate(self.categories)}\n",
        "\n",
        "        # Load JSON files.\n",
        "        print('Loading %s ...' % self.json_dir, end = '')\n",
        "        fdir = os.listdir(self.json_dir)\n",
        "        self.metadata = [(fname[:-5], json.load(open(os.path.join(self.json_dir, fname)))) \n",
        "                     for fname in sorted(fdir) if not fname.startswith('.')]\n",
        "        print(' finished')\n",
        "        print(self.metadata)\n",
        "        # Pre-tokenizing all sentences.\n",
        "        print('Tokenizing...', end = '')\n",
        "        self.tokenized_plots = list()\n",
        "        for i in range(0, len(self.metadata)):\n",
        "            text = self.metadata[i][1]['plot'][0]\n",
        "            encoded_text = self.tokenizer.encode_plus(\n",
        "                text, add_special_tokens = True, truncation = True, \n",
        "                max_length = 256, padding = 'max_length',\n",
        "                return_attention_mask = True,\n",
        "                return_tensors = 'pt')\n",
        "            self.tokenized_plots.append(encoded_text)\n",
        "        print(' finished')\n",
        "            \n",
        "    def __getitem__(self, index: int):\n",
        "        # Load images on the fly.\n",
        "        filename, movie_data = self.metadata[index]\n",
        "        img_path = os.path.join(self.image_dir, filename + '.jpeg')\n",
        "        image = Image.open(img_path).convert('RGB')\n",
        "        text = self.tokenized_plots[index]['input_ids'][0]\n",
        "        text_mask = self.tokenized_plots[index]['attention_mask'][0]\n",
        "        genres = movie_data['genres']\n",
        "\n",
        "        if self.image_transform: image = self.image_transform(image)\n",
        "\n",
        "        # Encode labels in a binary vector.\n",
        "        label_vector = torch.zeros((len(self.categories)))\n",
        "        label_ids = [self.categories2ids[cat] for cat in genres]\n",
        "        label_vector[label_ids] = 1\n",
        "\n",
        "        return image, text, text_mask, label_vector\n",
        "\n",
        "    def load_image_only(self, index: int):\n",
        "        filename, movie_data = self.metadata[index]\n",
        "        img_path = os.path.join(self.image_dir, filename + '.jpeg')\n",
        "        image = Image.open(img_path).convert('RGB')\n",
        "        return image\n",
        "\n",
        "    def get_metadata(self, index: int):\n",
        "        _, movie_data = self.metadata[index]\n",
        "        return movie_data\n",
        "\n",
        "    def __len__(self):\n",
        "        return len(self.metadata)\n",
        "\n",
        "val_data = MovieDataset(split = 'dev')\n",
        "print('Data size: %d samples' % len(val_data))"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "6705e37e708b4c4fae7920646dadffc7",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=231508.0, style=ProgressStyle(descripti…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n",
            "Loading data/mmimdb-256/dataset-resized-256max/dev/metadata ..."
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "IOPub data rate exceeded.\n",
            "The notebook server will temporarily stop sending output\n",
            "to the client in order to avoid crashing it.\n",
            "To change this limit, set the config variable\n",
            "`--NotebookApp.iopub_data_rate_limit`.\n",
            "\n",
            "Current values:\n",
            "NotebookApp.iopub_data_rate_limit=1000000.0 (bytes/sec)\n",
            "NotebookApp.rate_limit_window=3.0 (secs)\n",
            "\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            " finished\n",
            "Data size: 2608 samples\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IzfE0W7sK4h0",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "891c981f-c8a9-4a69-ef63-5f91f38b284d"
      },
      "source": [
        "# Let's setup the data loaders and preprocessing.\n",
        "import torchvision.transforms as transforms\n",
        "\n",
        "# Batch size is a liming factor on constrained resources.\n",
        "# only GPUs with a large memory can hold large batches.\n",
        "batch_size = 10\n",
        "\n",
        "image_transform = transforms.Compose([\n",
        "        transforms.Resize(256),\n",
        "        transforms.CenterCrop((224, 224)),\n",
        "        transforms.ToTensor(),\n",
        "        transforms.Normalize(mean=[0.485, 0.456, 0.406],\n",
        "                             std=[0.229, 0.224, 0.225])])\n",
        "\n",
        "trainset = MovieDataset(split = 'train', image_transform = image_transform)\n",
        "valset = MovieDataset(split = 'dev', image_transform = image_transform)\n",
        "\n",
        "train_loader = torch.utils.data.DataLoader(trainset, \n",
        "                                           batch_size = batch_size, \n",
        "                                           shuffle = True, \n",
        "                                           pin_memory = True,\n",
        "                                           num_workers = 4)\n",
        "val_loader = torch.utils.data.DataLoader(valset, \n",
        "                                         batch_size = batch_size, \n",
        "                                         shuffle = False)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Loading data/mmimdb-256/dataset-resized-256max/train/metadata ..."
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "IOPub data rate exceeded.\n",
            "The notebook server will temporarily stop sending output\n",
            "to the client in order to avoid crashing it.\n",
            "To change this limit, set the config variable\n",
            "`--NotebookApp.iopub_data_rate_limit`.\n",
            "\n",
            "Current values:\n",
            "NotebookApp.iopub_data_rate_limit=1000000.0 (bytes/sec)\n",
            "NotebookApp.rate_limit_window=3.0 (secs)\n",
            "\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            " finished\n",
            "Loading data/mmimdb-256/dataset-resized-256max/dev/metadata ..."
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "IOPub data rate exceeded.\n",
            "The notebook server will temporarily stop sending output\n",
            "to the client in order to avoid crashing it.\n",
            "To change this limit, set the config variable\n",
            "`--NotebookApp.iopub_data_rate_limit`.\n",
            "\n",
            "Current values:\n",
            "NotebookApp.iopub_data_rate_limit=1000000.0 (bytes/sec)\n",
            "NotebookApp.rate_limit_window=3.0 (secs)\n",
            "\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            " finished\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fL0v8hMWMCgE"
      },
      "source": [
        "gmu = Gated_MultiModal_Unit(image_model, text_model)"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kb1OlJaTbmIW"
      },
      "source": [
        "def top5_accuracy(predicted, labels):\n",
        "    sorted_vals, sorted_ids = \\\n",
        "        predicted.data.sigmoid().sort(dim = 1, descending = True)\n",
        "    pred_vals = sorted_vals[:, :5] > 0.5 # Anything with sigmoid > 0.5 is 1.\n",
        "    true_vals = labels.data.gather(1, sorted_ids[:, :5]) # Find true values.\n",
        "    return (pred_vals == true_vals).sum(dim = 1) / 5.0\n",
        "\n",
        "def f_score(predicted, labels):\n",
        "    #print(predicted)\n",
        "    #print(predicted.data.sigmoid())\n",
        "    pred_vals = predicted.data.sigmoid() > 0.5\n",
        "    #print(pred_vals)\n",
        "    #print(labels)\n",
        "    true_positive = ((pred_vals==1) & (pred_vals == labels)).sum(dim = 1)\n",
        "    false_positive = ((pred_vals==1) & (pred_vals != labels)).sum(dim = 1)\n",
        "    false_negative = ((pred_vals==0) & (pred_vals != labels)).sum(dim = 1)\n",
        "    f_score = true_positive/(true_positive+(false_positive+false_negative)/2)\n",
        "    return f_score"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cvRJD5lBK5zJ"
      },
      "source": [
        "from transformers import AdamW  # optimizer that comes with this library.\n",
        "from transformers import get_linear_schedule_with_warmup\n",
        "from livelossplot import PlotLosses\n",
        "import torch.nn as nn\n",
        "\n",
        "num_epochs = 4\n",
        "\n",
        "total_steps = len(train_loader) * num_epochs\n",
        "cost_function = nn.BCEWithLogitsLoss(reduction = 'none')\n",
        "best_accuracy = 0\n",
        "\n",
        "# Training Loop. \n",
        "def train_gmu(model, optimizer, criterion,num_epochs=4):\n",
        "  best_accuracy = 0\n",
        "  liveloss = PlotLosses(); current_step = 0\n",
        "  for epoch in range(0, num_epochs):\n",
        "      cumulative_accuracy = 0\n",
        "      cumulative_fscore = 0\n",
        "      cumulative_loss = 0\n",
        "      num_samples = 0\n",
        "      logs = {}\n",
        "      # model.train()\n",
        "      # training loop\n",
        "      for (batch_id, (imgs, texts, text_masks, labels)) in enumerate(train_loader):\n",
        "          optimizer.zero_grad()\n",
        "          # Move to GPU.\n",
        "          if(torch.cuda.is_available()):\n",
        "            imgs = imgs.cuda()\n",
        "            texts = texts.cuda()\n",
        "            text_masks = text_masks.cuda()\n",
        "            labels = labels.cuda()\n",
        "\n",
        "          # Compute predictions.\n",
        "          predicted = model.forward('train', imgs, texts, text_masks, labels)\n",
        "          loss = criterion(predicted, labels)\n",
        "          loss.mean().backward()\n",
        "          optimizer.step() \n",
        "          cumulative_accuracy += top5_accuracy(predicted, labels).sum().item()\n",
        "          cumulative_loss += loss.data.sum().item()\n",
        "          f_s = f_score(predicted, labels)\n",
        "          cumulative_fscore +=f_s.sum().item()\n",
        "          num_samples += texts.size(0)\n",
        "          # print(\"Multimodal Accuracy: \", cumulative_accuracy/num_samples)\n",
        "          # print(\"Multimodal Fscore: \", cumulative_fscore/num_samples)\n",
        "\n",
        "          if batch_id % 100 == 0:\n",
        "            print(epoch, batch_id, cumulative_accuracy / num_samples)\n",
        "            logs['loss'] = cumulative_loss / num_samples\n",
        "            logs['accuracy'] = cumulative_accuracy / num_samples\n",
        "            logs['f_score'] = cumulative_fscore/num_samples\n",
        "            liveloss.update(logs)\n",
        "            liveloss.send()\n",
        "            current_step += 1\n",
        "\n",
        "      # validation loop\n",
        "      cumulative_accuracy = 0\n",
        "      cumulative_fscore = 0\n",
        "      cumulative_loss = 0\n",
        "      num_samples = 0\n",
        "      for (batch_id, (imgs, texts, text_masks, labels)) in enumerate(train_loader):\n",
        "          optimizer.zero_grad()\n",
        "          # Move to GPU.\n",
        "          if(torch.cuda.is_available()):\n",
        "            imgs = imgs.cuda()\n",
        "            texts = texts.cuda()\n",
        "            text_masks = text_masks.cuda()\n",
        "            labels = labels.cuda()\n",
        "\n",
        "          # Compute predictions.\n",
        "          predicted = model.forward('validation', imgs, texts, text_masks, labels)\n",
        "          loss = criterion(predicted, labels)\n",
        "          cumulative_accuracy += top5_accuracy(predicted, labels).sum().item()\n",
        "          cumulative_loss += loss.data.sum().item()\n",
        "          f_s = f_score(predicted, labels)\n",
        "          cumulative_fscore +=f_s.sum().item()\n",
        "\n",
        "          num_samples += texts.size(0)\n",
        "          # print(\"Multimodal Accuracy: \", cumulative_accuracy/num_samples)\n",
        "          # print(\"Multimodal Fscore: \", cumulative_fscore/num_samples)\n",
        "          if (1 + batch_id) % 100 == 0:\n",
        "            logs['val_loss'] = cumulative_loss / num_samples\n",
        "            logs['val_accuracy'] = cumulative_accuracy / num_samples \n",
        "            logs['val_f_score']  = cumulative_fscore/num_samples\n",
        "            liveloss.update(logs, current_step)\n",
        "            liveloss.send()\n",
        "            current_step += 1\n",
        "\n",
        "\n",
        "      if logs['val_accuracy'] > best_accuracy:\n",
        "        best_accuracy = logs['val_accuracy']\n",
        "        torch.save({\n",
        "            'hv_gate_state_dict': model.hv_gate.state_dict(),\n",
        "            'ht_gate_state_dict': model.ht_gate.state_dict(),\n",
        "            'z_gate_state_dict': model.z_gate.state_dict(),\n",
        "            }, 'best_gmu.pth')\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "ofd12nHpWWo4",
        "outputId": "04779841-e94b-49fd-f929-deab2b922b49"
      },
      "source": [
        "from torch import optim \n",
        "\n",
        "learning_rate = 0.003\n",
        "\n",
        "optimizer = optim.Adam(list(gmu.ht_gate.parameters()) + list(gmu.z_gate.parameters()) + list(gmu.hv_gate.parameters()), lr=learning_rate)\n",
        "\n",
        "criterion = nn.BCEWithLogitsLoss(reduction = 'none')\n",
        "# scheduler = get_linear_schedule_with_warmup(optimizer, num_warmup_steps = 0, num_training_steps = total_steps) \n",
        "\n",
        "\n",
        "# train_gmu(gmu, criterion, optimizer, scheduler, num_epochs)\n",
        "train_gmu(gmu, optimizer, criterion, 4)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x864 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "accuracy\n",
            "\ttraining         \t (min:    0.797, max:    0.880, cur:    0.848)\n",
            "\tvalidation       \t (min:    0.840, max:    0.858, cur:    0.852)\n",
            "f_score\n",
            "\ttraining         \t (min:    0.290, max:    0.513, cur:    0.378)\n",
            "\tvalidation       \t (min:    0.352, max:    0.419, cur:    0.387)\n",
            "Loss\n",
            "\ttraining         \t (min:   10.111, max:   10.613, cur:   10.536)\n",
            "\tvalidation       \t (min:   10.459, max:   10.581, cur:   10.526)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6tW6DF5RBBOJ"
      },
      "source": [
        "# Load Train set"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "J9jqcf8IBF1Z",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4f1a6a38-5a10-4d9e-eff7-967f9f240d3d"
      },
      "source": [
        "checkpoint = torch.load('best_gmu.pth')\n",
        "gmu.hv_gate.load_state_dict(checkpoint['hv_gate_state_dict'])\n",
        "gmu.ht_gate.load_state_dict(checkpoint['ht_gate_state_dict'])\n",
        "gmu.z_gate.load_state_dict(checkpoint['z_gate_state_dict'])\n"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<All keys matched successfully>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3-YD93vfBF67",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d02f7389-d954-4209-a1ef-ad869e627622"
      },
      "source": [
        "# Let's setup the data loaders and preprocessing.\n",
        "import torchvision.transforms as transforms\n",
        "\n",
        "# Batch size is a liming factor on constrained resources.\n",
        "# only GPUs with a large memory can hold large batches.\n",
        "batch_size = 10\n",
        "\n",
        "image_transform = transforms.Compose([\n",
        "        transforms.Resize(256),\n",
        "        transforms.CenterCrop((224, 224)),\n",
        "        transforms.ToTensor(),\n",
        "        transforms.Normalize(mean=[0.485, 0.456, 0.406],\n",
        "                             std=[0.229, 0.224, 0.225])])\n",
        "\n",
        "testset = MovieDataset(split = 'test', image_transform = image_transform)\n",
        "\n",
        "test_loader = torch.utils.data.DataLoader(testset, \n",
        "                                           batch_size = batch_size, \n",
        "                                           shuffle = False, \n",
        "                                           pin_memory = False,)"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Loading data/mmimdb-256/dataset-resized-256max/test/metadata ..."
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "IOPub data rate exceeded.\n",
            "The notebook server will temporarily stop sending output\n",
            "to the client in order to avoid crashing it.\n",
            "To change this limit, set the config variable\n",
            "`--NotebookApp.iopub_data_rate_limit`.\n",
            "\n",
            "Current values:\n",
            "NotebookApp.iopub_data_rate_limit=1000000.0 (bytes/sec)\n",
            "NotebookApp.rate_limit_window=3.0 (secs)\n",
            "\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            " finished\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5_9emenGBfL_"
      },
      "source": [
        "from transformers import AdamW  # optimizer that comes with this library.\n",
        "from transformers import get_linear_schedule_with_warmup\n",
        "from livelossplot import PlotLosses\n",
        "import torch.nn as nn\n",
        "\n",
        "num_epochs = 4\n",
        "\n",
        "total_steps = len(test_loader) * num_epochs\n",
        "cost_function = nn.BCEWithLogitsLoss(reduction = 'none')\n",
        "best_accuracy = 0\n",
        "\n",
        "# Training Loop. \n",
        "def test_gmu(model, criterion, num_epochs=1):\n",
        "  best_accuracy = 0\n",
        "  liveloss = PlotLosses(); current_step = 0\n",
        "  for epoch in range(0, num_epochs):\n",
        "      cumulative_accuracy = 0\n",
        "      cumulative_fscore = 0\n",
        "      cumulative_loss = 0\n",
        "      num_samples = 0\n",
        "      logs = {}\n",
        "      # model.train()\n",
        "      # training loop\n",
        "      for (batch_id, (imgs, texts, text_masks, labels)) in enumerate(test_loader):\n",
        "          # Move to GPU.\n",
        "          if(torch.cuda.is_available()):\n",
        "            imgs = imgs.cuda()\n",
        "            texts = texts.cuda()\n",
        "            text_masks = text_masks.cuda()\n",
        "            labels = labels.cuda()\n",
        "\n",
        "          # Compute predictions.\n",
        "          predicted = model.forward('eval', imgs, texts, text_masks, labels) #no operation is performed with labels\n",
        "          loss = criterion(predicted, labels)\n",
        "          cumulative_accuracy += top5_accuracy(predicted, labels).sum().item()\n",
        "          cumulative_loss += loss.data.sum().item()\n",
        "          f_s = f_score(predicted, labels)\n",
        "          cumulative_fscore +=f_s.sum().item()\n",
        "          num_samples += texts.size(0)\n",
        "          # print(\"Multimodal Accuracy: \", cumulative_accuracy/num_samples)\n",
        "          # print(\"Multimodal Fscore: \", cumulative_fscore/num_samples)\n",
        "\n",
        "          if batch_id % 100 == 0:\n",
        "            print(epoch, batch_id, cumulative_accuracy / num_samples)\n",
        "            logs['loss'] = cumulative_loss / num_samples\n",
        "            logs['accuracy'] = cumulative_accuracy / num_samples\n",
        "            logs['f_score'] = cumulative_fscore/num_samples\n",
        "            liveloss.update(logs)\n",
        "            liveloss.send()\n",
        "            current_step += 1\n",
        "      print(\"MultiModal Accuracy: \", cumulative_accuracy/num_samples)\n",
        "      print(\"MultiModal Fscore: \", cumulative_fscore/num_samples)\n"
      ],
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "a3KIySvwBkn2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "ad3f1dae-6992-445d-e77f-8a6e95e82d89"
      },
      "source": [
        "\n",
        "criterion = nn.BCEWithLogitsLoss(reduction = 'none')\n",
        "test_gmu(gmu, criterion, 1)"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x864 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "accuracy\n",
            "\taccuracy         \t (min:    0.800, max:    0.846, cur:    0.846)\n",
            "f_score\n",
            "\tf_score          \t (min:    0.374, max:    0.457, cur:    0.383)\n",
            "Loss\n",
            "\tloss             \t (min:   10.098, max:   10.635, cur:   10.564)\n",
            "MultiModal Accuracy:  0.8474676241277349\n",
            "MultiModal Fscore:  0.3805149103498012\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}