pfamClass.ipynb

{
  "metadata": {
    "kernelspec": {
      "language": "python",
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python",
      "version": "3.7.12",
      "mimetype": "text/x-python",
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "pygments_lexer": "ipython3",
      "nbconvert_exporter": "python",
      "file_extension": ".py"
    },
    "colab": {
      "name": "pfamClass.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "6wLKiWHqZa0m",
        "5xGJ9OwcoQqT"
      ]
    },
    "accelerator": "GPU",
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "9665ca77c3194922a026ff34229969cf": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "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_12c55eb9cd7043779fa73b3ca4c59853",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_27dda3640926454ebbaf6b4c75fd527e",
              "IPY_MODEL_57e9f76e59ea4a728a8c3cd5698131b7",
              "IPY_MODEL_7d26c5f48d3d405ca85b85ac9b4c51b4"
            ]
          }
        },
        "12c55eb9cd7043779fa73b3ca4c59853": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "27dda3640926454ebbaf6b4c75fd527e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_929374a66d794f0facc166d5fda58a4b",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": "Downloading: 100%",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_e108e4b0f8804ada96e8b837cf8f6989"
          }
        },
        "57e9f76e59ea4a728a8c3cd5698131b7": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_731c3767a9e74856a4b8110955ff7878",
            "_dom_classes": [],
            "description": "",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 86,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 86,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_28c7d1af50484bb4bdf9e32435666a70"
          }
        },
        "7d26c5f48d3d405ca85b85ac9b4c51b4": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_2174ad9564084f6487265c80ef463e5b",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 86.0/86.0 [00:00&lt;00:00, 2.09kB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_7a626a9c9fa04b96b62b550cb008119d"
          }
        },
        "929374a66d794f0facc166d5fda58a4b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "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"
          }
        },
        "e108e4b0f8804ada96e8b837cf8f6989": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "731c3767a9e74856a4b8110955ff7878": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "",
            "_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"
          }
        },
        "28c7d1af50484bb4bdf9e32435666a70": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "2174ad9564084f6487265c80ef463e5b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "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"
          }
        },
        "7a626a9c9fa04b96b62b550cb008119d": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "f63d5d8f8efc4b2594dafb6bb46a694a": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "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_083ab680dae645b4993b4d47e4337117",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_1f689d10cdaf430bb60421c6ba514769",
              "IPY_MODEL_7ed9f595c03c40b5916282928c0aa7a0",
              "IPY_MODEL_f22e7f71b3324fe8974eae6507d34a60"
            ]
          }
        },
        "083ab680dae645b4993b4d47e4337117": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "1f689d10cdaf430bb60421c6ba514769": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_5da1828804624d79b6f38e777a0e4473",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": "Downloading: 100%",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_be3498de053d47fdb255f00ec3f8e23a"
          }
        },
        "7ed9f595c03c40b5916282928c0aa7a0": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_80daef55c452484c804f46f8e818a467",
            "_dom_classes": [],
            "description": "",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 81,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 81,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_a4544f1a14c34d24875799ff253288a6"
          }
        },
        "f22e7f71b3324fe8974eae6507d34a60": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_33909405108845f98e7e34a5b5a26172",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 81.0/81.0 [00:00&lt;00:00, 1.90kB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_2f83b47c9ea640c39a443f70d72ab6a7"
          }
        },
        "5da1828804624d79b6f38e777a0e4473": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "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"
          }
        },
        "be3498de053d47fdb255f00ec3f8e23a": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "80daef55c452484c804f46f8e818a467": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "",
            "_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"
          }
        },
        "a4544f1a14c34d24875799ff253288a6": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "33909405108845f98e7e34a5b5a26172": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "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"
          }
        },
        "2f83b47c9ea640c39a443f70d72ab6a7": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "1c9aeb9510ed42babd7cc181bc1c2e1e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "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_4a5a8270efd747ada938646da50404ef",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_ccc927c4e59044be8dd01c8bb709361e",
              "IPY_MODEL_b4b367b2d0bd4e989754588f7d5280aa",
              "IPY_MODEL_8da64c31495045d0a34cab69614619cc"
            ]
          }
        },
        "4a5a8270efd747ada938646da50404ef": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "ccc927c4e59044be8dd01c8bb709361e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_b95308c6445c45e4b05bd5195b615805",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": "Downloading: 100%",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_ec9b284a48034eb0b5aa7ed7cd724321"
          }
        },
        "b4b367b2d0bd4e989754588f7d5280aa": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_d49a09ae993844c6a254465c80ce3f45",
            "_dom_classes": [],
            "description": "",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 112,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 112,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_c0a23e15932b47e794ce2c6da43f63af"
          }
        },
        "8da64c31495045d0a34cab69614619cc": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_e404998de5014b8f9bc0fca415f9ac81",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 112/112 [00:00&lt;00:00, 1.70kB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_4e53920906bd4a36ad5c47d5b0bfef11"
          }
        },
        "b95308c6445c45e4b05bd5195b615805": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "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"
          }
        },
        "ec9b284a48034eb0b5aa7ed7cd724321": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "d49a09ae993844c6a254465c80ce3f45": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "",
            "_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"
          }
        },
        "c0a23e15932b47e794ce2c6da43f63af": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "e404998de5014b8f9bc0fca415f9ac81": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "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"
          }
        },
        "4e53920906bd4a36ad5c47d5b0bfef11": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "3be5e9bf107b4b3c838c4c1ec178fef7": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "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_f785d264b12843daa3389fb532ed571b",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_3313aca3c1c448cd8bab6e35d460a827",
              "IPY_MODEL_4c259618846d4c71bfcaedbb8a1df884",
              "IPY_MODEL_531280b73a954f5d891c28b0dcbdde5a"
            ]
          }
        },
        "f785d264b12843daa3389fb532ed571b": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "3313aca3c1c448cd8bab6e35d460a827": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_1d617358d87648f3bc1ea89e28386128",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": "Downloading: 100%",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_130d5fc49b4f4ebe99467120280b977d"
          }
        },
        "4c259618846d4c71bfcaedbb8a1df884": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_e7a32c3559fb4d80bdda881e8f3af906",
            "_dom_classes": [],
            "description": "",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 361,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 361,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_8d9cefd9e9e64659b9222fd81b9d28c0"
          }
        },
        "531280b73a954f5d891c28b0dcbdde5a": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_d69162b4a1534a15b70d5753bc68a834",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 361/361 [00:00&lt;00:00, 10.9kB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_3fbb3e490b5440279f5e796f3f01f451"
          }
        },
        "1d617358d87648f3bc1ea89e28386128": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "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"
          }
        },
        "130d5fc49b4f4ebe99467120280b977d": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "e7a32c3559fb4d80bdda881e8f3af906": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "",
            "_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"
          }
        },
        "8d9cefd9e9e64659b9222fd81b9d28c0": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "d69162b4a1534a15b70d5753bc68a834": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "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"
          }
        },
        "3fbb3e490b5440279f5e796f3f01f451": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "c9ccd2bbeadc486bb6ce728a48b05680": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "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_81788b1f4b644908b5f8928069f51dd4",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_04dedcc60efb4605b24fb8191dc4fd50",
              "IPY_MODEL_c21decc775da470b995d5e025d6e4fb8",
              "IPY_MODEL_029bbc7f7c4a48f8ad9cc4ccca0a8a10"
            ]
          }
        },
        "81788b1f4b644908b5f8928069f51dd4": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "04dedcc60efb4605b24fb8191dc4fd50": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_bc9ea8941b0248dc84a191f020934914",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": "Downloading: 100%",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_8de2c09232544678b25c24bff32c6e53"
          }
        },
        "c21decc775da470b995d5e025d6e4fb8": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_002f0156251c454c9f05f4a2d0e43091",
            "_dom_classes": [],
            "description": "",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 1684058277,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 1684058277,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_f1e8ab181ba64d14b117d19bef20954f"
          }
        },
        "029bbc7f7c4a48f8ad9cc4ccca0a8a10": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_614d238395284901be418686d1eaedc3",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 1.57G/1.57G [01:45&lt;00:00, 34.5MB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_6fcaa36637c6450c8dba1dbe2cef05a1"
          }
        },
        "bc9ea8941b0248dc84a191f020934914": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "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"
          }
        },
        "8de2c09232544678b25c24bff32c6e53": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "002f0156251c454c9f05f4a2d0e43091": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "",
            "_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"
          }
        },
        "f1e8ab181ba64d14b117d19bef20954f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        },
        "614d238395284901be418686d1eaedc3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "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"
          }
        },
        "6fcaa36637c6450c8dba1dbe2cef05a1": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "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
          }
        }
      }
    }
  },
  "nbformat_minor": 0,
  "nbformat": 4,
  "cells": [
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "id": "DYAd0P1OK09R",
        "outputId": "95a1e6ab-92f3-49e4-a8ae-c04b440aca5d",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!mkdir pfam\n",
        "!cp drive/MyDrive/archive.zip pfam"
      ],
      "metadata": {
        "id": "PDzzsP-PN4rs"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!unzip pfam/archive.zip -d pfam/"
      ],
      "metadata": {
        "id": "lQ48m730OJjB"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import glob\n",
        "import pandas as pd\n",
        "\n",
        "def read_files(dir):\n",
        "    \"\"\"\n",
        "    -Description:\n",
        "    \n",
        "    Read all files from directory matching substring\n",
        "\n",
        "    @param str dir: directory where the files are stored\n",
        "\n",
        "    @out DataFrame: pandas dataframe containing all\n",
        "    \"\"\"\n",
        "    files = glob.glob(dir+\"/*\")\n",
        "    dfs = []\n",
        "    for file in files:\n",
        "        df = pd.read_csv(file , index_col = None)\n",
        "        dfs.append(df)\n",
        "    return pd.concat(dfs)"
      ],
      "metadata": {
        "id": "IOZLxx07O_Fx",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:23:03.172208Z",
          "iopub.execute_input": "2022-02-17T23:23:03.172807Z",
          "iopub.status.idle": "2022-02-17T23:23:03.195668Z",
          "shell.execute_reply.started": "2022-02-17T23:23:03.172712Z",
          "shell.execute_reply": "2022-02-17T23:23:03.194928Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#Read training , dev and test data separately:\n",
        "\n",
        "train = read_files(\"pfam/random_split/train\")\n",
        "dev = read_files(\"pfam/random_split/dev\")\n",
        "test = read_files(\"pfam/random_split/test\")"
      ],
      "metadata": {
        "id": "xb4ZUqlzPDGl"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install pycodestyle pycodestyle_magic\n",
        "!pip install flake8"
      ],
      "metadata": {
        "id": "Yt5EYke38jGG"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Pycodestyle can be used as an alternative to **pylint** for ensuring the use of a correct coding standard."
      ],
      "metadata": {
        "id": "tleHAKR08kxm"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "%load_ext pycodestyle_magic"
      ],
      "metadata": {
        "id": "ALS0j4dU9Aiu"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "\n",
        "---\n",
        "\n",
        "\n",
        "# Dataset Analysis\n",
        "\n",
        "Before getting started with the modelling, I will perform exploratory data analysis to get a deeper understanding of the data and find potential caveats in the data. \n",
        "\n",
        "As the number of classes is very high, it is important to check that they are contained on the validation and test set."
      ],
      "metadata": {
        "id": "iINxordWs07c"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Unique classes on train set: \" , len(train[\"family_accession\"].unique()) )\n",
        "print(\"Unique classes on dev set: \" , len(dev[\"family_accession\"].unique()))\n",
        "print(\"Unique classes on test set: \" , len(test[\"family_accession\"].unique()))"
      ],
      "metadata": {
        "id": "gXlnQWvVsQ98",
        "outputId": "cd33ef32-390e-48fb-adaa-176ae9d08e9d",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:23:18.318518Z",
          "iopub.execute_input": "2022-02-17T23:23:18.318877Z",
          "iopub.status.idle": "2022-02-17T23:23:18.528495Z",
          "shell.execute_reply.started": "2022-02-17T23:23:18.318842Z",
          "shell.execute_reply": "2022-02-17T23:23:18.527125Z"
        },
        "trusted": true,
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Unique classes on train set:  17929\n",
            "Unique classes on dev set:  13071\n",
            "Unique classes on test set:  13071\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "unique_dev = dev[\"family_accession\"].unique().tolist()\n",
        "unique_test = test[\"family_accession\"].unique().tolist()\n",
        "print(\"Classes in dev and test are the same? \" , sorted(unique_dev) == sorted(unique_test))\n",
        "\n",
        "train = train[train.family_accession.isin(unique_dev)]"
      ],
      "metadata": {
        "id": "GyWetQohuKHT",
        "outputId": "66133350-8b17-4fdf-9860-f922a60920c7",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:23:18.529977Z",
          "iopub.execute_input": "2022-02-17T23:23:18.530300Z",
          "iopub.status.idle": "2022-02-17T23:23:18.828404Z",
          "shell.execute_reply.started": "2022-02-17T23:23:18.530237Z",
          "shell.execute_reply": "2022-02-17T23:23:18.827626Z"
        },
        "trusted": true,
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Classes in dev and test are the same?  True\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Since the length of unique values for the labels in the train and dev/test set differs, we should only retain those classes present on the test set. "
      ],
      "metadata": {
        "id": "TvrEp2vOtZDz"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "(Due to limitations on the access to GPU RAM I will slice a subset of the dataset)"
      ],
      "metadata": {
        "id": "6sqS_5R0u8o7"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "counts = train[\"family_accession\"].value_counts()"
      ],
      "metadata": {
        "id": "8w4rDpm6IRPw",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:23:18.830929Z",
          "iopub.execute_input": "2022-02-17T23:23:18.831427Z",
          "iopub.status.idle": "2022-02-17T23:23:19.005680Z",
          "shell.execute_reply.started": "2022-02-17T23:23:18.831379Z",
          "shell.execute_reply": "2022-02-17T23:23:19.004911Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#Obtain the classes with over 1000 data points on the training set\n",
        "usable_classes = counts[counts > 1000].index.values.tolist()"
      ],
      "metadata": {
        "id": "8GQ0j8naMlPV",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:23:19.006954Z",
          "iopub.execute_input": "2022-02-17T23:23:19.007252Z",
          "iopub.status.idle": "2022-02-17T23:23:19.013844Z",
          "shell.execute_reply.started": "2022-02-17T23:23:19.007212Z",
          "shell.execute_reply": "2022-02-17T23:23:19.012990Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train = train[train.family_accession.isin(usable_classes)]\n",
        "dev = dev[dev.family_accession.isin(usable_classes)]\n",
        "test = test[test.family_accession.isin(usable_classes)]"
      ],
      "metadata": {
        "id": "Of6DfyCTIsZd",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:23:19.015417Z",
          "iopub.execute_input": "2022-02-17T23:23:19.015690Z",
          "iopub.status.idle": "2022-02-17T23:23:19.305686Z",
          "shell.execute_reply.started": "2022-02-17T23:23:19.015655Z",
          "shell.execute_reply": "2022-02-17T23:23:19.304931Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "NUM_CLASSES = len(usable_classes)\n",
        "\n",
        "print(\"Number of classes \" , NUM_CLASSES)"
      ],
      "metadata": {
        "id": "CBm7Plk3Y8Hy",
        "outputId": "24bb6fab-6acc-42eb-be07-be3809c3b9aa",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:23:19.309102Z",
          "iopub.execute_input": "2022-02-17T23:23:19.309324Z",
          "iopub.status.idle": "2022-02-17T23:23:19.316108Z",
          "shell.execute_reply.started": "2022-02-17T23:23:19.309293Z",
          "shell.execute_reply": "2022-02-17T23:23:19.315456Z"
        },
        "trusted": true,
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Number of classes  26\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train.head(10)"
      ],
      "metadata": {
        "id": "VsEXZDq3PvxS",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 359
        },
        "outputId": "2156ea15-e854-45db-9628-1797a4e1fb94"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-7a8bc94c-7dfe-4115-8de0-0460bbfa8a49\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>family_id</th>\n",
              "      <th>sequence_name</th>\n",
              "      <th>family_accession</th>\n",
              "      <th>aligned_sequence</th>\n",
              "      <th>sequence</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>26</th>\n",
              "      <td>HTH_1</td>\n",
              "      <td>Q2YQJ6_BRUA2/15-73</td>\n",
              "      <td>PF00126.27</td>\n",
              "      <td>.DVLRTFVAIADTG.SFSSAAN.AVFR...TPSAVSMQIKKLEE.Q...</td>\n",
              "      <td>DVLRTFVAIADTGSFSSAANAVFRTPSAVSMQIKKLEEQLGVAVFD...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>48</th>\n",
              "      <td>WD40</td>\n",
              "      <td>SWD3_SCHPO/86-124</td>\n",
              "      <td>PF00400.32</td>\n",
              "      <td>FRLECT..L....F.....GH.....YR............GI...S...</td>\n",
              "      <td>FRLECTLFGHYRGISQVKWATGSKYLASASDDKTIRIWD</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>56</th>\n",
              "      <td>Methyltransf_25</td>\n",
              "      <td>K9YKX9_CYASC/84-176</td>\n",
              "      <td>PF13649.6</td>\n",
              "      <td>VLEIGSGR.G...G..GS...FF..IA.KY.......L....S......</td>\n",
              "      <td>VLEIGSGRGGGSFFIAKYLSPQEIIGIDISQENIALSKQLYQLPNL...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>87</th>\n",
              "      <td>Acetyltransf_7</td>\n",
              "      <td>C3MCY4_SINFN/54-142</td>\n",
              "      <td>PF13508.7</td>\n",
              "      <td>END.L..II...................V................A...</td>\n",
              "      <td>ENDLIIVAELEGRIVGLAGLHRHKGRRQHAADLGMSVHDDYRRRGI...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>92</th>\n",
              "      <td>Acetyltransf_7</td>\n",
              "      <td>G8SKR1_ACTS5/49-134</td>\n",
              "      <td>PF13508.7</td>\n",
              "      <td>RTG.L..AW...................V................A...</td>\n",
              "      <td>RTGLAWVAADPDDRPVAYLIADPVDGCLHIEQVSVDRRHARQGLGR...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>102</th>\n",
              "      <td>Ham1p_like</td>\n",
              "      <td>R7G4V4_9PROT/4-189</td>\n",
              "      <td>PF01725.16</td>\n",
              "      <td>LVIASNNKGKIRE...................F.........K......</td>\n",
              "      <td>LVIASNNKGKIREFKEILGEKFEIIPMSEAGFNADVEENGSTFQEN...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>169</th>\n",
              "      <td>Lipase_GDSL_2</td>\n",
              "      <td>Q4K496_PSEF5/169-357</td>\n",
              "      <td>PF13472.6</td>\n",
              "      <td>FVGDS.LMQ.......GVAPH............................</td>\n",
              "      <td>FVGDSLMQGVAPHMANTLRKRYNVKSLNLSKQSTGLAYPSFFNWPK...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>220</th>\n",
              "      <td>Lum_binding</td>\n",
              "      <td>K9Q2U0_9CYAN/101-193</td>\n",
              "      <td>PF00677.17</td>\n",
              "      <td>TGHVDSVGSL.E..KS...VQT.A.........................</td>\n",
              "      <td>TGHVDSVGSLEKSVQTAQAWEMWFTHNRENEDTWWENIAPCLVPKG...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>267</th>\n",
              "      <td>Methyltransf_25</td>\n",
              "      <td>Q5ZSX0_LEGPH/178-272</td>\n",
              "      <td>PF13649.6</td>\n",
              "      <td>VLDIGCGY.G...R..AL...IL..MA.KT.......Y...........</td>\n",
              "      <td>VLDIGCGYGRALILMAKTYPQSQFTGYDLCDETIAHAVHSAKKNKI...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>326</th>\n",
              "      <td>Prenyltransf</td>\n",
              "      <td>H2AM99_KAZAF/73-294</td>\n",
              "      <td>PF01255.19</td>\n",
              "      <td>IMDGNRRFAK..SS....NLSLQR.GHEAGGL....TLLH.....L...</td>\n",
              "      <td>IMDGNRRFAKSSNLSLQRGHEAGGLTLLHLVYICKKLGVKCVSAYA...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-7a8bc94c-7dfe-4115-8de0-0460bbfa8a49')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-7a8bc94c-7dfe-4115-8de0-0460bbfa8a49 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-7a8bc94c-7dfe-4115-8de0-0460bbfa8a49');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ],
            "text/plain": [
              "           family_id  ...                                           sequence\n",
              "26             HTH_1  ...  DVLRTFVAIADTGSFSSAANAVFRTPSAVSMQIKKLEEQLGVAVFD...\n",
              "48              WD40  ...            FRLECTLFGHYRGISQVKWATGSKYLASASDDKTIRIWD\n",
              "56   Methyltransf_25  ...  VLEIGSGRGGGSFFIAKYLSPQEIIGIDISQENIALSKQLYQLPNL...\n",
              "87    Acetyltransf_7  ...  ENDLIIVAELEGRIVGLAGLHRHKGRRQHAADLGMSVHDDYRRRGI...\n",
              "92    Acetyltransf_7  ...  RTGLAWVAADPDDRPVAYLIADPVDGCLHIEQVSVDRRHARQGLGR...\n",
              "102       Ham1p_like  ...  LVIASNNKGKIREFKEILGEKFEIIPMSEAGFNADVEENGSTFQEN...\n",
              "169    Lipase_GDSL_2  ...  FVGDSLMQGVAPHMANTLRKRYNVKSLNLSKQSTGLAYPSFFNWPK...\n",
              "220      Lum_binding  ...  TGHVDSVGSLEKSVQTAQAWEMWFTHNRENEDTWWENIAPCLVPKG...\n",
              "267  Methyltransf_25  ...  VLDIGCGYGRALILMAKTYPQSQFTGYDLCDETIAHAVHSAKKNKI...\n",
              "326     Prenyltransf  ...  IMDGNRRFAKSSNLSLQRGHEAGGLTLLHLVYICKKLGVKCVSAYA...\n",
              "\n",
              "[10 rows x 5 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 66
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Let's explore the dataset to gather some insights about the data and the classification task. From the dataframe itself, it is important to note two columns:\n",
        "\n",
        "\n",
        "\n",
        "*   **sequence:** containing the sequence of aminoacids determining the domain, it will be used as input to the model for the classification task.\n",
        "*   **family_accession:** label that the model will predict in the multi-label classification setting.\n",
        "\n",
        "\n",
        "\n",
        "---\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "*   Considering that the task involves multi-label classification, we will first investigate the label distribution to account for imbalance in the data.\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "z0Aje6qaQvEO"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "sns.set_style(\"darkgrid\")\n",
        "\n",
        "def histplot(df, column, title):\n",
        "    \"\"\"\n",
        "    -Description: Bar plot of occurences from a column in a pandas dataframe\n",
        "\n",
        "    @param DataFrame df: dataframe containing the class labels \n",
        "    @param str column: colun to access in the dataframe\n",
        "    @param str title: title of the plot\n",
        "\n",
        "    \"\"\"\n",
        "    plt.figure(figsize=(14, 6))\n",
        "    plt.hist(df[column].value_counts().values , bins = 20 , color = \"lightskyblue\")\n",
        "    plt.title(title)\n",
        "    plt.xlabel(\"Samples per Class\")\n",
        "    plt.ylabel(\"Counts\")\n",
        "    plt.xticks(rotation=45)\n"
      ],
      "metadata": {
        "id": "tDMfNDmtRs03"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "histplot(train , \"family_accession\" , \"Class label counts on training set\")"
      ],
      "metadata": {
        "id": "7rs3XrMtSrMg",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 451
        },
        "outputId": "5ff34de3-f7a2-4ee4-de8f-1c7c9feb5e88"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1008x432 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "histplot(dev , \"family_accession\" , \"Class label counts on dev set\")"
      ],
      "metadata": {
        "id": "vC6j-JWLVaxF",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 444
        },
        "outputId": "dcdca31e-c91a-4c55-86ed-2cecfd3c7bfe"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1008x432 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "histplot(test , \"family_accession\" , \"Class label counts on test set\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 444
        },
        "id": "8dVfjDwe1i2f",
        "outputId": "59fa7760-1068-4463-9f5a-35b46af1575e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1008x432 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The distribution of the classes labels is highly heterogeneous, with some classes overly represented in the dataset while other are barely represented on the training set. \n",
        "\n",
        "This must be taken into account when measuring the accuracy of the model, with the possibility to include a weighting factor as an evaluation metric."
      ],
      "metadata": {
        "id": "WsWyZDrzwfvi"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def plotChainLength(df , column , title):\n",
        "    \"\"\"\n",
        "    -Description: Plot the distribution of length of the protein sequences\n",
        "\n",
        "    @param DataFrame df: dataframe containing the class labels \n",
        "    @param str column: colun to access in the dataframe\n",
        "    @param str title: title of the plot\n",
        "\n",
        "    \"\"\"\n",
        "    lenSeq = df.apply(lambda row: len(row[column]), axis=1)\n",
        "    sns.kdeplot(data = lenSeq, shade=True , color = \"lightskyblue\")\n",
        "    plt.title(\"Sequence Length Density Distribution in \" + title + \" set\")\n",
        "    plt.xlabel(\"Sequence Length\")\n",
        "    plt.ylabel(\"Density\")\n",
        "    return "
      ],
      "metadata": {
        "id": "_uVo4G8EUBrD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plotChainLength(train , \"sequence\" , \"training\")"
      ],
      "metadata": {
        "id": "NN3Z4QNmV_3C",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 427
        },
        "outputId": "c3c271d8-ae67-4ecd-9054-428182020219"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAvoAAAGaCAYAAAB+A+cSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXgT1f4/8PdMmnSHspRFkIJKytKNtSxFFqkU2VqhgCy9LIKVH9ILioDoVdELKiAoiwiKILJDAREEAReQVdl6kYJYBIu1JbZ0S5e0yfz+qMm3IWmblLZZ+n49jw/2zJmZM8nM5JOTz5wjSJIkgYiIiIiInIpo6wYQEREREVHVY6BPREREROSEGOgTERERETkhBvpERERERE6IgT4RERERkRNioE9ERERE5IQY6BMRVSA+Ph7+/v44e/asrZtS4xzp2P39/TF37lyb79MW7bDlfksbP348+vXrZ9M2PIh+/fph/PjxlV7fka4Xqh1cbN0AopqQnJyMtWvX4qeffsJff/0FhUKBhg0bIigoCFFRUejWrZutm+iUxo8fjytXruDixYu2bkqFEhMTcfToUURFRaF58+ZVvv07d+7giSeeMPwtCAI8PT3RsGFDtGvXDk8++STCw8Ph4mL/t+Xqfq369euHP//80/C3h4cH6tati9atW6NXr16IjIxEnTp1qmx/R48eRWJiIl544YUq22Z1yM7OxsaNG9G1a1eEhobaujk1ylHeI2dV3dc8VR/7/0QhekD/+9//MH78eLi4uCAyMhKPPfYYCgoKcPv2bZw8eRKenp4M9AmJiYlYuXIlunbtWq0fZD179sSwYcMAAHl5eUhOTsb333+PgwcPon379li5ciUeeuihatu/tYYNG4ZBgwZBLpcbymritWrSpAlmzZoFANBoNEhLS8O5c+fw3//+F2vWrMHSpUvRvXt3o3USEhIgitb/UH306FHs2bOnUkFkZfdZGdnZ2Vi5ciWmT59uNtCvybaU5dNPP62W7T7Ie2SNQ4cOPdD65q4XZ1BT90eqegz0yemtWrUK+fn52LdvH9q0aWOyXKVS2aBVVFu1bNnSEOjrvfzyy9iwYQMWLVqE5557Dnv27LGbnn2ZTAaZTFbj+/X29jZ5nQDg3LlzeP755zFt2jTs3bsXfn5+hmWurq410raCggK4uLjAxcWlxvZpCXtoi0KhsHUTABi/R9Z40Pbb6nohKgtz9Mnp3bp1Cz4+PmaDfADw9fU1KTt16hQmTZqEzp07IzAwEEOGDMHWrVvNrr9jxw5EREQgICAA4eHh2LBhA3bv3m2Spzl37lz4+/ub3UZZubUHDx7EM888gw4dOiA4OBjR0dFme5z061+8eBHjxo1DSEgIQkNDMX/+fKjVapP6KpUKb7/9Np544gkEBASge/fumDhxIk6ePGlU79atW5g9ezbCwsIQEBCAfv364d1330VeXp7Z46gsSZKwZcsWPP300wgODkaHDh0wfvx4nDlzxqjenTt34O/vjxUrVuC7777D8OHDERgYiLCwMLz77rsoLi422fbhw4cxdOhQBAYGok+fPli5ciVOnToFf39/xMfHAwBWrFiBefPmAQBiYmLg7+9v9j3R6XT49NNP0b9/fwQEBGDAgAHYs2dPlbwGEyZMwJAhQ/Drr7/iwIEDRss0Gg3WrFmDQYMGITAwEJ07d0ZsbCyuXr1qVO/s2bOG49q9ezcGDRqEgIAA9O3bF+vWrTPZ54ULF/Dss8+iZ8+eCAwMRK9evTBlyhRcunTJUOf+nOPyXqsjR47A398fO3bsMHuMgwYNQnh4OCRJqvTr1LVrV8ydOxd5eXlYu3at0TJz79n333+PcePGITQ0FEFBQejTpw+mT5+O33//HUBJepn+PdQfS+lzQ3/dZmRkYN68eejRowdCQkKQmppa5j71Tp06hZEjRyI4OBg9e/bE22+/bXI9WnpfOHv2rCH1a+XKlYZ2ls6HL6stO3fuRFRUFIKCgtCpUydMmjQJP//8c5n7s/Q+Yo65HH19WVpaGmbNmoUuXbogODgYkydPNrwPFW3zQd6jzZs3Y9KkSejVqxcCAgIQFhaGl156CXfu3DHZl7kcfX1ZUlISpk6dig4dOqBTp06YMWOGSUeRuRx9fdnp06ctun9otVqsWrUKffv2NXz+HDx4ECtWrIC/v7/Zdt/PkmsbAHJycrB48WKEh4cjICAA3bp1w6xZs5CcnGyoY+n9keyTfXQZEVWjFi1a4Pfff8c333yDJ598ssL627dvx+uvv46QkBDExsbC3d0dp06dwhtvvIE//vgDc+bMMdTV98K2adMGs2bNQn5+PtavX48GDRo8cLuXLVuGNWvWoFevXoiLi4Moijhy5Aji4uLwn//8B2PHjjWqn5iYiNjYWDz99NMYPHgwzp07h127dkEURbz11luGenfu3MEzzzyD9PR0DBs2DAEBAcjPz8fly5dx6tQp9OzZEwBw5coV/Otf/0KdOnUwatQoNG7cGNeuXcOmTZtw8eJFbNq0qcp+np49ezYOHDiAAQMG4Omnn4ZGo8H+/fsxadIkrFixwii3HQB++OEHbNmyBaNHj8bw4cNx7NgxrF+/HnXr1kVsbKyh3sGDBzFr1iy0aNEC06dPh0wmw969e/Htt98abS88PBwqlQrbt29HbGwsHnnkEQAl505py5YtQ0FBAUaNGgWFQoGtW7di7ty5aNGiBTp16vTAr0N0dDT279+PH374wdCbXVRUhMmTJ+PixYsYNmwYxo4di9zcXOzYsQPPPPMMvvjiCwQGBhptZ9u2bfj7778xYsQI1KlTB19++SWWLFmCJk2aYMiQIQCAmzdvYtKkSWjYsCFiYmLQoEEDpKen4/z587h27RpCQkLMtrG81yowMBC+vr7YvXs3Ro4cabTepUuX8Ntvv2HmzJkQBOGBXqdhw4ZhwYIF+OGHH8qtp+/9b926NZ577jl4e3vj7t27OH36NP744w+0atUKsbGx0Ol0+Pnnn/Hee+8Z1u3YsaPRtiZOnIiGDRti2rRpyMvLg4eHR7n7/uWXX3D48GFER0dj2LBhOHv2LDZt2oQbN27gs88+szrF5tFHH8W8efOwaNEihIeHIzw8HADg6elZ7nqLFy/GJ598gqCgIMyaNctw7vzrX//C6tWr0bt3b6P6lt5HrJWXl4dx48YhODgYM2fOxJ07d/D5559j2rRp+Oqrr8rtBX/Q92j9+vUICQnB+PHj4ePjg19//RW7du3CmTNnsH//ftSrV6/C9qelpSEmJgb9+/fHyy+/jGvXrmH79u3Izc3F+vXrLXoNLL1/LFiwANu2bUNoaCgmTZqEjIwMvPnmm2jWrJlF+7H02s7JycHo0aORkpKC4cOHo3Xr1lCpVNiyZQuio6Oxe/duNGvWzOL7I9kpicjJXbhwQWrfvr2kVCqlJ598Upo7d660efNm6bfffjOpm5aWJgUEBEizZs0yWfbWW29Jbdq0kf744w9JkiQpKytLCg4OlgYOHCjl5eUZ6v31119SSEiIpFQqpTNnzhjK58yZIymVSrNtVCqV0pw5cwx/X7lyRVIqldLSpUtN6j7//PNShw4dpJycHKP1/f39pUuXLhnVnTJlitSuXTspNzfXUPbss89KSqVSOn78uMm2tVqt4f+HDBkiDRgwwGg/kiRJ33zzjaRUKqXdu3ebPZbSxo0bJ4WEhJRbR7+9bdu2GZUXFRVJUVFRUt++fSWdTidJkiQlJydLSqVSCg4OlpKTkw11dTqdNGjQIKlnz55G64eFhUndu3eXMjMzDeW5ublSv379TI5h9+7dJu/Z/cuGDRsmFRYWGspTU1Ol9u3bSzNnzqzwtdC3/c033yyzzr179ySlUilFRUUZyj777DOz71dOTo7Uu3dvady4cYayM2fOSEqlUurZs6eUnZ1tKM/Ly5NCQ0OlkSNHGso2btwoKZVK6fLly+W229zrUt5rtXTpUkmpVEo3btwwKp8/f77Utm1bKTU1tdz9SZIk9e3bVxo0aFC5dQYPHiwplUqT66D0dbRw4UJJqVRKf//9d7nbKu/a1C978cUXzS6/f5/6MqVSKR05csSo/K233pKUSqX01VdfWbTv+7etP4c+/PBDi+onJSVJ/v7+0ujRo03O206dOkl9+/aViouLjda39D5SlnHjxkl9+/Y1KVMqldLatWuNytetW1fmveh+D/IeqdVqk7JTp06ZbVPfvn2Nril9mVKplA4cOGBU/sYbb0hKpVJKSkoylJV3vVhy//j1118lpVIpTZo0yeh+fO3aNalNmzaSUqk0uveZY+m1/dZbb0mBgYFSYmKiUfmdO3ekDh06GJ1L5V3zZN+YukNOr0OHDti9ezeioqKQk5OD+Ph4vPnmm3jqqacwduxYo58oDx8+DI1GgxEjRiAjI8Pov379+kGn0+HUqVMAgB9//BH5+fkYO3Ys3N3dDdso3WtaWfv374cgCIiMjDTbDrVabfITbEhICIKDg43KunXrhuLiYsMIJpmZmThx4gR69eqFXr16mexX38t4/fp1XL9+HYMHD4ZGozHaf6dOneDh4WGS5lNZX375JTw9PdG/f3+j/WRnZxtGX7l165bROk888YTRA2GCICA0NBQqlcqQYvDLL7/g7t27iIqKQt26dQ11PT09MXr06Eq1dcyYMUY5vI0bN0arVq1M2ldZXl5eAIDc3FxD2ZdffolHHnkE7du3N3p9NBoNevTogfPnz6OgoMBoO8OHD4e3t7fhb3d3d4SEhBi1U7/82LFjKCwsrJL2AyW/SgiCgF27dhnK8vLycPDgQTz++ONo3LhxlexH/1qVl1KiP8bDhw+bTeuyxuTJk62q36pVK/Tv39+obOrUqQCAI0eOPFBbLHXs2DFIkoRnn33W5Lx9+umn8eeff5qkf1lyH6kMURQRExNjsl0AuH37dqW3W1pZ75G+Z1+n0yEnJwcZGRnw9/eHt7c3EhISLNp2o0aN8NRTTxmVWdt+S+4f3333HYCSFJnSv/r4+/sjLCzMov1Ycm1LkoT9+/ejS5cuaNSokdG9RX+/+PHHHy3aH9k3pu5QreDv74933nkHAPDnn3/ip59+ws6dO/Hzzz9j2rRp2L17NxQKBZKSkgCU5EuX5e+//wYAQ56k/mfM0h599NEHam9SUhIkScLAgQMrbIfeww8/bFLHx8cHQEmADwB//PEHJElCu3btKtw/UJKbuWLFCov2X1lJSUlQq9Xo0aNHmXXS09PRqlUrw98VHaunp6fh/Sm9np65MkuUtd8HCYBK0wf4+iAWKHl9CgoKTEaYKe3evXto2rSp4W9zo2L4+PgYzgOgJF/+yy+/xJo1a7BhwwYEBwcjLCwMgwYNsjhFwJyHH34YPXr0wL59+/Diiy9CLpfj66+/hlqtxogRIyq93fvpX6vyUlfGjh2LY8eO4c0338SSJUvQqVMn9OrVC4MHD0b9+vWt2l/Lli2tqm/uHtCoUSPUqVPHqHOhOumvgdatW5ss05clJycbpX5Zch+pjEaNGpk8LFwV2y2trPfo9OnTWL16NS5fvmwS+GZlZVm07ap4XSy5f5T3udKqVSscP368wv1Ycm1nZGQgMzMTP/74Y5n3FluP4ERVg4E+1TrNmjVDs2bNMGzYMIwZMwYXLlxAQkICOnfubHhI8N1330WjRo3Mrm/uZm2JsvKSzfU0SpIEQRCwbt26MnNXH3vsMaO/y8txlSr58KP+ATZzqmocc0mSUL9+fSxdurTMOvcHKtVxrJao7g++69evAzD+IiJJEpRKpeFhOHPuD1otGfVDoVDgs88+Q0JCAk6cOIGff/4ZH374IVauXImlS5cacsArY+TIkYiLi8O3336LAQMGYNeuXfD19UWfPn0qvc3SNBoNbt26BV9fX6MvRferV68edu3ahZ9//hmnTp3CTz/9hEWLFmHFihVYu3YtOnToYPE+S/9qV5WsuS/UhOq6tmrimjX3HiUkJGDy5Mlo0aIFXnzxRTRv3hxubm4QBAEzZ860eN9V0f6aCpwtubb1be7RowemTJlSI+0i22CgT7WWIAgIDg7GhQsXcPfuXQD/1yNUr169cnuYgf/rNb1586ZJj4i+R7w0ffpIZmamoScIgNnevZYtW+LEiRN46KGHHvjXgdJatGgBQRCQmJhYbj39kIWiKFb4OjwoPz8/3Lp1C8HBwRU+WGgNfc+VuVE9zJU96AOiVWHnzp0AYPSApJ+fH+7du4du3bpVS6AQFBSEoKAgAMBff/2FyMhILF++vNxAv6LX6oknnkCDBg2wa9cutG7dGhcuXMCUKVOqbMjQffv2QaPRWPTFQSaTITQ01DDu/LVr1zB8+HB89NFHhlF7quO9N3cPuHv3LrKzs406C6y5L1jbTv1+bty4YfLg5G+//WZUx95V9j366quvoNVqsW7dOqNjzcvLQ3Z2dlU1r8qU/ly5/72xZISi0sq7tuvXr486deogNzfXonu8PdwfqXL4uww5vZMnT5rtHSsoKDDkmeuD6YEDB0KhUGDFihUmec9AySgFGo0GQMnER25ubti8eTPy8/MNdVJTU7F//36TdfVfIvQ5/nqfffaZSd2hQ4cCAN5//31otVqT5ZVNm/Hx8cHjjz+O48ePm7QD+L+eqXbt2kGpVGLbtm1mA47i4uIq+7k9MjISOp0O77//vtnllT3WgIAA+Pr6Ys+ePUY/z6vVamzbts2kvj6P19Kf8qvaxo0bsX//fvj7+xvlAkdGRkKlUpk9T4DKvz4ZGRkmZU2aNEH9+vUrfA0qeq3kcjmioqLw448/YtWqVQBQZWk7586dwzvvvANPT08899xz5dY1d4yPPPIIXF1djdquP56qOqeBkqDs6NGjRmX6IU5L5+5bc1+w9hzt168fBEHAp59+iqKiIkP53bt3ER8fj2bNmlWYxmcvKvseldUT//HHH0On0z1wu6pa3759AQCff/65UfuuX79ucc68Jde2KIoYMmQIEhISypwkLD093fD/tr4/UuWxR5+c3qJFi5CZmYl+/fpBqVTCzc3NEIzfunULkZGRhnGsmzRpgjfeeAOvvvoqnnrqKQwdOhTNmjVDRkYGfv31Vxw9ehQHDhxA8+bNUbduXcTFxeHdd9/F6NGjERkZifz8fGzbtg0tW7Y0echt8ODBWLZsGf7zn//g5s2b8PHxwYkTJ3Dv3j2TNgcFBeGFF17AihUrEBkZiQEDBqBx48a4e/cufvnlFxw/fhxXrlyp1Ovx2muv4erVq5gyZQoiIyPRvn17FBYW4vLly2jWrBlmz54NQRDw3nvv4V//+heGDh2K4cOHG80ofOTIEcyaNQtPP/10hfsrKirC6tWrzS578sknERERgaeffhpffPEFfvnlF/Tt2xf16tVDamoqLl26hNu3b+PYsWNWH6eLiwvmzJmDl156CdHR0RgxYgRkMhn27NkDHx8f3Llzx6iXKjAwEKIoYs2aNcjKyoKHhweaN29u8mDig7p16xb27dsHoOTL5h9//IHvv/8ev/32G9q3b4/Vq1cb9XzHxMTg1KlTeO+993DmzBl069YNXl5eSElJwZkzZ6BQKLBp0yar2/HRRx/h5MmT6NOnD5o3bw5JkvDdd9/h5s2bePbZZ8td15LXauTIkfj000/x1VdfoWvXrlbnuOfk5BheJ41Gg7t37+Ls2bM4d+4cGjRogPfff7/C3ujXXnsNqampCAsLw0MPPYSCggLD8wKlJ+MKDg7GF198gTfffBO9e/eGXC5HUFDQA/V2K5VKzJ49G9HR0fDz88PZs2dx+PBhdO3a1eiLnDX3hXr16sHPzw8HDhzAww8/jIYNG8Ld3d1k3Hq9Rx55BJMnT8Ynn3yCcePGYeDAgVCr1dixYwfy8vKwZMkSh5ncqbLvUf/+/bFhwwZMmTIFo0aNglwux8mTJ3H9+nWLhtWsaa1bt8aoUaOwfft2TJgwAeHh4cjIyMCWLVvQtm1b/PLLLxX2rlt6bc+cORMXLlzAv//9bwwcOBDBwcGQy+VISUnB8ePH0b59e8OzbTV1f6Sqx0CfnN7cuXNx7NgxnD9/HocPH0ZOTg68vb2hVCoxZcoUk2B1+PDhaNmyJdavX4/t27cjJycHPj4+aNWqFeLi4owm2Jo0aRI8PDzw2WefYenSpWjatCkmTZoEb29vvPLKK0bb9fLywtq1a7Fo0SJ8/PHH8PDwwJNPPonFixejS5cuJu2ePn06AgICsGnTJnz++efIy8tDgwYN0Lp1a8yfP7/Sr8fDDz+M3bt3Y9WqVTh+/Dj27duHOnXqoE2bNhg1apShXtu2bbFnzx58/PHH+Pbbb7Ft2zZ4enqiWbNmiIqKKvfh0NKKiorwwQcfmF3m5+eHxx57DIsWLUJoaCh27NiBjz/+GEVFRfD19UW7du3w4osvVvpYhwwZAhcXF6xevRoffvghGjZsiBEjRsDf3x/Tp083ejjwoYcewsKFC7Fu3Tq8+eabKCoqQlRUVJV/kJ08eRInT56EIAjw8PAwHOf06dMRHh5ukt4il8vx8ccfY8uWLdi3b5/h4ehGjRohMDAQUVFRlWpH//79oVKpcOjQIfz9999wc3ODn58f3n777Qp73y15rfz8/BAaGoozZ85g+PDhVrcvNTUVL7/8MgDAzc0NPj4+aN26NV555RVERkZa9IzIsGHDEB8fjz179iAjIwNeXl547LHH8OGHH2LAgAGGeoMHD0ZiYiIOHDiAQ4cOQafTYdGiRQ8U6Ldv3x7z5s3DsmXLsG3bNnh5eWHcuHGYOXOmUQqWtfeFJUuWYOHChVi2bBny8/PRrFmzMgN9oGSOCj8/P2zZsgVLly6FXC5HcHAwli5dis6dO1f6+GpaZd+jTp06YcWKFVi9ejU++OADuLq6okePHvjiiy8wbty4Gmq9dV5//XU0atQIu3btwrvvvotWrVrh9ddfx//+9z/88ssvcHNzK3d9S69tb29vbN26FevXr8ehQ4dw7NgxyGQyNGnSBJ06dUJ0dLShbk3dH6nqCVJ1PrlGVEvFx8dj3rx5+Pzzzw25wWQ/1q9fj3fffRfbt28vc2IoenD6mThPnDhRYXBCROWLjY3FmTNncP78eYf5JYZsjzn6ROS0NBqNyTMOarUamzdvho+Pj8PkJzui27dv48cff8TQoUMZ5BNZwdzzYdeuXcPx48fRrVs3BvlkFabuEJHTSk5OxpQpUzBo0CA0b94cKpUKe/bswZ07d/DGG28YTV5DVePy5ctISkrCpk2bIJfLMXHiRFs3icih7NmzB/v27UPv3r1Rv3593Lx5Ezt27IBcLseMGTNs3TxyMAz0ichp1a9fHyEhIdi/fz/S09Ph4uICpVKJF1980WSWS6oaW7duxd69e/Hwww9jyZIlZifvIqKytW/fHkePHsWmTZuQlZUFT09PhIaGYvr06fwVkqzGHH0iIiIiIifEHH0iIiIiIifEQJ+IiIiIyAkxR78a3bunhk7neJlRDRp4IT0919bNIAfB84WswfOFrMHzhaxRW88XURRQr56n2WUM9KuRTic5ZKAPwGHbTbbB84WswfOFrMHzhazB88UYU3eIiIiIiJwQA30iIiIiIifEQJ+IiIiIyAkx0CciIiIickIM9ImIiIiInBADfSIiIiIiJ8RAn4iIiIjICTHQJyIiIiJyQgz0iYiIiIicEAN9IiIiIiInxECfiIiIiMgJMdAnIiIiInJCDPSJnJROAlQawdbNICIiIhthoE/khIp1wNF0Ob68q0ByPi9zIiKi2ogRAJGTydcCX6kU0ElA1zrFOH5PDnWxrVtFRERENY2BPpETKdQBX95VoL5chyAvLRoqJLRw0+LbjJLAn4iIiGoPBvpETiQpT4Y6LhJae+gg/JOe/6i7DjoJuJDtYtvGERERUY1ioE/kRK6rZWjmqjMqEwQg2LsYv+TKUKgrY0UiIiJyOgz0iZzEvSIBeVoBDeWmOTquIlBfrkNKAS95IiKi2oKf+kRO4oZahodctYaUnfs1kEtIZqBPRERUa/BTn8gJ6CTgtzwZmrmW/cStr1yHPwtkkPhQLhERUa3AQJ/ICfxVKEIhSvB2KTuK95SV/JtVzEm0iIiIagMG+kRO4FczD+HeTxCABgod7jB9h4iIqFbgJz6Rg9PogOQCEU0rCPQBoKGcgT4REVFtwU98IgeXUijCx0WCqwVXc0O5hDSNiGLm6RMRETk9BvpEDi6lQER9hWUD5MtFoI5MQlohL30iIiJnx097IgeXUiiiQTkP4d6vPvP0iYiIagV+2hM5sHwtkKcVUNeKQN9XLjHQJyIiqgX4aU/kwFIKRTSQ68qcJMucui4ScrUCCrTV1y4iIiKyPQb6RA4spUBEPbl1T9aKAlDHRcK9Il7+REREzoyf9EQOrKRH3/ohdLxkEjI4cRYREZFTY6BP5KByiwGNToC3zPpA31smIUPDQJ+IiMiZMdAnclB/FYpooLAuP1/P20VCBlN3iIiInBo/6Ykc1J8FMtS3YrSd0rxlEjKLBUicOIuIiMhpMdAnckCS9H89+pUhFwG5AORomb5DRETkrBjoEzmgHK0AHQDPB7iC67jokFHEQJ+IiMhZ2TTQ12g0WLx4McLCwhAUFISRI0fi9OnTFq2blpaGuLg4dO7cGR07dsS0adOQnJxsUu+jjz7C888/j549e8Lf3x8rVqwwqaPT6bB7927Exsaid+/eCAkJweDBg7FmzRpoNJoHPk6iqpZaifHz7+fJB3KJiIicmk0D/blz52Ljxo0YOnQo5s+fD1EUMWXKFFy8eLHc9dRqNWJiYnD+/HnExsZixowZuHr1KmJiYpCVlWVUd/ny5UhISEDbtm3L3F5+fj5eeeUV3Lt3D6NHj8Yrr7yCwMBAfPDBB5g6dWqVHCtRVUottG42XHO8ZeADuURERE7MxVY7TkhIwIEDBzBv3jxMmDABABAZGYnBgwdjyZIl2Lx5c5nrbtmyBbdv30Z8fDzatWsHAOjVqxeGDBmCDRs2IC4uzlD32LFjaN68ObKzs9GlSxez25PL5di6dSs6duxoKBs5ciSaNWuGFStW4OzZswgNDa2CoyaqGmkaEYFeDza1bR0XCTfzGegTERE5K5t9yh86dAhyuSajGywAACAASURBVBzR0dGGMldXV4wYMQLnz5/H3bt3y1z38OHDCAkJMQT5APDoo4+ie/fu+Prrr43qNm/evMK2KBQKoyBfLzw8HACQlJRU4TaIakqhDsjTVm78/NI8ZRLytAKKOfIOERGRU7JZoJ+YmIhWrVrB09PTqDwoKAiSJCExMdHsejqdDtevX0dAQIDJssDAQNy6dQv5+flV0sa///4bAFCvXr0q2R5RVUgrFOHjIkF8wPR6UQA8XSRk8oFcIiIip2SzQF+lUqFRo0Ym5b6+vgBQZo9+ZmYmNBqNod7960qSBJVKVSVt/OSTT+Dt7Y2wsLAq2R5RVUgrFOEjr9ywmvfzlnHiLCIiImdlsxz9goICyOVyk3JXV1cAQGFhodn19OUKhaLMdQsKCh64fWvWrMGpU6ewYMECeHt7V2obDRp4PXA7bMXXt3LHTNUv414xHq0roK7ng/fEN5R0yBcBX98HuxXwfCFr8Hwha/B8IWvwfDFms0Dfzc0NRUVFJuX6QF4ftN9PX25u2Ev9um5ubg/UtoMHD2L58uUYNWoURo0aVentpKfnQqdzvARoX19vqFQ5tm4GmaGTgNQ8V7R1LUKW6eVjNXmRgDsFIlSulU934/lC1uD5Qtbg+ULWqK3niygKZXYu2+w3e19fX7PpOfq0G3NpPQDg4+MDhUJhNj1HpVJBEASzaT2WOnnyJF5++WX07dsXr7/+eqW3Q1Qd0osEeMgkyKvoyq0jk3CvmKk7REREzshmn/Bt2rTB77//DrVabVR++fJlw3JzRFGEUqnElStXTJYlJCTAz88P7u7ulWrT5cuXMX36dAQGBmLZsmWQyWSV2g5Rdbn7z4O4VcVVLPmVIP/BRuokIiIiO2SzQD8iIgJFRUXYuXOnoUyj0SA+Ph4dO3ZE48aNAQApKSkmw1sOGDAAly5dwtWrVw1lN2/exJkzZxAREVGp9iQlJWHq1Klo1qwZ1qxZ88DpP0TVIVVTtYG+IABeMgnZxRx5h4iIyNnYLEc/ODgYERERWLJkCVQqFVq0aIE9e/YgJSUFixYtMtSbM2cOzp07h+vXrxvKxowZg507d2Lq1KmYOHEiZDIZNmzYAF9fX8PkW3p79+5FSkqKIX//p59+wurVqwEA48ePh7e3N3JzczF58mRkZ2dj8uTJ+P7774224e/vX+YvDEQ16W6hCL+6VZCcX4rHP4F+Y1fHe56EiIiIymazQB8A3nvvPSxfvhz79u1DVlYW/P39sXbtWnTq1Knc9by8vLBp0yYsXLgQq1evhk6nQ2hoKObPn28y5v3u3btx7tw5w99nz57F2bNnAQBDhw6Ft7c3MjMz8ddffwEAli5darK/6dOnM9Anm8stBrQS4FHFv8O5yyRkF4sAqmbITiIiIrIPgiRJ7MarJhx1h6rSzTwRibkydKxTtQn1dwpE5GgF9GtQuV8KeL6QNXi+kDV4vpA1auv5Ypej7hCRdVILRdStwvx8PQ+ZhBzm6BMRETkdBvpEDiJNI6KevOoDfU8+jEtEROSUGOgTOYAiHZBVLFRLj75CKMnOL2SKPhERkVNhoE/kAP7WiKgjkyCrho53QSjp1Wf6DhERkXNhoE/kANI0Anzk1dfl7sH0HSIiIqfDQJ/IAaRW8Yy493MXGegTERE5Gwb6RHZOkgCVRkS9agz0PWQlzwAQERGR82CgT2TnsooFuAiAm6z69uEh6ifNIiIiImfBT3YiO3dXI6JeNebnA3wYl4iIyBkx0Ceyc6mF1TOsZmluYsnwmsUcYpOIiMhpMNAnsnPVNVFWafohNrO17NUnIiJyFgz0iexYoQ7I0wrwllVvoA+UDLHJ9B0iIiLnwUCfyI6l/TOsplgD8bc7x9InIiJyKgz0iexYWqFYrRNlleYhgoE+ERGRE2GgT2THUqt5/PzSPGQSx9InIiJyIgz0ieyUVgIyioRqnRG3NOboExERORcG+kR2Kl0jwEMmQV5DV6mHWPLgr7ZmvlcQERFRNWOgT2Sn0mowbQcARKFkPP1c9uoTERE5BQb6RHYq9Z8Rd2qSh0xCLsfSJyIicgoM9InskCQBdzUi6tXQiDt6biIDfSIiImfBQJ/IDuX8E2y71/AV6iaTmLpDRETkJBjoE9mhtEIB9eQ6CDUcc7uL4Mg7REREToKBPpEdSi2s2Qdx9dxFyfBrAhERETk2BvpEdihNI6Ke3AaBvkyCWlvjuyUiIqJqwECfyM4U6gC1VoC3rOYDfTcRyNcK0HEsfSIiIofHQJ/IzqT9M6ymaIMMGpkAKEQgj736REREDo+BPpGduasR4VPDw2qW5s4hNomIiJwCA30iO5NWKNT4RFmluXPSLCIiIqfAQJ/IjugkIL2o5mfELc1NlKDmEJtEREQOj4E+kR3JLBbgKkpQ2PDKdBeBbAb6REREDo+BPpEdUWls25sPMHWHiIjIWTDQJ7IjaYUC6to60BfBQJ+IiMgJMNAnsiMlI+7YNtB3EyWotQIkjqVPRETk0Gwa6Gs0GixevBhhYWEICgrCyJEjcfr0aYvWTUtLQ1xcHDp37oyOHTti2rRpSE5ONqn30Ucf4fnnn0fPnj3h7++PFStWlLnNpKQkTJ48GR06dEDXrl0xZ84cZGRkVPr4iKxRqCvpSbfFRFmlycWSG0Oh7Ub4JCIioipg00B/7ty52LhxI4YOHYr58+dDFEVMmTIFFy9eLHc9tVqNmJgYnD9/HrGxsZgxYwauXr2KmJgYZGVlGdVdvnw5EhIS0LZt23K3mZqairFjxyI5ORkzZ87EpEmT8N1332Hy5MkoKip64GMlqsjfGttNlHU/D+bpExEROTwXW+04ISEBBw4cwLx58zBhwgQAQGRkJAYPHowlS5Zg8+bNZa67ZcsW3L59G/Hx8WjXrh0AoFevXhgyZAg2bNiAuLg4Q91jx46hefPmyM7ORpcuXcrc5po1a1BYWIhNmzahcePGAICgoCBMnDgR+/btw4gRI6rgqInKdrdQQF0X++hG10+a1RDM3yEiInJUNuvRP3ToEORyOaKjow1lrq6uGDFiBM6fP4+7d++Wue7hw4cREhJiCPIB4NFHH0X37t3x9ddfG9Vt3ry5Re355ptv0K9fP0OQDwA9evRAy5YtTbZJVB3S7GDEHT03mYRcDrFJRETk0GwW6CcmJqJVq1bw9PQ0Kg8KCoIkSUhMTDS7nk6nw/Xr1xEQEGCyLDAwELdu3UJ+fr5VbUlLS0N6errZbQYFBZXZFqKqIkn/DK1p4wdx9dwEIIeBPhERkUOzWaCvUqnQqFEjk3JfX18AKLNHPzMzExqNxlDv/nUlSYJKpbKqLfp9lbXN9PR0aLVaq7ZJZI3sYgEuAuBmJ+Ngucsk5DBHn4iIyKHZLEe/oKAAcrncpNzV1RUAUFhYaHY9fblCoShz3YKCAqvaYuk27//1oSINGnhZVd+e+Pp627oJtUpahg6NPHWo6+Nu66YAAIoLJNz5WwdfXzeL6vN8IWvwfCFr8Hwha/B8MWazQN/Nzc3saDb6oFsfYN9PX67RaMpc183NsuCkOrcJAOnpudDp7CMVwxq+vt5QqXJs3YxaJemeC9wkICvTPh7G1eqArEK5RecBzxeyBs8XsgbPF7JGbT1fRFEos3PZZokCvr6+ZtNz9Gk35tJ6AMDHxwcKhcJseo5KpYIgCGZTcMqj31dZ22zQoAFkMplV2ySyRnqRiDp28iAuACgEQCsBRfbxvYOIiIgqwWaBfps2bfD7779DrVYblV++fNmw3BxRFKFUKnHlyhWTZQkJCfDz84O7u3XpD40bN0b9+vXL3GZFY/ATPQhJAu4VCXYV6AtCSZ4+x9InIiJyXDYL9CMiIlBUVISdO3cayjQaDeLj49GxY0fDMJcpKSlISkoyWnfAgAG4dOkSrl69aii7efMmzpw5g4iIiEq158knn8S3336LtLQ0Q9np06dx69atSm+TyBI52pIHcRV28iCunrsIqBnoExEROSyb5egHBwcjIiICS5YsgUqlQosWLbBnzx6kpKRg0aJFhnpz5szBuXPncP36dUPZmDFjsHPnTkydOhUTJ06ETCbDhg0b4Ovra5h8S2/v3r1ISUkx5Nr/9NNPWL16NQBg/Pjx8PYueWgjNjYWhw4dQkxMDMaNG4e8vDx8+umnaNOmDYYNG1bNrwbVZhlFAurYyURZpXEsfSIiIsdms0AfAN577z0sX74c+/btQ1ZWFvz9/bF27Vp06tSp3PW8vLywadMmLFy4EKtXr4ZOp0NoaCjmz5+PevXqGdXdvXs3zp07Z/j77NmzOHv2LABg6NChhkC/adOm+OKLL/DOO+9g6dKlkMvl6NOnD+bNm2d2NB6iqpKuEeFtR2k7eq4iU3eIiIgcmSBJkv1FGE6Co+6QJQ6r5Gik0KGJq32dK8kFIvJ0AvrUNx0dqzSeL2QNni9kDZ4vZI3aer7Y5ag7RFQiw85G3NFzEyWoi23dCiIiIqosBvpENlSgBTRSyYOv9sZdlPgwLhERkQOzw/CCqPbIKBJR10WCYIfxtJsMyNMKYHIfERGRY2KgT2RD6UUCvO1wxB0AcBEAUQAK7bN5REREVAEG+kQ29LdGhLcdT7rMSbOIiIgcFwN9IhtKt7MZce/HPH0iIiLHxUCfyEaKJSC3WICXzH4DfTfOjktEROSwGOgT2UhmkQBPFwkyO46jXUXOjktEROSoGOgT2Uh6kYg6dtybD5QM+8kcfSIiIsfEQJ/IRtI19p22AwBufBiXiIjIYTHQJ7KRzGIBXnb8IC5Q8jBuHgN9IiIih8RAn8hGsotFeNp5j76rCORrAZ19N5OIiIjMYKBPZAPFElCgBTzs/AqUCYBcBPI5aRYREZHDsfMwg8g5ZRcJ8JBJEBwgK8ZdlKDmyDtEREQOh4E+kQ1kFpcMrekI3GScNIuIiMgRMdAnsoGsIgEeooME+pwdl4iIyCEx0CeygcxiEV4yW7fCMm4CkMPUHSIiIofDQJ/IBjKLBbsfcUfPTcZJs4iIiBwRA32iGiZJQLYDBfruTN0hIiJySAz0iWpYvq7kwlM4yNXnxkmziIiIHJKDhBpEziOrSLT7GXFLcxOBQh2gdZwmExERERjoE9W4LAdK2wEAQSiZIZe9+kRERI6FgT5RDbtXJMDTQYbW1HOXSVBrbd0KIiIisgYDfaIa5kgj7uhxLH0iIiLHw0CfqIY50og7eq6ihFyOpU9ERORQGOgT1SCtVJLr7uEgk2XpuYscS5+IiMjRMNAnqkHZxQI8ZBJEB4uZ3dijT0RE5HAY6BPVIEcbcUfPXQbm6BMRETkYBvpENSiryDEDfT6MS0RE5HgY6BPVoMxix8vPBwCFABRLQLHO1i0hIiIiSzHQJ6pB2cUiPBxsDH2gZNIsdxkfyCUiInIkDPSJapBaK8DdAVN3AMCd6TtEREQOhYE+UQ3RSUC+tmSoSkfEPH0iIiLHYnXIMWHCBBw8eBAajeaBd67RaLB48WKEhYUhKCgII0eOxOnTpy1aNy0tDXFxcejcuTM6duyIadOmITk52WzdnTt3YuDAgQgMDMSAAQOwefNms/VOnTqF8ePHIzQ0FF26dMGoUaNw8ODBSh8fUWlqrQBXEQ43tKaeq0xCbrGtW0FERESWsjrQT0xMxIsvvohevXrhv//9L65fv17pnc+dOxcbN27E0KFDMX/+fIiiiClTpuDixYvlrqdWqxETE4Pz588jNjYWM2bMwNWrVxETE4OsrCyjutu2bcOrr74KpVKJ1157DcHBwViwYAHWr19vVO+7777DpEmTUFxcjBdeeAFxcXEQRREzZ87Ezp07K32MRHo5/4yh76jcBeboExERORJBkiSrIg+NRoMjR45g165dOHPmDAAgICAA0dHRGDRoEDw9PS3aTkJCAqKjozFv3jxMmDABAFBYWIjBgwejUaNGZfa6A8C6deuwdOlSxMfHo127dgCApKQkDBkyBM899xzi4uIAAAUFBejduzc6deqE1atXG9Z/6aWX8O233+KHH36At7c3AODZZ5/F9evXcezYMSgUCsOxPvHEE/Dz88MXX3xhzcsEAEhPz4VO53iBna+vN1SqHFs3w+n8qpbh9zwRQd5aWzelUu5qBPxZKOIp3yKjcp4vZA2eL2QNni9kjdp6voiigAYNvMwvs3ZjCoUCgwYNwmeffYajR4/i+eefR3p6Ov7zn/8gLCwMr7zyCs6fP1/hdg4dOgS5XI7o6GhDmaurK0aMGIHz58/j7t27Za57+PBhhISEGIJ8AHj00UfRvXt3fP3114ays2fPIjMzE2PGjDFaf+zYsVCr1Th+/LihLDc3F3Xr1jUE+fpjrVu3LlxdXSs8HqKKZBcLcHPkHn2Rk2YRERE5kgd6LLBZs2aYMWMGjh07hk8++QShoaHYs2cPxo0bh6eeegobNmyAWq02u25iYiJatWpl8gtAUFAQJElCYmKi2fV0Oh2uX7+OgIAAk2WBgYG4desW8vPzAQBXr14FAJO67du3hyiKhuUA0LVrV9y4cQPLly/HH3/8gT/++APLly/HrVu3MGnSJMtfFKIy5BQL8HDQB3GBkodx87QCrPsNkIiIiGylSsKOxMREfPvttzh//jwkSUKLFi0giiLeeecdDBgwABcuXDBZR6VSoVGjRiblvr6+AFBmj35mZiY0Go2h3v3rSpIElUpl2IdCoYCPj49RPX1Z6X3ExsZi4MCBWLNmDcLDwxEeHo6NGzdi9erV6Nmzp+UvBlEZchx4aE0AkP9zt9A47iEQERHVKi6VXTE7Oxv79+/Hrl27cO3aNbi4uKB///4YOXIkunfvDgA4ffo0XnvtNSxYsAB79+41Wr+goAByudxku/o0mcLCQrP71ZeXTrG5f92CgoJy96GvW3ofCoUCLVu2REREBMLDw6HVarFjxw78+9//xoYNGxAUFFTu62FOWflSjsDX19vWTXA6ealFaOzjAk+546a/eGVroajjBV9342Pg+ULW4PlC1uD5Qtbg+WLM6kD/9OnT2LVrF44ePYrCwkK0bNkSs2fPRlRUFOrVq2dUt3v37pg6dSoWLFhgsh03NzcUFRWZlOuD77Ly4vXl5ob31K/r5uZm+LesYUALCwuN9vHWW2/hf//7H3bt2gVRLOm6HDhwIAYPHoyFCxdi27ZtZrdTHj6MS3paCcgvdkVRbgGyHDfOhwIuuKPKg+CuM5TxfCFr8Hwha/B8IWvU1vOlvIdxrQ70J06cCIVCgfDwcIwaNQpdu3Ytt36LFi3QoUMHk3JfX1+z6Tn6tBtzaT0A4OPjA4VCYah3/7qCIBjSenx9fVFUVITMzEyj9B2NRoPMzEzDPjQaDXbt2oXnnnvOEOQDgFwuR69evbB161YUFxfDxaXSP4BQLZdbLMBdBggOHOQDgKsocYhNIiIiB2F15Dp37lxERkaa5L2XpVu3bujWrZtJeZs2bbBp0yao1WqjB3IvX75sWG6OKIpQKpW4cuWKybKEhAT4+fnB3d0dANC2bVsAwJUrVxAWFmaod+XKFeh0OsPyzMxMFBcXQ6s1HfawuLgYxcXFsHIUUiIjOVoB7qLjn0NuooTcYgb6REREjsDqh3Fzc3PLHfryxo0bWLlyZYXbiYiIQFFRkdFkVBqNBvHx8ejYsSMaN24MAEhJSUFSUpLRugMGDMClS5eMRs25efMmzpw5g4iICENZt27d4OPjgy1bthitv3XrVnh4eODxxx8HADRo0AB16tTBkSNHjNKJ1Go1vvvuOyiVyjJz/YksUdKj7/iBvrvISbOIiIgchdU9+qtWrYKfnx+USqXZ5Tdu3MCqVaswffr0crcTHByMiIgILFmyBCqVCi1atMCePXuQkpKCRYsWGerNmTMH586dM5qBd8yYMdi5cyemTp2KiRMnQiaTYcOGDfD19TVMvgWU5OjPmDEDCxYsQFxcHMLCwvDzzz/jyy+/xEsvvYQ6deoAAGQyGSZNmoTly5dj1KhRGDp0KHQ6HXbt2oXU1FTMmTPH2peJyEhOsfP06KdpGOgTERE5AqsD/YpSWAoLCyGTySza1nvvvYfly5dj3759yMrKgr+/P9auXYtOnTqVu56Xlxc2bdqEhQsXYvXq1dDpdAgNDcX8+fNNHggeO3Ys5HI51q9fj2PHjqFp06aYP38+YmJijOo9//zzaN68OT7//HOsWrUKGo0G/v7+WLlyJcLDwy06HqKyZGsF1HWCHn03mcRJs4iIiByEIFmQfJ6bm4vs7GwAQL9+/TB//nw88cQTJvWysrKwePFiJCcn48iRI1XfWgfDUXdIb1+aAo95aFFf7njnQ2laCTiSLseEZoWGB4t5vpA1eL6QNXi+kDVq6/nywKPubNiwAatWrQIACIKAhQsXYuHChWbrSpKE2bNnV7KpRM4pVyvAwwlSd2QC4CIABTrA3bIf7oiIiMhGLAr09UNoSpKEVatWITw8HP7+/ib1PD09ERwcjI4dO1ZtK4kcWLEO0OgA1yqZh9r23GUlQ2w6w8PFREREzsziQF8f7KekpGD06NEIDg6u1oYROYtcrQAPmeTwY+jruYslefq+YKBPRERkz6x+GLf0iDhEVLGcfybLchZuMo6lT0RE5AgqDPRTUlIAAA899JDR3xXR1yeq7Zxlsiw9N6HkywsRERHZtwoD/X79+kEURVy6dAkKhQL9+vWDYEEOQmJiYpU0kMjROcsY+nruMgl/FznJAwdEREROrMJA///9v/8HQRDg4uJi9DcRWSanWEA9Bx9WszR3EUzdISIicgAVBvovvPBCuX8TUfnUWgEPueps3Ywq485Js4iIiBwCf38nqmZqrQA3J0rdUQglE2cVOc93FyIiIqdkdaB/+/ZtHD9+3Kjs8uXLiI2NxejRo7F9+/YqaxyRo9NJQKETjaEPAIIAePwzlj4RERHZL6uH11yyZAkyMzPx+OOPAwAyMjIwZcoU5OXlwdXVFW+88QYaNGiA/v37V3ljiRyNWlsS5ItOFhO7y0ry9J3p2QMiIiJnY3U/45UrV9CjRw/D3wcOHEBubi7i4+Nx+vRpBAcHY+PGjVXaSCJH5WxpO3puooQc9ugTERHZNasD/YyMDDRq1Mjw94kTJ9CxY0colUooFAo89dRTSEpKqtJGEjkqtVaAu8w5A32OvENERGTfrA703d3dkZOTAwDQarU4f/48OnfubFju5uaG3NzcqmshkQNTawW4OmGPvrvISbOIiIjsndWBfuvWrbF3717cu3cPO3bsQF5eHnr27GlY/ueff6J+/fpV2kgiR5VbLMDNiR7E1XPnw7hERER2z+qHcSdPnoxp06YZ8vTbtm1r1KN/8uRJtGvXrupaSOTAcrUCGsqdbxxKd5Fj6RMREdk7qwP9Pn36YOPGjTh27Bi8vLwwbtw4w0y59+7dQ5MmTRAZGVnlDSVyROpiAQ+72roVVc9NLBk2tNj5spKIiIichtWBPgB06dIFXbp0MSmvV68eVq5c+cCNInIWap0ANyd8GFcQADdZyRcZIiIisk9OmD1MZB+K/5k91tVJY2EPkXn6RERE9qxSPfoXL17EF198gdu3byMzMxOSZNxjKQgCjh49WiUNJHJUedqSB3EFJ42F3RjoExER2TWrA/29e/di3rx5cHFxQcuWLdG0adPqaBeRw8stds4x9PXcZBxLn4iIyJ5ZHeh/9NFHaNWqFT777DM0bty4OtpE5BTUWjjlrLh67iKQzUCfiIjIblmdo5+SkoJnnnmGQT5RBZx1siw9d6buEBER2TWrA/0mTZpAo9FUR1uInEqu1jkny9Jzl3EsfSIiIntmdRgyevRo7N+/H1qttjraQ+Q0cosFuDt1jz6QrwV0kvMeIxERkSOzOke/ffv2+OabbxAdHY0xY8agefPmkMlkJvXMjbNPVJuotQLcTC8NpyEKgKsI5BTZuiVERERkjtWB/oQJEwz//+qrrxpmxdWTJAmCICAxMfGBG0fkyEqG13Tu3m53mYQsjQR3WzeEiIiITFgd6C9atKg62kHkVIp0JRNmKZw8hd1dlJCtAQN9IiIiO2R1oB8VFVUd7SByKmptyRj6zjpZlp6bTMK9QgmNKzX1HhEREVUnJx4ThMh21FoB7rXg6vIUgYxC505PIiIiclSVCkX++usvzJs3D48//jgCAgJw+vRpAEBGRgbmzZuHhISEKm0kkaNR14L8fADwkEm4V+D8x0lEROSIrA70k5OTMXz4cHzzzTdo3bq10TCb9evXx5UrV7Br164qbSSRo8ktBlxlzh8Ae8okZHJaDSIiIrtkdWbt8uXLIYoivvrqK7i6uqJHjx5Gy3v37o3vvvuuyhpI5IhytQLcnTw/Hyh52LhYB2h0gKIWpCoRERE5Eqs/mk+dOoVnnnkGTZs2NRlaEwAeeughpKamWrQtjUaDxYsXIywsDEFBQRg5cqQhDagiaWlpiIuLQ+fOndGxY0dMmzYNycnJZuvu3LkTAwcORGBgIAYMGIDNmzeXud39+/djxIgRCAkJQdeuXTFu3DimIpHVcrUC3GpBj74gAHUUQHZxLfhWQ0RE5GCs7tHPzc1Fo0aNylxeVFRk8ay5c+fOxTfffIOYmBj4+flhz549mDJlCjZt2oQOHTqUuZ5arUZMTAzUajViY2Ph4uKCDRs2ICYmBnv37kXdunUNdbdt24bXX38dERERmDhxIn7++WcsWLAAhYWFmDRpktF2ly1bhk8++QRDhw7FqFGjkJeXh2vXrkGlUll0PER6JWPo27oVNcNLXhLoN1Q4/xcbIiIiR2J1oN+0aVPcuHGjzOWXL19GixYtKtxOQkICDhw4gHnz5hkm4YqMjMTgwYOxZMmScnvdt2zZgtu3byM+Ph7t2rUDAPTq1QtDhgzBhg0bEBcXBwAoKCjAsmXL8MQTT+CDDz4AAIwcORI6nQ4rV65EdHQ0vL29AQAXLlzAxx9/jBUrViA8PNyi14KoLHlaAe614GFc/26V5AAAIABJREFUAPCUC+zRJyIiskNW9zmGh4dj9+7d+PXXXw1l+hSew4cP49ChQxg4cGCF2zl06BDkcjmio6MNZa6urhgxYgTOnz+Pu3fvlrnu4cOHERISYgjyAeDRRx9F9+7d8fXXXxvKzp49i8zMTIwZM8Zo/bFjx0KtVuP48eOGss8//xyBgYEIDw+HTqeDWq2u8BiIzNHoAB0Al1oS+3q5AFkM9ImIiOyO1YH+888/jyZNmmDkyJGYPXs2BEHAunXrMGrUKPz73/9GmzZtTFJizElMTESrVq3g6elpVB4UFARJkpCYmGh2PZ1Oh+vXryMgIMBkWWBgIG7duoX8/HwAwNWrVwHApG779u0hiqJhOQCcPn0agYGBeP/999GpUyd07NgR/fr1w5dfflnhsRCVpv6nN9/ZJ8vS85ILyC6uJXlKREREDsTq1B0vLy9s374dy5cvx1dffQVJknDy5EnUqVMHY8aMwcyZM+Hq6lrhdlQqFRo3bmxS7uvrCwBl9uhnZmZCo9EY6t2/riRJUKlUaNGiBVQqFRQKBXx8fIzq6cv0+8jKykJmZiYOHDgAmUyGl156CT4+Pti8eTNmz54Nd3f3SqXzNGjgZfU69sLX19vWTXBYudk6eOfoUNdHbuum1AiXIglqnchzhizGc4WswfOFrMHzxVilJq738vLCq6++ildffRUZGRmQJAn169c3OwpPWQoKCiCXmwZC+i8JhYWFZtfTlysUijLXLSgoKHcf+rr6beXl5QEo+RKxY8cOBAcHAyhJUwoPD8eqVasqFeinp+dCp3O8PG1fX2+oVDm2bobDSlHLINOKyMq07KF0R1enrhvyiyWkpOVAzo59qgDvL2QNni9kjdp6voiiUGbnstWB/oULF/DDDz/g999/h1qthqenJx555BH06dMHISEhFm/Hzc0NRUVFJuX64LusXwX05RqN6Sw9+nXd3NwM/5qrp6+r35b+3+bNmxuCfKDky8SAAQPw+eefG46VqCK1ZbIsPUEQ4CmTkKMVUL+WPIBMRETkCCwO9HNzczFr1iycOHECkmT6Yf7xxx+jd+/eWLJkCby8Kk5Z8fX1NZueox/KsqwhPH18fKBQKMwOealSqSAIgiGtx9fXF0VFRcjMzDRK39FoNMjMzDTsQ7/Nhg0bmmyzYcOGkCQJubm5DPTJIrlaAW61JD9fz0MmIbtYQH05A30iIiJ7YXGgP2PGDJw6dQqdOnXCiBEj4O/vDy8vL+Tm5uL69evYuXMnvv/+e8ycORPr1q2rcHtt2rTBpk2bTHrKL1++bFhujiiKUCqVuHLlismyhIQE+Pn5wd3dHQDQtm1bAMCVK1cQFhZmqHflyhXodDrDclEU0bZtW6SlpZlsMzU1FTKZzGhsfqLyqLUC6rrqbN2MGuUhk5DDkXeIiIjsikUZtSdOnMCpU6cwceJEbN68GVFRUWjXrh1atGiBdu3aISoqClu2bMHEiRPx448/4uTJkxVuMyIiAkVFRdi5c6ehTKPRID4+Hh07djQ8qJuSkoKkpCSjdQcMGIBLly4ZjZpz8+ZNnDlzBhEREYaybt26wcfHB1u2bDFaf+vW/8/evYdHVd2NHv/uy1xyT4AEEAiikqDcL5Wrvq3wClhvqGgFoWrVWntOqT7tAx7qeVv7tr5Vq/Z46Xm1pahHaYuiiFZBsdVWRASBGIhYEATkkkkgt8nc9t7r/DFkSkgCuc9k5vd5njzKmrX37D1Z2fPba//WWitIT0/n4osvbnQ8hw8fbnTsdXV1vPnmm4wdOzaWDiTEmfhTaLGsBmk6VEUk0BdCCCESifHTn/70p2eq9Nvf/paamhqefvrp0w64nTJlCq+88gr19fXMmDHjtPvs168fu3fv5oUXXsDv93Pw4EEeeOAB9uzZw0MPPcRZZ50FwF133cWDDz7I//yf/zO2bXFxMW+++SavvPIKSilKSkr42c9+Rnp6Ov/1X/8V69E3TZP09HSWL1/O7t27qaur47nnnmP16tUsWrSIKVOmxPY5bNgw1q1bx8qVKwmFQpSVlfHzn/+co0eP8uCDD8aOpy0CgTDNZDklvIwMD/X1zY9tEGe2udrkvHQbI0XiXq/XRV3AwhfWKcpIrScZou3k+iLaQtqLaItUbS+appGe3nSSGmhl6s6OHTuYMWPGGWfV0XWdGTNm8OGHH7bqwB588EEee+wxVq9eTXV1NcXFxTz99NOMHz/+tNtlZmby/PPP88tf/pKnnnoKx3GYOHEiS5cuJS8vr1Hd+fPn43K5WLZsGevXr6d///4sXbqUhQsXNqqXlpbGc889x4MPPsj/+3//j2AwyPDhw/nDH/5wxuMRokHEAUeBK0WC/AbphqLWTrHHGEIIIUSC01RzI2tPceGFF3L33Xdz4403nnGHK1as4JFHHuHjjz/ulAPsyWR6zdRTFdFYW+Hi4jwr3ofSbXJy0zh+PMC6ShcLB4RS5kmGaB+5voi2kPYi2iJV28vpptdsVRdcW2acycjIiM1LL0SqScX8fABdgzQDGZArhBBCJJBWhSSO47RpMSzHkTxdkZqigX7Pe4rTGTJOTLEphBBCiMTQ6uk133vvPSoqKs5Yr7lpL4VIFX4bPCka6KfpEugLIYQQiaTVgf7rr7/O66+/3qq6ben9FyKZ1FmpmboDkGkojssUm0IIIUTCaFWg/9xzz3X1cQiRFPy2Rj93aqauZZmK3fVGvA9DCCGEECe0KtC/8MILu/o4hEgKflvDm6KxbqahqLI0lAJ5qCeEEELEX4omGQjRNepTeDCuWwdTgzpbonwhhBAiEUigL0QniThgKXCncJybbUqefrKosTQOBvUeubq3EEKIqFYPxhVCnF69rZFmpHbaSobhcDyiUZgW7yMR7WU5sK3WpKzOwKMr0gyYnBuhj1sifiGE6Gkk0Beik6TyHPoNsgyojOiAHe9DEe1QEdZ4p9JNjukwNTeCV4cDIY23KtwMy7CZkJM6Kz4LIUQykNQdITqJ30YCfUnd6bHCDqyvdHFems2YLDv2dKrQq7goN8JnfkN+t0II0cNIoC9EJ/HbWsoultUg01DUWhpOan8MPdLGKpM8l2KAt+n0sG4dhqTZbK6Wh8BCCNGTSKAvRCeps1N3sawGhgZphqJaVsjtUfYFdA6FDM5PbznlarDXoTys4wvL71YIIXqKFA9LhOg8dZZGWor36ANkSvpOj1JvwwfHXYzKtDBP841gaHBuusXH0qsvhBA9hgT6QnSS+hReLOtkmYbiWEQuLT3FR1UuBnhs8lxnvkkd5FHUWDqHg/L7FUKInkCu1kJ0kjpbevQBsgzFMenR7xHKQxqHQjrnpjfNy2+OrsF5aTZbauSOVgghegIJ9IXoBGEHHMAl8S1ZhqTu9ARKwcYqF0PTbcw2/Lr6exyORXTqZQZVIYRIeBLoC9EJ6myNdF2l9GJZDdINCNgakdZ1Eos42RfQCSsY6GnbL0rXoMDt8GVAevWFECLRSaAvRCeos6Kr4opoIJhpKqpk5p2EZSn4qNqkON1u181pgdthb0C+PoQQItHJlVqITlAnq+I2kmkojsuA3IS1o9Yg01D0cbevzea7Fb6wTkie2gghREKTb2IhOkGdJYH+ybIMRYXMt56Q/DaU1JoMy2h/kr2pQW+XwwGZfUcIIRKaXKWF6AS1tkaa/DXF5LkUR0LygSSiTVUuCtNsMjqYalbgVuyrl3w1IYRIZPJNLEQniOboS49+g2xTUWNrhCW1I6EcDmkcDumck9bxX0xft8OhkI4lzV4IIRKWBPpCdAK/zKHfiKFBrqkoD8slJlE4CjYcdzEsw2rTdJotceuQYyq+kvQdIYRIWHKFFqKDbAUhBzzy19RIrsvhqKTvJIzSWgNTg37tHIDbnAK3wz6ZfUcIIRKWXKGF6KB6W8OrR6eVFP+SZyoOh+RDSQRHQxoldSYjMq1OXeuhr9vhQNBAycMsIYRISBLoC9FBkp/fvDxTURnRceSjiat6G9ZXuhmZaZHeyWNn04xomla1rJkghBAJSQJ9ITqozkam1myGS4d0Q1EZkSAwXmwF71S6Gei1KejElJ2T9TIdDkuKlhBCJCS5OgvRQXWWhld69JuVZ8o0m/FiOfDeMRc6ivM6YZadluS5lAT6QgiRoOTqLEQHyRz6LZP59OOjKqKxutxNyIHRWXan5uWfqpfL4UhIlzx9IYRIQPINLEQH1VoytWZL8kyHo2EJArtL2IGyOoPXy6PpOqMz7U6ZSvN00nVQRG94hRBCJBYz3gcgRE8nc+i3LM0AA6ixNHJc8hl1hbADR0I6e+oNDgR1erscJuRY5Jjd83lrGvQ+0aufbdrd8p5CCCFaJ649+uFwmIceeohp06YxatQorr/+ej788MNWbXv06FEWLVrEhAkTGDduHHfddRcHDhxotu7KlSuZPXs2I0eOZObMmbzwwgtn3P/tt99OcXExv/jFL9p0TiK1KHVies1Ons0kmeS5or36ovMEbNhabbD6qJsVhz18UmPiNRwuzoswLtvutiC/Qa6pOCQpWkIIkXDiemVesmQJzz77LFdeeSVLly5F13Vuv/12tm7detrt/H4/CxcuZMuWLdx555384Ac/YOfOnSxcuJDq6upGdf/4xz/yk5/8hKKiIu677z5Gjx7N/fffz7Jly1rc/9/+9jc2b97cKecoklvQiU4v2NXpET1ZnktxQFZP7RS1Fmw4bvLSEQ8VEZ1z0mym94owMcfibK+K26Jt0R59+SMQQohEE7fUnZKSEt544w3uvfdebr75ZgCuvvpqLr/8ch5++OHT9rq/+OKLfPnll6xatYoLLrgAgIsuuogrrriC5cuXs2jRIgCCwSCPPvoo06dP5ze/+Q0A119/PY7j8MQTTzB37lyysrIa7TscDvPAAw/wne98h8cff7wLzlwkkzpbI11m3Dmtvm6HXcddWA6YEu+321dBnb9WuhjotZmWF8GbQJ9lhgGWo1FnQaYkhAohRMKI21fFW2+9hcvlYu7cubEyj8fDddddx5YtWygvL29x27Vr1zJmzJhYkA9w7rnnMnnyZN58881Y2UcffURVVRXz5s1rtP38+fPx+/28//77Tfb93HPPEQwG+c53vtOR0xMpQqbWPDOPHk3tkF799ttbr/PXYy7GZlsUZzgJFeRDNE+/l9uRGZaEECLBxO2qXFZWxpAhQ8jIyGhUPmrUKJRSlJWVNbud4zjs2rWLESNGNHlt5MiR7Nu3j0AgAMDOnTsBmtQdPnw4uq7HXm/g8/l46qmnuPvuu0lLS2v3uYnUUScDcVuln8dhT70MZGiPz+oMNlS5+Fq2Ra8EHtCcJ3n6QgiRcOJ2Vfb5fBQUFDQpz8/PB2ixR7+qqopwOByrd+q2Sil8Pl/sPdxuN7m5uY3qNZSd+h6PPPIIQ4YM4aqrrmrXOYnUU2tpCde7moj6uh0OhXTCXbduU1I6GtLYXGMyMSdCdjcPsG2rXrJmghBCJJy4ZVMGg0FcLleTco/HA0AoFGp2u4Zyt9vd4rbBYPC079FQ9+T3KCkp4dVXX+X5559H66TVZXr3zuyU/cRDfn7WmSsJwjUW/dM1cjJTeyBiTu6Zn4AVhGyqXBkM7yXBYGuEbcVLn1l8rUDnrB6Q+J6lFJtqHDLyMkk/w+h0ub6ItpD2ItpC2ktjcfv28Hq9RCKRJuUNwXdD0H6qhvJwONzitl6vN/bf5uo11G3Yl1KKX/ziF1x66aVMmDChjWfSssrKOhwnsXvhmpOfn4XPVxvvw+gRjtW76avZVFs97/fcWXJy06iuCpyxXh9NZ+sRmwK76d+9aOqD4yZZmka2Faa6Kt5H0zrZhslnh+oZlNbyoxu5voi2kPYi2iJV24uuay12Lsetay0/P7/Z9JyGtJvm0noAcnNzcbvdsXqnbqtpWiytJz8/n0gkQlVV42/JcDhMVVVV7D3efvttSkpKuPHGGzl48GDsB6Curo6DBw/GnhII0UCp6GqgGTIYt1UK3NH59EOSvnNGB4M6XwYMzs/oWQtQ5bgcymXNBCGESBhxuyIPGzaMvXv34vf7G5Vv37499npzdF2nqKiI0tLSJq+VlJQwePDg2EDa888/H6BJ3dLSUhzHib1+6NAhHMfh29/+NtOnT4/9AKxatYrp06ezadOmDpytSEZBJ/oH5JK4plVcOvR2O3wZkEG5p2M58PdjLkZmWj2ubeUaiqPh1E5jE0KIRBK3r5FZs2YRiURYuXJlrCwcDrNq1SrGjRtH3759gWgQvmfPnkbbzpw5k23btjWaNeeLL75g48aNzJo1K1Y2adIkcnNzefHFFxttv2LFCtLT07n44osBuOSSS3jyySeb/AB84xvf4Mknn2T48OGd+wGIHq/a0siU3vw2OcvtUFZnoORja1FpnUGOy6GPu+d9SLkuRUVYl9+vEEIkiLjl6I8ePZpZs2bx8MMP4/P5KCws5JVXXuHQoUM88MADsXqLFy9m06ZN7Nq1K1Y2b948Vq5cyR133MEtt9yCYRgsX76c/Pz82OJbEM3R/8EPfsD999/PokWLmDZtGps3b+a1117jRz/6EdnZ2QAUFhZSWFjY7HEOGjSIGTNmdM2HIHq0akuXxbLaqK9b8Xk9fBXSGeiVHJ5ThRworY3OstMTefTok5tqSyM3gacCFUKIVBHXqRwefPBBHnvsMVavXk11dTXFxcU8/fTTjB8//rTbZWZm8vzzz/PLX/6Sp556CsdxmDhxIkuXLiUvL69R3fnz5+NyuVi2bBnr16+nf//+LF26lIULF3blqYkUUB2RVXHbStPgvHSHT6pNBnjCdNIEV0lje41Jgcfp0avL5pnRPP1cV88aXyCEEMlIU0oesnYVmXUnub1d4aK3y6G/p+f9jjtTa2fdaaAU/L3KZFqexQDp1Y+pt+HlIx6m5kZI68HDGPYFdGwF03pZzb4u1xfRFtJeRFukantJyFl3hOjpqi2NjB4ckMWLpsG5aQ5bakzJ5T7J1hqTAV67Rwf5ALmmkpl3hBAiQcjVWIh2UArqLEndaa+zPA4BGw7JSqoA1FrwRb3BuaeZf76nyDIVNbZGpOefihBC9HjyLStEO/jt6KDDMywAKlrQ0Kv/cbVJD8xu63QltSaDvDbuJLgiGxpkG4qKiPxxCCFEvCXB14oQ3a/a0mVqzQ46y+OggLK6Hp6r0kH1drQ3/+wk6M1vkOtyKJenNUIIEXdyJRaiHaolbafDNA2GZ9hsrTHxp/AELZ/WmpzltfEk0dU4x4SjkqcvhBBxJ1diIdqhOqKRrkug31FZpmKg12ZjlSvehxIXQRs+9xsMSbLZh3JNB58snCWEEHEngb4Q7RCdcUeimM5wXno0KDwYTL3L0Y46k74ep8fPtHOqNB0UpPSTGiGESASp980qRCeokUC/0xgaXJBh8cFxEyuFPtKwA2V+g3PSki8a1rR/LZwlhBAifuQqLEQbOQr8tpZ0vbDxlO9WZBiK0trU+VB31Br0cTlJuxZDtkvm0xdCiHiTq7AQbVRraXiNaE+06DznZ9h8WmtS1/yCqkkl5MAOv8m56cnXm98g11Ay844QQsSZXIWFaCPJz+8a6QYUptl8VJ38A3N31Brkuxwyk7Q3H6Ir5B63NGz5UxFCiLiRQF+INqqxZMadrnJumsPRkM7hJB6YG3Kig3CTuTcfwNQh3VAck4WzhBAibpL321SILiI9+l3H0GBYhsWHVWbSTs1YWmtS4E7e3PyT5ZgKn+TpCyFE3MgVWIg2qopIoN+V+rkVNnAgCXv1gzbsrDOSvje/Qa6pOCp5+kIIETdyBRaijWosXVbF7UKaBuekOWyrSb5e/Y+rTfp7UqM3H6KBvi8sqTtCCBEvEuiLhOMoqLPgSEhjT71ObQLNwhKwIawgXf5yulR/t0O9A0eSKEj0hTX2Bw2KUqQ3HyDTUAQcjWDqnLIQQiQUM94HIMTJLAde97mj89TrCo+u+NByMTk3wrnpTrwPj4qwTq6p0JIn/kxIJ/fq98+PxPtwOsxR8I/jLorSbVwpdJOoaQ29+jqD0uL/9yuEEKlGAn2RMJSC94678OqKC7OtWDBdbTlsrjY5EFBMyYvgjmOg5AtrZJsSsHSHszwOu+tdVIQ1+rh7dg7PZ3XRXJ0BntRrOzknVsiVQF8IIbpfCvUtiUS3vdagKqIxItNu1GOeYyqm5FoEHVhX4Y7rvNzlYZ0cs2cHnT2FocHZaTZba3p2f0S9DZ/UmlyQYafkk6BcU1GeRClYQgjRk0igLxLClwGdHXUm47KsZlecNTUYlWkDir8fc8VlkKZSUBGJpu6I7jHI63A4pONPoHEabWErWF/pptBrk52i7SbXpagI60k3sFoIIXoCCfRF3DkKNhx3MSbLwnua2Ug0DUZn2VRENLbVdv+0Jf4TAwq98lfTbUwN+nscPvf3zGlqNlaZaCjOS+G0FY8eXTyrypJefSGE6G4Ssoi4+zKgk2YoernO3OVnajA+y6KszuSL+u5tvr6wTq7ppGT6RTwN9Djsqjd6XI/wZ3UGB4M6ozJTM2XnZL1MR+bTF0KIOJArr4i70jqTQm/r59/zGjA+22LDcReV3Zj7Wx7WUzb9Ip5yTIWhwaEeFCgeDOpsrommoqXSLDstyXMpDveg358QQiQLufKKuKoIa9RaGn3bOKtKtqk4P9PmnUoXoW7KivCFNRmIGweaFu3V/6yHpO/s9uv87ZiLsVkWmT17HHGnyTMdjoTl60YIIbqbXHlFXO2oMyhMs9Hb0TF/lsch3+3wbqULp4vjb6XgmAzEjZsBHoevgnrCL7z0aY3BpmoXF2ZbrUpFSxUZBthOdCE8IYQQ3UcCfRE3ARu+DBgM6sDc4sXpDmFH4+Pqru06rbY0XDpxncM/lbl0KHA7/LM+MXv1LQX/OG5S5jeYlBMhS24IG9E0yHNLr74QQnQ3ueqKuPmszqC/x+lQ8KxrMCbLYm9Ap6yu64JAX1gjVxbKiqtBXodd/sQblFtraawpd1Md0ZiUY5GWmPcicZdnKo4E5StHCCG6k1x1RVwoBbvqTQZ5Ox48u3WYkG3xSY3JgUDXNGlfWCfbSLAIM8XkmYqI0qiMJM4UNl8FdVaXu+nrdhibZcvA29PIM5X06AshRDeTq66Ii2iwpjoteM4wYGyWxXvHXFR0wUw8vrBOjuRcx5WmwVkem90Jkr7zeZ3BX08Muh2SJtOunkm2qfDbGgFL/o6EEKK7SKAv4mJfwKCvu3ODozyXYnimxboKN1Wd2OsbdqKL/eRIj37cneV2+KLe6PLB16ejFGypNthSazIxOyKDbltJ16K9+gfq5PMSQojuIoG+iIt9Ab3NU2q2Rj+PoijD4i++aM50Z9gXMOjjcjDlryXuMk1w6/Gdk31TtcnegMHknIhMn9lGuS6HA3Uy1kUIIbqLhC6i21VHNEKO1mVTVQ7wKM5Ljwb7tVbHg/3P/Tr9OzAzkOhc/T0Ou7t5VeQGn9UZ7A3ofC3bwiNXzzbrJT36QgjRreSrSnS7aG9+1+Y0D/IqhqRbvNHBYL/OguMRnYIuePog2ucst8OXAYPuTvU+dGK12/HZlkyz2k65LkVFECJy3yyEEN0irl9X4XCYhx56iGnTpjFq1Ciuv/56Pvzww1Zte/ToURYtWsSECRMYN24cd911FwcOHGi27sqVK5k9ezYjR45k5syZvPDCC03qrFu3jh/+8IdccskljB49mlmzZvGrX/2K2traDp2jaGpfwKDA3fXf9IO9irPTLF73ualpZ7C/p96gn8fBkIGWCcNrQK6pumyGpeZURzTePeZiTKZFZmKMBe6RDA16e4lr6pUQQqSSuF5tlyxZwrPPPsuVV17J0qVL0XWd22+/na1bt552O7/fz8KFC9myZQt33nknP/jBD9i5cycLFy6kurq6Ud0//vGP/OQnP6GoqIj77ruP0aNHc//997Ns2bJG9e677z727NnDVVddxU9+8hOmTZvG888/z4033kgoFOr0c09Vfju6+FTvbhrAONirOCfN4o3ytgf7SsE/6w3OkrSdhNPf032LZzkK/nrMxbnpFr3lyU6H9U/XOCDz6QshRLeI21CykpIS3njjDe69915uvvlmAK6++mouv/xyHn744WZ73Ru8+OKLfPnll6xatYoLLrgAgIsuuogrrriC5cuXs2jRIgCCwSCPPvoo06dP5ze/+Q0A119/PY7j8MQTTzB37lyysrIA+D//5/8wceLERu8zYsQIFi9ezBtvvME111zT2R9BSvryRG++3o095IVehYbF6+VuLssPk9vKm4zKiEbE0ciTVU4TTl+3Q5nfRdCO9vB3pdJaAw0Y7JF20Bn6p2u8X6WjFDIlqRBCdLG4dau89dZbuFwu5s6dGyvzeDxcd911bNmyhfLy8ha3Xbt2LWPGjIkF+QDnnnsukydP5s0334yVffTRR1RVVTFv3rxG28+fPx+/38/7778fKzs1yAeYMWMGAHv27Gn7CYpm7a2P5ud3t0He6Gw8b/jcVLZynv3dfoP+XluCkQTk0qHA7fBFoGuj/BpLY3utyfBMS9pBJ8lxg6M0qjthoLwQQojTi1ugX1ZWxpAhQ8jIyGhUPmrUKJRSlJWVNbud4zjs2rWLESNGNHlt5MiR7Nu3j0AgAMDOnTsBmtQdPnw4uq7HXm9JRUUFAHl5ea07KXFaYQcqIjp94pT+MMCjOD/D4s0KN74zBPthB/YEDAbE4aZEtM5ZHofP/V0X6CsF7x8zOSfdJkPy8juNpmnkux32d+MYCyGESFVxu9L6fD4KCgqalOfn5wO02KNfVVVFOByO1Tt1W6UUPp8v9h5ut5vc3NxG9RrKTvfUAOCZZ57BMAwuvfTSVp2TOL3DIZ1cU2HGsSOvv0cxItNirc/Nrrrmo7eIA2sr3PR1OzJPegLr44qutNqZi6OdbJffIORoDPHKzV5ny3c5kqcvhBDdIG5hTDAYxOXFocnDAAAgAElEQVRyNSn3eDwALQ6AbSh3u90tbhsMBk/7Hg11TzfIds2aNbz00kt897vfpbCw8DRn0rLevTPbtV0iyM/P6vR9bt5vMSgbcnKb/510lxygb47ig6MaxzU3MwcZuE9MqxNxFCv32GR64MICDU1r2s5EUzm5aXF53yGOw0HlZWh+517KgrZi6xGLi/rr5Hri216T0ZB8DyX7HLJ7efDIlFbiDLri+0gkL2kvjcUt0Pd6vUQikSblDcF3Q9B+qobycDjc4rZerzf23+bqNdRt6T02b97M0qVL+frXvx4b2NselZV1OE7PG8CXn5+Fz9e504oqBburPIzLsqiuSozPZGIW7Kgz+G2pQ2+XQ7apOGZp6MDwTJua6jPuQhAN8qurAnF57z5KY3OFyQWu2k4d4P1RlUlvU8MIhKmOz6klrZzcNOprguQaJiVf1XN2mjwxES3riu8jkbxStb3outZi53Lcnp3m5+c3mzrTkHbTXFoPQG5uLm63O1bv1G01TYul9eTn5xOJRKiqqmpULxwOU1VV1ex7fPbZZ3zve9+juLiYRx99FMOQ5NzOUG1pOAoyjcQI8gFMDUZl2kzIjlDgdnAUZJuKUZkyALenyDIVHl1xqBPnZa+1ND73GwxNszttn6KpPm6nW9dCEEKIVBS3q+ywYcPYu3cvfr+/Ufn27dtjrzdH13WKioooLS1t8lpJSQmDBw8mLS2aRnD++ecDNKlbWlqK4zix1xvs37+f2267jV69evHf//3fpKent+/kRBMHgzr5XbwabntoGmSZ0M+jOCfd4by07p36U3TcWV6bf3bioNxN1SaD0+wun7Yz1UXz9A1U4tz7CyFE0olboD9r1iwikQgrV66MlYXDYVatWsW4cePo27cvAIcOHWoyveXMmTPZtm1bo1lzvvjiCzZu3MisWbNiZZMmTSI3N5cXX3yx0fYrVqwgPT2diy++OFbm8/m49dZb0TSN3//+9/Tq1atTzzfV7Q/q9HHJI3rR+c5yKw4EdUKd0LyOhjSOhnTOkXSSLpdhgK5BRRcNphZCCBHHHP3Ro0cza9YsHn74YXw+H4WFhbzyyiscOnSIBx54IFZv8eLFbNq0iV27dsXK5s2bx8qVK7njjju45ZZbMAyD5cuXk5+fH1t8C6I5+j/4wQ+4//77WbRoEdOmTWPz5s289tpr/OhHPyI7OztW97bbbuPAgQPcdtttbNmyhS1btsReKywsZOzYsV37gSQxywFfWGdEhqRCiM7n1k8soFVnMCa7/W1MKfiwykVRuo2MD+16mgZneWx2+Q3y3Va8D0cIIZJSXCcPfPDBB3nsscdYvXo11dXVFBcX8/TTTzN+/PjTbpeZmcnzzz/PL3/5S5566ikcx2HixIksXbq0yZz38+fPx+VysWzZMtavX0///v1ZunQpCxcubFTvs88+A+B3v/tdk/ebM2eOBPodcDikk2MqXJKOK7rI2WkOm2tMRmTZ7Z6+dXe9jqOi8/OL7jHQ4/BBlYuJOZZcH4QQogtoSkmGZFeRWXeiNhw3cYBzJR0iKcVz1p2Tba4xKcqwKW7Hk6OwAy8d8TA2yyLX1fP+ZnuSU9vLlhqDogyHInniJ5qRqrOoiPZJ1faSkLPuiNRxIKiTL8GT6GJne21Kato3uHN7jUlvlyNBfhwM9Dh81sLidUIIITpG1v0UXao6omEpjawEmlZTJKfeLoWmRQd+D27D06MaS+Mzv8G03Kbreoiul+9W7PBrHI9o5HXSjZZScCSks8tvUGNrFLgcCjwO/T0OaXJPIYRIIdKjL7rUgQSdVlMkH02DIWkOJbWt779QKppadrZMpxk3ugYDPHan9ep/Ua/z5yNu/n7cxK0rhnhtIsDOOoNVRz0cCcnFSAiROiTQF11qf1AnX6bVFN2kn9uh3o4Ge63xaa1Bva3JdJpxNtDrsCdgYHegQ99R0RWNP6pyMSLTZlquxZA0hz5uxXlpDuOzbUZmWrxT6eaffvnqE0KkBrnaiS4TOTGtZh/JexbdRNdgdJbNhioXNdbpe26PhjRK6kzGZFmySFqcZRiQazqU1ravVz/kwNoKF0dCOlNyI/Q6kcZ1qny34sJsi83VLrbVyCMcIUTyk0BfdJlDIZ1cU2FKKxPdKMdUnJtusb7S1WIPcdCGdyvdjMiwJGc7QZyfYVNSa1LdxgW0/DasKXfj1hUTsi3cZ7jeZJmKybkRdtaZHAzKxUkIkdzkKie6zP6ATh+3pESI7jfYo3Brik1VTfP1qyIa6yrc9PXY9PXI06ZEkW7Auek2fz/uavXMSdURjTXlbvp5HC7IcFr9ZMajw6hMi/ePuQjIrJ5CiCQmgb7oEkrBwaBBgeTnizjQNBiZafNlUOe1o2521hnUWvBxlcmacje93Q7F6dI2E83ZXoeQA5/5z/yYpSKs8YbPzZA0u11rdPR2KwZ4bf52rPU3FkII0dNIoC+6xPGIhqZFc2+FiAeXDhflWgxOs9kf0Fl11ENFRGNaboRz0lrf+yu6T8MN2pZqk2MtpPAoFR1E/abPzbAMi0Jv+6P089IcArbGp+0cGyCEEIlO5tEXXWJ/UKePS6bVFPGla1DgVhS4bZSypT32AFmmYlimxRvlbooybMZlW7j0aIBfY2lsqDIJ2BpTciOkdzA+jw7etthQ5eLsdIdsU7r2hRDJRQJ90SUOBA0GeyX5VSQOCfJ7jgEeRR9XhM/8Bi8d8ZBhKqoiGoYGg7w2ozLtTnsik27A2ek2H1aZzOwji6YJIZKLBPqi0wXsaOrO2CzpHRNCtI9Hj06VWhVxsBVkZaozzqjTXkO8Dv+oMtkf0CmUNRWEEElEAn3R6b4MGOS7HQzpQRVCdFBuN6zDYWhwQUa0V/8sbxhTrl1dpiKs4Qv/644ty1QM8EiapxBdRQJ90en21Ouc5ZFeMSFEz5HvVhwIKj6tMRibI2mHnclRsC+gU1pr4re1RtMuV1satoLhmTZFGXaXPbURIlVJoC86VcCGyojO6Cz5ohRC9CzDMqKrKg/NsMmUb8dO4bfg7Uo3Cjjba1PgVo3GVygFVZbGl0GdklqTS3qH6SfrWwjRaeTeWXSqfQGDAknbEUL0QOkGDE6z2VTtivehJIXDIY1Xyz30cTtMzLbo51FNBlFrGuS5FGOybIZnWrxT6aa01pC1DYToJBLoi061p16nn6yGK4TooYakORwJ6RwJSW9FR5TVGayvdDMy0+LctNbl4Be4FZNzorMt/e2YC0eCfSE6TAJ90WnqbTgW0enjlquzEKJnMjUozrD5sEoCzfb6rM5ga43JpJwI+W38Pkg3YFKORa2l8ddjLhzp2heiQyTQF51G0naEEMmg/4mnkv/0y4q5bfV5ncGWGpMLcyLtXhnd0GBctkW9Da/uteWGS4gOkEBfdJov6g1J2xFC9HiaBudn2GyuMQnJJa3Vdvt1Pq4xuTC7/UF+A0ODsVk2fkud6NnvnGMUItVIoC86Ra2lcTyiSdqOECIp5JiKfh6bjVUy/U5rHAzqbKx28bVsq9NmLDI0mNpPx29p/OO4KQN0hWgHCfRFp9hZZzDQa0vajhAiaRSlOXwVNDgUlK/K0/GFNf5W6WJslkWW2bnRuKFpjM22qAjrfFQtwb4QbSVXL9FhEQc+9xsUeuUZtxAieZg6XJBp8Y/jJpYEmM2qjmisq3AzItOiVxetYmxqMD7bYn9AZ1utjJsQoi0k0BcdtqfeoJfLIV2uv0KIJNPXrcg0FVurJYXnVH4L3qxwMTTdom8XL3Ll1uFrORa7/AalEuwL0WoS6IsOUQpK6wwGS2++ECJJXZBhs6vekLn1TxKw4S8VbgZ5HQZ5u+dxh1eHC7MtPq012SnBvhCtIoG+6JDDIR2l6LJHtkIIEW8eHUZlWrxb6abejvfRxF/Ygbcq3BS4Hc5J695OnjQDLsyJsK3WZFedBPtCnIkE+qJDSusMClu56qEQQvRU+W7FQK/Nu5XulJ7qMeTAWz43WYZiaDcH+Q3SDfhaToQtNSY7pGdfiNOSQF+027GwRnlIZ4BH0naEEMnvvDQHR8HHKZqvH7DhjXI3GabDBRl2XDt4Mg2YmBPh0zqDLTIbjxAtkkBftItS8I8qF0MzLJlSUwiREjQNRmdZ7A0YfFqTWj3JdRasKXfTx+1wfnpiPMVNN2BSjsXegM6GKjOln7QI0RIJ9EW7/LPeIOLAoC6eaUEIIRKJW4cLsyPs8BtsS5Fg/2hI47VyDwO9DkMTJMhv4DkxQLcyrPMXn4yhEOJUEuiLNgs50UfX8X50K4QQ8ZBmwMTs6FSPn1QbSZs2ohSU1hq8XelmeKbFkDjl5J+JS4cJ2RbZpsOrRz0cltmRhIhJzURD0SEfV5v0dTvkykw7QogU5TVgYo7F5hqTyojORXkRvEnUwR+w4YPjLqosjck5kYRfJ0XTYGi6Q66peLfSzTlpNuNzLNzSnSlSnPwJiDbZH9D5MmBQlC7PR4UQqc2jR3PETU2x6qiHA4Ge/5XqKNhZa/DyUQ+GppiUYyV8kH+yfLdiWm6EWlvj5SMe9tbrSfvERYjWiOtVKRwO89BDDzFt2jRGjRrF9ddfz4cfftiqbY8ePcqiRYuYMGEC48aN46677uLAgQPN1l25ciWzZ89m5MiRzJw5kxdeeKHD+0xFh4M67x93MS7LwtXzv8+EEKLDDA2GZTiMyrT4x3EXb/lcHO2BqSNKRTtyXit3s6ve4GvZFsMynB452YJbh5GZNiOzLD6uNlld7mZ/QAJ+kZo0peLX9O+55x7WrVvHwoULGTx4MK+88gqlpaU8//zzjB07tsXt/H4/11xzDX6/n5tvvhnTNFm+fDmapvHqq6+Sk5MTq/vHP/6R//iP/2DWrFlMnTqVzZs3s3r1ahYvXsytt97arn22VmVlHU4PnAYgPz8Ln6+2UZkvrLG2ws2YTIve7p53TqLr5OSmUV0ViPdhiB4imduLreCroM4XQYMsQ3F+psUgr5PQ6SOWgn31OttrTRRwTppDP3fiDLjtaHtRCo6ENfYEDAzggkybc9LthP6d9GRKQUOEoEG3t6Pm4pdUoOsavXtnNvta3AL9kpIS5s6dy7333svNN98MQCgU4vLLL6egoKDFXneAZ555hl//+tesWrWKCy64AIA9e/ZwxRVX8N3vfpdFixYBEAwG+bd/+zfGjx/PU089Fdv+Rz/6Ee+++y7vvfceWVlZbdpnWyRLoP9VUOdvx1wMz7DoK7PsiFMkc+AmOl8qtBdHRVcNPxLWOBbR6edxKPQ69PM45Jgq7kG0o+BISGd3fTQVM8dUnJ1m08cV/2M7VWe1F6XAF9E4GNSpjOgM8joMSbc5y5PYN2KJKORAZVjnuKVxLKxRY2kEHI2ArRE+ESI0NCOXBmmGIk1X5LoUeS5Fnqno7e6az10C/abiNhj3rbfewuVyMXfu3FiZx+Phuuuu49FHH6W8vJyCgoJmt127di1jxoyJBeQA5557LpMnT+bNN9+MBeUfffQRVVVVzJs3r9H28+fPZ82aNbz//vt885vfbNM+U0nQho1VLg6FdEZmWuRLT74QQpyRrsEAr8MAL0Qcm6NhnX0Bna01JjaQ73Lo7Vb0djnkuRRZpurSFBlLnVjgMKzzVVDnaFgnw1D08zhMy02uQcQt0TQocCsK3DYhx+ZwSKekxuQ9S6O3y6G/R1Hgdsh3OynxebSWraDyRNspD+v4whpBRyPHVGQaigxDMcjr4NXBoytc2r968ZWCsIKQoxFywG9rfBXU2WVHbw7SdUUft6Kvx6HA7dDLpdAT7EYzGcQt0C8rK2PIkCFkZGQ0Kh81ahRKKcrKypoN9B3HYdeuXdxwww1NXhs5ciQffPABgUCAtLQ0du7cCcCIESMa1Rs+fDi6rrNz506++c1vtmmfqaAiqNhSbfKZ36C/x+aivAim/PEJIUSbuXQY6HUY6AWwCdhQbWnU2BpHQgZ+W6Pe1kgzFFmGIsOELEORbii8usKrg1tXuHRwaf+6IdCJpkjYKhrIW0oj6EDQie6v1ooGU9WWRp2lkWUqss1oYDUsI5LSvdgeHc5Oczg7zcFSUBnRqIpoHAyaVFkaLg2yzX/dhGUY0aA23QCvrjC17k9J6Q5hJ9o2K8M6FWENX0Sn2tLINBS5piLHVAzMssk0Wnf+mgYeLXoDAJDPvzoLHQV1drR9HgzqlNYaBByNPJci3+WQ74kG/jldfBOcCuIW6Pt8Pvr27dukPD8/H4Dy8vJmt6uqqiIcDsfqnbqtUgqfz0dhYSE+nw+3201ubm6jeg1lDe/Rln22hZ7gt6bOibvteluj2tI5HtYoD2uEKy36uDSm9rLJNBT/eggnRFOGBi65EotWSvX24jIg2w3RMD0a+CgFQQUBWzvR+6lx3NaJWBBx/hXIOwrshq1OxEyGFg36DS16M+DWojcXXl0xIE0x1FBk6Kem5PScz7+r24sLGGjCwDQF2CgFoRPfi/W2Rr3SqAprBG0IOxphANXwOSvcejQ9JfrZR28CXCduBkwdTKL/b2hgnrhRi/5Ee68Nok+AdLru5sFREDnRhsIOBG2NoNII2FBnafhtjVo7+lqGocgwINOl6J/mkGmcGmh33kF6TOjt+dffgeVA7Ymb1ENhk3/WQ72jkaErMs3osWWZ0TQgj3HSTfApn1+ix15d4XTnHLdAPxgM4nK5mpR7PB4gmq/fnIZyt9vd4rbBYPC079FQt2FfbdlnW+TlZZy5UsJq/nMTolmZ6fE+AtGTSHsRbSHtRbRBS7nqqSpuD++8Xi+RSKRJeUPQ3RBgn6qhPBwOt7it1+uN/be5eg11G/bVln0KIYQQQgjRE8Qt0M/Pz282Pcfn8wG0OBA3NzcXt9sdq3fqtpqmxVJw8vPziUQiVFVVNaoXDoepqqqKvUdb9imEEEIIIURPELdAf9iwYezduxe/39+ofPv27bHXm6PrOkVFRZSWljZ5raSkhMGDB8cGzZ5//vkATeqWlpbiOE7s9bbsUwghhBBCiJ4gboH+rFmziEQirFy5MlYWDodZtWoV48aNiw3UPXToEHv27Gm07cyZM9m2bVtsVh2AL774go0bNzJr1qxY2aRJk8jNzeXFF19stP2KFStIT0/n4osvbvM+hRBCCCGE6AniujLuokWLWL9+Pd/+9rcpLCyMrYz77LPPMn78eAAWLFjApk2b2LVrV2y7uro65syZQyAQ4JZbbsEwDJYvX45SildffZW8vLxY3RdeeIH777+fWbNmMW3aNDZv3syrr77Kj370I26//fZ27VMIIYQQQohEF9dAPxQK8dhjj7FmzRqqq6spLi7mnnvuYcqUKbE6zQX6AEeOHOGXv/wlH3zwAY7jMHHiRJYuXcqgQYOavM+f//xnli1bxsGDB+nfvz8LFixg4cKFTeq1ZZ9CCCGEEEIksrgG+kIIIYQQQoiukcJr4wkhhBBCCJG8JNAXQgghhBAiCUmgL4QQQgghRBKSQF8A0alNH3roIaZNm8aoUaO4/vrr+fDDD+N9WKIblZeX8/DDD7NgwQLGjh1LcXExH330UbN1169fz5w5cxg5ciRf//rXeeKJJ7Asq0m9mpoa7rvvPiZNmsSYMWNYuHAhZWVlXX0qohuUlJTws5/9jMsuu4wxY8bw9a9/nbvvvpsvv/yySd1PPvmEG2+8kdGjRzN16lT+8z//k0Ag0KSeXIeS16effsr3v/99vvGNbzBq1CimTp3Kd77zHT755JMmdaW9iFM988wzFBcXc9VVVzV5TdrL6UmgLwBYsmQJzz77LFdeeSVLly5F13Vuv/12tm7dGu9DE91k7969PPPMMxw9epTi4uIW67333nt8//vfJycnh/vuu48ZM2bw5JNP8sADDzSq5zgOd9xxB2+88QY33XQTP/7xj6msrGTBggXs37+/q09HdLHf/e53vP3220yZMoWlS5dy/fXXs2nTJq6++upGa5+UlZVx8803EwqFWLJkCddddx1/+tOfuPvuu5vsU65DyevAgQPYts3cuXO57777+M53vsOxY8e46aab+OCDD2L1pL2IU/l8Pn7729+Snp7e5DVpL62gRMrbvn27KioqUn/4wx9iZcFgUM2YMUPNmzcvfgcmulVtba06duyYUkqpt99+WxUVFamNGzc2qXfZZZepOXPmKMuyYmWPPPKIGjZsmNq7d2+s7I033lBFRUXq7bffjpVVVlaqCRMmqB//+MdddyKiW2zZskWFQqFGZXv37lUjRoxQixcvjpXddttt6qKLLlJ1dXWxsj//+c+qqKhIbdiwIVYm16HUU19fr6ZMmaLuuOOOWJm0F3GqxYsXqwULFqibbrpJXXnllY1ek/ZyZtKjL3jrrbdwuVzMnTs3VubxeLjuuuvYsmUL5eXlcTw60V0yMzPPuDDc7t272b17NzfccAOGYcTK582bh+M4rFu3Lla2du1aCgoKmD59eqysV69ezJ49m3feeYdIJNL5JyG6zbhx43C73Y3Kzj77bIYOHRrr0a+rq2PDhg1cffXVZGRkxOpdddVVpKen8+abb8bK5DqUetLS0ujVqxc1NTWAtBfRVElJCa+99hr33ntvk9ekvbSOBPqCsrIyhgwZ0ugPBWDUqFEopSSnWsTs3LkTgBEjRjQq79u3L/369Yu9DtF2NXz4cDRNa1R35MiR+P1+Sd9JQkopKioqYjeMu3btwrKsJu3F7XZz/vnnN7q2yHUoNdTV1XHs2DG++OILHnnkET7//HMmT54MSHsRjSml+PnPf87VV1/N+eef3+R1aS+tI4G+wOfzUVBQ0KQ8Pz8fIKnvdEXb+Hw+4F9t42T5+fmN2kpL7aqhTNpV8nnttdc4evQos2fPBjqnvch1KLn8r//1v5g8eTKzZ89m2bJlfOtb3+LOO+8EpL2Ixl599VV2797ND3/4w2Zfl/bSOma8D0DEXzAYxOVyNSn3eDwAhEKh7j4kkaCCwSBAk5QNiLaXk2c6CAaDzdZrKGvYl0gOe/bs4f7772f8+PGxmTHO1F5ObgNyHUoN3//+97nhhhs4cuQIq1evJhwOE4lEcLvd0l5ETF1dHb/+9a+54447mg3QQa4vrSU9+gKv19tsvnRDw2/4QxDC6/UC0WnKThUKhWKvN9Rtrl5D2cl1Rc/m8/n47ne/S05ODr/5zW/Q9ehXS1vbi1yHkl9xcTFTp07l2muv5fe//z07duyI5V9LexENfvvb3+JyubjllltarCPtpXUk0BdNHnE1aHgs1tLdtEg9DY85G9rGyU59NNpSu2ook3aVHGpra7n99tupra3ld7/7XaPH6J3RXuQ6lLxcLhfTp09n3bp1BINBaS8CiH5HPPvss8ybN4+KigoOHjzIwYMHCYVCRCIRDh48SHV1tbSXVpJAXzBs2DD27t2L3+9vVL59+/bY60IAsQFRpaWljcqPHj3KkSNHGg2YGjZsGDt27EAp1ahuSUkJ6enpFBYWdv0Biy4VCoW488472bdvH//93//NOeec0+j1oqIiTNNs0l7C4TBlZWVN2otch1JPMBhEKYXf75f2IgCorKwkEonw8MMPM3369NjP9u3b2bNnD9OnT+eZZ56R9tJKEugLZs2aRSQSYeXKlbGycDjMqlWrGDduHH379o3j0YlEMnToUM455xz+9Kc/Ydt2rHzFihXous6ll14aK5s1axbl5eWsX78+Vnbs2DHeeustpk+f3my+pOg5bNvmhz/8Idu2beM3v/kNY8aMaVInKyuLyZMns3r16kZfsKtXr6a+vp5Zs2bFyuQ6lNyOHTvWpKyuro61a9fSv39/evfuLe1FADBw4ECefPLJJj9Dhw5lwIABPPnkk1x99dXSXlpJU6d2t4mUtGjRItavX8+3v/1tCgsLeeWVVygtLeXZZ59l/Pjx8T480U2eeuopIDqw8vXXX+faa69l4MCBZGdnc9NNNwHw17/+le9973tMmjSJyy67jM8//5wXXniBG264gZ/+9Kexfdm2zbx58/jnP//JrbfeSl5eHitWrODw4cOsWrWKwYMHx+MURSf5xS9+wXPPPcc3vvGN2Cw7DTIyMpgxYwYAO3bs4Fvf+hZDhw5l7ty5HDlyhD/84Q9MnDiRZ555ptF2ch1KXgsXLsTj8TB27Fjy8/Nj14EjR47wyCOPcNlllwHSXkTLFixYQE1NDatXr46VSXs5Mwn0BRB9BP/YY4+xZs0aqqurKS4u5p577mHKlCnxPjTRjYqLi5stHzBgAO+++27s3++88w5PPPEEe/bsoVevXlx77bXcddddmGbjibyqq6t58MEHeeeddwiFQowcOZIlS5YwfPjwLj0P0fUWLFjApk2bmn3t1PayefNmHn74YXbu3ElmZiaXXXYZ99xzT5Ml7eU6lLxeeuklVq9eze7du6mpqSErK4sxY8Zw6623cuGFFzaqK+1FNKe5QB+kvZyJBPpCCCGEEEIkIcnRF0IIIYQQIglJoC+EEEIIIUQSkkBfCCGEEEKIJCSBvhBCCCGEEElIAn0hhBBCCCGSkAT6QgghhBBCJCEJ9IUQQgghhEhCEugLIYQQCWrVqlUUFxfz0UcfxftQhBA9kHnmKkIIIdrrwIEDPP3003z88cccPnwYt9tNnz59GDVqFHPmzGHSpEnxPsSktGDBAkpLS9m6dWu8D+WMysrKeOedd5gzZw4DBw6M9+EIIZKIBPpCCNFFPv30UxYsWIBpmlx99dWcd955BINBvvzySz744AMyMjIk0BeUlZXxxBNPcOGFF0qgL4ToVBLoCyFEF3nyyScJBAKsXr2aYcOGNXnd5/PF4aiEEEKkCsnRF0KILrJv3z5yc3ObDfIB8vPzm5Rt2LCBW2+9lQkTJjBy5EiuuOIKVqxY0ez2f/7zn5k1axYjRozg3//931m+fDkvv/xyk5zuJUuWUFxc3Ow+iouLWbJkSZPyv/zlL9x4442MHTuW0aNHM3fuXN566/AhAm8AAArxSURBVK0Wt9+6dSs33XQTY8aMYeLEiSxduhS/39+kvs/n4z//8z+ZPn06I0aMYPLkydxyyy188MEHjert27ePH//4x0ybNo0RI0ZwySWX8Ktf/Yr6+vpmz6O9lFK8+OKLXHPNNYwePZqxY8eyYMECNm7c2KjewYMHKS4u5vHHH+evf/0r1157LSNHjmTatGn86le/wrKsJvteu3YtV155JSNHjuTrX/86TzzxBBs2bKC4uJhVq1YB8Pjjj3PvvfcCsHDhQoqLi5v9nTiOw+9//3tmzJjBiBEjmDlzJq+88kqnfhZCiOQjPfpCCNFFCgsL2bt3L+vWrePSSy89Y/0//elP/Md//AdjxozhzjvvJC0tjQ0bNvDTn/6U/fv3s3jx4ljd5cuX88ADDzBs2DDuueceAoEAy5Yto3fv3h0+7kcffZT/+3//LxdddBGLFi1C13XefvttFi1axP/+3/+b+fPnN6pfVlbGnXfeyTXXXMPll1/Opk2beOmll9B1nZ///OexegcPHuTGG2+ksrKSq666ihEjRhAIBNi+fTsbNmxg6tSpAJSWlvLtb3+b7OxsbrjhBvr27ctnn33G888/z9atW3n++edxuVwdPk+AH//4x7zxxhvMnDmTa665hnA4zJo1a7j11lt5/PHHmT59eqP67733Hi+++CLf+ta3uPbaa1m/fj3Lli0jJyeHO++8M1bvL3/5C/fccw+FhYX8j//xPzAMg1dffZV333230f7+/d//HZ/Px5/+9CfuvPNOzjnnHCDadk726KOPEgwGueGGG3C73axYsYIlS5ZQWFjI+PHjO+WzEEIkISWEEKJLfPLJJ2r48OGqqKhIXXrppWrJkiXqhRdeULt3725S9+jRo2rEiBHqnnvuafLaz3/+czVs2DC1f/9+pZRS1dXVavTo0Wr27Nmqvr4+Vu/w4cNqzJgxqqioSG3cuDFWvnjxYlVUVNTsMRYVFanFixfH/l1aWqqKiorUr3/96yZ1v/e976mxY8eq2traRtsXFxerbdu2Nap7++23qwsuuEDV1dXFym677TZVVFSk3n///Sb7tm079v9XXHGFmjlzZqP3UUqpdevWqaKiIvXyyy83ey4nu+mmm9SYMWNOW6dhf3/84x8blUciETVnzhz1jW98QzmOo5RS6sCBA6qoqEiNHj1aHThwIFbXcRz1zW9+U02dOrXR9tOmTVOTJ09WVVVVsfK6ujp1ySWXNDmHl19+ucnv7NTXrrrqKhUKhWLlR44cUcOHD1d33333GT8LIUTqktQdIYToImPHjuXll19mzpw51NbWsmrVKn72s59x2WWXMX/+fA4cOBCru3btWsLhMNdddx3Hjh1r9HPJJZfgOA4bNmwA4B//+AeBQID58+eTlpYW20e/fv244oorOnTMa9asQdM0rr766maPw+/3s23btkbbjBkzhtGjRzcqmzRpEpZl8dVXXwFQVVXF3//+dy666CIuuuiiJu+r69Gvo127drFr1y4uv/xywuFwo/cfP3486enpTdJ82uu1114jIyODGTNmNHqfmpoaLrnkEr766iv27dvXaJvp06c3GjCraRoTJ07E5/PFUpV27NhBeXk5c+bMIScnJ1Y3IyODb33rW+061nnz5uF2u2P/7tu3L0OGDGlyfEIIcTJJ3RFCiC5UXFzMf/3XfwHw1Vdf8fHHH7Ny5Uo2b97MXXfdxcsvv4zb7WbPnj0A3HzzzS3uq6KiAoimwACxNI+TnXvuuR063j179qCUYvbs2Wc8jgaDBg1qUic3NxeIBvgA+/fvRynFBRdccMb3h2ju+uOPP96q92+vPXv24Pf7mTJlSot1KisrGTJkSOzfZzrXjIyM2O/n5O0aNFfWGi29b8ONlBBCNEcCfSGE6CYDBgxgwIABXHXVVcybN49PPvmEkpISJkyYgFIKgF/96lcUFBQ0u31zwV5raJrWbHlzA0iVUmiaxjPPPINhGM1ud9555zX6d0v1GvbXHrfeemuzPf8A2dnZ7drnqZRS9OrVi1//+tct1hk6dGijf3fFubZGwxMPIYRoCwn0hRCim2maxujRo/nkk08oLy8H4OyzzwYgLy/vtD3MQCx15IsvvmDy5MmNXmvoET9ZQ/pIVVVVrPcZaJQ61ODss8/m73//O2eddVaHnw6crLCwEE3TKCsrO229wYMHA9HA9kyfQ0cNHjyYffv2MXr0aDIyMjptvwMGDABg7969TV5rrqylGzEhhOgo6SIQQogu8sEHHzTbax4MBmN55g3B9OzZs3G73Tz++OMEg8Em29TW1hIOhwGYOnUqXq+XF154gUAgEKtz5MgR1qxZ02TbhpuIhhz/Bn/4wx+a1L3yyisBeOSRR7Btu8nr7U2byc3N5f+3d2+h7L9xHMDfflY0xhyWwyZTzFm7kTvFBXJY3znEnVK7I2XkLG4shBAZRTnUpliSCzlczLWicOFCZBcih7UQk9//bv/2n3+/fn781f7v1+Wep+3zfK/ee/o83yc7OxtWq9WjDuDv3fCUlBSoVCqYTKZ3/4i8vr662oH+lCAIeHt7w9DQ0LvjH11rWloaZDIZLBYL7Ha76/OHhweYTCaP+WKxGADc5hIRfQbu6BMRfRGDwYD7+3vk5uZCpVLB39/fFcbPzs4gCILr/faRkZHo7u5GR0cHCgsLodFoIJfLcXt7i5OTE2xtbWF9fR0KhQLBwcGor69HX18fqqqqIAgCnp6eYDKZoFQqcXx87FZHcXExhoeH0dXVhdPTU0ilUuzu7uLu7s6j5oyMDNTV1WFsbAyCICA/Px8RERG4urrC0dERrFYrDg8PP/Q8Ojs7cXx8DJ1OB0EQkJqaiufnZxwcHEAul6OpqQk+Pj7o7+9HdXU1NBoNysrK3G4U3tzcRENDA0pLS3/5e06nExMTE++O5eXloaCgAKWlpVhYWMDR0RFycnIQEhKCy8tL7O/v4/z8HNvb27+9TpFIhObmZjQ2NqKiogLl5eXw9fWFxWKBVCqFzWZz28VPT0/Hjx8/MDk5CbvdDrFYDIVC4XHAmYjodzHoExF9kZaWFmxvb2Nvbw8bGxtwOByQSCRQqVTQ6XQeYbWsrAxKpRIzMzMwm81wOByQSqWIi4tDfX292wVbNTU1EIvFmJ2dxeDgIKKiolBTUwOJRIK2tja37w0MDMTU1BQMBgOMRiPEYjHy8vIwMDCAzMxMj7pra2uRlpaG+fl5zM3N4fHxEWFhYUhISEB7e/uHn0dMTAyWl5cxPj4Oq9WK1dVVBAUFISkpCZWVla55ycnJsFgsMBqN2NnZgclkQkBAAORyObRarUe70r9xOp0YGRl5dyw2Nhbx8fEwGAzIysrC0tISjEYjnE4nZDIZUlJSoNfrP7zWkpISiEQiTExMYHR0FOHh4SgvL0diYiJqa2vh5+fnmhsdHY3e3l5MT0+jp6cHTqcTWq2WQZ+I/pjPz688PURERP+plZUVtLa2Ym5uDllZWd9dDv3DzMwM+vr6YDaboVarv7scIvJy7NEnIiL6ZC8vLx5nHB4eHrC4uAipVPrL14wSEX0Gtu4QERF9souLC+h0OhQVFUGhUOD6+hoWiwU2mw3d3d1ul18REX0VBn0iIqJPFhoaCrVajbW1Ndzc3EAkEkGlUkGv16OwsPC7yyOi/wn26BMREREReSH26BMREREReSEGfSIiIiIiL8SgT0RERETkhRj0iYiIiIi8EIM+EREREZEXYtAnIiIiIvJCfwHt6NQbLwCNQQAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plotChainLength(dev , \"sequence\" , \"dev\")"
      ],
      "metadata": {
        "id": "3aEuB9ctVNC-",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 427
        },
        "outputId": "2b515d0f-3b33-4876-a56d-7b1df19ca1a9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plotChainLength(test , \"sequence\" , \"test\")"
      ],
      "metadata": {
        "id": "stho0s9OVSlp",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 427
        },
        "outputId": "43ccf3a8-7e91-41ca-b77a-7e3e2d892337"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "*   Since the problem will require to encode the sequences in some way, it is important to define a **maximum length** for the input. Based on the distribution of characters in all sets, it would be reasonable to have a maximum length set around 200 characters. \n",
        "After running into GPU RAM limitations, it was necessary to reduce the maximum length to 128.\n",
        "*   Another design decision would be to discard long or extremely short sequences, but I will try to show how the proposed models are robust with respect to this type of data.\n",
        "\n"
      ],
      "metadata": {
        "id": "H2cv-UpRoXp3"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "MAX_LEN = 128"
      ],
      "metadata": {
        "id": "tjHmUno5v5yK",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:23:22.063977Z",
          "iopub.execute_input": "2022-02-17T23:23:22.064394Z",
          "iopub.status.idle": "2022-02-17T23:23:22.068259Z",
          "shell.execute_reply.started": "2022-02-17T23:23:22.064342Z",
          "shell.execute_reply": "2022-02-17T23:23:22.067567Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from collections import Counter\n",
        "\n",
        "def countAminoacids(df , seqCol):\n",
        "    \"\"\"\n",
        "    -Description: Bar plot of occurences of aminoacids in the training set\n",
        "\n",
        "    @param DataFrame df: dataframe containing the sequences\n",
        "    @param str column: column to access in the dataframe\n",
        "\n",
        "    \"\"\"\n",
        "    counter = Counter(aa for seq in df[seqCol].values.tolist() for aa in seq )\n",
        "    y = [count for tag, count in counter.most_common()]\n",
        "    x = [tag for tag, count in counter.most_common()]\n",
        "    #Obtain values sorted for the bar plot\n",
        "    sns.barplot(x=x,y=y, palette=(\"Blues_d\"))\n",
        "    plt.xlabel(\"Aminoacid counts\")\n",
        "    plt.ylabel(\"Aminoacid codes\")\n",
        "    plt.title(\"Aminoacid Counts in the training set\")\n",
        "    \n",
        "    return "
      ],
      "metadata": {
        "id": "sqgrc9R7WNsi"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "countAminoacids(train , \"sequence\")"
      ],
      "metadata": {
        "id": "GXZTr9-dXS-v",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 427
        },
        "outputId": "68085ec9-6529-4860-a01a-536316d11f33"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The occurences of some aminoacids are very limited. This is important to notice and usually in NLP applications the least occuring words are **mapped into a common token**. This is again a design question that I will consider. \n",
        "\n",
        "\n",
        "\n",
        "---\n",
        "\n"
      ],
      "metadata": {
        "id": "39VyBuV54zuW"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Method Explanation:\n",
        "\n",
        "\n",
        "\n",
        "*   Given the task of multi-class classification, it is necessary to find a model with the right **inductive bias** given the sequential nature of the input data. For this purpose, it is interesting to read recent papers on **NLP**, assuming that those methods that perform well for natural language would work in the case of protein sequences.\n",
        "---\n",
        "The first model that I will implement is based on the work of Jennifer M. Johnson *et al.* (2020) **Journal of Chemical Information and Modeling** : [Deep Dive into Machine Learning for Protein Engineering](https://pubs.acs.org/doi/10.1021/acs.jcim.0c00073).\n",
        "\n",
        "\n",
        "*   The proposed model is a convolutional NN with dense layers on top, with the (1-dimensional) convolutional layers would be able to extract relevant features and learn the underlying patterns present in each protein family. Thus, these features extracted by the first layers would be relevant for the classification layers to perform the task.\n",
        "*   Related to inductive biases, it is important to note how this architecture would be shift invariant (translational invariance), thanks to the max pooling layers as well.\n",
        "\n",
        "![picture](https://drive.google.com/uc?export=view&id=1YCmmJ30kkwD8-h8erAadAK89uBlZVTst)\n",
        "\n",
        "\n",
        "---\n",
        "\n",
        "\n",
        "Given the success of pre-trained models in NLP, my second solution would be geared towards leveraging transfer learning. In particular, I will be using the model from Burkhard Rost *et al.* (2020) **IEEE Transactions on Pattern Analysis and Machine Intelligence** [ProtTrans: Towards Cracking the Language of Life’s Code Through Self-Supervised Learning](https://doi.org/10.1101/2020.07.12.199554). The model hosted on HuggingFace API [ProtBert](https://huggingface.co/Rostlab/prot_bert) was trained on a self-supervised task, namely MLM (Masked-Language Model). \n",
        "\n",
        "\n",
        "\n",
        "*   After downloading the model with the trained parameters, I will add additional layers for the classifier and then **fine-tune** the whole resulting model in our downstream task. \n",
        "*   Pre-training on a MLM objective also allows to learn an internal representation which can be used by the classification layers to incorporate useful features. Transformer-based models capitlize on the **attention mechanism**, such that a certain element of the sequence can attend on another element, thus in a sense it is able to generate contextual embeddings.\n",
        "\n",
        "\n",
        "![picture](https://drive.google.com/uc?export=view&id=15jY6eLQJ_K3nwi_kb5_niNqLPdeXKugz)\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "---\n",
        "\n",
        "\n",
        "It is also important to note how the paper which produced the dataset by *Bileschi et al.* (2019) **BioArxiv** [Can Deep Learning Classify the Protein Universe](https://research.google/pubs/pub48390/#:~:text=Our%20model%20co%2Dlocates%20sequences,purpose%20protein%20function%20prediction%20tools.) is based on a 1-D CNN model as well. "
      ],
      "metadata": {
        "id": "6wLKiWHqZa0m"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Experiment Description\n",
        "\n",
        "*   The hypothesis that I want to test in this experiment is if fine-tuning a **pre-trained LM on the classification task performs better than convolutional based models**.\n",
        "*   The evaluation metric for this experiment will be accuracy on the hold-out set, the test set. With more domain knowledge it could also be interesting to understand performance for a particularly important class, where more insightful metrics could be employed (Precision, recall and F1 for example). \n",
        "\n"
      ],
      "metadata": {
        "id": "D6Hz7jQCohtB"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#Slice the datasets such that only relevant columns remain\n",
        "train = train[[\"sequence\" , \"family_accession\"]]\n",
        "dev = dev[[\"sequence\" , \"family_accession\"]]\n",
        "test = test[[\"sequence\" , \"family_accession\"]]"
      ],
      "metadata": {
        "id": "2ML_d4lhYqHh",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:23:27.821202Z",
          "iopub.execute_input": "2022-02-17T23:23:27.821748Z",
          "iopub.status.idle": "2022-02-17T23:23:27.876508Z",
          "shell.execute_reply.started": "2022-02-17T23:23:27.821708Z",
          "shell.execute_reply": "2022-02-17T23:23:27.875626Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "For the CNN model, I will implement the model in Keras, thus it is necessary to convert the target labels into a onehot encoded representation."
      ],
      "metadata": {
        "id": "KdlN2PcPn1Nn"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.preprocessing import OneHotEncoder\n",
        "import numpy as np\n",
        "enc = OneHotEncoder(sparse=False, handle_unknown = 'ignore')\n",
        "y_train = np.expand_dims(train[\"family_accession\"].values, -1)\n",
        "y_dev = np.expand_dims(dev[\"family_accession\"].values, -1)\n",
        "y_test = np.expand_dims(test[\"family_accession\"].values, -1)\n",
        "\n",
        "enc.fit(y_train)\n",
        "\n",
        "y_train = enc.transform(y_train)\n",
        "y_dev = enc.transform(y_dev)\n",
        "y_test = enc.transform(y_test)"
      ],
      "metadata": {
        "id": "e0HU1_8bbWZx"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense, Conv1D, MaxPooling1D,AveragePooling1D, Flatten ,Dropout,BatchNormalization\n",
        "from keras.layers.embeddings import Embedding\n",
        "from keras.callbacks import EarlyStopping\n",
        "\n",
        "from keras.preprocessing import text, sequence\n",
        "from keras.preprocessing.text import Tokenizer\n",
        "from sklearn.model_selection import train_test_split"
      ],
      "metadata": {
        "id": "Upb_g_GjSYer"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "In order to input the protein sequences, it is necesary to implement a tokenizer to map character occurences into numbers. Then, the sequences which contain less characters than the maximum length, need to be padded to ensure that every data point has the same dimensions."
      ],
      "metadata": {
        "id": "KjkRCE3a6hiQ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "tokenizer = Tokenizer(char_level=True)\n",
        "tokenizer.fit_on_texts(train[\"sequence\"])\n",
        "#represent input data as word rank number sequences\n",
        "X_train = tokenizer.texts_to_sequences(train[\"sequence\"])\n",
        "X_train = sequence.pad_sequences(X_train, maxlen=MAX_LEN)\n",
        "\n",
        "X_dev = tokenizer.texts_to_sequences(dev[\"sequence\"])\n",
        "X_dev = sequence.pad_sequences(X_dev, maxlen=MAX_LEN)\n",
        "\n",
        "X_test = tokenizer.texts_to_sequences(test[\"sequence\"])\n",
        "X_test = sequence.pad_sequences(X_test, maxlen=MAX_LEN)"
      ],
      "metadata": {
        "id": "hyQ6KnIESwD2"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "The first layer of the model is going to be an Embedding layer, which matches every character in our vocabulary into a dense vector. This mapping is also learned as the other parameters in the network, leading to a more appropiate representation where then the convolutional layers can operate. \n",
        "This leads to another design decision, which is the **dimensionality of the embeddings**.\n",
        "In fact, this can be tuned by resorting to the validation set."
      ],
      "metadata": {
        "id": "ILJHyhxD63WH"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from functools import partial\n",
        "def cnnModel(output_shape , max_len , tokenizer , embedding_dim):\n",
        "        '''\n",
        "        @param int output_shape: output shape corresponding to the number of classes\n",
        "        @param int max_len: maximum length for the sequences\n",
        "        @param Tokenizer: keras tokenizer used as input for the embedding layer\n",
        "        @param int embedding_dim: \n",
        "\n",
        "        @output keras.Sequential model: final model without compulation\n",
        "        '''\n",
        "\n",
        "        model = Sequential([ \n",
        "            Embedding(len(tokenizer.word_index)+1, embedding_dim, input_length=max_len),\n",
        "            Conv1D(filters=4, kernel_size=5, padding='same', activation=partial(tf.nn.leaky_relu, alpha=0.02)),\n",
        "            BatchNormalization(),   \n",
        "            MaxPooling1D(pool_size=2, padding='valid'),\n",
        "            Conv1D(filters=8, kernel_size=3, padding='same', activation=partial(tf.nn.leaky_relu, alpha=0.02)),\n",
        "            BatchNormalization(), \n",
        "            MaxPooling1D(pool_size=2, padding='valid'),\n",
        "            Flatten(),\n",
        "            Dense(16, activation=partial(tf.nn.leaky_relu, alpha=0.02)),\n",
        "            Dropout(.5),\n",
        "            Dense(output_shape, activation='softmax')\n",
        "        ])\n",
        "        return model"
      ],
      "metadata": {
        "id": "pEMqG_G-TTi6"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "EMBED_DIM = 64"
      ],
      "metadata": {
        "id": "TlR61MG-m9Ek"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "cnn_model = cnnModel(NUM_CLASSES , MAX_LEN , tokenizer , EMBED_DIM)\n",
        "cnn_model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
        "ES = EarlyStopping(monitor='val_accuracy', verbose=1, patience=3)"
      ],
      "metadata": {
        "id": "e8KiMeC8VCbC"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(cnn_model.summary())"
      ],
      "metadata": {
        "id": "PgtjuyIDVprb",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e83aef49-466e-4e65-d02f-cb91c662d815"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " embedding (Embedding)       (None, 128, 64)           1472      \n",
            "                                                                 \n",
            " conv1d (Conv1D)             (None, 128, 4)            1284      \n",
            "                                                                 \n",
            " batch_normalization (BatchN  (None, 128, 4)           16        \n",
            " ormalization)                                                   \n",
            "                                                                 \n",
            " max_pooling1d (MaxPooling1D  (None, 64, 4)            0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " conv1d_1 (Conv1D)           (None, 64, 8)             104       \n",
            "                                                                 \n",
            " batch_normalization_1 (Batc  (None, 64, 8)            32        \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " max_pooling1d_1 (MaxPooling  (None, 32, 8)            0         \n",
            " 1D)                                                             \n",
            "                                                                 \n",
            " flatten (Flatten)           (None, 256)               0         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 16)                4112      \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 16)                0         \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 26)                442       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 7,462\n",
            "Trainable params: 7,438\n",
            "Non-trainable params: 24\n",
            "_________________________________________________________________\n",
            "None\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "history = cnn_model.fit(X_train, y_train,  batch_size=128, verbose=1, validation_data = (X_dev , y_dev),callbacks=[ES],epochs=50)"
      ],
      "metadata": {
        "id": "-upIa2DyVkq-",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4402dd80-8647-4df6-e92c-964445fde57f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/50\n",
            "269/269 [==============================] - 15s 15ms/step - loss: 2.3775 - accuracy: 0.3017 - val_loss: 2.5697 - val_accuracy: 0.2990\n",
            "Epoch 2/50\n",
            "269/269 [==============================] - 3s 12ms/step - loss: 1.3316 - accuracy: 0.5732 - val_loss: 0.5474 - val_accuracy: 0.9207\n",
            "Epoch 3/50\n",
            "269/269 [==============================] - 3s 12ms/step - loss: 1.0443 - accuracy: 0.6546 - val_loss: 0.2547 - val_accuracy: 0.9558\n",
            "Epoch 4/50\n",
            "269/269 [==============================] - 3s 12ms/step - loss: 0.9175 - accuracy: 0.6890 - val_loss: 0.2120 - val_accuracy: 0.9546\n",
            "Epoch 5/50\n",
            "269/269 [==============================] - 3s 12ms/step - loss: 0.8557 - accuracy: 0.7041 - val_loss: 0.1423 - val_accuracy: 0.9719\n",
            "Epoch 6/50\n",
            "269/269 [==============================] - 3s 12ms/step - loss: 0.8092 - accuracy: 0.7255 - val_loss: 0.1407 - val_accuracy: 0.9726\n",
            "Epoch 7/50\n",
            "269/269 [==============================] - 3s 12ms/step - loss: 0.7728 - accuracy: 0.7367 - val_loss: 0.1486 - val_accuracy: 0.9673\n",
            "Epoch 8/50\n",
            "269/269 [==============================] - 3s 12ms/step - loss: 0.7470 - accuracy: 0.7490 - val_loss: 0.0997 - val_accuracy: 0.9773\n",
            "Epoch 9/50\n",
            "269/269 [==============================] - 3s 12ms/step - loss: 0.7097 - accuracy: 0.7617 - val_loss: 0.1003 - val_accuracy: 0.9755\n",
            "Epoch 10/50\n",
            "269/269 [==============================] - 3s 12ms/step - loss: 0.6803 - accuracy: 0.7753 - val_loss: 0.1005 - val_accuracy: 0.9745\n",
            "Epoch 11/50\n",
            "269/269 [==============================] - 3s 13ms/step - loss: 0.6543 - accuracy: 0.7855 - val_loss: 0.0909 - val_accuracy: 0.9771\n",
            "Epoch 11: early stopping\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Early stopping** is applied based on the loss function evaluated on the validation set, preventing the model from overfitting to the training set. Other hyperparameters of the network can be tuned as well using the validation set."
      ],
      "metadata": {
        "id": "z6ECUJwp8EKx"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "sns.set_style(\"darkgrid\")"
      ],
      "metadata": {
        "id": "q6IRD6NG680d"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "sns.set(font_scale=1.5)\n",
        "plt.rcParams[\"figure.figsize\"] = (12,6)\n",
        "\n",
        "# Plot the learning curve.\n",
        "plt.plot(history.history[\"accuracy\"], '-o', label=\"Training Accuracy\" , color = \"lightskyblue\")\n",
        "plt.plot(history.history[\"val_accuracy\"], '-o', label=\"Validation Acuracy\" , color = \"indigo\")\n",
        "\n",
        "# Label the plot.\n",
        "plt.title(\"Training & Validation Accuracies\")\n",
        "plt.xlabel(\"Epoch\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.legend()\n",
        "plt.xticks(list(range(17)))\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "Zn6T7YTJ8_4e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 427
        },
        "outputId": "1004dc23-3b09-4ca1-a2cc-8fcea2b7713f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "sns.set(font_scale=1.5)\n",
        "plt.rcParams[\"figure.figsize\"] = (12,6)\n",
        "\n",
        "# Plot the learning curve.\n",
        "plt.plot(history.history[\"loss\"], '-o', label=\"Training Loss\" , color = \"lightskyblue\")\n",
        "plt.plot(history.history[\"val_loss\"], '-o', label=\"Validation Loss\" , color = \"indigo\")\n",
        "\n",
        "# Label the plot.\n",
        "plt.title(\"Training & Validation Losses\")\n",
        "plt.xlabel(\"Epoch\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.legend()\n",
        "plt.xticks(list(range(17)))\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "IMPDCOMw9KFz",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 427
        },
        "outputId": "116b00ba-c67c-41c4-c561-efd8040dfbdc"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "cnn_eval = cnn_model.evaluate(X_test, y_test)\n",
        "\n",
        "print(\"Test Set Accuracy: {:.2f}\".format(cnn_eval[1]*100))"
      ],
      "metadata": {
        "id": "Yb5uBLSubWcs",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "aa028c45-499f-470c-cc71-83f405d78af1"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "134/134 [==============================] - 1s 8ms/step - loss: 0.1071 - accuracy: 0.9759\n",
            "Test Set Accuracy: 97.59\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "For this second model, I will do the implementation in PyTorch, which would require a different encoding for the labels of the class. "
      ],
      "metadata": {
        "id": "jRHnI90-xYhN"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.preprocessing import LabelEncoder\n",
        "label_encoder = LabelEncoder()\n",
        "label_encoder.fit(train[\"family_accession\"].values)\n",
        "\n",
        "train[\"family_accession\"] = label_encoder.transform(train[\"family_accession\"])\n",
        "dev[\"family_accession\"] = label_encoder.transform(dev[\"family_accession\"])\n",
        "test[\"family_accession\"] = label_encoder.transform(test[\"family_accession\"])"
      ],
      "metadata": {
        "id": "GC9hStKRbGjV",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:23:34.364077Z",
          "iopub.execute_input": "2022-02-17T23:23:34.364356Z",
          "iopub.status.idle": "2022-02-17T23:23:35.198421Z",
          "shell.execute_reply.started": "2022-02-17T23:23:34.364324Z",
          "shell.execute_reply": "2022-02-17T23:23:35.197701Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Considering the preprocessing described in [ProtBert](https://huggingface.co/Rostlab/prot_bert) , it is necessary that every character in the sequence is white-space separated, and the less common aminoacids UZOB should be mapped to X as input for the tokenizer."
      ],
      "metadata": {
        "id": "1zDx8RynxjfZ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import re\n",
        "train[\"sequence\"] = train[\"sequence\"].apply(lambda x : \" \".join(x))\n",
        "dev[\"sequence\"] = dev[\"sequence\"].apply(lambda x : \" \".join(x))\n",
        "test[\"sequence\"] = test[\"sequence\"].apply(lambda x : \" \".join(x))\n",
        "train[\"sequence\"] = list(map(lambda seq : re.sub(r\"[UZOB]\", \"X\", seq ) , train[\"sequence\"].values.tolist()))\n",
        "\n",
        "dev[\"sequence\"] = list(map(lambda seq : re.sub(r\"[UZOB]\", \"X\", seq ) , dev[\"sequence\"].values.tolist()))\n",
        "\n",
        "test[\"sequence\"] = list(map(lambda seq : re.sub(r\"[UZOB]\", \"X\", seq ) , test[\"sequence\"].values.tolist()))"
      ],
      "metadata": {
        "id": "el48vJhfZFBU",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:23:36.102484Z",
          "iopub.execute_input": "2022-02-17T23:23:36.103116Z",
          "iopub.status.idle": "2022-02-17T23:23:36.482869Z",
          "shell.execute_reply.started": "2022-02-17T23:23:36.103079Z",
          "shell.execute_reply": "2022-02-17T23:23:36.482147Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#Checking the result of the preprocessing stage\n",
        "train.iloc[0].sequence"
      ],
      "metadata": {
        "id": "5kyWir6kYY9c"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install transformers"
      ],
      "metadata": {
        "id": "tMkSnh0zaF8J",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:23:39.125823Z",
          "iopub.execute_input": "2022-02-17T23:23:39.126080Z",
          "iopub.status.idle": "2022-02-17T23:23:47.704179Z",
          "shell.execute_reply.started": "2022-02-17T23:23:39.126051Z",
          "shell.execute_reply": "2022-02-17T23:23:47.703349Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from transformers import BertModel, BertTokenizer\n",
        "tokenizer = BertTokenizer.from_pretrained(\"Rostlab/prot_bert\", do_lower_case=False )\n",
        "pretrained_model = BertModel.from_pretrained(\"Rostlab/prot_bert\")"
      ],
      "metadata": {
        "id": "KPV7JzziaC0d",
        "outputId": "4ffdb134-4297-4fb8-a133-a43fa4c8fd26",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:23:47.706448Z",
          "iopub.execute_input": "2022-02-17T23:23:47.706739Z",
          "iopub.status.idle": "2022-02-17T23:24:38.659766Z",
          "shell.execute_reply.started": "2022-02-17T23:23:47.706704Z",
          "shell.execute_reply": "2022-02-17T23:24:38.658996Z"
        },
        "trusted": true,
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 251,
          "referenced_widgets": [
            "9665ca77c3194922a026ff34229969cf",
            "12c55eb9cd7043779fa73b3ca4c59853",
            "27dda3640926454ebbaf6b4c75fd527e",
            "57e9f76e59ea4a728a8c3cd5698131b7",
            "7d26c5f48d3d405ca85b85ac9b4c51b4",
            "929374a66d794f0facc166d5fda58a4b",
            "e108e4b0f8804ada96e8b837cf8f6989",
            "731c3767a9e74856a4b8110955ff7878",
            "28c7d1af50484bb4bdf9e32435666a70",
            "2174ad9564084f6487265c80ef463e5b",
            "7a626a9c9fa04b96b62b550cb008119d",
            "f63d5d8f8efc4b2594dafb6bb46a694a",
            "083ab680dae645b4993b4d47e4337117",
            "1f689d10cdaf430bb60421c6ba514769",
            "7ed9f595c03c40b5916282928c0aa7a0",
            "f22e7f71b3324fe8974eae6507d34a60",
            "5da1828804624d79b6f38e777a0e4473",
            "be3498de053d47fdb255f00ec3f8e23a",
            "80daef55c452484c804f46f8e818a467",
            "a4544f1a14c34d24875799ff253288a6",
            "33909405108845f98e7e34a5b5a26172",
            "2f83b47c9ea640c39a443f70d72ab6a7",
            "1c9aeb9510ed42babd7cc181bc1c2e1e",
            "4a5a8270efd747ada938646da50404ef",
            "ccc927c4e59044be8dd01c8bb709361e",
            "b4b367b2d0bd4e989754588f7d5280aa",
            "8da64c31495045d0a34cab69614619cc",
            "b95308c6445c45e4b05bd5195b615805",
            "ec9b284a48034eb0b5aa7ed7cd724321",
            "d49a09ae993844c6a254465c80ce3f45",
            "c0a23e15932b47e794ce2c6da43f63af",
            "e404998de5014b8f9bc0fca415f9ac81",
            "4e53920906bd4a36ad5c47d5b0bfef11",
            "3be5e9bf107b4b3c838c4c1ec178fef7",
            "f785d264b12843daa3389fb532ed571b",
            "3313aca3c1c448cd8bab6e35d460a827",
            "4c259618846d4c71bfcaedbb8a1df884",
            "531280b73a954f5d891c28b0dcbdde5a",
            "1d617358d87648f3bc1ea89e28386128",
            "130d5fc49b4f4ebe99467120280b977d",
            "e7a32c3559fb4d80bdda881e8f3af906",
            "8d9cefd9e9e64659b9222fd81b9d28c0",
            "d69162b4a1534a15b70d5753bc68a834",
            "3fbb3e490b5440279f5e796f3f01f451",
            "c9ccd2bbeadc486bb6ce728a48b05680",
            "81788b1f4b644908b5f8928069f51dd4",
            "04dedcc60efb4605b24fb8191dc4fd50",
            "c21decc775da470b995d5e025d6e4fb8",
            "029bbc7f7c4a48f8ad9cc4ccca0a8a10",
            "bc9ea8941b0248dc84a191f020934914",
            "8de2c09232544678b25c24bff32c6e53",
            "002f0156251c454c9f05f4a2d0e43091",
            "f1e8ab181ba64d14b117d19bef20954f",
            "614d238395284901be418686d1eaedc3",
            "6fcaa36637c6450c8dba1dbe2cef05a1"
          ]
        }
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "9665ca77c3194922a026ff34229969cf",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "Downloading:   0%|          | 0.00/86.0 [00:00<?, ?B/s]"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "f63d5d8f8efc4b2594dafb6bb46a694a",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "Downloading:   0%|          | 0.00/81.0 [00:00<?, ?B/s]"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "1c9aeb9510ed42babd7cc181bc1c2e1e",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "Downloading:   0%|          | 0.00/112 [00:00<?, ?B/s]"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "3be5e9bf107b4b3c838c4c1ec178fef7",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "Downloading:   0%|          | 0.00/361 [00:00<?, ?B/s]"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "c9ccd2bbeadc486bb6ce728a48b05680",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "Downloading:   0%|          | 0.00/1.57G [00:00<?, ?B/s]"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Some weights of the model checkpoint at Rostlab/prot_bert were not used when initializing BertModel: ['cls.predictions.decoder.weight', 'cls.predictions.transform.dense.weight', 'cls.seq_relationship.weight', 'cls.predictions.transform.LayerNorm.bias', 'cls.predictions.transform.LayerNorm.weight', 'cls.predictions.decoder.bias', 'cls.seq_relationship.bias', 'cls.predictions.bias', 'cls.predictions.transform.dense.bias']\n",
            "- This IS expected if you are initializing BertModel 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 BertModel from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import torch\n",
        "import numpy as np\n",
        "from torch import nn\n",
        "import torch.utils.data\n",
        "import torch.utils.data.distributed\n",
        "from torch.utils.data import Dataset, DataLoader, RandomSampler, TensorDataset\n",
        "\n",
        "class proteinDataset(Dataset):\n",
        "    def __init__(self, sequence, targets, tokenizer, max_len):\n",
        "        '''\n",
        "        @param list[str] sequence: protein sequence\n",
        "        @param list[str] targets: list containing the labels\n",
        "        @param custom tokenizer: huggingface tokenizer\n",
        "        @param int max_len: maximum length to truncate the sequences \n",
        "        '''\n",
        "        self.sequence = sequence\n",
        "        self.targets = targets\n",
        "        self.tokenizer = tokenizer\n",
        "        self.max_len = max_len\n",
        "    def __len__(self):\n",
        "        return len(self.sequence)\n",
        "\n",
        "    def __getitem__(self, item):\n",
        "        sequence = str(self.sequence[item])\n",
        "        target = self.targets[item]\n",
        "        encoding = self.tokenizer.encode_plus(\n",
        "            sequence,\n",
        "            truncation=True,\n",
        "            add_special_tokens=True,\n",
        "            max_length=self.max_len,\n",
        "            return_token_type_ids=False,\n",
        "            padding='max_length',\n",
        "            return_attention_mask=True,\n",
        "            return_tensors='pt',\n",
        "        )\n",
        "        return {\n",
        "          'protein_sequence': sequence,\n",
        "          'input_ids': encoding['input_ids'].flatten(),\n",
        "          'attention_mask': encoding['attention_mask'].flatten(),\n",
        "          'target': torch.tensor(target, dtype=torch.long)\n",
        "        }"
      ],
      "metadata": {
        "id": "fsEw2zESYgoR",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:24:38.661276Z",
          "iopub.execute_input": "2022-02-17T23:24:38.661557Z",
          "iopub.status.idle": "2022-02-17T23:24:38.674684Z",
          "shell.execute_reply.started": "2022-02-17T23:24:38.661519Z",
          "shell.execute_reply": "2022-02-17T23:24:38.670281Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_set = proteinDataset(train[\"sequence\"].values.tolist() , train[\"family_accession\"].values.tolist() , tokenizer , MAX_LEN)\n",
        "dev_set = proteinDataset(dev[\"sequence\"].values.tolist() , dev[\"family_accession\"].values.tolist() , tokenizer , MAX_LEN)\n",
        "test_set = proteinDataset(test[\"sequence\"].values.tolist() , test[\"family_accession\"].values.tolist() , tokenizer , MAX_LEN)"
      ],
      "metadata": {
        "id": "GWI9g_rjZVxC",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:24:38.676496Z",
          "iopub.execute_input": "2022-02-17T23:24:38.676956Z",
          "iopub.status.idle": "2022-02-17T23:24:43.240237Z",
          "shell.execute_reply.started": "2022-02-17T23:24:38.676918Z",
          "shell.execute_reply": "2022-02-17T23:24:43.239329Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "class pfamClassifier(nn.Module):\n",
        "    def __init__(self, n_classes , model):\n",
        "        '''\n",
        "        @param int n_classes: number of unique classes present in the dataset\n",
        "        @param custom model: huggingface downloaded model\n",
        "        '''\n",
        "        super().__init__()\n",
        "        self.model = model\n",
        "        self.classifier = nn.Sequential(nn.Dropout(p=0.2),\n",
        "                                        nn.Linear(self.model.config.hidden_size, n_classes),\n",
        "                                        nn.Tanh())\n",
        "        \n",
        "    def forward(self, input_ids, attention_mask):\n",
        "        '''\n",
        "        @param Tensor input_ids: tokenized inputs for the pre-trained model\n",
        "        @param Tensor attention_mask\n",
        "\n",
        "        @out Tensor: output from the classifier\n",
        "        '''\n",
        "        output = self.model(\n",
        "          input_ids=input_ids,\n",
        "          attention_mask=attention_mask\n",
        "        )\n",
        "        return self.classifier(output.pooler_output)"
      ],
      "metadata": {
        "id": "ukzv0k-sZxoR",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:24:43.241415Z",
          "iopub.execute_input": "2022-02-17T23:24:43.244426Z",
          "iopub.status.idle": "2022-02-17T23:24:43.252129Z",
          "shell.execute_reply.started": "2022-02-17T23:24:43.244384Z",
          "shell.execute_reply": "2022-02-17T23:24:43.251449Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "if torch.cuda.is_available():    \n",
        "\n",
        "    # Tell PyTorch to use the GPU.    \n",
        "    device = torch.device(\"cuda\")\n",
        "\n",
        "    print('There are %d GPU(s) available.' % torch.cuda.device_count())\n",
        "\n",
        "    print('GPU:', torch.cuda.get_device_name(0))\n",
        "else:\n",
        "    device = 'cpu'\n",
        "    print('No GPU available')"
      ],
      "metadata": {
        "id": "ezrrhnR-rwJF",
        "outputId": "889bae1e-da94-4058-d94d-74ce6ca62869",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:24:43.254674Z",
          "iopub.execute_input": "2022-02-17T23:24:43.254898Z",
          "iopub.status.idle": "2022-02-17T23:24:43.308313Z",
          "shell.execute_reply.started": "2022-02-17T23:24:43.254866Z",
          "shell.execute_reply": "2022-02-17T23:24:43.307545Z"
        },
        "trusted": true,
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "There are 1 GPU(s) available.\n",
            "GPU: Tesla K80\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model = pfamClassifier(NUM_CLASSES ,  pretrained_model)\n",
        "model.to(device)"
      ],
      "metadata": {
        "id": "dkjkjYOCaNlC",
        "outputId": "c92e4cd8-16d5-4ba2-86fb-1a554f81ed2f",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:24:43.310013Z",
          "iopub.execute_input": "2022-02-17T23:24:43.310350Z",
          "iopub.status.idle": "2022-02-17T23:24:46.575745Z",
          "shell.execute_reply.started": "2022-02-17T23:24:43.310281Z",
          "shell.execute_reply": "2022-02-17T23:24:46.573176Z"
        },
        "trusted": true,
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "pfamClassifier(\n",
              "  (model): BertModel(\n",
              "    (embeddings): BertEmbeddings(\n",
              "      (word_embeddings): Embedding(30, 1024, padding_idx=0)\n",
              "      (position_embeddings): Embedding(40000, 1024)\n",
              "      (token_type_embeddings): Embedding(2, 1024)\n",
              "      (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "      (dropout): Dropout(p=0.0, inplace=False)\n",
              "    )\n",
              "    (encoder): BertEncoder(\n",
              "      (layer): ModuleList(\n",
              "        (0): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (1): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (2): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (3): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (4): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (5): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (6): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (7): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (8): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (9): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (10): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (11): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (12): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (13): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (14): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (15): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (16): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (17): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (18): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (19): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (20): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (21): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (22): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (23): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (24): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (25): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (26): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (27): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (28): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "        (29): BertLayer(\n",
              "          (attention): BertAttention(\n",
              "            (self): BertSelfAttention(\n",
              "              (query): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (key): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (value): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "            (output): BertSelfOutput(\n",
              "              (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "              (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "              (dropout): Dropout(p=0.0, inplace=False)\n",
              "            )\n",
              "          )\n",
              "          (intermediate): BertIntermediate(\n",
              "            (dense): Linear(in_features=1024, out_features=4096, bias=True)\n",
              "          )\n",
              "          (output): BertOutput(\n",
              "            (dense): Linear(in_features=4096, out_features=1024, bias=True)\n",
              "            (LayerNorm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)\n",
              "            (dropout): Dropout(p=0.0, inplace=False)\n",
              "          )\n",
              "        )\n",
              "      )\n",
              "    )\n",
              "    (pooler): BertPooler(\n",
              "      (dense): Linear(in_features=1024, out_features=1024, bias=True)\n",
              "      (activation): Tanh()\n",
              "    )\n",
              "  )\n",
              "  (classifier): Sequential(\n",
              "    (0): Dropout(p=0.2, inplace=False)\n",
              "    (1): Linear(in_features=1024, out_features=26, bias=True)\n",
              "    (2): Tanh()\n",
              "  )\n",
              ")"
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "params = list(model.named_parameters())\n",
        "\n",
        "print('The pretrained model has {:} different named parameters.\\n'.format(len(params)))\n",
        "\n",
        "print('==== Embedding Layer ====\\n')\n",
        "\n",
        "for p in params[0:5]:\n",
        "    print(\"{:<55} {:>12}\".format(p[0], str(tuple(p[1].size()))))\n",
        "\n",
        "print('\\n==== Pre-trained Layers ====\\n')\n",
        "\n",
        "for p in params[5:21]:\n",
        "    print(\"{:<55} {:>12}\".format(p[0], str(tuple(p[1].size()))))\n",
        "\n",
        "print('\\n==== Output Layer ====\\n')\n",
        "\n",
        "for p in params[-4:]:\n",
        "    print(\"{:<55} {:>12}\".format(p[0], str(tuple(p[1].size()))))"
      ],
      "metadata": {
        "id": "TB0NYio6e88B",
        "outputId": "d39a503c-f1fc-4fd5-9f1d-02f8dadf45fe",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:24:46.580832Z",
          "iopub.execute_input": "2022-02-17T23:24:46.581298Z",
          "iopub.status.idle": "2022-02-17T23:24:46.608630Z",
          "shell.execute_reply.started": "2022-02-17T23:24:46.581255Z",
          "shell.execute_reply": "2022-02-17T23:24:46.607981Z"
        },
        "trusted": true,
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The pretrained model has 489 different named parameters.\n",
            "\n",
            "==== Embedding Layer ====\n",
            "\n",
            "model.embeddings.word_embeddings.weight                   (30, 1024)\n",
            "model.embeddings.position_embeddings.weight             (40000, 1024)\n",
            "model.embeddings.token_type_embeddings.weight              (2, 1024)\n",
            "model.embeddings.LayerNorm.weight                            (1024,)\n",
            "model.embeddings.LayerNorm.bias                              (1024,)\n",
            "\n",
            "==== Pre-trained Layers ====\n",
            "\n",
            "model.encoder.layer.0.attention.self.query.weight       (1024, 1024)\n",
            "model.encoder.layer.0.attention.self.query.bias              (1024,)\n",
            "model.encoder.layer.0.attention.self.key.weight         (1024, 1024)\n",
            "model.encoder.layer.0.attention.self.key.bias                (1024,)\n",
            "model.encoder.layer.0.attention.self.value.weight       (1024, 1024)\n",
            "model.encoder.layer.0.attention.self.value.bias              (1024,)\n",
            "model.encoder.layer.0.attention.output.dense.weight     (1024, 1024)\n",
            "model.encoder.layer.0.attention.output.dense.bias            (1024,)\n",
            "model.encoder.layer.0.attention.output.LayerNorm.weight      (1024,)\n",
            "model.encoder.layer.0.attention.output.LayerNorm.bias        (1024,)\n",
            "model.encoder.layer.0.intermediate.dense.weight         (4096, 1024)\n",
            "model.encoder.layer.0.intermediate.dense.bias                (4096,)\n",
            "model.encoder.layer.0.output.dense.weight               (1024, 4096)\n",
            "model.encoder.layer.0.output.dense.bias                      (1024,)\n",
            "model.encoder.layer.0.output.LayerNorm.weight                (1024,)\n",
            "model.encoder.layer.0.output.LayerNorm.bias                  (1024,)\n",
            "\n",
            "==== Output Layer ====\n",
            "\n",
            "model.pooler.dense.weight                               (1024, 1024)\n",
            "model.pooler.dense.bias                                      (1024,)\n",
            "classifier.1.weight                                       (26, 1024)\n",
            "classifier.1.bias                                              (26,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "\n",
        "\n",
        "---\n",
        "\n",
        "\n",
        "*   Epochs: when finetuning a LM, only a few epochs suffice to achieve great performances on a dowstream task. In this case, I will use only 1 due to computational power constraints.\n",
        "*   Learning rate and batch size are crucial hyperparameters when fine-tuning. A large learning rate would cause **catastrophic forgetting** of the weights learned during MLM objective. There are some guidelines on the [Bert Repository](https://github.com/google-research/bert).\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "---\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "8SuvDy2fzF-V"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "EPOCHS = 1\n",
        "BATCH_SIZE = 8\n",
        "LEARNING_RATE = 1e-5\n",
        "output_path = \".\"\n",
        "EVAL_BATCH = 1072\n",
        "#Set a checkpoint to evaluate the model after training on EVAL_BATCH number of batches"
      ],
      "metadata": {
        "id": "xxP2kagmcEvU",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:24:46.611937Z",
          "iopub.execute_input": "2022-02-17T23:24:46.612597Z",
          "iopub.status.idle": "2022-02-17T23:24:46.617306Z",
          "shell.execute_reply.started": "2022-02-17T23:24:46.612503Z",
          "shell.execute_reply": "2022-02-17T23:24:46.616654Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_loader  = DataLoader(dataset = train_set , batch_size = BATCH_SIZE ,drop_last=True)\n",
        "dev_loader = DataLoader(dataset = dev_set , batch_size = BATCH_SIZE  ,drop_last=True)\n",
        "test_loader = DataLoader(dataset = test_set , batch_size = BATCH_SIZE ,drop_last=True)"
      ],
      "metadata": {
        "id": "eYSGDAtWbAib",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:24:46.618750Z",
          "iopub.execute_input": "2022-02-17T23:24:46.619030Z",
          "iopub.status.idle": "2022-02-17T23:24:46.627248Z",
          "shell.execute_reply.started": "2022-02-17T23:24:46.618958Z",
          "shell.execute_reply": "2022-02-17T23:24:46.626584Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "len(train_loader)"
      ],
      "metadata": {
        "id": "aQS9pB2oa0hd",
        "execution": {
          "iopub.status.busy": "2022-02-17T22:39:32.617314Z",
          "iopub.execute_input": "2022-02-17T22:39:32.617793Z",
          "iopub.status.idle": "2022-02-17T22:39:32.623067Z",
          "shell.execute_reply.started": "2022-02-17T22:39:32.617751Z",
          "shell.execute_reply": "2022-02-17T22:39:32.622330Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "criterion = nn.CrossEntropyLoss()\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=LEARNING_RATE)"
      ],
      "metadata": {
        "id": "HbF4a2Hla6-O",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:24:46.628667Z",
          "iopub.execute_input": "2022-02-17T23:24:46.628837Z",
          "iopub.status.idle": "2022-02-17T23:24:46.641557Z",
          "shell.execute_reply.started": "2022-02-17T23:24:46.628816Z",
          "shell.execute_reply": "2022-02-17T23:24:46.640723Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from tqdm import tqdm"
      ],
      "metadata": {
        "id": "sFNzL2bKciW5",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:24:46.643379Z",
          "iopub.execute_input": "2022-02-17T23:24:46.643957Z",
          "iopub.status.idle": "2022-02-17T23:24:46.650146Z",
          "shell.execute_reply.started": "2022-02-17T23:24:46.643880Z",
          "shell.execute_reply": "2022-02-17T23:24:46.649440Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def multi_acc(y_pred, y_test):\n",
        "    '''\n",
        "    @param Tensor y_pred: tensor containig the actual labels \n",
        "    @param Tensor y_test: output from the model, it is still necessary to convert those into a probability via softmax.\n",
        "\n",
        "    @output Float acc: accuracy on the batch\n",
        "    '''\n",
        "    \n",
        "    y_pred_softmax = torch.log_softmax(y_pred, dim = 1)\n",
        "    _, y_pred_tags = torch.max(y_pred_softmax, dim = 1)    \n",
        "    correct_pred = (y_pred_tags == y_test).int()\n",
        "    acc = correct_pred.sum()*1.0 / len(correct_pred)*100\n",
        "    return acc"
      ],
      "metadata": {
        "id": "aYh7Y51JgN0T",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:24:46.651531Z",
          "iopub.execute_input": "2022-02-17T23:24:46.651793Z",
          "iopub.status.idle": "2022-02-17T23:24:46.659692Z",
          "shell.execute_reply.started": "2022-02-17T23:24:46.651759Z",
          "shell.execute_reply": "2022-02-17T23:24:46.658834Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "losses = {\n",
        "    'train' : [],\n",
        "    'val' : [] \n",
        "}\n",
        "\n",
        "accuracies = {\n",
        "    'train' : [],\n",
        "    'val' : [] \n",
        "}"
      ],
      "metadata": {
        "id": "BHc6rYOzhZC1",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:24:46.660739Z",
          "iopub.execute_input": "2022-02-17T23:24:46.661065Z",
          "iopub.status.idle": "2022-02-17T23:24:46.668657Z",
          "shell.execute_reply.started": "2022-02-17T23:24:46.661029Z",
          "shell.execute_reply": "2022-02-17T23:24:46.667841Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import shutil\n",
        "\n",
        "def save_checkpoint(path , state, filename):\n",
        "    torch.save(state, filename)\n",
        "    #shutil.copyfile(path, filename)"
      ],
      "metadata": {
        "id": "zs1QWOFdi6f_",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:24:46.670105Z",
          "iopub.execute_input": "2022-02-17T23:24:46.670376Z",
          "iopub.status.idle": "2022-02-17T23:24:46.676977Z",
          "shell.execute_reply.started": "2022-02-17T23:24:46.670342Z",
          "shell.execute_reply": "2022-02-17T23:24:46.676164Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!ln -sf /opt/bin/nvidia-smi /usr/bin/nvidia-smi\n",
        "!pip install gputil\n",
        "!pip install psutil\n",
        "!pip install humanize\n",
        "import psutil\n",
        "import humanize\n",
        "import os\n",
        "import GPUtil as GPU\n"
      ],
      "metadata": {
        "id": "6Mmrsxze1c5g",
        "execution": {
          "iopub.status.busy": "2022-02-17T22:17:09.669357Z",
          "iopub.execute_input": "2022-02-17T22:17:09.669631Z",
          "iopub.status.idle": "2022-02-17T22:17:34.900507Z",
          "shell.execute_reply.started": "2022-02-17T22:17:09.669596Z",
          "shell.execute_reply": "2022-02-17T22:17:34.899674Z"
        },
        "trusted": true
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "GPUs = GPU.getGPUs()\n",
        "gpu = GPUs[0]\n",
        "def printm():\n",
        " process = psutil.Process(os.getpid())\n",
        " print(\"Gen RAM Free: \" + humanize.naturalsize( psutil.virtual_memory().available ), \n",
        "       \" | Proc size: \" + humanize.naturalsize( process.memory_info().rss))\n",
        " print(\"GPU RAM Free: {0:.0f}MB | Used: {1:.0f}MB | Util {2:3.0f}% | Total {3:.0f}MB\"\n",
        "        .format(gpu.memoryFree, gpu.memoryUsed, gpu.memoryUtil*100, gpu.memoryTotal))\n",
        "printm()"
      ],
      "metadata": {
        "id": "3OKnWEMq4-4C",
        "outputId": "29c69438-028e-4292-8f19-9d6ed8644eaf",
        "execution": {
          "iopub.status.busy": "2022-02-17T22:17:34.902187Z",
          "iopub.execute_input": "2022-02-17T22:17:34.902448Z",
          "iopub.status.idle": "2022-02-17T22:17:35.008064Z",
          "shell.execute_reply.started": "2022-02-17T22:17:34.902413Z",
          "shell.execute_reply": "2022-02-17T22:17:35.007239Z"
        },
        "trusted": true,
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Gen RAM Free: 8.6 GB  | Proc size: 6.8 GB\n",
            "GPU RAM Free: 9317MB | Used: 2124MB | Util  19% | Total 11441MB\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Initialize losses and loss histories\n",
        "train_loss = 0.0\n",
        "val_loss = 0.0\n",
        "best_valid_loss = float('Inf')\n",
        "train_acc = 0.0\n",
        "val_acc = 0.0\n",
        "step = 0\n",
        "\n",
        "model.train()\n",
        "\n",
        "# Train loop\n",
        "for epoch in range(EPOCHS):\n",
        "    for j,train_batch in enumerate(tqdm(train_loader)):\n",
        "        train_mask = train_batch['attention_mask'].to(device)\n",
        "        train_input = train_batch['input_ids'].to(device)\n",
        "        train_targets = train_batch['target'].to(device)\n",
        "\n",
        "        train_y_pred = model(input_ids=train_input,  \n",
        "                        attention_mask=train_mask)\n",
        "        loss = criterion(train_y_pred, train_targets)\n",
        "        \n",
        "        loss.backward()\n",
        "\n",
        "        optimizer.step()    \n",
        "            \n",
        "        optimizer.zero_grad()\n",
        "        \n",
        "        # Update train loss and global step\n",
        "        train_loss += loss.item()\n",
        "        train_acc += multi_acc(train_y_pred , train_targets)\n",
        "        step += 1\n",
        "\n",
        "        # Validation loop. Save progress and evaluate model performance.\n",
        "\n",
        "        if (step % EVAL_BATCH) == 0:       \n",
        "          with torch.no_grad():     \n",
        "              model.eval()               \n",
        "              for i,val_batch in enumerate(dev_loader):\n",
        "                  val_mask = val_batch['attention_mask'].to(device)\n",
        "                  val_input = val_batch['input_ids'].to(device)\n",
        "                  val_targets = val_batch['target'].to(device)\n",
        "\n",
        "\n",
        "                  val_y_pred = model(input_ids=val_input,  \n",
        "                                attention_mask=val_mask)\n",
        "\n",
        "                  \n",
        "                  loss = criterion(val_y_pred, val_targets)\n",
        "                  \n",
        "                  val_loss += loss.item()\n",
        "                  val_acc += multi_acc(val_y_pred , val_targets)\n",
        "              # Store train and validation loss history\n",
        "              train_loss /= step\n",
        "              val_loss /=  len(dev_loader)  \n",
        "              train_acc /= step\n",
        "              val_acc /= len(dev_loader)\n",
        "              losses['train'].append(train_loss)\n",
        "              losses['val'].append(val_loss)\n",
        "              accuracies['train'].append(train_acc)\n",
        "              accuracies['val'].append(val_acc)\n",
        "\n",
        "              print('BATCH [{}/{}]  | Train Loss: {:.4f}, Valid Loss: {:.4f} , Train Acc: {:4f}, Val Acc: {:4f}'\n",
        "                    .format(j + 1, len(train_loader),\n",
        "                            train_loss, val_loss , train_acc , val_acc))\n",
        "              \n",
        "              # checkpoint\n",
        "              \n",
        "              if best_valid_loss > val_loss:\n",
        "                  best_valid_loss = val_loss\n",
        "                  save_checkpoint(\"/\" +output_path ,\n",
        "                      {\n",
        "                    'epoch': epoch + 1,\n",
        "                    'state_dict': model.state_dict(),\n",
        "                    'best_prec1': val_acc,\n",
        "                    'optimizer' : optimizer.state_dict(),\n",
        "                    }, 'finetuned_model.pt')\n",
        "\n",
        "              train_loss = 0.0                \n",
        "              val_loss = 0.0\n",
        "              step = 0\n",
        "              model.train()\n",
        "\n",
        "print('Training done!')"
      ],
      "metadata": {
        "id": "wEjj9WTRpCnI",
        "outputId": "7dbd9a07-7c8e-4371-d24a-f5ea90559e40",
        "execution": {
          "iopub.status.busy": "2022-02-17T23:24:46.678583Z",
          "iopub.execute_input": "2022-02-17T23:24:46.678836Z",
          "iopub.status.idle": "2022-02-18T00:03:12.308850Z",
          "shell.execute_reply.started": "2022-02-17T23:24:46.678804Z",
          "shell.execute_reply": "2022-02-18T00:03:12.308127Z"
        },
        "trusted": true,
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            " 25%|██▍       | 1071/4292 [29:40<1:28:37,  1.65s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "BATCH [1072/4292]  | Train Loss: 2.3809, Valid Loss: 1.8529 , Train Acc: 75.664642, Val Acc: 99.602058\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            " 50%|████▉     | 2143/4292 [1:05:00<59:56,  1.67s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "BATCH [2144/4292]  | Train Loss: 1.6851, Valid Loss: 1.5819 , Train Acc: 99.137741, Val Acc: 99.882217\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            " 75%|███████▍  | 3215/4292 [1:40:12<30:01,  1.67s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "BATCH [3216/4292]  | Train Loss: 1.5492, Valid Loss: 1.5199 , Train Acc: 99.637726, Val Acc: 99.952965\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|█████████▉| 4287/4292 [2:15:26<00:08,  1.67s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "BATCH [4288/4292]  | Train Loss: 1.5094, Valid Loss: 1.4975 , Train Acc: 99.813095, Val Acc: 100.046730\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 4292/4292 [2:20:55<00:00,  1.97s/it]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training done!\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import seaborn as sns\n",
        "import matplotlib.pyplot as plt\n",
        "sns.set_style(\"darkgrid\")\n",
        "sns.set(font_scale=1.5)\n",
        "plt.rcParams[\"figure.figsize\"] = (12,6)\n",
        "\n",
        "# Plot the learning curve.\n",
        "plt.plot(np.array([1, 2, 3, 4])*EVAL_BATCH , losses[\"train\"], '-o', label=\"Training Loss\" , color = \"lightskyblue\")\n",
        "plt.plot(np.array([1, 2, 3, 4])*EVAL_BATCH , losses[\"val\"], '-o', label=\"Validation Loss\" , color = \"indigo\")\n",
        "\n",
        "# Label the plot.\n",
        "plt.title(\"Training & Validation Loss\")\n",
        "plt.xlabel(\"Batches\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.legend()\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "IGKBe8LP2rYq",
        "execution": {
          "iopub.status.busy": "2022-02-18T00:05:34.452351Z",
          "iopub.execute_input": "2022-02-18T00:05:34.453102Z",
          "iopub.status.idle": "2022-02-18T00:05:34.715840Z",
          "shell.execute_reply.started": "2022-02-18T00:05:34.453059Z",
          "shell.execute_reply": "2022-02-18T00:05:34.715227Z"
        },
        "trusted": true,
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 427
        },
        "outputId": "e9fbd847-ac53-4fb3-cd14-5e02b1fd7ae3"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "sns.set(font_scale=1.5)\n",
        "plt.rcParams[\"figure.figsize\"] = (12,6)\n",
        "\n",
        "# Plot the learning curve.\n",
        "plt.plot(np.array([1, 2, 3, 4])*EVAL_BATCH , accuracies[\"train\"], '-o', label=\"Training Accuracy\" , color = \"lightskyblue\")\n",
        "plt.plot(np.array([1, 2, 3, 4])*EVAL_BATCH , accuracies[\"val\"], '-o', label=\"Validation Acuracy\" , color = \"indigo\")\n",
        "\n",
        "# Label the plot.\n",
        "plt.title(\"Training & Validation Accuracies\")\n",
        "plt.xlabel(\"Batches\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.legend()\n",
        "\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "Zd6mDwtf8J2l",
        "execution": {
          "iopub.status.busy": "2022-02-18T00:11:11.117218Z",
          "iopub.execute_input": "2022-02-18T00:11:11.117783Z",
          "iopub.status.idle": "2022-02-18T00:11:11.338826Z",
          "shell.execute_reply.started": "2022-02-18T00:11:11.117745Z",
          "shell.execute_reply": "2022-02-18T00:11:11.338170Z"
        },
        "trusted": true,
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 427
        },
        "outputId": "cba59d93-6a31-450d-f447-40c85c73c7c5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x432 with 1 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test_loss = 0.0\n",
        "test_acc = 0.0\n",
        "with torch.no_grad():   \n",
        "    print(' Starting Evaluation on Test Set ')  \n",
        "    model.eval()               \n",
        "    for i,test_batch in enumerate(tqdm(test_loader)):\n",
        "        test_mask = test_batch['attention_mask'].to(device)\n",
        "        test_input = test_batch['input_ids'].to(device)\n",
        "        test_targets = test_batch['target'].to(device)\n",
        "\n",
        "\n",
        "        test_y_pred = model(input_ids=test_input,  \n",
        "                      attention_mask=test_mask)\n",
        "\n",
        "        \n",
        "        loss = criterion(test_y_pred, test_targets)\n",
        "        \n",
        "        test_loss += loss.item()\n",
        "        test_acc += multi_acc(test_y_pred , test_targets)\n",
        "        torch.cuda.empty_cache()\n",
        "    # Store train and validation loss history\n",
        "    test_loss /= len(test_loader)  \n",
        "    test_acc /= len(test_loader)\n",
        "    \n",
        "print('Test Loss: {:.4f} , Test Accuracy: {:4f}'\n",
        "    .format(test_loss , test_acc))"
      ],
      "metadata": {
        "id": "Yk1zG8u16BA5",
        "execution": {
          "iopub.status.busy": "2022-02-18T00:03:29.884418Z",
          "iopub.execute_input": "2022-02-18T00:03:29.884895Z",
          "iopub.status.idle": "2022-02-18T00:04:54.671700Z",
          "shell.execute_reply.started": "2022-02-18T00:03:29.884857Z",
          "shell.execute_reply": "2022-02-18T00:04:54.671013Z"
        },
        "trusted": true,
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d3cb3cc3-a278-4125-dd71-1b4399b388dc"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            " Starting Evaluation on Test Set \n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 534/534 [05:09<00:00,  1.72it/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test Loss: 1.4969 , Test Accuracy: 99.882957\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Results Analysis\n",
        "\n",
        "\n",
        "\n",
        "*   CNN-based model is able to achieve 97% accuracy on the held-out set, thus proving that a convolutional model is able to capture the dependencies and underlying patterns in the protein sequence. Using an embedding layer and batch normalization provide the right inductive bias for learning representations of the protein sequences. Also, running the model on a GPU allows for an extremely fast learning with the subset of the data that I was using.\n",
        "*   The fine-tuned transformer model is able to achieve 99% accuracy on the test set, showcasing the power of transfer learning for downstream tasks. The main drawback would be both training time and inference time, which would require a lot more computational power. This implementation allows further interpretability of the model by visualizing attention weights in the transformer architecture.\n",
        "\n",
        "\n",
        "*   It is also remarkable how both models are robust with respect to short sequences and truncation set up to only 128 characters as well. \n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "---\n",
        "\n",
        "\n",
        "Further work:\n",
        "\n",
        "\n",
        "\n",
        "*   Test the same models trained on the whole dataset, given access to more computational power.\n",
        "*   Another interesting point would be related to few-shot learning. Siamese networks have shown to perform well without many labelled examples. More precisely, it would be interesting to leverage the same transformer architecture to obtain the learned representation and compute similarity metrics among samples to perform the classification. This type of siamese transformer architectures have been shown to perform well in NLP in the work by Shuai Gao *et al.* (2021) [TBSN: Sparse-Transformer Based Siamese Network for Few-Shot Action Recognition](https://ieeexplore.ieee.org/document/9441568).\n",
        "\n"
      ],
      "metadata": {
        "id": "5xGJ9OwcoQqT"
      }
    }
  ]
}