TensorFlow_Milestone_Project_2_SkimLit_NLP.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "toc_visible": true,
      "gpuType": "T4"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# Milestone Project 2: SkimLit 📄🔥\n",
        "\n",
        "The purpose of this notebook is to build an NLP model to make reading medical abstracts easier.\n",
        "\n",
        "The paper we're replicating (the source of the dataset that we'll be using) is available here : https://arxiv.org/abs/1710.06071\n",
        "\n",
        "And reading through the paper above, we see that the model architecture that they used to achieve their best results is available here : https://arxiv.org/abs/1612.05251"
      ],
      "metadata": {
        "id": "iVjYcTosRsot"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Install tensorflow version 2.15\n",
        "!pip install tensorflow==2.15"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "APKdwK4TSjzF",
        "outputId": "f10270ef-9c1f-4906-fbb3-4e3cd92593c2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting tensorflow==2.15\n",
            "  Downloading tensorflow-2.15.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (4.4 kB)\n",
            "Requirement already satisfied: absl-py>=1.0.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.15) (1.4.0)\n",
            "Requirement already satisfied: astunparse>=1.6.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.15) (1.6.3)\n",
            "Requirement already satisfied: flatbuffers>=23.5.26 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.15) (24.3.25)\n",
            "Requirement already satisfied: gast!=0.5.0,!=0.5.1,!=0.5.2,>=0.2.1 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.15) (0.6.0)\n",
            "Requirement already satisfied: google-pasta>=0.1.1 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.15) (0.2.0)\n",
            "Requirement already satisfied: h5py>=2.9.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.15) (3.11.0)\n",
            "Requirement already satisfied: libclang>=13.0.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.15) (18.1.1)\n",
            "Collecting ml-dtypes~=0.2.0 (from tensorflow==2.15)\n",
            "  Downloading ml_dtypes-0.2.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (20 kB)\n",
            "Requirement already satisfied: numpy<2.0.0,>=1.23.5 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.15) (1.26.4)\n",
            "Requirement already satisfied: opt-einsum>=2.3.2 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.15) (3.3.0)\n",
            "Requirement already satisfied: packaging in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.15) (24.1)\n",
            "Requirement already satisfied: protobuf!=4.21.0,!=4.21.1,!=4.21.2,!=4.21.3,!=4.21.4,!=4.21.5,<5.0.0dev,>=3.20.3 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.15) (3.20.3)\n",
            "Requirement already satisfied: setuptools in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.15) (71.0.4)\n",
            "Requirement already satisfied: six>=1.12.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.15) (1.16.0)\n",
            "Requirement already satisfied: termcolor>=1.1.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.15) (2.4.0)\n",
            "Requirement already satisfied: typing-extensions>=3.6.6 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.15) (4.12.2)\n",
            "Collecting wrapt<1.15,>=1.11.0 (from tensorflow==2.15)\n",
            "  Downloading wrapt-1.14.1-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (6.7 kB)\n",
            "Requirement already satisfied: tensorflow-io-gcs-filesystem>=0.23.1 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.15) (0.37.1)\n",
            "Requirement already satisfied: grpcio<2.0,>=1.24.3 in /usr/local/lib/python3.10/dist-packages (from tensorflow==2.15) (1.64.1)\n",
            "Collecting tensorboard<2.16,>=2.15 (from tensorflow==2.15)\n",
            "  Downloading tensorboard-2.15.2-py3-none-any.whl.metadata (1.7 kB)\n",
            "Collecting tensorflow-estimator<2.16,>=2.15.0 (from tensorflow==2.15)\n",
            "  Downloading tensorflow_estimator-2.15.0-py2.py3-none-any.whl.metadata (1.3 kB)\n",
            "Collecting keras<2.16,>=2.15.0 (from tensorflow==2.15)\n",
            "  Downloading keras-2.15.0-py3-none-any.whl.metadata (2.4 kB)\n",
            "Requirement already satisfied: wheel<1.0,>=0.23.0 in /usr/local/lib/python3.10/dist-packages (from astunparse>=1.6.0->tensorflow==2.15) (0.44.0)\n",
            "Requirement already satisfied: google-auth<3,>=1.6.3 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.16,>=2.15->tensorflow==2.15) (2.27.0)\n",
            "Requirement already satisfied: google-auth-oauthlib<2,>=0.5 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.16,>=2.15->tensorflow==2.15) (1.2.1)\n",
            "Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.16,>=2.15->tensorflow==2.15) (3.7)\n",
            "Requirement already satisfied: requests<3,>=2.21.0 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.16,>=2.15->tensorflow==2.15) (2.32.3)\n",
            "Requirement already satisfied: tensorboard-data-server<0.8.0,>=0.7.0 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.16,>=2.15->tensorflow==2.15) (0.7.2)\n",
            "Requirement already satisfied: werkzeug>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.16,>=2.15->tensorflow==2.15) (3.0.4)\n",
            "Requirement already satisfied: cachetools<6.0,>=2.0.0 in /usr/local/lib/python3.10/dist-packages (from google-auth<3,>=1.6.3->tensorboard<2.16,>=2.15->tensorflow==2.15) (5.5.0)\n",
            "Requirement already satisfied: pyasn1-modules>=0.2.1 in /usr/local/lib/python3.10/dist-packages (from google-auth<3,>=1.6.3->tensorboard<2.16,>=2.15->tensorflow==2.15) (0.4.0)\n",
            "Requirement already satisfied: rsa<5,>=3.1.4 in /usr/local/lib/python3.10/dist-packages (from google-auth<3,>=1.6.3->tensorboard<2.16,>=2.15->tensorflow==2.15) (4.9)\n",
            "Requirement already satisfied: requests-oauthlib>=0.7.0 in /usr/local/lib/python3.10/dist-packages (from google-auth-oauthlib<2,>=0.5->tensorboard<2.16,>=2.15->tensorflow==2.15) (1.3.1)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.21.0->tensorboard<2.16,>=2.15->tensorflow==2.15) (3.3.2)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.21.0->tensorboard<2.16,>=2.15->tensorflow==2.15) (3.8)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.21.0->tensorboard<2.16,>=2.15->tensorflow==2.15) (2.0.7)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.21.0->tensorboard<2.16,>=2.15->tensorflow==2.15) (2024.7.4)\n",
            "Requirement already satisfied: MarkupSafe>=2.1.1 in /usr/local/lib/python3.10/dist-packages (from werkzeug>=1.0.1->tensorboard<2.16,>=2.15->tensorflow==2.15) (2.1.5)\n",
            "Requirement already satisfied: pyasn1<0.7.0,>=0.4.6 in /usr/local/lib/python3.10/dist-packages (from pyasn1-modules>=0.2.1->google-auth<3,>=1.6.3->tensorboard<2.16,>=2.15->tensorflow==2.15) (0.6.0)\n",
            "Requirement already satisfied: oauthlib>=3.0.0 in /usr/local/lib/python3.10/dist-packages (from requests-oauthlib>=0.7.0->google-auth-oauthlib<2,>=0.5->tensorboard<2.16,>=2.15->tensorflow==2.15) (3.2.2)\n",
            "Downloading tensorflow-2.15.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (475.2 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m475.2/475.2 MB\u001b[0m \u001b[31m2.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading keras-2.15.0-py3-none-any.whl (1.7 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.7/1.7 MB\u001b[0m \u001b[31m24.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading ml_dtypes-0.2.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.0 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.0/1.0 MB\u001b[0m \u001b[31m42.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading tensorboard-2.15.2-py3-none-any.whl (5.5 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m5.5/5.5 MB\u001b[0m \u001b[31m23.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading tensorflow_estimator-2.15.0-py2.py3-none-any.whl (441 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m442.0/442.0 kB\u001b[0m \u001b[31m18.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading wrapt-1.14.1-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl (77 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m77.9/77.9 kB\u001b[0m \u001b[31m7.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hInstalling collected packages: wrapt, tensorflow-estimator, ml-dtypes, keras, tensorboard, tensorflow\n",
            "  Attempting uninstall: wrapt\n",
            "    Found existing installation: wrapt 1.16.0\n",
            "    Uninstalling wrapt-1.16.0:\n",
            "      Successfully uninstalled wrapt-1.16.0\n",
            "  Attempting uninstall: ml-dtypes\n",
            "    Found existing installation: ml-dtypes 0.4.0\n",
            "    Uninstalling ml-dtypes-0.4.0:\n",
            "      Successfully uninstalled ml-dtypes-0.4.0\n",
            "  Attempting uninstall: keras\n",
            "    Found existing installation: keras 3.4.1\n",
            "    Uninstalling keras-3.4.1:\n",
            "      Successfully uninstalled keras-3.4.1\n",
            "  Attempting uninstall: tensorboard\n",
            "    Found existing installation: tensorboard 2.17.0\n",
            "    Uninstalling tensorboard-2.17.0:\n",
            "      Successfully uninstalled tensorboard-2.17.0\n",
            "  Attempting uninstall: tensorflow\n",
            "    Found existing installation: tensorflow 2.17.0\n",
            "    Uninstalling tensorflow-2.17.0:\n",
            "      Successfully uninstalled tensorflow-2.17.0\n",
            "\u001b[31mERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.\n",
            "tensorstore 0.1.64 requires ml-dtypes>=0.3.1, but you have ml-dtypes 0.2.0 which is incompatible.\n",
            "tf-keras 2.17.0 requires tensorflow<2.18,>=2.17, but you have tensorflow 2.15.0 which is incompatible.\u001b[0m\u001b[31m\n",
            "\u001b[0mSuccessfully installed keras-2.15.0 ml-dtypes-0.2.0 tensorboard-2.15.2 tensorflow-2.15.0 tensorflow-estimator-2.15.0 wrapt-1.14.1\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "## Confirm access to a GPU\n",
        "!nvidia-smi -L"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Zfy4Pm4GVs6Z",
        "outputId": "d08c7b90-4e9e-4359-e03c-b451ca1ee3dc"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "GPU 0: Tesla T4 (UUID: GPU-aee4f111-630d-78e0-e5b1-7746e3085dd2)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Get the data\n",
        "\n",
        "Since we'll be replicating the paper above (PubMed 200k RCT), let's download the dataset they used.\n",
        "\n",
        "We can do so from the author's github page here : https://github.com/Franck-Dernoncourt/pubmed-rct"
      ],
      "metadata": {
        "id": "T1UIRAG1V6Xj"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!git clone https://github.com/Franck-Dernoncourt/pubmed-rct.git\n",
        "!ls pubmed-rct"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "F1PT5lz-Wkah",
        "outputId": "333107b2-9709-4c98-8c8c-6f8afd300ff8"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Cloning into 'pubmed-rct'...\n",
            "remote: Enumerating objects: 39, done.\u001b[K\n",
            "remote: Counting objects: 100% (14/14), done.\u001b[K\n",
            "remote: Compressing objects: 100% (9/9), done.\u001b[K\n",
            "remote: Total 39 (delta 8), reused 5 (delta 5), pack-reused 25 (from 1)\u001b[K\n",
            "Receiving objects: 100% (39/39), 177.08 MiB | 12.40 MiB/s, done.\n",
            "Resolving deltas: 100% (15/15), done.\n",
            "Updating files: 100% (13/13), done.\n",
            "PubMed_200k_RCT\t\t\t\t       PubMed_20k_RCT_numbers_replaced_with_at_sign\n",
            "PubMed_200k_RCT_numbers_replaced_with_at_sign  README.md\n",
            "PubMed_20k_RCT\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Check what files are there in the PubMed-20K dataset\n",
        "!ls pubmed-rct/PubMed_20k_RCT_numbers_replaced_with_at_sign/"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pNgXA6htWq_5",
        "outputId": "c31d8196-f122-4299-dad3-73abf8961d15"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "dev.txt  test.txt  train.txt\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Start our experiments using the 20K datasets with numbers replaced with \"@\" sign\n",
        "data_dir = \"/content/pubmed-rct/PubMed_20k_RCT_numbers_replaced_with_at_sign\""
      ],
      "metadata": {
        "id": "arOwj15xZ9qA"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Check all the filenames in the target directory\n",
        "import os\n",
        "filenames = [data_dir + filename for filename in os.listdir(data_dir)]\n",
        "filenames"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KOKup5b5bfOB",
        "outputId": "bd5778af-540a-4124-cf2a-19a387391e23"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['/content/pubmed-rct/PubMed_20k_RCT_numbers_replaced_with_at_signtest.txt',\n",
              " '/content/pubmed-rct/PubMed_20k_RCT_numbers_replaced_with_at_signdev.txt',\n",
              " '/content/pubmed-rct/PubMed_20k_RCT_numbers_replaced_with_at_signtrain.txt']"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Preprocess Data\n",
        "\n",
        "Now we've got some text data, it's time to become one with it.\n",
        "\n",
        "And one of the best ways to become one with the data is to visualize it.\n",
        "\n",
        "Keeping that in mind, let's write a function to read in all of the lines of a target text file."
      ],
      "metadata": {
        "id": "s9br72kcbuF3"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Create a function to read the lines of a document\n",
        "def get_lines(filename):\n",
        "  \"\"\"\n",
        "  Reads filename (a text filename) and returns the lines of text as a list.\n",
        "\n",
        "  Args:\n",
        "    filename: a string containing the target filepath.\n",
        "\n",
        "  Returns:\n",
        "    list: A list of strings with one string per line from the target filename.\n",
        "  \"\"\"\n",
        "  with open(filename, \"r\") as f:\n",
        "    return f.readlines()"
      ],
      "metadata": {
        "id": "yDmuwhYKdBEH"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Let's read in the training lines\n",
        "train_lines = get_lines(data_dir + \"/train.txt\") # read the lines within the training file\n",
        "train_lines[:20] # show the first 20 lines"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Hjr-kthvfRE2",
        "outputId": "453cb4a3-53d0-46ed-a6be-3ac4968e09e9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['###24293578\\n',\n",
              " 'OBJECTIVE\\tTo investigate the efficacy of @ weeks of daily low-dose oral prednisolone in improving pain , mobility , and systemic low-grade inflammation in the short term and whether the effect would be sustained at @ weeks in older adults with moderate to severe knee osteoarthritis ( OA ) .\\n',\n",
              " 'METHODS\\tA total of @ patients with primary knee OA were randomized @:@ ; @ received @ mg/day of prednisolone and @ received placebo for @ weeks .\\n',\n",
              " 'METHODS\\tOutcome measures included pain reduction and improvement in function scores and systemic inflammation markers .\\n',\n",
              " 'METHODS\\tPain was assessed using the visual analog pain scale ( @-@ mm ) .\\n',\n",
              " 'METHODS\\tSecondary outcome measures included the Western Ontario and McMaster Universities Osteoarthritis Index scores , patient global assessment ( PGA ) of the severity of knee OA , and @-min walk distance ( @MWD ) .\\n',\n",
              " 'METHODS\\tSerum levels of interleukin @ ( IL-@ ) , IL-@ , tumor necrosis factor ( TNF ) - , and high-sensitivity C-reactive protein ( hsCRP ) were measured .\\n',\n",
              " 'RESULTS\\tThere was a clinically relevant reduction in the intervention group compared to the placebo group for knee pain , physical function , PGA , and @MWD at @ weeks .\\n',\n",
              " 'RESULTS\\tThe mean difference between treatment arms ( @ % CI ) was @ ( @-@ @ ) , p < @ ; @ ( @-@ @ ) , p < @ ; @ ( @-@ @ ) , p < @ ; and @ ( @-@ @ ) , p < @ , respectively .\\n',\n",
              " 'RESULTS\\tFurther , there was a clinically relevant reduction in the serum levels of IL-@ , IL-@ , TNF - , and hsCRP at @ weeks in the intervention group when compared to the placebo group .\\n',\n",
              " 'RESULTS\\tThese differences remained significant at @ weeks .\\n',\n",
              " 'RESULTS\\tThe Outcome Measures in Rheumatology Clinical Trials-Osteoarthritis Research Society International responder rate was @ % in the intervention group and @ % in the placebo group ( p < @ ) .\\n',\n",
              " 'CONCLUSIONS\\tLow-dose oral prednisolone had both a short-term and a longer sustained effect resulting in less knee pain , better physical function , and attenuation of systemic inflammation in older patients with knee OA ( ClinicalTrials.gov identifier NCT@ ) .\\n',\n",
              " '\\n',\n",
              " '###24854809\\n',\n",
              " 'BACKGROUND\\tEmotional eating is associated with overeating and the development of obesity .\\n',\n",
              " 'BACKGROUND\\tYet , empirical evidence for individual ( trait ) differences in emotional eating and cognitive mechanisms that contribute to eating during sad mood remain equivocal .\\n',\n",
              " 'OBJECTIVE\\tThe aim of this study was to test if attention bias for food moderates the effect of self-reported emotional eating during sad mood ( vs neutral mood ) on actual food intake .\\n',\n",
              " 'OBJECTIVE\\tIt was expected that emotional eating is predictive of elevated attention for food and higher food intake after an experimentally induced sad mood and that attentional maintenance on food predicts food intake during a sad versus a neutral mood .\\n',\n",
              " 'METHODS\\tParticipants ( N = @ ) were randomly assigned to one of the two experimental mood induction conditions ( sad/neutral ) .\\n']"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "len(train_lines)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "77dVPJEJfyY_",
        "outputId": "c76e7fe9-40bf-409f-a9c8-32a4ca69a36d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "210040"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Let's think about how we want our data to look...\n",
        "\n",
        "How I think our data would be best represented...\n",
        "\n",
        "```\n",
        " [{'line_number':0,\n",
        "   'target':'BACKGROUND',\n",
        "   'text': \"Emotional eating is associated with overeating and the development of obesity.\\n\",\n",
        "   'total_lines': 11},\n",
        "  ...]\n",
        "```\n",
        "Let's write a function which turns each of our datasets into the above format so we can continue to prepare our data for modelling."
      ],
      "metadata": {
        "id": "8GJcD7BeimNd"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def preprocess_text_with_line_numbers(filename):\n",
        "  \"\"\"\n",
        "  Returns a list of dictionaries of abstract line data.\n",
        "\n",
        "  Takes in filename, reads its contents and sorts through each line,\n",
        "  extracting things like the target label, the text of the sentence,\n",
        "  how many sentences are in the current abstract and what sentence number\n",
        "  the target line is.\n",
        "  \"\"\"\n",
        "  input_lines = get_lines(filename)\n",
        "  abstract_lines = \"\" # create an empty abstract\n",
        "  abstract_samples = [] # create an empty list of abstracts\n",
        "\n",
        "  # Loop through each line in the target line\n",
        "  for line in input_lines:\n",
        "    if line.startswith(\"###\"): # check to see if the line is an ID line\n",
        "      abstract_id = line\n",
        "      abstract_lines = \"\" # reset the abstract string if the line is an ID line\n",
        "    elif line.isspace(): # check to see if the line is a new line\n",
        "      abstract_line_split = abstract_lines.split(\"\\n\") # split the abstract into seperate lines\n",
        "\n",
        "\n",
        "      # Iterate through each line in a single abstract and count them at the same time\n",
        "      for abstract_line_number, abstract_line in enumerate(abstract_line_split):\n",
        "        line_data = {} # create an empty dictionary for each line\n",
        "        target_text_split = abstract_line.split(\"\\t\") # split the target from the text\n",
        "        # Handle cases where the line does not contain a tab\n",
        "        if len(target_text_split) > 1:\n",
        "            line_data[\"target\"] = target_text_split[0] # get the target label\n",
        "            line_data[\"text\"] = target_text_split[1].lower() # get the target text and lower it\n",
        "            line_data[\"line_number\"] = abstract_line_number # get the line number\n",
        "            line_data[\"total_lines\"] = len(abstract_line_split) - 1 # get the total number of lines in the target abstract\n",
        "            abstract_samples.append(line_data) # add line data to the empty list of abstracts\n",
        "        # else:\n",
        "        #     print(f\"Skipping line without tab delimiter: {abstract_line}\") # Optional: Print a message for skipped lines\n",
        "\n",
        "    else: # if the line is not a new line or an ID line, the line contains a labelled sentence\n",
        "      abstract_lines += line # add the line to the abstract string\n",
        "\n",
        "  return abstract_samples"
      ],
      "metadata": {
        "id": "AA9KUouGRRSt"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Get data from file and preprocess it\n",
        "%%time\n",
        "train_samples = preprocess_text_with_line_numbers(data_dir + \"/train.txt\")\n",
        "val_samples = preprocess_text_with_line_numbers(data_dir + \"/dev.txt\")\n",
        "test_samples = preprocess_text_with_line_numbers(data_dir + \"/test.txt\")\n",
        "len(train_samples), len(val_samples), len(test_samples)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wLZ6dAudQH7Z",
        "outputId": "c46fa6ee-5683-421f-e147-5031c68ea2e3"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "CPU times: user 737 ms, sys: 125 ms, total: 862 ms\n",
            "Wall time: 1.28 s\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(180040, 30212, 30135)"
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Check the first abstract of our training data\n",
        "train_samples[:13]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ceEaIM0wQanx",
        "outputId": "1aac737d-0105-4fb8-cfa4-c280be56f785"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[{'target': 'OBJECTIVE',\n",
              "  'text': 'to investigate the efficacy of @ weeks of daily low-dose oral prednisolone in improving pain , mobility , and systemic low-grade inflammation in the short term and whether the effect would be sustained at @ weeks in older adults with moderate to severe knee osteoarthritis ( oa ) .',\n",
              "  'line_number': 0,\n",
              "  'total_lines': 12},\n",
              " {'target': 'METHODS',\n",
              "  'text': 'a total of @ patients with primary knee oa were randomized @:@ ; @ received @ mg/day of prednisolone and @ received placebo for @ weeks .',\n",
              "  'line_number': 1,\n",
              "  'total_lines': 12},\n",
              " {'target': 'METHODS',\n",
              "  'text': 'outcome measures included pain reduction and improvement in function scores and systemic inflammation markers .',\n",
              "  'line_number': 2,\n",
              "  'total_lines': 12},\n",
              " {'target': 'METHODS',\n",
              "  'text': 'pain was assessed using the visual analog pain scale ( @-@ mm ) .',\n",
              "  'line_number': 3,\n",
              "  'total_lines': 12},\n",
              " {'target': 'METHODS',\n",
              "  'text': 'secondary outcome measures included the western ontario and mcmaster universities osteoarthritis index scores , patient global assessment ( pga ) of the severity of knee oa , and @-min walk distance ( @mwd ) .',\n",
              "  'line_number': 4,\n",
              "  'total_lines': 12},\n",
              " {'target': 'METHODS',\n",
              "  'text': 'serum levels of interleukin @ ( il-@ ) , il-@ , tumor necrosis factor ( tnf ) - , and high-sensitivity c-reactive protein ( hscrp ) were measured .',\n",
              "  'line_number': 5,\n",
              "  'total_lines': 12},\n",
              " {'target': 'RESULTS',\n",
              "  'text': 'there was a clinically relevant reduction in the intervention group compared to the placebo group for knee pain , physical function , pga , and @mwd at @ weeks .',\n",
              "  'line_number': 6,\n",
              "  'total_lines': 12},\n",
              " {'target': 'RESULTS',\n",
              "  'text': 'the mean difference between treatment arms ( @ % ci ) was @ ( @-@ @ ) , p < @ ; @ ( @-@ @ ) , p < @ ; @ ( @-@ @ ) , p < @ ; and @ ( @-@ @ ) , p < @ , respectively .',\n",
              "  'line_number': 7,\n",
              "  'total_lines': 12},\n",
              " {'target': 'RESULTS',\n",
              "  'text': 'further , there was a clinically relevant reduction in the serum levels of il-@ , il-@ , tnf - , and hscrp at @ weeks in the intervention group when compared to the placebo group .',\n",
              "  'line_number': 8,\n",
              "  'total_lines': 12},\n",
              " {'target': 'RESULTS',\n",
              "  'text': 'these differences remained significant at @ weeks .',\n",
              "  'line_number': 9,\n",
              "  'total_lines': 12},\n",
              " {'target': 'RESULTS',\n",
              "  'text': 'the outcome measures in rheumatology clinical trials-osteoarthritis research society international responder rate was @ % in the intervention group and @ % in the placebo group ( p < @ ) .',\n",
              "  'line_number': 10,\n",
              "  'total_lines': 12},\n",
              " {'target': 'CONCLUSIONS',\n",
              "  'text': 'low-dose oral prednisolone had both a short-term and a longer sustained effect resulting in less knee pain , better physical function , and attenuation of systemic inflammation in older patients with knee oa ( clinicaltrials.gov identifier nct@ ) .',\n",
              "  'line_number': 11,\n",
              "  'total_lines': 12},\n",
              " {'target': 'BACKGROUND',\n",
              "  'text': 'emotional eating is associated with overeating and the development of obesity .',\n",
              "  'line_number': 0,\n",
              "  'total_lines': 11}]"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now that our data is in the format of a list of dictionaries, how about we turn it into a DataFrame to further visualize it ?"
      ],
      "metadata": {
        "id": "94ZYV_TlTlBV"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "train_df = pd.DataFrame(train_samples)\n",
        "val_df = pd.DataFrame(val_samples)\n",
        "test_df = pd.DataFrame(test_samples)\n",
        "train_df.head(14)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 488
        },
        "id": "Uxfi1iR8Rx7g",
        "outputId": "866ed4a5-3f6a-4d6a-c82c-ebabd602864d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "         target                                               text  \\\n",
              "0     OBJECTIVE  to investigate the efficacy of @ weeks of dail...   \n",
              "1       METHODS  a total of @ patients with primary knee oa wer...   \n",
              "2       METHODS  outcome measures included pain reduction and i...   \n",
              "3       METHODS  pain was assessed using the visual analog pain...   \n",
              "4       METHODS  secondary outcome measures included the wester...   \n",
              "5       METHODS  serum levels of interleukin @ ( il-@ ) , il-@ ...   \n",
              "6       RESULTS  there was a clinically relevant reduction in t...   \n",
              "7       RESULTS  the mean difference between treatment arms ( @...   \n",
              "8       RESULTS  further , there was a clinically relevant redu...   \n",
              "9       RESULTS  these differences remained significant at @ we...   \n",
              "10      RESULTS  the outcome measures in rheumatology clinical ...   \n",
              "11  CONCLUSIONS  low-dose oral prednisolone had both a short-te...   \n",
              "12   BACKGROUND  emotional eating is associated with overeating...   \n",
              "13   BACKGROUND  yet , empirical evidence for individual ( trai...   \n",
              "\n",
              "    line_number  total_lines  \n",
              "0             0           12  \n",
              "1             1           12  \n",
              "2             2           12  \n",
              "3             3           12  \n",
              "4             4           12  \n",
              "5             5           12  \n",
              "6             6           12  \n",
              "7             7           12  \n",
              "8             8           12  \n",
              "9             9           12  \n",
              "10           10           12  \n",
              "11           11           12  \n",
              "12            0           11  \n",
              "13            1           11  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-ddf7a329-6204-4d69-81b3-3eb099736ccf\" 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>target</th>\n",
              "      <th>text</th>\n",
              "      <th>line_number</th>\n",
              "      <th>total_lines</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>OBJECTIVE</td>\n",
              "      <td>to investigate the efficacy of @ weeks of dail...</td>\n",
              "      <td>0</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>METHODS</td>\n",
              "      <td>a total of @ patients with primary knee oa wer...</td>\n",
              "      <td>1</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>METHODS</td>\n",
              "      <td>outcome measures included pain reduction and i...</td>\n",
              "      <td>2</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>METHODS</td>\n",
              "      <td>pain was assessed using the visual analog pain...</td>\n",
              "      <td>3</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>METHODS</td>\n",
              "      <td>secondary outcome measures included the wester...</td>\n",
              "      <td>4</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>METHODS</td>\n",
              "      <td>serum levels of interleukin @ ( il-@ ) , il-@ ...</td>\n",
              "      <td>5</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>there was a clinically relevant reduction in t...</td>\n",
              "      <td>6</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>the mean difference between treatment arms ( @...</td>\n",
              "      <td>7</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>further , there was a clinically relevant redu...</td>\n",
              "      <td>8</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>these differences remained significant at @ we...</td>\n",
              "      <td>9</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>the outcome measures in rheumatology clinical ...</td>\n",
              "      <td>10</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>CONCLUSIONS</td>\n",
              "      <td>low-dose oral prednisolone had both a short-te...</td>\n",
              "      <td>11</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>BACKGROUND</td>\n",
              "      <td>emotional eating is associated with overeating...</td>\n",
              "      <td>0</td>\n",
              "      <td>11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>BACKGROUND</td>\n",
              "      <td>yet , empirical evidence for individual ( trai...</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-ddf7a329-6204-4d69-81b3-3eb099736ccf')\"\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 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\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",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\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-ddf7a329-6204-4d69-81b3-3eb099736ccf 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-ddf7a329-6204-4d69-81b3-3eb099736ccf');\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",
              "\n",
              "\n",
              "<div id=\"df-6d3f82c1-1d95-4799-a690-b09f283d5ffc\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-6d3f82c1-1d95-4799-a690-b09f283d5ffc')\"\n",
              "            title=\"Suggest charts\"\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",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-6d3f82c1-1d95-4799-a690-b09f283d5ffc button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "train_df"
            }
          },
          "metadata": {},
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Distribution of labels\n",
        "train_df.target.value_counts()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 272
        },
        "id": "rNU8bRdfUC9u",
        "outputId": "3b94498f-9117-4ad5-bb9b-7008f357b8b6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "target\n",
              "METHODS        59353\n",
              "RESULTS        57953\n",
              "CONCLUSIONS    27168\n",
              "BACKGROUND     21727\n",
              "OBJECTIVE      13839\n",
              "Name: count, dtype: int64"
            ],
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>target</th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>METHODS</th>\n",
              "      <td>59353</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>RESULTS</th>\n",
              "      <td>57953</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>CONCLUSIONS</th>\n",
              "      <td>27168</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>BACKGROUND</th>\n",
              "      <td>21727</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>OBJECTIVE</th>\n",
              "      <td>13839</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div><br><label><b>dtype:</b> int64</label>"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Let's check the length of different lines\n",
        "train_df.total_lines.plot.hist();"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 431
        },
        "id": "GjcsXkKFUYYf",
        "outputId": "3c9e9cd2-60e8-49e0-88d9-0f2dfe592f46"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Get lists of sentences"
      ],
      "metadata": {
        "id": "cpO84ebBUuJ8"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Convert abstract text lines to lists\n",
        "train_sentences = train_df[\"text\"].to_list()\n",
        "val_sentences = val_df[\"text\"].to_list()\n",
        "test_sentences = test_df[\"text\"].to_list()\n",
        "len(train_sentences), len(val_sentences), len(test_sentences)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "68Z1eeaeU5Oz",
        "outputId": "3120843b-c95b-4c3a-813a-3ee02c49d9e2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(180040, 30212, 30135)"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# View the first 10 lines of the training sentences\n",
        "train_sentences[:10]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8eM6EwtrVOdI",
        "outputId": "2332e68a-5a72-4b4b-c6be-10dbf1ae74a2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['to investigate the efficacy of @ weeks of daily low-dose oral prednisolone in improving pain , mobility , and systemic low-grade inflammation in the short term and whether the effect would be sustained at @ weeks in older adults with moderate to severe knee osteoarthritis ( oa ) .',\n",
              " 'a total of @ patients with primary knee oa were randomized @:@ ; @ received @ mg/day of prednisolone and @ received placebo for @ weeks .',\n",
              " 'outcome measures included pain reduction and improvement in function scores and systemic inflammation markers .',\n",
              " 'pain was assessed using the visual analog pain scale ( @-@ mm ) .',\n",
              " 'secondary outcome measures included the western ontario and mcmaster universities osteoarthritis index scores , patient global assessment ( pga ) of the severity of knee oa , and @-min walk distance ( @mwd ) .',\n",
              " 'serum levels of interleukin @ ( il-@ ) , il-@ , tumor necrosis factor ( tnf ) - , and high-sensitivity c-reactive protein ( hscrp ) were measured .',\n",
              " 'there was a clinically relevant reduction in the intervention group compared to the placebo group for knee pain , physical function , pga , and @mwd at @ weeks .',\n",
              " 'the mean difference between treatment arms ( @ % ci ) was @ ( @-@ @ ) , p < @ ; @ ( @-@ @ ) , p < @ ; @ ( @-@ @ ) , p < @ ; and @ ( @-@ @ ) , p < @ , respectively .',\n",
              " 'further , there was a clinically relevant reduction in the serum levels of il-@ , il-@ , tnf - , and hscrp at @ weeks in the intervention group when compared to the placebo group .',\n",
              " 'these differences remained significant at @ weeks .']"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Make numeric labels (ML models require numeric data, they also require numeric labels)"
      ],
      "metadata": {
        "id": "7z5HSO4ZVXuu"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# One hot encode labels\n",
        "from sklearn.preprocessing import OneHotEncoder\n",
        "one_hot_encoder = OneHotEncoder(sparse_output=False)\n",
        "train_labels_one_hot = one_hot_encoder.fit_transform(train_df[\"target\"].to_numpy().reshape(-1, 1))\n",
        "val_labels_one_hot = one_hot_encoder.transform(val_df[\"target\"].to_numpy().reshape(-1, 1))\n",
        "test_labels_one_hot = one_hot_encoder.transform(test_df[\"target\"].to_numpy().reshape(-1, 1))\n",
        "\n",
        "# Check what one hot encoded labels look like\n",
        "train_labels_one_hot"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HHoSXHMkWMd_",
        "outputId": "60300de6-48b4-4da9-818d-a16655399afe"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0., 0., 0., 1., 0.],\n",
              "       [0., 0., 1., 0., 0.],\n",
              "       [0., 0., 1., 0., 0.],\n",
              "       ...,\n",
              "       [0., 0., 0., 0., 1.],\n",
              "       [0., 1., 0., 0., 0.],\n",
              "       [0., 1., 0., 0., 0.]])"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Label encode labels"
      ],
      "metadata": {
        "id": "UxmkI7yeXSVj"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Extract labels (\"target\" columns) and encode them into integers\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "label_encoder = LabelEncoder()\n",
        "train_labels_encoded = label_encoder.fit_transform(train_df[\"target\"].to_numpy())\n",
        "val_labels_encoded = label_encoder.transform(val_df[\"target\"].to_numpy())\n",
        "test_labels_encoded = label_encoder.transform(test_df[\"target\"].to_numpy())\n",
        "\n",
        "# Check what training labels look like\n",
        "train_labels_encoded"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Vi7usFStW_mu",
        "outputId": "6989d43e-3763-4bbf-eba6-0ef797711143"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([3, 2, 2, ..., 4, 1, 1])"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Get class names and number of classes from label encoder instance\n",
        "num_classes = len(label_encoder.classes_)\n",
        "class_names = label_encoder.classes_\n",
        "num_classes, class_names"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zTPCZBDiYMb-",
        "outputId": "b36cb54d-6352-467a-c868-14a52c3f54c8"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(5,\n",
              " array(['BACKGROUND', 'CONCLUSIONS', 'METHODS', 'OBJECTIVE', 'RESULTS'],\n",
              "       dtype=object))"
            ]
          },
          "metadata": {},
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Starting a series of modelling experiments...\n",
        "\n",
        "As usual, we're going to be trying out a bunch of different models and seeing which one works best.\n",
        "\n",
        "And as always, we're going to start with a baseline(TF-IDF Multinomial Naive Bayes classifier)."
      ],
      "metadata": {
        "id": "pe26Y07WZBTi"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model 0: Getting a baseline"
      ],
      "metadata": {
        "id": "RijdarchYawl"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.feature_extraction.text import TfidfVectorizer\n",
        "from sklearn.naive_bayes import MultinomialNB\n",
        "from sklearn.pipeline import Pipeline\n",
        "\n",
        "# Create a pipeline\n",
        "model_0 = Pipeline([\n",
        "    (\"tf-idf\",TfidfVectorizer()),\n",
        "    (\"clf\",MultinomialNB())\n",
        "])\n",
        "\n",
        "# Fit the pipeline to the training data\n",
        "model_0.fit(X=train_sentences,\n",
        "            y=train_labels_encoded)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 125
        },
        "id": "Y9qplshsZlSk",
        "outputId": "5a268b79-c529-4468-f6db-963011cffa04"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(steps=[('tf-idf', TfidfVectorizer()), ('clf', MultinomialNB())])"
            ],
            "text/html": [
              "<style>#sk-container-id-1 {color: black;}#sk-container-id-1 pre{padding: 0;}#sk-container-id-1 div.sk-toggleable {background-color: white;}#sk-container-id-1 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-1 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-1 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-1 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-1 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-1 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-1 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-1 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-1 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-1 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-1 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-1 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-1 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-1 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-1 div.sk-item {position: relative;z-index: 1;}#sk-container-id-1 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-1 div.sk-item::before, #sk-container-id-1 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-1 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-1 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-1 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-1 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-1 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-1 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-1 div.sk-label-container {text-align: center;}#sk-container-id-1 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-1 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-1\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>Pipeline(steps=[(&#x27;tf-idf&#x27;, TfidfVectorizer()), (&#x27;clf&#x27;, MultinomialNB())])</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-1\" type=\"checkbox\" ><label for=\"sk-estimator-id-1\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">Pipeline</label><div class=\"sk-toggleable__content\"><pre>Pipeline(steps=[(&#x27;tf-idf&#x27;, TfidfVectorizer()), (&#x27;clf&#x27;, MultinomialNB())])</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-2\" type=\"checkbox\" ><label for=\"sk-estimator-id-2\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">TfidfVectorizer</label><div class=\"sk-toggleable__content\"><pre>TfidfVectorizer()</pre></div></div></div><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-3\" type=\"checkbox\" ><label for=\"sk-estimator-id-3\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">MultinomialNB</label><div class=\"sk-toggleable__content\"><pre>MultinomialNB()</pre></div></div></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate baseline model on validation dataset\n",
        "model_0.score(X=val_sentences,\n",
        "              y=val_labels_encoded)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eenh22b8akHD",
        "outputId": "39968ec7-26e1-411e-bbe4-fa859c66e753"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.7218323844829869"
            ]
          },
          "metadata": {},
          "execution_count": 22
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Make predictions using our baseline model\n",
        "baseline_preds = model_0.predict(val_sentences)\n",
        "baseline_preds"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EWd90K6qa1Tt",
        "outputId": "188d3169-2cc6-443e-e48c-a014ca232773"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([4, 1, 3, ..., 4, 4, 1])"
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Download helper functions script\n",
        "In the previous module, we wrote a function to compare predictions across different metrics (accuracy, precision, recall, F1-score) and rather than rewriting it here, let's download it from our helper functions script."
      ],
      "metadata": {
        "id": "wy4p_Wr6a_Dc"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!wget https://raw.githubusercontent.com/mrdbourke/tensorflow-deep-learning/main/extras/helper_functions.py"
      ],
      "metadata": {
        "id": "8HaHDx3Fbi8O",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "fb896ca2-4f65-4593-9109-1055ff17a4f0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2024-09-02 14:15:07--  https://raw.githubusercontent.com/mrdbourke/tensorflow-deep-learning/main/extras/helper_functions.py\n",
            "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.108.133, 185.199.111.133, 185.199.109.133, ...\n",
            "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.108.133|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 10246 (10K) [text/plain]\n",
            "Saving to: ‘helper_functions.py’\n",
            "\n",
            "helper_functions.py 100%[===================>]  10.01K  --.-KB/s    in 0s      \n",
            "\n",
            "2024-09-02 14:15:08 (77.8 MB/s) - ‘helper_functions.py’ saved [10246/10246]\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from helper_functions import calculate_results"
      ],
      "metadata": {
        "id": "IcAie6MvbxjN"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Calculate baseline results\n",
        "baseline_results = calculate_results(y_true=val_labels_encoded,\n",
        "                                    y_pred=baseline_preds)\n",
        "baseline_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bT3Jqu2lb3bc",
        "outputId": "9e746330-b6e9-445f-d806-f01c38038816"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'accuracy': 72.1832384482987,\n",
              " 'precision': 0.7186466952323352,\n",
              " 'recall': 0.7218323844829869,\n",
              " 'f1': 0.6989250353450294}"
            ]
          },
          "metadata": {},
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Preparing our data (the text) for fdeep sequence models\n",
        "\n",
        "Before we start building our deeper models, we've got to create vectorizaton and embedding layers."
      ],
      "metadata": {
        "id": "G6FLBryfcEeh"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras import layers"
      ],
      "metadata": {
        "id": "n5hGVkfed5sk"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# How long is each sentence on average ?\n",
        "sent_lens = [len(sentence.split()) for sentence in train_sentences]\n",
        "avg_sent_len = np.mean(sent_lens)\n",
        "avg_sent_len"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6Pd3NHv5d_7J",
        "outputId": "c871069b-58d3-4790-c42a-ff0dbd37308c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "26.338269273494777"
            ]
          },
          "metadata": {},
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# What's the distribution look like ?\n",
        "import matplotlib.pyplot as plt\n",
        "plt.hist(sent_lens, bins=20);"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 430
        },
        "id": "Y3KaW3qvhuKC",
        "outputId": "a6b8b686-e2e3-491c-f968-75278b70db23"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# How long of a sentence covers 95% of examples?\n",
        "output_seq_len = int(np.percentile(sent_lens, 95))\n",
        "output_seq_len"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ct2CtKVviUMh",
        "outputId": "85a08b77-dc0f-48e6-83c2-226169f2289c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "55"
            ]
          },
          "metadata": {},
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Maximum sequence length in the training set\n",
        "max(sent_lens)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ahLaBI9ojNaj",
        "outputId": "8d7bb948-57f1-493b-dc10-cba9aa0af02c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "296"
            ]
          },
          "metadata": {},
          "execution_count": 31
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Create text vectorizer layer\n",
        "\n",
        "We want to make a layer which maps our texts from words to numbers."
      ],
      "metadata": {
        "id": "-UcPY9xMjYyL"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# How many words are in our vocabulary ? (taken from table 2 in pdf)\n",
        "max_tokens = 68000"
      ],
      "metadata": {
        "id": "flQSrzdKkVq2"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Create a text vectorizer\n",
        "from tensorflow.keras.layers import TextVectorization\n",
        "\n",
        "text_vectorizer = TextVectorization(max_tokens=max_tokens, # number of words in vocabulary\n",
        "                                    output_sequence_length=output_seq_len) # desired output length of vectorized sequences"
      ],
      "metadata": {
        "id": "GvZl3n4rlV7g"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Adapt text_vectorizer to training sentences\n",
        "text_vectorizer.adapt(train_sentences)"
      ],
      "metadata": {
        "id": "B7eKmdM4mt8X"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Test out text vectorizer on random sentences\n",
        "import random\n",
        "target_sentence = random.choice(train_sentences)\n",
        "print(f\"Text:\\n {target_sentence}\")\n",
        "print(f\"\\nLength of text:\\n {len(target_sentence)}\")\n",
        "print(f\"\\nVectorized text:\\n {text_vectorizer(target_sentence.split())}\")\n",
        "print(f\"\\nLength of vectorized text:\\n {len(text_vectorizer([target_sentence]))}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Br3UO5vnm2Mg",
        "outputId": "73b19953-97c5-4b08-b1ff-9786c48af922"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Text:\n",
            " patients with stage iii/iv resectable head and neck squamous cell carcinoma were randomized to surgery followed by rt or ccrt .\n",
            "\n",
            "Length of text:\n",
            " 127\n",
            "\n",
            "Vectorized text:\n",
            " [[  12    0    0 ...    0    0    0]\n",
            " [   7    0    0 ...    0    0    0]\n",
            " [ 531    0    0 ...    0    0    0]\n",
            " ...\n",
            " [  16    0    0 ...    0    0    0]\n",
            " [7255    0    0 ...    0    0    0]\n",
            " [   0    0    0 ...    0    0    0]]\n",
            "\n",
            "Length of vectorized text:\n",
            " 1\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# How many words in our training vocabulary\n",
        "rct_20k_text_vocab = text_vectorizer.get_vocabulary()\n",
        "print(f\"Number of words in vocab: {len(rct_20k_text_vocab)}\")\n",
        "print(f\"Top 5 most common words: {rct_20k_text_vocab[:5]}\")\n",
        "print(f\"Bottom 5 least common words: {rct_20k_text_vocab[-5:]}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Y8zymw7Kny7m",
        "outputId": "8f3509e7-a72c-4750-e029-5e77ab5c19f0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Number of words in vocab: 64841\n",
            "Top 5 most common words: ['', '[UNK]', 'the', 'and', 'of']\n",
            "Bottom 5 least common words: ['aainduced', 'aaigroup', 'aachener', 'aachen', 'aaacp']\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Get the config of our text_vectorizer\n",
        "text_vectorizer.get_config()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6xrkKxZpoh8X",
        "outputId": "73cb46d1-963b-475b-9ad6-5185036101f2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'name': 'text_vectorization',\n",
              " 'trainable': True,\n",
              " 'dtype': 'string',\n",
              " 'batch_input_shape': (None,),\n",
              " 'max_tokens': 68000,\n",
              " 'standardize': 'lower_and_strip_punctuation',\n",
              " 'split': 'whitespace',\n",
              " 'ngrams': None,\n",
              " 'output_mode': 'int',\n",
              " 'output_sequence_length': 55,\n",
              " 'pad_to_max_tokens': False,\n",
              " 'sparse': False,\n",
              " 'ragged': False,\n",
              " 'vocabulary': None,\n",
              " 'idf_weights': None,\n",
              " 'encoding': 'utf-8',\n",
              " 'vocabulary_size': 64841}"
            ]
          },
          "metadata": {},
          "execution_count": 37
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Create custom text embedding"
      ],
      "metadata": {
        "id": "wiFf8XrQor5n"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Create token embedding layer\n",
        "token_embed = layers.Embedding(input_dim=len(rct_20k_text_vocab), # length of vocabulary\n",
        "                               output_dim = 128, # Note: different embedding sizes result in drastically different sizes of parameters to train\n",
        "                               mask_zero = True, # use masking to handle variable sequence lengths (save space),\n",
        "                               name=\"token_embedding\")"
      ],
      "metadata": {
        "id": "F1dlgo_vpMAV"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Show an example embedding\n",
        "print(f\"Sentence before vectorization:\\n {target_sentence}\\n\")\n",
        "vectorized_sentence = text_vectorizer([target_sentence])\n",
        "print(f\"Sentence after vectorization (before embedding):\\n {vectorized_sentence}\\n\")\n",
        "embedded_sentence = token_embed(vectorized_sentence)\n",
        "print(f\"Sentence after embedding:\\n {embedded_sentence}\\n\")\n",
        "print(f\"Shape of embedded sentence:\\n {embedded_sentence.shape}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "l8UIdwEkqpAv",
        "outputId": "dac6c483-f0ef-415e-fd25-aa2f7484ba98"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Sentence before vectorization:\n",
            " patients with stage iii/iv resectable head and neck squamous cell carcinoma were randomized to surgery followed by rt or ccrt .\n",
            "\n",
            "Sentence after vectorization (before embedding):\n",
            " [[  12    7  531 6185 7300 1331    3  932 4619  507 1440    9   29    6\n",
            "   115  284   22 1095   16 7255    0    0    0    0    0    0    0    0\n",
            "     0    0    0    0    0    0    0    0    0    0    0    0    0    0\n",
            "     0    0    0    0    0    0    0    0    0    0    0    0    0]]\n",
            "\n",
            "Sentence after embedding:\n",
            " [[[-0.02454619  0.00023599 -0.02444463 ... -0.022552   -0.04585155\n",
            "   -0.04712247]\n",
            "  [-0.00879901 -0.04993704 -0.00141471 ...  0.03293983 -0.04913882\n",
            "   -0.03328526]\n",
            "  [ 0.03600181  0.02443949 -0.00563819 ...  0.01455407 -0.01336281\n",
            "    0.00085812]\n",
            "  ...\n",
            "  [ 0.03022117  0.03243425 -0.00968224 ... -0.03236888 -0.02527928\n",
            "    0.04837836]\n",
            "  [ 0.03022117  0.03243425 -0.00968224 ... -0.03236888 -0.02527928\n",
            "    0.04837836]\n",
            "  [ 0.03022117  0.03243425 -0.00968224 ... -0.03236888 -0.02527928\n",
            "    0.04837836]]]\n",
            "\n",
            "Shape of embedded sentence:\n",
            " (1, 55, 128)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Creating datasets (making sure our data loads as fast as possible)\n",
        "\n",
        "We're going to setup our data to run as fast as possible with the TensorFlow tf.data API, many of the steps here are discussed at length in these two resources :\n",
        "* https://www.tensorflow.org/guide/data_performance\n",
        "* https://www.tensorflow.org/guide/data"
      ],
      "metadata": {
        "id": "CiSe8EHLr7o2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Turn our data into tensorflow datasets\n",
        "train_dataset = tf.data.Dataset.from_tensor_slices((train_sentences, train_labels_one_hot))\n",
        "val_dataset = tf.data.Dataset.from_tensor_slices((val_sentences, val_labels_one_hot))\n",
        "test_dataset = tf.data.Dataset.from_tensor_slices((test_sentences, test_labels_one_hot))\n",
        "\n",
        "train_dataset, val_dataset, test_dataset"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bNItEiw1DEmq",
        "outputId": "4b68e01f-e747-46c5-b4ba-2c619c7859cb"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(<_TensorSliceDataset element_spec=(TensorSpec(shape=(), dtype=tf.string, name=None), TensorSpec(shape=(5,), dtype=tf.float64, name=None))>,\n",
              " <_TensorSliceDataset element_spec=(TensorSpec(shape=(), dtype=tf.string, name=None), TensorSpec(shape=(5,), dtype=tf.float64, name=None))>,\n",
              " <_TensorSliceDataset element_spec=(TensorSpec(shape=(), dtype=tf.string, name=None), TensorSpec(shape=(5,), dtype=tf.float64, name=None))>)"
            ]
          },
          "metadata": {},
          "execution_count": 40
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Take the TensorSliceDataset's and turn them into prefetched datasets\n",
        "train_dataset = train_dataset.batch(32).prefetch(tf.data.AUTOTUNE)\n",
        "val_dataset = val_dataset.batch(32).prefetch(tf.data.AUTOTUNE)\n",
        "test_dataset = test_dataset.batch(32).prefetch(tf.data.AUTOTUNE)\n",
        "\n",
        "train_dataset, val_dataset, test_dataset"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lz2sGitgDchb",
        "outputId": "71b4fb1a-3e58-4517-c634-d0d3778e2693"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(<_PrefetchDataset element_spec=(TensorSpec(shape=(None,), dtype=tf.string, name=None), TensorSpec(shape=(None, 5), dtype=tf.float64, name=None))>,\n",
              " <_PrefetchDataset element_spec=(TensorSpec(shape=(None,), dtype=tf.string, name=None), TensorSpec(shape=(None, 5), dtype=tf.float64, name=None))>,\n",
              " <_PrefetchDataset element_spec=(TensorSpec(shape=(None,), dtype=tf.string, name=None), TensorSpec(shape=(None, 5), dtype=tf.float64, name=None))>)"
            ]
          },
          "metadata": {},
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model 1: Conv1D with token embeddings"
      ],
      "metadata": {
        "id": "xoYnfK1gEv3q"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Create 1D conv model to process sequences\n",
        "inputs = layers.Input(shape=(1,), dtype=tf.string)\n",
        "text_vectors = text_vectorizer(inputs) # vectorize text inputs\n",
        "token_embeddings = token_embed(text_vectors) # create embedding\n",
        "x = layers.Conv1D(filters=64, kernel_size=5, padding=\"same\", activation=\"relu\")(token_embeddings)\n",
        "x = layers.GlobalAveragePooling1D()(x) # condense the output of our feature vector form conv layer\n",
        "outputs = layers.Dense(num_classes, activation=\"softmax\")(x)\n",
        "model_1 = tf.keras.Model(inputs, outputs, name=\"model_1_conv1d\")\n",
        "\n",
        "# Compile\n",
        "model_1.compile(loss=\"categorical_crossentropy\",\n",
        "                optimizer=tf.keras.optimizers.Adam(),\n",
        "                metrics=[\"accuracy\"])"
      ],
      "metadata": {
        "id": "vKs-8Rz0HaO3"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model_1.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GUAQFOCUIpCp",
        "outputId": "b044ca63-c8dd-4138-8f12-f289150dd231"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_1_conv1d\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_1 (InputLayer)        [(None, 1)]               0         \n",
            "                                                                 \n",
            " text_vectorization (TextVe  (None, 55)                0         \n",
            " ctorization)                                                    \n",
            "                                                                 \n",
            " token_embedding (Embedding  (None, 55, 128)           8299648   \n",
            " )                                                               \n",
            "                                                                 \n",
            " conv1d (Conv1D)             (None, 55, 64)            41024     \n",
            "                                                                 \n",
            " global_average_pooling1d (  (None, 64)                0         \n",
            " GlobalAveragePooling1D)                                         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 5)                 325       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 8340997 (31.82 MB)\n",
            "Trainable params: 8340997 (31.82 MB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Fit the model\n",
        "history_model_1 = model_1.fit(train_dataset,\n",
        "                              steps_per_epoch = int(0.1 * len(train_dataset)),\n",
        "                              epochs=3,\n",
        "                              validation_data=val_dataset,\n",
        "                              validation_steps = int(0.1 * len(val_dataset))) # only validate on 10% of the data"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cwOVKaHlI_iv",
        "outputId": "66085587-c310-492d-c4df-319cbee35d6f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/3\n",
            "562/562 [==============================] - 21s 32ms/step - loss: 0.9096 - accuracy: 0.6400 - val_loss: 0.6864 - val_accuracy: 0.7340\n",
            "Epoch 2/3\n",
            "562/562 [==============================] - 5s 9ms/step - loss: 0.6606 - accuracy: 0.7519 - val_loss: 0.6390 - val_accuracy: 0.7680\n",
            "Epoch 3/3\n",
            "562/562 [==============================] - 5s 10ms/step - loss: 0.6213 - accuracy: 0.7727 - val_loss: 0.6001 - val_accuracy: 0.7816\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate on whole validation dataset\n",
        "model_1.evaluate(val_dataset)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VJYevnCaJxRo",
        "outputId": "aaefe216-940e-41e3-827c-bd71ac1eae0a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "945/945 [==============================] - 3s 3ms/step - loss: 0.6011 - accuracy: 0.7849\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.6011207103729248, 0.7848536968231201]"
            ]
          },
          "metadata": {},
          "execution_count": 45
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Make predictions (our model predicts prediction probabilities for each class)\n",
        "model_1_pred_probs = model_1.predict(val_dataset)\n",
        "model_1_pred_probs, model_1_pred_probs.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NPNTcr3GKNCG",
        "outputId": "e9b64e17-98f8-4933-b817-4de69efb4732"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "945/945 [==============================] - 2s 2ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([[4.0099218e-01, 2.1697499e-01, 7.0379697e-02, 2.7427101e-01,\n",
              "         3.7382200e-02],\n",
              "        [4.3288320e-01, 3.2385138e-01, 1.1183839e-02, 2.2280152e-01,\n",
              "         9.2799822e-03],\n",
              "        [1.8008643e-01, 1.0721195e-02, 2.2310009e-03, 8.0691308e-01,\n",
              "         4.8248174e-05],\n",
              "        ...,\n",
              "        [6.3462489e-06, 6.5897976e-04, 1.0119893e-03, 3.5571911e-06,\n",
              "         9.9831909e-01],\n",
              "        [5.2907035e-02, 4.8389488e-01, 8.1045456e-02, 7.1251571e-02,\n",
              "         3.1090108e-01],\n",
              "        [1.7327568e-01, 6.6476977e-01, 4.5381915e-02, 4.5210611e-02,\n",
              "         7.1362056e-02]], dtype=float32),\n",
              " (30212, 5))"
            ]
          },
          "metadata": {},
          "execution_count": 46
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Convert pred probs to classes\n",
        "model_1_preds = tf.argmax(model_1_pred_probs, axis=1)\n",
        "model_1_preds"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cBNf9s6vKias",
        "outputId": "308552f4-3ff0-46c6-fa65-ad470497b960"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(30212,), dtype=int64, numpy=array([0, 0, 3, ..., 4, 1, 1])>"
            ]
          },
          "metadata": {},
          "execution_count": 47
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Calculate model_1 results\n",
        "model_1_results = calculate_results(y_true=val_labels_encoded,\n",
        "                                    y_pred=model_1_preds)\n",
        "model_1_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EKDFIr5mK4qW",
        "outputId": "54160ee3-6f3c-4571-b6f7-d0b27667b935"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'accuracy': 78.48537005163512,\n",
              " 'precision': 0.7813028708243923,\n",
              " 'recall': 0.7848537005163512,\n",
              " 'f1': 0.7821852447612673}"
            ]
          },
          "metadata": {},
          "execution_count": 48
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "baseline_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZGVHnCXULDmP",
        "outputId": "06e5d689-66a5-4fca-b329-727b9608d265"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'accuracy': 72.1832384482987,\n",
              " 'precision': 0.7186466952323352,\n",
              " 'recall': 0.7218323844829869,\n",
              " 'f1': 0.6989250353450294}"
            ]
          },
          "metadata": {},
          "execution_count": 49
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Here, we can see that we've already beat the baseline model, let's build more models further to improve the results even more."
      ],
      "metadata": {
        "id": "oz9paYSCLJk0"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model 2: Feature extraction with pretrained token embeddings\n",
        "\n",
        "Now lets use pretrained word embeddings from TensorFlow Hub, more specifically the universal sentence encoder (USE) : https://tfhub.dev/google/universal-sentence-encoder/4\n",
        "\n",
        "The paper originally used GloVe embeddings, however, we're going to stick with the later created USE pretrained embeddings."
      ],
      "metadata": {
        "id": "9Xk4ZEaJLwS3"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Download pretrained TensorFlow Hub USE\n",
        "import tensorflow_hub as hub\n",
        "tf_hub_embedding_layer = hub.KerasLayer(\"https://tfhub.dev/google/universal-sentence-encoder/4\",\n",
        "                                        trainable=False,\n",
        "                                        name=\"universal_sentence_encoder\")"
      ],
      "metadata": {
        "id": "0bGiV4gdN-Kc"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Test out the pretrained embedding on a random sentence\n",
        "random_train_sentence = random.choice(train_sentences)\n",
        "print(f\"Random Sentence:\\n {random_train_sentence}\\n\")\n",
        "use_embedded_sentence = tf_hub_embedding_layer([random_train_sentence])\n",
        "print(f\"Sentence after embedding:\\n {use_embedded_sentence[0][:30]}\\n\")\n",
        "print(f\"Length of sentence embedding:\\n {len(use_embedded_sentence[0])}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Cf2iYpURO8WF",
        "outputId": "ea10ff70-25a9-4de6-a08f-f17f31883a56"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Random Sentence:\n",
            " objective pco assessment resulted in a mean pco score ( scale @-@ ) of @ was found for the corneal j@d group and a score of @ was found for the corneal a@d group ( p > @ ) .\n",
            "\n",
            "Sentence after embedding:\n",
            " [-0.01802811 -0.04011983  0.039552    0.00698578  0.05330688  0.05695429\n",
            " -0.03917615  0.01665497 -0.03382353 -0.01957732  0.07120212  0.04649854\n",
            " -0.03488247  0.05165797  0.06297681  0.0356151  -0.01818537 -0.00346021\n",
            "  0.02188957  0.01222319 -0.06479126  0.02106623 -0.04631111 -0.00378186\n",
            "  0.04827685  0.02641159 -0.00514895  0.00234413 -0.02602232 -0.01054108]\n",
            "\n",
            "Length of sentence embedding:\n",
            " 512\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Building and fitting an NLP feature extraction model using pretrained embeddings from TensorFlow Hub"
      ],
      "metadata": {
        "id": "Ywul7ikVPvp2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Define feature extraction model using TF Hub layer\n",
        "inputs = layers.Input(shape=[], dtype=tf.string)\n",
        "pretrained_embedding = tf_hub_embedding_layer(inputs) # tokenize text and create embedding of each sequence (512 long vector)\n",
        "x = layers.Dense(128, activation=\"relu\")(pretrained_embedding)\n",
        "# Note : We could add more layers here if we wanted to\n",
        "outputs = layers.Dense(num_classes, activation=\"softmax\")(x) # create the output layer\n",
        "model_2 = tf.keras.Model(inputs, outputs, name=\"model_2_use\")\n",
        "\n",
        "# Compile\n",
        "model_2.compile(loss=\"categorical_crossentropy\",\n",
        "                optimizer=tf.keras.optimizers.Adam(),\n",
        "                metrics=[\"accuracy\"])"
      ],
      "metadata": {
        "id": "GiF51Am9Qy79"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model_2.summary()"
      ],
      "metadata": {
        "id": "oodZFM8zSDqV",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d6ee25e5-292e-4931-ab32-467301d6eafe"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_2_use\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_2 (InputLayer)        [(None,)]                 0         \n",
            "                                                                 \n",
            " universal_sentence_encoder  (None, 512)               256797824 \n",
            "  (KerasLayer)                                                   \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 128)               65664     \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 5)                 645       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 256864133 (979.86 MB)\n",
            "Trainable params: 66309 (259.02 KB)\n",
            "Non-trainable params: 256797824 (979.61 MB)\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Fit model 2 to the data\n",
        "history_model_2 = model_2.fit(train_dataset,\n",
        "                              steps_per_epoch = int(0.1 * len(train_dataset)),\n",
        "                              epochs=3,\n",
        "                              validation_data=val_dataset,\n",
        "                              validation_steps = int(0.1 * len(val_dataset)))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UTqrVV7uWPtv",
        "outputId": "e8d3b778-2d2e-4322-acf1-ae460ed7741b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/3\n",
            "562/562 [==============================] - 12s 16ms/step - loss: 0.9144 - accuracy: 0.6505 - val_loss: 0.7967 - val_accuracy: 0.6905\n",
            "Epoch 2/3\n",
            "562/562 [==============================] - 7s 13ms/step - loss: 0.7679 - accuracy: 0.7021 - val_loss: 0.7552 - val_accuracy: 0.7058\n",
            "Epoch 3/3\n",
            "562/562 [==============================] - 8s 15ms/step - loss: 0.7507 - accuracy: 0.7139 - val_loss: 0.7372 - val_accuracy: 0.7181\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate on the whole validation dataset\n",
        "model_2.evaluate(val_dataset)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "q20q-AnTWcqu",
        "outputId": "6a2f40c3-2497-48cf-a6e3-b59b659cbb52"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "945/945 [==============================] - 11s 11ms/step - loss: 0.7399 - accuracy: 0.7142\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.7399210929870605, 0.7141532897949219]"
            ]
          },
          "metadata": {},
          "execution_count": 55
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Make predictions with our feature extraction dataset\n",
        "model_2_pred_probs = model_2.predict(val_dataset)\n",
        "model_2_pred_probs"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6UFIUdlHWvtX",
        "outputId": "7e278c4a-2310-4456-9be1-9831b24757e4"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "945/945 [==============================] - 10s 11ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.43030038, 0.35709623, 0.00233151, 0.20195688, 0.00831493],\n",
              "       [0.33296695, 0.52269596, 0.00457564, 0.13702261, 0.00273883],\n",
              "       [0.2180269 , 0.1602005 , 0.01676266, 0.5595123 , 0.04549755],\n",
              "       ...,\n",
              "       [0.00156528, 0.00492699, 0.04685218, 0.00096908, 0.9456865 ],\n",
              "       [0.00395061, 0.04681797, 0.22008188, 0.00150024, 0.7276493 ],\n",
              "       [0.17044207, 0.24031566, 0.5048252 , 0.01051558, 0.0739015 ]],\n",
              "      dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 56
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Convert the prediction probabilities found with the feature extraction model to labels\n",
        "model_2_preds = tf.argmax(model_2_pred_probs, axis=1)\n",
        "model_2_preds"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "I8YNPFodW63u",
        "outputId": "2b42f21c-655a-4da4-ac61-5e27ab50251e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(30212,), dtype=int64, numpy=array([0, 1, 3, ..., 4, 4, 2])>"
            ]
          },
          "metadata": {},
          "execution_count": 57
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Calculate results from TF Hub pretrained embeddings results on val set\n",
        "model_2_results = calculate_results(y_true=val_labels_encoded,\n",
        "                                    y_pred=model_2_preds)\n",
        "model_2_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3HqVfwEZXTum",
        "outputId": "2b7bc3a9-4534-4e90-c6ca-83a3b9475d90"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'accuracy': 71.4153316562955,\n",
              " 'precision': 0.7148443855134143,\n",
              " 'recall': 0.7141533165629551,\n",
              " 'f1': 0.7111865032848426}"
            ]
          },
          "metadata": {},
          "execution_count": 58
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model 3: Conv1D with character embeddings\n",
        "\n",
        "The paper which we're replicating states that they used a combination of token and character-level embeddings.\n",
        "\n",
        "Previously we've used token-level embeddings but we'll need to do similar steps for characters if we want to use char-level embeddings."
      ],
      "metadata": {
        "id": "ZWC6bGHzYmg-"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Creating a character level tokenizer\n"
      ],
      "metadata": {
        "id": "U9zN4SArenas"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_sentences[:5]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OTnf9cxofS-2",
        "outputId": "622850c1-1e29-47cd-f2e4-1dd2e8cd0877"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['to investigate the efficacy of @ weeks of daily low-dose oral prednisolone in improving pain , mobility , and systemic low-grade inflammation in the short term and whether the effect would be sustained at @ weeks in older adults with moderate to severe knee osteoarthritis ( oa ) .',\n",
              " 'a total of @ patients with primary knee oa were randomized @:@ ; @ received @ mg/day of prednisolone and @ received placebo for @ weeks .',\n",
              " 'outcome measures included pain reduction and improvement in function scores and systemic inflammation markers .',\n",
              " 'pain was assessed using the visual analog pain scale ( @-@ mm ) .',\n",
              " 'secondary outcome measures included the western ontario and mcmaster universities osteoarthritis index scores , patient global assessment ( pga ) of the severity of knee oa , and @-min walk distance ( @mwd ) .']"
            ]
          },
          "metadata": {},
          "execution_count": 59
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Make function to split sentences into characters\n",
        "def split_chars(text):\n",
        "  return \" \".join(list(text))\n",
        "\n",
        "# Text splitting non-character-level sequence into characters\n",
        "split_chars(random_train_sentence)"
      ],
      "metadata": {
        "id": "XwM3xfvafWvH",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 70
        },
        "outputId": "294a3714-2cf0-45f0-fc5d-f70c51768887"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'o b j e c t i v e   p c o   a s s e s s m e n t   r e s u l t e d   i n   a   m e a n   p c o   s c o r e   (   s c a l e   @ - @   )   o f   @   w a s   f o u n d   f o r   t h e   c o r n e a l   j @ d   g r o u p   a n d   a   s c o r e   o f   @   w a s   f o u n d   f o r   t h e   c o r n e a l   a @ d   g r o u p   (   p   >   @   )   .'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 60
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Split sequence-level data splits into character-level data splits\n",
        "train_chars = [split_chars(sentence) for sentence in train_sentences]\n",
        "val_chars = [split_chars(sentence) for sentence in val_sentences]\n",
        "test_chars = [split_chars(sentence) for sentence in test_sentences]"
      ],
      "metadata": {
        "id": "yypRGoIkfn2g"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# What's the average character length ?\n",
        "char_lens = [len(sentence) for sentence in train_sentences]\n",
        "mean_char_len = np.mean(char_lens)\n",
        "mean_char_len"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "g9A4bh-nfukC",
        "outputId": "c13786da-265a-41e2-a2e2-1013af73197a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "149.3662574983337"
            ]
          },
          "metadata": {},
          "execution_count": 62
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Check the distribution of our sequences at a character-level\n",
        "import matplotlib.pyplot as plt\n",
        "plt.hist(char_lens, bins=20);"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 430
        },
        "id": "Pr_tWMuviMkh",
        "outputId": "c1a0b6d4-bfce-4251-9cf4-132b88378663"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Find what character length covers 95% of our sequences\n",
        "output_seq_char_len = int(np.percentile(char_lens, 95))\n",
        "output_seq_char_len"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jzxF1LokipmY",
        "outputId": "ad17fdd8-a8b6-462c-c082-3281518a1bb5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "290"
            ]
          },
          "metadata": {},
          "execution_count": 64
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        " # Get all keyboard characters\n",
        " import string\n",
        " alphabet = string.ascii_lowercase + string.digits + string.punctuation\n",
        " alphabet"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        },
        "id": "yW_LzDXFjCNZ",
        "outputId": "f68f91da-7712-4417-efe3-4479b93f56c5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'abcdefghijklmnopqrstuvwxyz0123456789!\"#$%&\\'()*+,-./:;<=>?@[\\\\]^_`{|}~'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 65
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Create char-level token vectorizer instance\n",
        "NUM_CHAR_TOKENS = len(alphabet) + 2 # add 2 for space and OOV tokens\n",
        "char_vectorizer = TextVectorization(max_tokens = NUM_CHAR_TOKENS,\n",
        "                                    output_sequence_length = output_seq_char_len,\n",
        "                                    standardize=\"lower_and_strip_punctuation\",\n",
        "                                    name=\"char_vectorizer\")"
      ],
      "metadata": {
        "id": "vB00FiiglyaH"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Adapt character vectorizer to our training character\n",
        "char_vectorizer.adapt(train_chars)"
      ],
      "metadata": {
        "id": "MVen_VN2mwJY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Check character vocab stats\n",
        "char_vocab = char_vectorizer.get_vocabulary()\n",
        "print(f\"Number of characters in character vocab: {len(char_vocab)}\")\n",
        "print(f\"Top 5 most common characters in vocab: {char_vocab[:5]}\")\n",
        "print(f\"Bottom 5 least common characters in vocab: {char_vocab[-5:]}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "k03-iSrvm32n",
        "outputId": "e91cccbb-625e-473e-eb7d-a69e8fdf23e0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Number of characters in character vocab: 28\n",
            "Top 5 most common characters in vocab: ['', '[UNK]', 'e', 't', 'i']\n",
            "Bottom 5 least common characters in vocab: ['k', 'x', 'z', 'q', 'j']\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Test our character vectorizer\n",
        "random_train_chars = random.choice(train_chars)\n",
        "print(f\"Charified text:\\n {random_train_chars}\\n\")\n",
        "print(f\"Length of random_train_chars:\\n {len(random_train_chars.split())}\\n\")\n",
        "vectorized_chars = char_vectorizer([random_train_chars])\n",
        "print(f\"Vectorized chars:\\n {vectorized_chars}\\n\")\n",
        "print(f\"Length of vectorized text:\\n {len(vectorized_chars[0])}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Jih0JAdSnLMv",
        "outputId": "1c0cc44e-e7e7-4dba-b40d-5bd372ba4b04"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Charified text:\n",
            " o u t c o m e s   w e r e   t h r e e - m o n t h   c h a n g e s   i n   t h e   s h o r t   f o r m   h e a l t h   s u r v e y   (   s f - @   )   m e n t a l   h e a l t h   c o m p o n e n t   s u m m a r y   s c o r e   (   m c s   )   a n d   p h y s i c a l   h e a l t h   c o m p o n e n t   s u m m a r y   s c o r e   (   p c s   )   .\n",
            "\n",
            "Length of random_train_chars:\n",
            " 144\n",
            "\n",
            "Vectorized chars:\n",
            " [[ 7 16  3 11  7 15  2  9 20  2  8  2  3 13  8  2  2 15  7  6  3 13 11 13\n",
            "   5  6 18  2  9  4  6  3 13  2  9 13  7  8  3 17  7  8 15 13  2  5 12  3\n",
            "  13  9 16  8 21  2 19  9 17 15  2  6  3  5 12 13  2  5 12  3 13 11  7 15\n",
            "  14  7  6  2  6  3  9 16 15 15  5  8 19  9 11  7  8  2 15 11  9  5  6 10\n",
            "  14 13 19  9  4 11  5 12 13  2  5 12  3 13 11  7 15 14  7  6  2  6  3  9\n",
            "  16 15 15  5  8 19  9 11  7  8  2 14 11  9  0  0  0  0  0  0  0  0  0  0\n",
            "   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
            "   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
            "   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
            "   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
            "   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
            "   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
            "   0  0]]\n",
            "\n",
            "Length of vectorized text:\n",
            " 290\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Creating a character-level embedding"
      ],
      "metadata": {
        "id": "vUjMBS6moe1t"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Create char embedding layer\n",
        "char_embed = layers.Embedding(input_dim=len(char_vocab), # number of different characters\n",
        "                              output_dim=25, # this is the size of the char embedding in the paper\n",
        "                              mask_zero=True,\n",
        "                              name=\"char_embed\")"
      ],
      "metadata": {
        "id": "jnvkVKBmrder"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Test our character embedding layer\n",
        "print(f\"Charified text:\\n {random_train_chars}\\n\")\n",
        "char_embed_example = char_embed(char_vectorizer([random_train_chars]))\n",
        "print(f\"Embedded chars (after vectorization and embedding):\\n {char_embed_example}\\n\")\n",
        "print(f\"Shape of embedded chars:\\n {char_embed_example.shape}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hLJ1KX-lsL0W",
        "outputId": "bb05a801-7ed7-4a03-fffc-1d7292573a97"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Charified text:\n",
            " o u t c o m e s   w e r e   t h r e e - m o n t h   c h a n g e s   i n   t h e   s h o r t   f o r m   h e a l t h   s u r v e y   (   s f - @   )   m e n t a l   h e a l t h   c o m p o n e n t   s u m m a r y   s c o r e   (   m c s   )   a n d   p h y s i c a l   h e a l t h   c o m p o n e n t   s u m m a r y   s c o r e   (   p c s   )   .\n",
            "\n",
            "Embedded chars (after vectorization and embedding):\n",
            " [[[-0.00356125 -0.04174126 -0.02171342 ... -0.01290495  0.01534958\n",
            "   -0.03712969]\n",
            "  [-0.04422226 -0.02800444 -0.02616677 ...  0.00196229 -0.02469349\n",
            "    0.03971403]\n",
            "  [ 0.04985894 -0.04337765  0.04349054 ...  0.02425935 -0.01033183\n",
            "   -0.01801559]\n",
            "  ...\n",
            "  [-0.0005231  -0.02790444 -0.00159536 ... -0.039543   -0.04994811\n",
            "   -0.04296265]\n",
            "  [-0.0005231  -0.02790444 -0.00159536 ... -0.039543   -0.04994811\n",
            "   -0.04296265]\n",
            "  [-0.0005231  -0.02790444 -0.00159536 ... -0.039543   -0.04994811\n",
            "   -0.04296265]]]\n",
            "\n",
            "Shape of embedded chars:\n",
            " (1, 290, 25)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Building a Conv1D model to fit on character embeddings"
      ],
      "metadata": {
        "id": "VRK3FysstQTt"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Make Conv1D on chars only\n",
        "inputs = layers.Input(shape=(1,), dtype=tf.string)\n",
        "char_vectors = char_vectorizer(inputs) # vectorize text inputs\n",
        "char_embeddings = char_embed(char_vectors) # create embedding\n",
        "x = layers.Conv1D(filters=64, kernel_size=5, padding=\"same\", activation=\"relu\")(char_embeddings)\n",
        "x = layers.GlobalMaxPool1D()(x) # condense the output of our feature vector form conv layer\n",
        "outputs = layers.Dense(num_classes, activation=\"softmax\")(x)\n",
        "model_3 = tf.keras.Model(inputs, outputs, name=\"model_3_conv1D_char_embeddings\")\n",
        "\n",
        "# Compile\n",
        "model_3.compile(loss=\"categorical_crossentropy\",\n",
        "                optimizer=tf.keras.optimizers.Adam(),\n",
        "                metrics=[\"accuracy\"])"
      ],
      "metadata": {
        "id": "48A0ZaoxuIfT"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Summary of our model\n",
        "model_3.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qSZkOm-Puueb",
        "outputId": "b72fab68-aabe-4160-d83f-02f95a3a1de7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_3_conv1D_char_embeddings\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_3 (InputLayer)        [(None, 1)]               0         \n",
            "                                                                 \n",
            " char_vectorizer (TextVecto  (None, 290)               0         \n",
            " rization)                                                       \n",
            "                                                                 \n",
            " char_embed (Embedding)      (None, 290, 25)           700       \n",
            "                                                                 \n",
            " conv1d_1 (Conv1D)           (None, 290, 64)           8064      \n",
            "                                                                 \n",
            " global_max_pooling1d (Glob  (None, 64)                0         \n",
            " alMaxPooling1D)                                                 \n",
            "                                                                 \n",
            " dense_3 (Dense)             (None, 5)                 325       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 9089 (35.50 KB)\n",
            "Trainable params: 9089 (35.50 KB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Create char level datasets\n",
        "train_char_dataset = tf.data.Dataset.from_tensor_slices((train_chars, train_labels_one_hot)).batch(32).prefetch(tf.data.AUTOTUNE)\n",
        "val_char_dataset = tf.data.Dataset.from_tensor_slices((val_chars, val_labels_one_hot)).batch(32).prefetch(tf.data.AUTOTUNE)\n",
        "test_char_dataset = tf.data.Dataset.from_tensor_slices((test_chars, test_labels_one_hot)).batch(32).prefetch(tf.data.AUTOTUNE)\n",
        "\n",
        "train_char_dataset, val_char_dataset, test_char_dataset"
      ],
      "metadata": {
        "id": "VY6Ul3bTuw_8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "04fc429a-5ec7-433a-91dd-23f81513fe81"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(<_PrefetchDataset element_spec=(TensorSpec(shape=(None,), dtype=tf.string, name=None), TensorSpec(shape=(None, 5), dtype=tf.float64, name=None))>,\n",
              " <_PrefetchDataset element_spec=(TensorSpec(shape=(None,), dtype=tf.string, name=None), TensorSpec(shape=(None, 5), dtype=tf.float64, name=None))>,\n",
              " <_PrefetchDataset element_spec=(TensorSpec(shape=(None,), dtype=tf.string, name=None), TensorSpec(shape=(None, 5), dtype=tf.float64, name=None))>)"
            ]
          },
          "metadata": {},
          "execution_count": 74
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Fit the model on chars only\n",
        "model_3_history = model_3.fit(train_char_dataset,\n",
        "                              steps_per_epoch = int(0.1 * len(train_char_dataset)),\n",
        "                              epochs=3,\n",
        "                              validation_data=val_char_dataset,\n",
        "                              validation_steps = int(0.1 * len(val_char_dataset)))"
      ],
      "metadata": {
        "id": "bTvaSe8jvbUM",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "9d751c6a-e11e-4cc8-efbd-df698375cfda"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/3\n",
            "562/562 [==============================] - 5s 7ms/step - loss: 1.2655 - accuracy: 0.4786 - val_loss: 1.0303 - val_accuracy: 0.6031\n",
            "Epoch 2/3\n",
            "562/562 [==============================] - 3s 5ms/step - loss: 0.9872 - accuracy: 0.6079 - val_loss: 0.9350 - val_accuracy: 0.6360\n",
            "Epoch 3/3\n",
            "562/562 [==============================] - 3s 5ms/step - loss: 0.9118 - accuracy: 0.6433 - val_loss: 0.8527 - val_accuracy: 0.6759\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Make predictions with character model only\n",
        "model_3_pred_probs = model_3.predict(val_char_dataset)\n",
        "model_3_pred_probs"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CTKB08xOv0Ib",
        "outputId": "b35d457c-4e55-4582-8280-d845e1c5590e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "945/945 [==============================] - 3s 3ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.12567368, 0.55209416, 0.08688297, 0.19915816, 0.03619099],\n",
              "       [0.1999659 , 0.3853983 , 0.0350528 , 0.3305815 , 0.04900153],\n",
              "       [0.11672918, 0.1764305 , 0.15693064, 0.5274434 , 0.02246627],\n",
              "       ...,\n",
              "       [0.02547557, 0.02516029, 0.21492286, 0.02278581, 0.7116555 ],\n",
              "       [0.03637941, 0.03779868, 0.39697987, 0.064648  , 0.46419412],\n",
              "       [0.23691571, 0.48893374, 0.16949433, 0.07756282, 0.02709344]],\n",
              "      dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 76
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Convert prediction probabilities to class labels\n",
        "model_3_preds = tf.argmax(model_3_pred_probs, axis=1)\n",
        "model_3_preds"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "A3R1Xv2CwTtF",
        "outputId": "a7310c2b-bccc-4f52-fe60-8d8e166daff9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(30212,), dtype=int64, numpy=array([1, 1, 3, ..., 4, 4, 1])>"
            ]
          },
          "metadata": {},
          "execution_count": 77
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Calculate results for Conv1D model chars only\n",
        "model_3_results = calculate_results(y_true=val_labels_encoded,\n",
        "                                    y_pred=model_3_preds)\n",
        "model_3_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qYKO8aMiwekU",
        "outputId": "d9833f60-a0e1-4fd6-9fa9-5d2ecfb0c775"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'accuracy': 66.06315371375612,\n",
              " 'precision': 0.6544577397132115,\n",
              " 'recall': 0.6606315371375613,\n",
              " 'f1': 0.6521810036635403}"
            ]
          },
          "metadata": {},
          "execution_count": 78
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model 4: Combining pretrained token embeddings + character embeddings (hybrid embedding layer)\n",
        "\n",
        "1. Create a token-level embedding model (similar to `model_1`)\n",
        "2. Create a character-level model (similar to `model_3` with a slight modification)\n",
        "3. Combine 1 & 2 with a concatenate layer (`layers.concatenate`)\n",
        "4. Build a series of output layers on top of 3 similar to Figure 1 and section 4.2 of the paper.\n",
        "5. Construct a model which takes token and character-level sequences as input and produces sequence label probabilites as output."
      ],
      "metadata": {
        "id": "DkjPDOtlwk3L"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 1. Setup the token inputs/model\n",
        "token_inputs = layers.Input(shape=[], dtype=tf.string, name=\"token_input\")\n",
        "token_embeddings = tf_hub_embedding_layer(token_inputs)\n",
        "token_output = layers.Dense(128, activation=\"relu\")(token_embeddings)\n",
        "token_model = tf.keras.Model(token_inputs, token_output)\n",
        "\n",
        "# 2. Setup char inputs/models\n",
        "char_inputs = layers.Input(shape=(1,), dtype=tf.string, name=\"char_input\")\n",
        "char_vectors = char_vectorizer(char_inputs)\n",
        "char_embeddings = char_embed(char_vectors)\n",
        "char_bi_lstm = layers.Bidirectional(layers.LSTM(24))(char_embeddings) # bi-LSTM layer\n",
        "char_model = tf.keras.Model(char_inputs, char_bi_lstm)\n",
        "\n",
        "# 3. Concatenate token and char inputs (create a hybrid token embedding)\n",
        "token_char_concat = layers.Concatenate(name=\"token_char_hybrid\")([token_model.output,\n",
        "                                                                  char_model.output])\n",
        "\n",
        "# 4. Create output layers - adding in dropout (discussed in section 4.2 of the paper)\n",
        "combined_dropout = layers.Dropout(0.5)(token_char_concat)\n",
        "combined_dense = layers.Dense(128, activation=\"relu\")(combined_dropout)\n",
        "final_dropout = layers.Dropout(0.5)(combined_dense)\n",
        "output_layer = layers.Dense(num_classes, activation=\"softmax\")(final_dropout)\n",
        "\n",
        "# 5. Construct model with char and token inputs\n",
        "model_4 = tf.keras.Model(inputs=[token_model.input, char_model.input],\n",
        "                      outputs=output_layer,\n",
        "                      name=\"model_4_token_and_char_embeddings\")"
      ],
      "metadata": {
        "id": "Po7nRCmMzZiN"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Get a summary of our model\n",
        "model_4.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "M0H2jaKL3rbY",
        "outputId": "5729386c-6a89-48dd-a1f8-ab055351b154"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_4_token_and_char_embeddings\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                Output Shape                 Param #   Connected to                  \n",
            "==================================================================================================\n",
            " char_input (InputLayer)     [(None, 1)]                  0         []                            \n",
            "                                                                                                  \n",
            " token_input (InputLayer)    [(None,)]                    0         []                            \n",
            "                                                                                                  \n",
            " char_vectorizer (TextVecto  (None, 290)                  0         ['char_input[0][0]']          \n",
            " rization)                                                                                        \n",
            "                                                                                                  \n",
            " universal_sentence_encoder  (None, 512)                  2567978   ['token_input[0][0]']         \n",
            "  (KerasLayer)                                            24                                      \n",
            "                                                                                                  \n",
            " char_embed (Embedding)      (None, 290, 25)              700       ['char_vectorizer[1][0]']     \n",
            "                                                                                                  \n",
            " dense_4 (Dense)             (None, 128)                  65664     ['universal_sentence_encoder[1\n",
            "                                                                    ][0]']                        \n",
            "                                                                                                  \n",
            " bidirectional (Bidirection  (None, 48)                   9600      ['char_embed[1][0]']          \n",
            " al)                                                                                              \n",
            "                                                                                                  \n",
            " token_char_hybrid (Concate  (None, 176)                  0         ['dense_4[0][0]',             \n",
            " nate)                                                               'bidirectional[0][0]']       \n",
            "                                                                                                  \n",
            " dropout (Dropout)           (None, 176)                  0         ['token_char_hybrid[0][0]']   \n",
            "                                                                                                  \n",
            " dense_5 (Dense)             (None, 128)                  22656     ['dropout[0][0]']             \n",
            "                                                                                                  \n",
            " dropout_1 (Dropout)         (None, 128)                  0         ['dense_5[0][0]']             \n",
            "                                                                                                  \n",
            " dense_6 (Dense)             (None, 5)                    645       ['dropout_1[0][0]']           \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 256897089 (979.98 MB)\n",
            "Trainable params: 99265 (387.75 KB)\n",
            "Non-trainable params: 256797824 (979.61 MB)\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Plot hybrid token an character model\n",
        "from keras.utils import plot_model\n",
        "plot_model(model_4, show_shapes=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 976
        },
        "id": "Q_jW5ZuqXqrL",
        "outputId": "8cb73cb9-d28f-4e15-96ee-14690736ed2a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 81
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Compile token char model\n",
        "model_4.compile(loss=\"categorical_crossentropy\",\n",
        "                optimizer=tf.keras.optimizers.Adam(), #section 4.2 in the paper used SGD, but we're sticking with adam\n",
        "                metrics=[\"accuracy\"])"
      ],
      "metadata": {
        "id": "avFvxePyYCW1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Combining token and character into a tf.data dataset"
      ],
      "metadata": {
        "id": "AuDctBf5ZaA-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Combine chars and tokens into a dataset\n",
        "train_char_token_data = tf.data.Dataset.from_tensor_slices((train_sentences, train_chars)) # make data\n",
        "train_char_token_labels = tf.data.Dataset.from_tensor_slices(train_labels_one_hot) # make labels\n",
        "train_char_token_dataset = tf.data.Dataset.zip((train_char_token_data, train_char_token_labels)) # combine data and labels\n",
        "\n",
        "# Prefetch and batch train data\n",
        "train_char_token_dataset = train_char_token_dataset.batch(32).prefetch(tf.data.AUTOTUNE)"
      ],
      "metadata": {
        "id": "TU_6zo9fbwQ3"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Repeat the above steps for our validation data\n",
        "val_char_token_data = tf.data.Dataset.from_tensor_slices((val_sentences, val_chars))\n",
        "val_char_token_labels = tf.data.Dataset.from_tensor_slices(val_labels_one_hot)\n",
        "val_char_token_dataset = tf.data.Dataset.zip((val_char_token_data, val_char_token_labels))\n",
        "\n",
        "# Prefetch and batch validation data\n",
        "val_char_token_dataset = val_char_token_dataset.batch(32).prefetch(tf.data.AUTOTUNE)"
      ],
      "metadata": {
        "id": "WGAs3wmGc44x"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Check out our training char and token embedding dataset\n",
        "train_char_token_dataset, val_char_token_dataset"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_vceB7SmdbbA",
        "outputId": "41a2a496-cd88-4a6c-8ac6-448366bca64a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(<_PrefetchDataset element_spec=((TensorSpec(shape=(None,), dtype=tf.string, name=None), TensorSpec(shape=(None,), dtype=tf.string, name=None)), TensorSpec(shape=(None, 5), dtype=tf.float64, name=None))>,\n",
              " <_PrefetchDataset element_spec=((TensorSpec(shape=(None,), dtype=tf.string, name=None), TensorSpec(shape=(None,), dtype=tf.string, name=None)), TensorSpec(shape=(None, 5), dtype=tf.float64, name=None))>)"
            ]
          },
          "metadata": {},
          "execution_count": 85
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Fitting a model on token and character-level sequences"
      ],
      "metadata": {
        "id": "lrFjkLvAdnQo"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Fit the model on tokens and chars\n",
        "history_model_4 = model_4.fit(train_char_token_dataset,\n",
        "                              steps_per_epoch = int(0.1 * len(train_char_token_dataset)),\n",
        "                              epochs=3,\n",
        "                              validation_data=val_char_token_dataset,\n",
        "                              validation_steps = int(0.1 * len(val_char_token_dataset)))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KkeJQ6LZeJjC",
        "outputId": "7c1448f4-0dca-45c2-9b4d-b7572704d33b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/3\n",
            "562/562 [==============================] - 42s 53ms/step - loss: 0.9846 - accuracy: 0.6116 - val_loss: 0.7689 - val_accuracy: 0.6998\n",
            "Epoch 2/3\n",
            "562/562 [==============================] - 26s 47ms/step - loss: 0.8009 - accuracy: 0.6921 - val_loss: 0.7071 - val_accuracy: 0.7317\n",
            "Epoch 3/3\n",
            "562/562 [==============================] - 24s 43ms/step - loss: 0.7734 - accuracy: 0.6993 - val_loss: 0.6897 - val_accuracy: 0.7477\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate on the whole validation dataset\n",
        "model_4.evaluate(val_char_token_dataset)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "egtDLcU1epLm",
        "outputId": "7e00c5b0-8557-4405-8f9a-5e6343d8ceb2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "945/945 [==============================] - 20s 22ms/step - loss: 0.6986 - accuracy: 0.7361\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.698586106300354, 0.7360982298851013]"
            ]
          },
          "metadata": {},
          "execution_count": 87
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Make predictions using the token-character model hybrid\n",
        "model_4_pred_probs = model_4.predict(val_char_token_dataset)\n",
        "model_4_pred_probs"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xozzKglhfN9g",
        "outputId": "40302eaf-480a-40f0-a896-ed89d274c9f1"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "945/945 [==============================] - 21s 19ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[3.9955631e-01, 3.8651645e-01, 3.4716188e-03, 2.0379913e-01,\n",
              "        6.6564591e-03],\n",
              "       [3.6199307e-01, 3.7493423e-01, 1.5884436e-03, 2.5974387e-01,\n",
              "        1.7403899e-03],\n",
              "       [3.1491762e-01, 1.4328654e-01, 6.0326759e-02, 4.4818357e-01,\n",
              "        3.3285566e-02],\n",
              "       ...,\n",
              "       [7.6489267e-04, 8.8162944e-03, 7.6739408e-02, 5.2728789e-04,\n",
              "        9.1315210e-01],\n",
              "       [6.3268822e-03, 4.8787631e-02, 2.2795756e-01, 4.4388003e-03,\n",
              "        7.1248919e-01],\n",
              "       [3.3232075e-01, 4.1615418e-01, 1.5646978e-01, 4.5230616e-02,\n",
              "        4.9824689e-02]], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 88
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Format pred_probs into pred_labels\n",
        "model_4_preds = tf.argmax(model_4_pred_probs, axis=1)\n",
        "model_4_preds"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BTZGWUJJfyge",
        "outputId": "9face10d-f01d-4bda-c654-48400af8e343"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(30212,), dtype=int64, numpy=array([0, 1, 3, ..., 4, 4, 1])>"
            ]
          },
          "metadata": {},
          "execution_count": 89
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Get the results of token-char-hybrid model\n",
        "model_4_results = calculate_results(y_true=val_labels_encoded,\n",
        "                                    y_pred=model_4_preds)\n",
        "model_4_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lEZEE8g-gB82",
        "outputId": "e559aef1-579c-4a26-ef32-3f96fffe20a1"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'accuracy': 73.60982391102873,\n",
              " 'precision': 0.7371389748493671,\n",
              " 'recall': 0.7360982391102873,\n",
              " 'f1': 0.7338690547268851}"
            ]
          },
          "metadata": {},
          "execution_count": 90
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model 5 : Transfer Learning with pretrained token embeddings + character embeddings + positional embeddings"
      ],
      "metadata": {
        "id": "LHu2_BXGgKrT"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_df.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "CK_wnSzulGS4",
        "outputId": "2a938dc2-971f-431a-8c8a-4c53e39048cf"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "      target                                               text  line_number  \\\n",
              "0  OBJECTIVE  to investigate the efficacy of @ weeks of dail...            0   \n",
              "1    METHODS  a total of @ patients with primary knee oa wer...            1   \n",
              "2    METHODS  outcome measures included pain reduction and i...            2   \n",
              "3    METHODS  pain was assessed using the visual analog pain...            3   \n",
              "4    METHODS  secondary outcome measures included the wester...            4   \n",
              "\n",
              "   total_lines  \n",
              "0           12  \n",
              "1           12  \n",
              "2           12  \n",
              "3           12  \n",
              "4           12  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-9bc0d79c-7379-4cb4-99d4-e716071e8892\" 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>target</th>\n",
              "      <th>text</th>\n",
              "      <th>line_number</th>\n",
              "      <th>total_lines</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>OBJECTIVE</td>\n",
              "      <td>to investigate the efficacy of @ weeks of dail...</td>\n",
              "      <td>0</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>METHODS</td>\n",
              "      <td>a total of @ patients with primary knee oa wer...</td>\n",
              "      <td>1</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>METHODS</td>\n",
              "      <td>outcome measures included pain reduction and i...</td>\n",
              "      <td>2</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>METHODS</td>\n",
              "      <td>pain was assessed using the visual analog pain...</td>\n",
              "      <td>3</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>METHODS</td>\n",
              "      <td>secondary outcome measures included the wester...</td>\n",
              "      <td>4</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-9bc0d79c-7379-4cb4-99d4-e716071e8892')\"\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 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\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",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\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-9bc0d79c-7379-4cb4-99d4-e716071e8892 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-9bc0d79c-7379-4cb4-99d4-e716071e8892');\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",
              "\n",
              "\n",
              "<div id=\"df-5e60f2df-d63c-47fc-afd1-a87c7b487cf1\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-5e60f2df-d63c-47fc-afd1-a87c7b487cf1')\"\n",
              "            title=\"Suggest charts\"\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",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-5e60f2df-d63c-47fc-afd1-a87c7b487cf1 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "train_df"
            }
          },
          "metadata": {},
          "execution_count": 91
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        ">🔑**Note:** Any engineered features used to train a model need to be available at test time. In our case, line numbers and total lines are available."
      ],
      "metadata": {
        "id": "QKxlLNjXjIw2"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Create positional embeddings"
      ],
      "metadata": {
        "id": "mDnM_N8rkuiR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# How many different line numbers are there ?\n",
        "train_df[\"line_number\"].value_counts()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "qqcTM5l9k5GZ",
        "outputId": "07686985-25b4-4834-b4e3-44b9b2eaf564"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "line_number\n",
              "0     15000\n",
              "1     15000\n",
              "2     15000\n",
              "3     15000\n",
              "4     14992\n",
              "5     14949\n",
              "6     14758\n",
              "7     14279\n",
              "8     13346\n",
              "9     11981\n",
              "10    10041\n",
              "11     7892\n",
              "12     5853\n",
              "13     4152\n",
              "14     2835\n",
              "15     1861\n",
              "16     1188\n",
              "17      751\n",
              "18      462\n",
              "19      286\n",
              "20      162\n",
              "21      101\n",
              "22       66\n",
              "23       33\n",
              "24       22\n",
              "25       14\n",
              "26        7\n",
              "27        4\n",
              "28        3\n",
              "29        1\n",
              "30        1\n",
              "Name: count, dtype: int64"
            ],
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>line_number</th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>15000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>15000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>15000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>15000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>14992</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>14949</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>14758</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>14279</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>13346</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>11981</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>10041</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>7892</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>5853</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>4152</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>2835</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>1861</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>1188</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>751</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>462</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>286</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>162</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>21</th>\n",
              "      <td>101</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>22</th>\n",
              "      <td>66</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>23</th>\n",
              "      <td>33</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>24</th>\n",
              "      <td>22</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25</th>\n",
              "      <td>14</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>26</th>\n",
              "      <td>7</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>27</th>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>28</th>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29</th>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>30</th>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div><br><label><b>dtype:</b> int64</label>"
            ]
          },
          "metadata": {},
          "execution_count": 92
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Check the distribution of \"line_number\" column\n",
        "train_df.line_number.plot.hist()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 447
        },
        "id": "7YvGkzdck_xl",
        "outputId": "8a7d52ac-7e73-47cd-ced9-c817e917d1ee"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<Axes: ylabel='Frequency'>"
            ]
          },
          "metadata": {},
          "execution_count": 93
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Use TensorFlow to create one-hot encoded tensors of our \"line_number\" column\n",
        "train_line_number_one_hot = tf.one_hot(train_df[\"line_number\"].to_numpy(), depth=15)\n",
        "val_line_number_one_hot = tf.one_hot(val_df[\"line_number\"].to_numpy(), depth=15)\n",
        "test_line_number_one_hot = tf.one_hot(test_df[\"line_number\"].to_numpy(), depth=15)\n",
        "train_line_number_one_hot[:10], train_line_number_one_hot.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nSWBrsQellWz",
        "outputId": "0173409d-7c34-4b4a-ba60-5b50efeabb24"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(<tf.Tensor: shape=(10, 15), dtype=float32, numpy=\n",
              " array([[1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0.]],\n",
              "       dtype=float32)>,\n",
              " TensorShape([180040, 15]))"
            ]
          },
          "metadata": {},
          "execution_count": 94
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now we've encoded our line_number feature , let's do the same for our total_lines feature..."
      ],
      "metadata": {
        "id": "CPTJTfIamam0"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# How many different numbers of lines are there ?\n",
        "train_df[\"total_lines\"].value_counts()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 962
        },
        "id": "L5-zEV5FnjOI",
        "outputId": "d0209d4a-0e13-4cea-8b04-62bf6e425630"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "total_lines\n",
              "12    24468\n",
              "11    23639\n",
              "13    22113\n",
              "10    19400\n",
              "14    18438\n",
              "15    14610\n",
              "9     12285\n",
              "16    10768\n",
              "8      7464\n",
              "17     7429\n",
              "18     5202\n",
              "7      3353\n",
              "19     3344\n",
              "20     2480\n",
              "21     1281\n",
              "6      1146\n",
              "22      770\n",
              "23      759\n",
              "24      264\n",
              "5       215\n",
              "25      200\n",
              "26      182\n",
              "27       81\n",
              "29       58\n",
              "4        32\n",
              "31       31\n",
              "28       28\n",
              "Name: count, dtype: int64"
            ],
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>count</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>total_lines</th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>24468</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>23639</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>22113</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>19400</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>18438</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>14610</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>12285</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>10768</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>7464</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>7429</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>5202</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>3353</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>3344</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>2480</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>21</th>\n",
              "      <td>1281</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>1146</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>22</th>\n",
              "      <td>770</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>23</th>\n",
              "      <td>759</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>24</th>\n",
              "      <td>264</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>215</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25</th>\n",
              "      <td>200</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>26</th>\n",
              "      <td>182</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>27</th>\n",
              "      <td>81</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29</th>\n",
              "      <td>58</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>32</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>31</th>\n",
              "      <td>31</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>28</th>\n",
              "      <td>28</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div><br><label><b>dtype:</b> int64</label>"
            ]
          },
          "metadata": {},
          "execution_count": 95
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Check the distributions of total lines\n",
        "train_df.total_lines.plot.hist()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 448
        },
        "id": "vTPaCCV2nr-k",
        "outputId": "81997a57-74d5-4f70-fc45-e4847b78b4fd"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<Axes: ylabel='Frequency'>"
            ]
          },
          "metadata": {},
          "execution_count": 96
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Check the coverage of a \"total_lines\" value of 20\n",
        "np.percentile(train_df.total_lines, 97)"
      ],
      "metadata": {
        "id": "cqmJEeJToqHD",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c0edd097-3001-4cfe-c629-13b25397e05c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "20.0"
            ]
          },
          "metadata": {},
          "execution_count": 97
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Use TensorFlow to create one-hot-encoded tensors of our \"total_lines\" feature\n",
        "train_total_lines_one_hot = tf.one_hot(train_df[\"total_lines\"].to_numpy(), depth=20)\n",
        "val_total_lines_one_hot = tf.one_hot(val_df[\"total_lines\"].to_numpy(), depth=20)\n",
        "test_total_lines_one_hot = tf.one_hot(test_df[\"total_lines\"].to_numpy(), depth=20)\n",
        "train_total_lines_one_hot[:10], train_total_lines_one_hot.shape"
      ],
      "metadata": {
        "id": "81Mjr0IOpFjK",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "39ab78d1-6362-434d-a9ec-1fadc8b91a00"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(<tf.Tensor: shape=(10, 20), dtype=float32, numpy=\n",
              " array([[0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0.,\n",
              "         0., 0., 0., 0.]], dtype=float32)>,\n",
              " TensorShape([180040, 20]))"
            ]
          },
          "metadata": {},
          "execution_count": 98
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Building a tribrid embedding model\n",
        "\n",
        "1. Create a token-level model\n",
        "2. Create a character-level model\n",
        "3. Create a model for the \"line_number\" feature\n",
        "4. Create a model for the \"total_lines\" feature\n",
        "5. Combine the outputs of 1 & 2 using tf.keras.layers.Concatenate\n",
        "6. Combine the outputs of 3, 4 & 5 using tf.keras.layers.Concatenate\n",
        "7. Create an output layer to accept the tribrid embedding and output label probabilities\n",
        "8. Combine the inputs for 1, 2, 3, 4, and outputs of 7 into a tf.keras.Model"
      ],
      "metadata": {
        "id": "1m1IfgYGp6HT"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 1. Token inputs\n",
        "token_inputs = layers.Input(shape=[], dtype=tf.string, name=\"token_input\")\n",
        "token_embeddings = tf_hub_embedding_layer(token_inputs)\n",
        "token_output = layers.Dense(128, activation=\"relu\")(token_embeddings)\n",
        "token_model = tf.keras.Model(inputs=token_inputs,\n",
        "                             outputs=token_output)\n",
        "\n",
        "# 2. Char inputs\n",
        "char_inputs = layers.Input(shape=(1,), dtype=tf.string, name=\"char_input\")\n",
        "char_vectors = char_vectorizer(char_inputs)\n",
        "char_embeddings = char_embed(char_vectors)\n",
        "char_bi_lstm = layers.Bidirectional(layers.LSTM(24))(char_embeddings)\n",
        "char_model = tf.keras.Model(inputs=char_inputs,\n",
        "                            outputs=char_bi_lstm)\n",
        "\n",
        "# 3. Line numbers model\n",
        "# inputs & shape\n",
        "line_number_inputs = layers.Input(shape=(15,), dtype=tf.float32, name=\"line_number_input\")\n",
        "# dense layer with 32 units and relu activation\n",
        "x = layers.Dense(32, activation=\"relu\")(line_number_inputs)\n",
        "# combine inputs and dense layer into model\n",
        "line_number_model = tf.keras.Model(inputs=line_number_inputs,\n",
        "                                   outputs=x)\n",
        "\n",
        "# 4. Total lines model\n",
        "# inputs & shape\n",
        "total_lines_inputs = layers.Input(shape=(20,), dtype=tf.float32, name=\"total_lines_input\")\n",
        "# dense layer with 32 units and relu activation\n",
        "y = layers.Dense(32, activation=\"relu\")(total_lines_inputs)\n",
        "# combine inputs and dense layer into model\n",
        "total_lines_model = tf.keras.Model(inputs=total_lines_inputs,\n",
        "                                   outputs=y)\n",
        "\n",
        "# 5. Combine token and char embeddings into a hybrid embedding\n",
        "combined_embeddings = layers.Concatenate(name=\"combined_embeddings\")([token_model.output,\n",
        "                                                                      char_model.output])\n",
        "\n",
        "z = layers.Dense(256, activation=\"relu\")(combined_embeddings)\n",
        "z = layers.Dropout(0.5)(z)\n",
        "\n",
        "# 6. Combine positional with combined token and char embeddings\n",
        "tribrid_embeddings = layers.Concatenate(name=\"combined_positional\")([line_number_model.output,\n",
        "                                                                     total_lines_model.output,\n",
        "                                                                     z])\n",
        "\n",
        "# 7. Create output layer\n",
        "output_layer = layers.Dense(5, activation=\"softmax\", name=\"output_layer\")(tribrid_embeddings)\n",
        "\n",
        "# 8. Put together model with all kind of inputs\n",
        "model_5 = tf.keras.Model(inputs=[token_model.input,\n",
        "                                 char_model.input,\n",
        "                                 line_number_model.input,\n",
        "                                 total_lines_model.input],\n",
        "                         outputs=output_layer)"
      ],
      "metadata": {
        "id": "q_qLt2b5rUG_"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Get a summary of our tribrid embedding model\n",
        "model_5.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hNTA_GnhZf_l",
        "outputId": "406dbb42-5596-4d16-eeed-cea8f7454b4a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_6\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                Output Shape                 Param #   Connected to                  \n",
            "==================================================================================================\n",
            " char_input (InputLayer)     [(None, 1)]                  0         []                            \n",
            "                                                                                                  \n",
            " token_input (InputLayer)    [(None,)]                    0         []                            \n",
            "                                                                                                  \n",
            " char_vectorizer (TextVecto  (None, 290)                  0         ['char_input[0][0]']          \n",
            " rization)                                                                                        \n",
            "                                                                                                  \n",
            " universal_sentence_encoder  (None, 512)                  2567978   ['token_input[0][0]']         \n",
            "  (KerasLayer)                                            24                                      \n",
            "                                                                                                  \n",
            " char_embed (Embedding)      (None, 290, 25)              700       ['char_vectorizer[2][0]']     \n",
            "                                                                                                  \n",
            " dense_7 (Dense)             (None, 128)                  65664     ['universal_sentence_encoder[2\n",
            "                                                                    ][0]']                        \n",
            "                                                                                                  \n",
            " bidirectional_1 (Bidirecti  (None, 48)                   9600      ['char_embed[2][0]']          \n",
            " onal)                                                                                            \n",
            "                                                                                                  \n",
            " combined_embeddings (Conca  (None, 176)                  0         ['dense_7[0][0]',             \n",
            " tenate)                                                             'bidirectional_1[0][0]']     \n",
            "                                                                                                  \n",
            " line_number_input (InputLa  [(None, 15)]                 0         []                            \n",
            " yer)                                                                                             \n",
            "                                                                                                  \n",
            " total_lines_input (InputLa  [(None, 20)]                 0         []                            \n",
            " yer)                                                                                             \n",
            "                                                                                                  \n",
            " dense_10 (Dense)            (None, 256)                  45312     ['combined_embeddings[0][0]'] \n",
            "                                                                                                  \n",
            " dense_8 (Dense)             (None, 32)                   512       ['line_number_input[0][0]']   \n",
            "                                                                                                  \n",
            " dense_9 (Dense)             (None, 32)                   672       ['total_lines_input[0][0]']   \n",
            "                                                                                                  \n",
            " dropout_2 (Dropout)         (None, 256)                  0         ['dense_10[0][0]']            \n",
            "                                                                                                  \n",
            " combined_positional (Conca  (None, 320)                  0         ['dense_8[0][0]',             \n",
            " tenate)                                                             'dense_9[0][0]',             \n",
            "                                                                     'dropout_2[0][0]']           \n",
            "                                                                                                  \n",
            " output_layer (Dense)        (None, 5)                    1605      ['combined_positional[0][0]'] \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 256921889 (980.08 MB)\n",
            "Trainable params: 124065 (484.63 KB)\n",
            "Non-trainable params: 256797824 (979.61 MB)\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Plot model_5 to explore it visually...\n",
        "from keras.utils import plot_model\n",
        "plot_model(model_5, show_shapes=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 835
        },
        "id": "SYG8buAEjiNj",
        "outputId": "fc4d6623-90f0-419c-b604-9081ea5851eb"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 101
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "What is label smoothing ?\n",
        "\n",
        "For example , if our model gets too confident on a single class (e.g. its prediction probability is really high), it may get stuck on that class and not consider other classes...\n",
        "\n",
        "Really confident: `[0.0, 0.0, 1.0, 0.0, 0.0]`\n",
        "\n",
        "What label smoothing does is it assigns some of the value from the highest pred prob to other classes, in turn, hopefully improving generalization: `[0.01, 0.01, 0.96, 0.01, 0.01]`\n",
        "\n",
        ">📖**Resource:** For more on label smoothing , see the blog post from PyImageSearch"
      ],
      "metadata": {
        "id": "XjYsWGhylvKC"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Compile token, char and positional embedding model\n",
        "model_5.compile(loss=tf.keras.losses.CategoricalCrossentropy(label_smoothing=0.2),\n",
        "                optimizer=\"Adam\",\n",
        "                metrics=[\"accuracy\"])"
      ],
      "metadata": {
        "id": "DirA8DUhj2Ij"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Create tribrid embedding datasets using tf.data\n"
      ],
      "metadata": {
        "id": "TnE4uh7jnnS6"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Create training and validation datasets (with all four kinds of input data)\n",
        "train_char_token_pos_data = tf.data.Dataset.from_tensor_slices((train_sentences, train_chars, train_line_number_one_hot, train_total_lines_one_hot))\n",
        "train_char_token_pos_labels = tf.data.Dataset.from_tensor_slices(train_labels_one_hot)\n",
        "train_char_token_pos_dataset = tf.data.Dataset.zip((train_char_token_pos_data, train_char_token_pos_labels))\n",
        "train_char_token_pos_dataset = train_char_token_pos_dataset.batch(32).prefetch(tf.data.AUTOTUNE)\n",
        "\n",
        "# Repeat for validation data\n",
        "val_char_token_pos_data = tf.data.Dataset.from_tensor_slices((val_sentences, val_chars, val_line_number_one_hot, val_total_lines_one_hot))\n",
        "val_char_token_pos_labels = tf.data.Dataset.from_tensor_slices(val_labels_one_hot)\n",
        "val_char_token_pos_dataset = tf.data.Dataset.zip((val_char_token_pos_data, val_char_token_pos_labels))\n",
        "val_char_token_pos_dataset = val_char_token_pos_dataset.batch(32).prefetch(tf.data.AUTOTUNE)"
      ],
      "metadata": {
        "id": "tBXa8yqyn6gK"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Check input shapes\n",
        "train_char_token_pos_dataset, val_char_token_pos_dataset"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Zi9fFPeAsA_w",
        "outputId": "205581e0-4fda-4aa4-d062-f62f4addf71c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(<_PrefetchDataset element_spec=((TensorSpec(shape=(None,), dtype=tf.string, name=None), TensorSpec(shape=(None,), dtype=tf.string, name=None), TensorSpec(shape=(None, 15), dtype=tf.float32, name=None), TensorSpec(shape=(None, 20), dtype=tf.float32, name=None)), TensorSpec(shape=(None, 5), dtype=tf.float64, name=None))>,\n",
              " <_PrefetchDataset element_spec=((TensorSpec(shape=(None,), dtype=tf.string, name=None), TensorSpec(shape=(None,), dtype=tf.string, name=None), TensorSpec(shape=(None, 15), dtype=tf.float32, name=None), TensorSpec(shape=(None, 20), dtype=tf.float32, name=None)), TensorSpec(shape=(None, 5), dtype=tf.float64, name=None))>)"
            ]
          },
          "metadata": {},
          "execution_count": 104
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Fitting, evaluating and making predictions with our tribrid model"
      ],
      "metadata": {
        "id": "ea9qKx4CsShY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Fit our tribrid embedding model\n",
        "history_model_5 = model_5.fit(train_char_token_pos_dataset,\n",
        "                              steps_per_epoch = int(0.1 * len(train_char_token_pos_dataset)),\n",
        "                              epochs=3,\n",
        "                              validation_data=val_char_token_pos_dataset,\n",
        "                              validation_steps = int(0.1 * len(val_char_token_pos_dataset)))"
      ],
      "metadata": {
        "id": "BhM24uQhsr6_",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2b2ee425-1dc7-42e7-ccc9-956a3d53ea86"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/3\n",
            "562/562 [==============================] - 41s 56ms/step - loss: 1.0917 - accuracy: 0.7258 - val_loss: 0.9782 - val_accuracy: 0.8019\n",
            "Epoch 2/3\n",
            "562/562 [==============================] - 28s 50ms/step - loss: 0.9667 - accuracy: 0.8153 - val_loss: 0.9490 - val_accuracy: 0.8271\n",
            "Epoch 3/3\n",
            "562/562 [==============================] - 25s 45ms/step - loss: 0.9505 - accuracy: 0.8218 - val_loss: 0.9386 - val_accuracy: 0.8328\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Make predictions with the char_token_pos model\n",
        "model_5_pred_probs = model_5.predict(val_char_token_pos_dataset, verbose=1)\n",
        "model_5_pred_probs"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LiVVZZXLtIT3",
        "outputId": "b46a5361-c460-43b0-b8be-7a8f29032e5e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "945/945 [==============================] - 22s 20ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.48825917, 0.11443482, 0.01278997, 0.3645956 , 0.01992048],\n",
              "       [0.5005966 , 0.10633911, 0.06062087, 0.3200213 , 0.0124221 ],\n",
              "       [0.2732376 , 0.12667266, 0.15233418, 0.3815087 , 0.06624694],\n",
              "       ...,\n",
              "       [0.0340605 , 0.12036058, 0.04545934, 0.02760639, 0.7725132 ],\n",
              "       [0.02834765, 0.3504329 , 0.08919345, 0.0269989 , 0.5050271 ],\n",
              "       [0.14714898, 0.60980463, 0.13387753, 0.03803249, 0.07113636]],\n",
              "      dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 106
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Convert pred probs to pred labels\n",
        "model_5_preds = tf.argmax(model_5_pred_probs, axis=1)\n",
        "model_5_preds"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7heix-qPtXyX",
        "outputId": "3243a2d9-5158-476c-8dd6-ea7f000b7aa3"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(30212,), dtype=int64, numpy=array([0, 0, 3, ..., 4, 4, 1])>"
            ]
          },
          "metadata": {},
          "execution_count": 107
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Calculate results of char token pos model\n",
        "model_5_results = calculate_results(y_true=val_labels_encoded,\n",
        "                                    y_pred=model_5_preds)\n",
        "model_5_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kNWv-ynVtcfW",
        "outputId": "316a9c5a-2a48-42a4-8e21-36b2a5f99a18"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'accuracy': 83.4072553952072,\n",
              " 'precision': 0.8329354601792406,\n",
              " 'recall': 0.834072553952072,\n",
              " 'f1': 0.8326912166954082}"
            ]
          },
          "metadata": {},
          "execution_count": 108
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Compare model results"
      ],
      "metadata": {
        "id": "z0bC45wDtnAV"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Combine model results into a DataFrame\n",
        "all_model_results = pd.DataFrame({\"model_0_baseline\": baseline_results,\n",
        "                                   \"model_1_custom_token_embedding\": model_1_results,\n",
        "                                   \"model_2_pretrained_token_embedding\": model_2_results,\n",
        "                                   \"model_3_custom_char_embedding\": model_3_results,\n",
        "                                   \"model_4_hybrid_char_token_embedding\": model_4_results,\n",
        "                                   \"model_5_pos_char_token_embedding\": model_5_results})\n",
        "all_model_results = all_model_results.transpose()\n",
        "all_model_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "id": "Ij8d0GafvG45",
        "outputId": "9f29b876-1901-418d-9423-25e0a3c9f611"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                      accuracy  precision    recall        f1\n",
              "model_0_baseline                     72.183238   0.718647  0.721832  0.698925\n",
              "model_1_custom_token_embedding       78.485370   0.781303  0.784854  0.782185\n",
              "model_2_pretrained_token_embedding   71.415332   0.714844  0.714153  0.711187\n",
              "model_3_custom_char_embedding        66.063154   0.654458  0.660632  0.652181\n",
              "model_4_hybrid_char_token_embedding  73.609824   0.737139  0.736098  0.733869\n",
              "model_5_pos_char_token_embedding     83.407255   0.832935  0.834073  0.832691"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-a51ba0aa-ec1b-40d9-867e-c4477b6b03d5\" 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>accuracy</th>\n",
              "      <th>precision</th>\n",
              "      <th>recall</th>\n",
              "      <th>f1</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>model_0_baseline</th>\n",
              "      <td>72.183238</td>\n",
              "      <td>0.718647</td>\n",
              "      <td>0.721832</td>\n",
              "      <td>0.698925</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>model_1_custom_token_embedding</th>\n",
              "      <td>78.485370</td>\n",
              "      <td>0.781303</td>\n",
              "      <td>0.784854</td>\n",
              "      <td>0.782185</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>model_2_pretrained_token_embedding</th>\n",
              "      <td>71.415332</td>\n",
              "      <td>0.714844</td>\n",
              "      <td>0.714153</td>\n",
              "      <td>0.711187</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>model_3_custom_char_embedding</th>\n",
              "      <td>66.063154</td>\n",
              "      <td>0.654458</td>\n",
              "      <td>0.660632</td>\n",
              "      <td>0.652181</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>model_4_hybrid_char_token_embedding</th>\n",
              "      <td>73.609824</td>\n",
              "      <td>0.737139</td>\n",
              "      <td>0.736098</td>\n",
              "      <td>0.733869</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>model_5_pos_char_token_embedding</th>\n",
              "      <td>83.407255</td>\n",
              "      <td>0.832935</td>\n",
              "      <td>0.834073</td>\n",
              "      <td>0.832691</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-a51ba0aa-ec1b-40d9-867e-c4477b6b03d5')\"\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 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\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",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\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-a51ba0aa-ec1b-40d9-867e-c4477b6b03d5 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-a51ba0aa-ec1b-40d9-867e-c4477b6b03d5');\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",
              "\n",
              "\n",
              "<div id=\"df-1878cd62-8d54-4341-a8e3-d5d402dc1813\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-1878cd62-8d54-4341-a8e3-d5d402dc1813')\"\n",
              "            title=\"Suggest charts\"\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",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-1878cd62-8d54-4341-a8e3-d5d402dc1813 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "  <div id=\"id_09fbbc8e-c99a-4885-8439-d0c1436dc8d4\">\n",
              "    <style>\n",
              "      .colab-df-generate {\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-generate: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-generate {\n",
              "        background-color: #3B4455;\n",
              "        fill: #D2E3FC;\n",
              "      }\n",
              "\n",
              "      [theme=dark] .colab-df-generate: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",
              "    <button class=\"colab-df-generate\" onclick=\"generateWithVariable('all_model_results')\"\n",
              "            title=\"Generate code using this dataframe.\"\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=\"M7,19H8.4L18.45,9,17,7.55,7,17.6ZM5,21V16.75L18.45,3.32a2,2,0,0,1,2.83,0l1.4,1.43a1.91,1.91,0,0,1,.58,1.4,1.91,1.91,0,0,1-.58,1.4L9.25,21ZM18.45,9,17,7.55Zm-12,3A5.31,5.31,0,0,0,4.9,8.1,5.31,5.31,0,0,0,1,6.5,5.31,5.31,0,0,0,4.9,4.9,5.31,5.31,0,0,0,6.5,1,5.31,5.31,0,0,0,8.1,4.9,5.31,5.31,0,0,0,12,6.5,5.46,5.46,0,0,0,6.5,12Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "    <script>\n",
              "      (() => {\n",
              "      const buttonEl =\n",
              "        document.querySelector('#id_09fbbc8e-c99a-4885-8439-d0c1436dc8d4 button.colab-df-generate');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      buttonEl.onclick = () => {\n",
              "        google.colab.notebook.generateWithVariable('all_model_results');\n",
              "      }\n",
              "      })();\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "all_model_results",
              "summary": "{\n  \"name\": \"all_model_results\",\n  \"rows\": 6,\n  \"fields\": [\n    {\n      \"column\": \"accuracy\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 6.02521393326319,\n        \"min\": 66.06315371375612,\n        \"max\": 83.4072553952072,\n        \"num_unique_values\": 6,\n        \"samples\": [\n          72.1832384482987,\n          78.48537005163512,\n          83.4072553952072\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"precision\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.06124913638143175,\n        \"min\": 0.6544577397132115,\n        \"max\": 0.8329354601792406,\n        \"num_unique_values\": 6,\n        \"samples\": [\n          0.7186466952323352,\n          0.7813028708243923,\n          0.8329354601792406\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"recall\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.060252139332631884,\n        \"min\": 0.6606315371375613,\n        \"max\": 0.834072553952072,\n        \"num_unique_values\": 6,\n        \"samples\": [\n          0.7218323844829869,\n          0.7848537005163512,\n          0.834072553952072\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"f1\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0.0640290440978558,\n        \"min\": 0.6521810036635403,\n        \"max\": 0.8326912166954082,\n        \"num_unique_values\": 6,\n        \"samples\": [\n          0.6989250353450294,\n          0.7821852447612673,\n          0.8326912166954082\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 109
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Reduce the accuracy to same scale as other metrics\n",
        "all_model_results[\"accuracy\"] = all_model_results[\"accuracy\"]/100"
      ],
      "metadata": {
        "id": "N1NYNR-evqGW"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Plot and compare all model results\n",
        "all_model_results.plot(kind=\"bar\", figsize=(10,7)).legend(bbox_to_anchor=(1.0, 1.0))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 882
        },
        "id": "JizzRAO7wHgj",
        "outputId": "2d12b265-18b8-4ab6-c72f-b8e9323530dd"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7b90bc5f8b80>"
            ]
          },
          "metadata": {},
          "execution_count": 111
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x700 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Sort the models results according to f1-score\n",
        "all_model_results.sort_values(\"f1\", ascending=True)[\"f1\"].plot(kind=\"bar\", figsize=(10,7))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 882
        },
        "id": "feT4j9FzwYZd",
        "outputId": "e8c5af96-ad3b-4707-a1a3-46e58f5d34bb"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<Axes: >"
            ]
          },
          "metadata": {},
          "execution_count": 112
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x700 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Save and load model"
      ],
      "metadata": {
        "id": "-VpXMc5gw1ou"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Save the best performing model to the SavedModel format (default)\n",
        "model_5.save(\"SkimLit_tribrid_embedding_model\")"
      ],
      "metadata": {
        "id": "t1vBYLy5xKm6"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Load in best performing model\n",
        "loaded_model = tf.keras.models.load_model(\"SkimLit_tribrid_embedding_model\")"
      ],
      "metadata": {
        "id": "z8qXFR73x-XN"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Make predictions with loaded model on the validation set\n",
        "loaded_model_pred_probs = loaded_model.predict(val_char_token_pos_dataset)\n",
        "loaded_preds = tf.argmax(loaded_model_pred_probs, axis=1)\n",
        "loaded_preds[:10]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xpJOqcMjyQvV",
        "outputId": "36198260-c136-4f48-e9ca-7341afb4777f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "945/945 [==============================] - 23s 21ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(10,), dtype=int64, numpy=array([0, 0, 3, 2, 2, 4, 4, 4, 4, 1])>"
            ]
          },
          "metadata": {},
          "execution_count": 115
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Calculate the results of our loaded model\n",
        "loaded_model_results = calculate_results(y_true=val_labels_encoded,\n",
        "                                         y_pred=loaded_preds)\n",
        "loaded_model_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "m3sDqU3NynHB",
        "outputId": "f460ec89-1694-43d4-d4ba-3ea4ae17b5d7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'accuracy': 83.4072553952072,\n",
              " 'precision': 0.8329354601792406,\n",
              " 'recall': 0.834072553952072,\n",
              " 'f1': 0.8326912166954082}"
            ]
          },
          "metadata": {},
          "execution_count": 116
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model_5_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wqKoxdm5yWxo",
        "outputId": "c2676359-5351-4adf-d4b2-039f658825b4"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'accuracy': 83.4072553952072,\n",
              " 'precision': 0.8329354601792406,\n",
              " 'recall': 0.834072553952072,\n",
              " 'f1': 0.8326912166954082}"
            ]
          },
          "metadata": {},
          "execution_count": 117
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "assert model_5_results == loaded_model_results # Would give error if it wasn't true"
      ],
      "metadata": {
        "id": "1kcHM7CqyYkd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Checking the loaded model summary\n",
        "loaded_model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BdoI1Im2zDk-",
        "outputId": "97306444-6333-4a7d-980c-b077a7863360"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_6\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                Output Shape                 Param #   Connected to                  \n",
            "==================================================================================================\n",
            " char_input (InputLayer)     [(None, 1)]                  0         []                            \n",
            "                                                                                                  \n",
            " token_input (InputLayer)    [(None,)]                    0         []                            \n",
            "                                                                                                  \n",
            " char_vectorizer (TextVecto  (None, 290)                  0         ['char_input[0][0]']          \n",
            " rization)                                                                                        \n",
            "                                                                                                  \n",
            " universal_sentence_encoder  (None, 512)                  2567978   ['token_input[0][0]']         \n",
            "  (KerasLayer)                                            24                                      \n",
            "                                                                                                  \n",
            " char_embed (Embedding)      (None, 290, 25)              700       ['char_vectorizer[0][0]']     \n",
            "                                                                                                  \n",
            " dense_7 (Dense)             (None, 128)                  65664     ['universal_sentence_encoder[0\n",
            "                                                                    ][0]']                        \n",
            "                                                                                                  \n",
            " bidirectional_1 (Bidirecti  (None, 48)                   9600      ['char_embed[0][0]']          \n",
            " onal)                                                                                            \n",
            "                                                                                                  \n",
            " combined_embeddings (Conca  (None, 176)                  0         ['dense_7[0][0]',             \n",
            " tenate)                                                             'bidirectional_1[0][0]']     \n",
            "                                                                                                  \n",
            " line_number_input (InputLa  [(None, 15)]                 0         []                            \n",
            " yer)                                                                                             \n",
            "                                                                                                  \n",
            " total_lines_input (InputLa  [(None, 20)]                 0         []                            \n",
            " yer)                                                                                             \n",
            "                                                                                                  \n",
            " dense_10 (Dense)            (None, 256)                  45312     ['combined_embeddings[0][0]'] \n",
            "                                                                                                  \n",
            " dense_8 (Dense)             (None, 32)                   512       ['line_number_input[0][0]']   \n",
            "                                                                                                  \n",
            " dense_9 (Dense)             (None, 32)                   672       ['total_lines_input[0][0]']   \n",
            "                                                                                                  \n",
            " dropout_2 (Dropout)         (None, 256)                  0         ['dense_10[0][0]']            \n",
            "                                                                                                  \n",
            " combined_positional (Conca  (None, 320)                  0         ['dense_8[0][0]',             \n",
            " tenate)                                                             'dense_9[0][0]',             \n",
            "                                                                     'dropout_2[0][0]']           \n",
            "                                                                                                  \n",
            " output_layer (Dense)        (None, 5)                    1605      ['combined_positional[0][0]'] \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 256921889 (980.08 MB)\n",
            "Trainable params: 124065 (484.63 KB)\n",
            "Non-trainable params: 256797824 (979.61 MB)\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# END OF PROJECT"
      ],
      "metadata": {
        "id": "xcGileg-4jUL"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**This later part is for experimental purposes only.**"
      ],
      "metadata": {
        "id": "oct77yAM4nyR"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Your Challenges\n",
        "\n",
        "1. Turn the test data samples into a tf.data.Dataset (fast loading) and then evaluate (make predictions) the best performing model on the test samples.\n",
        "2. Find the most wrong predictions from 1 (these are the samples where the model has predicted the wrong label with the highest prediction probabiltiy).\n",
        "3. Make example predictions (on RCT abstracts from the wild), we can go to PubMed to find these : https://pubmed.ncbi.nlm.nih.gov/, find and use our model to predict on 3-4 abstracts from the wild.\n",
        " * Some examples: https://github.com/mrdbourke/tensorflow-deep-learning/blob/main/extras/skimlit_example_abstracts.json\n",
        "4. See the exercises and extra-curriculum on GitHub , and try to solve them."
      ],
      "metadata": {
        "id": "6HsqK94506Kw"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Evaluating on test data"
      ],
      "metadata": {
        "id": "7YvvMW-Ig30c"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Creating test_char_token_pos_Dataset\n",
        "test_char_token_pos_data = tf.data.Dataset.from_tensor_slices((test_sentences, test_chars, test_line_number_one_hot, test_total_lines_one_hot))\n",
        "test_char_token_pos_labels = tf.data.Dataset.from_tensor_slices(test_labels_one_hot)\n",
        "test_char_token_pos_dataset = tf.data.Dataset.zip((test_char_token_pos_data, test_char_token_pos_labels))\n",
        "test_char_token_pos_dataset = test_char_token_pos_dataset.batch(32).prefetch(tf.data.AUTOTUNE)"
      ],
      "metadata": {
        "id": "9Vi2LNmPKUcg"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate the test data with our best performing models\n",
        "model_5.evaluate(test_char_token_pos_dataset)"
      ],
      "metadata": {
        "id": "HRz4wDM448Pu",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "1c93c921-7d48-4b25-86a4-67b3145ea57d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "942/942 [==============================] - 21s 22ms/step - loss: 0.9439 - accuracy: 0.8277\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.94388347864151, 0.827741801738739]"
            ]
          },
          "metadata": {},
          "execution_count": 121
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Make predictions with the best_model on validation data\n",
        "model_5_pred_probs = model_5.predict(val_char_token_pos_dataset)\n",
        "model_5_preds = (tf.round(model_5_pred_probs))\n",
        "model_5_preds[:10] # these should be in label format"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "N0ianKdQL2XD",
        "outputId": "088e0319-075a-4268-e83d-98d3e6a7f727"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "945/945 [==============================] - 19s 20ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(10, 5), dtype=float32, numpy=\n",
              "array([[0., 0., 0., 0., 0.],\n",
              "       [1., 0., 0., 0., 0.],\n",
              "       [0., 0., 0., 0., 0.],\n",
              "       [0., 0., 1., 0., 0.],\n",
              "       [0., 0., 1., 0., 0.],\n",
              "       [0., 0., 0., 0., 1.],\n",
              "       [0., 0., 0., 0., 1.],\n",
              "       [0., 0., 0., 0., 1.],\n",
              "       [0., 0., 0., 0., 1.],\n",
              "       [0., 1., 0., 0., 0.]], dtype=float32)>"
            ]
          },
          "metadata": {},
          "execution_count": 124
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Finding out the most wrong predictions"
      ],
      "metadata": {
        "id": "y65KNvn_gwiw"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Create a DataFrame with validation sentences, validation labels and best performing model prediction labels + probabilities\n",
        "val_df = pd.DataFrame({\"text\": np.array(val_sentences).flatten(),\n",
        "                       \"target\": val_labels_encoded,\n",
        "                       \"pred\": np.argmax(model_5_preds, axis=1), # Get the index of the class with highest probability\n",
        "                       \"pred_prob\": tf.squeeze(tf.reduce_max(model_5_pred_probs, axis=1))}) # Get the highest probability for each sample\n",
        "val_df.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "LbuZjvd2bQoD",
        "outputId": "b2cb1f57-1ac4-45ff-c0a9-82a59b4bb2a7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                                text  target  pred  pred_prob\n",
              "0  ige sensitization to aspergillus fumigatus and...       0     0   0.488259\n",
              "1  it is not clear whether these patients would b...       0     0   0.500597\n",
              "2  we sought to determine whether a @-month cours...       3     0   0.381509\n",
              "3  asthmatic patients who were ige sensitized to ...       2     2   0.721598\n",
              "4  primary outcomes were improvement in quality o...       2     2   0.693473"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-36d4ebf0-1084-4fe0-be30-b3618a2fd7ae\" 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>text</th>\n",
              "      <th>target</th>\n",
              "      <th>pred</th>\n",
              "      <th>pred_prob</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>ige sensitization to aspergillus fumigatus and...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.488259</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>it is not clear whether these patients would b...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.500597</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>we sought to determine whether a @-month cours...</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0.381509</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>asthmatic patients who were ige sensitized to ...</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>0.721598</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>primary outcomes were improvement in quality o...</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>0.693473</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-36d4ebf0-1084-4fe0-be30-b3618a2fd7ae')\"\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 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\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",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\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-36d4ebf0-1084-4fe0-be30-b3618a2fd7ae 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-36d4ebf0-1084-4fe0-be30-b3618a2fd7ae');\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",
              "\n",
              "\n",
              "<div id=\"df-250d6abc-24e5-4348-955c-1625d66c1d18\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-250d6abc-24e5-4348-955c-1625d66c1d18')\"\n",
              "            title=\"Suggest charts\"\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",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-250d6abc-24e5-4348-955c-1625d66c1d18 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "val_df",
              "summary": "{\n  \"name\": \"val_df\",\n  \"rows\": 30212,\n  \"fields\": [\n    {\n      \"column\": \"text\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 29870,\n        \"samples\": [\n          \"a reduction in pigment was observed with both lasers during the study period .\",\n          \"ten paramedics with field experience were trained with an ultrasound machine in the performance of the fast scan .\",\n          \"we examined the impact of milk proteins and combined exercise training ( cet ) on bp , arterial function , and muscle strength ( one-repetition maximum ( @-rm ) ) .\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"target\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1,\n        \"min\": 0,\n        \"max\": 4,\n        \"num_unique_values\": 5,\n        \"samples\": [\n          3,\n          1,\n          2\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"pred\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1,\n        \"min\": 0,\n        \"max\": 4,\n        \"num_unique_values\": 5,\n        \"samples\": [\n          2,\n          3,\n          4\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"pred_prob\",\n      \"properties\": {\n        \"dtype\": \"float32\",\n        \"num_unique_values\": 30035,\n        \"samples\": [\n          0.5088409781455994,\n          0.5409425497055054,\n          0.4790516197681427\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 168
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Find the wrong predictions and sort by prediction probabilities\n",
        "most_wrong = val_df[val_df[\"target\"] != val_df[\"pred\"]].sort_values(\"pred_prob\", ascending=False)\n",
        "most_wrong[:10] # these are false positives"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 363
        },
        "id": "kIPu-RXRb3JE",
        "outputId": "11ed141c-5b84-4dac-8060-9249b9bffaae"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                                    text  target  pred  \\\n",
              "28193  primary outcome measures were symptom self-rat...       4     2   \n",
              "15279  to assess the short-term efficacy and safety o...       0     3   \n",
              "18115  to test the hypothesis that the administration...       0     3   \n",
              "25081  the resource consumption , costs , and health-...       3     2   \n",
              "10034  secondary outcomes included time to tb treatme...       4     2   \n",
              "28492                     clinicaltrials.gov id : nct@ .       1     0   \n",
              "16173  a qol questionnaire including a pain scale wil...       1     2   \n",
              "9498                       ( clinicaltrials.gov nct@ ) .       2     0   \n",
              "28639  no difference was found between children with ...       1     4   \n",
              "892                          clinicaltrials.gov : nct@ .       2     0   \n",
              "\n",
              "       pred_prob  \n",
              "28193   0.933388  \n",
              "15279   0.924459  \n",
              "18115   0.923645  \n",
              "25081   0.923115  \n",
              "10034   0.920944  \n",
              "28492   0.918967  \n",
              "16173   0.915548  \n",
              "9498    0.908425  \n",
              "28639   0.906175  \n",
              "892     0.905295  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-64825b92-26b0-458c-bfe2-6c58faa01aa0\" 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>text</th>\n",
              "      <th>target</th>\n",
              "      <th>pred</th>\n",
              "      <th>pred_prob</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>28193</th>\n",
              "      <td>primary outcome measures were symptom self-rat...</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>0.933388</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15279</th>\n",
              "      <td>to assess the short-term efficacy and safety o...</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0.924459</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18115</th>\n",
              "      <td>to test the hypothesis that the administration...</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0.923645</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25081</th>\n",
              "      <td>the resource consumption , costs , and health-...</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>0.923115</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10034</th>\n",
              "      <td>secondary outcomes included time to tb treatme...</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>0.920944</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>28492</th>\n",
              "      <td>clinicaltrials.gov id : nct@ .</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0.918967</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16173</th>\n",
              "      <td>a qol questionnaire including a pain scale wil...</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>0.915548</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9498</th>\n",
              "      <td>( clinicaltrials.gov nct@ ) .</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>0.908425</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>28639</th>\n",
              "      <td>no difference was found between children with ...</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>0.906175</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>892</th>\n",
              "      <td>clinicaltrials.gov : nct@ .</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>0.905295</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-64825b92-26b0-458c-bfe2-6c58faa01aa0')\"\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 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\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",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\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-64825b92-26b0-458c-bfe2-6c58faa01aa0 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-64825b92-26b0-458c-bfe2-6c58faa01aa0');\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",
              "\n",
              "\n",
              "<div id=\"df-df51bcf7-9c06-4d17-bedf-122f0f84d5cb\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-df51bcf7-9c06-4d17-bedf-122f0f84d5cb')\"\n",
              "            title=\"Suggest charts\"\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",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-df51bcf7-9c06-4d17-bedf-122f0f84d5cb button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "summary": "{\n  \"name\": \"most_wrong[:10] # these are false positives\",\n  \"rows\": 10,\n  \"fields\": [\n    {\n      \"column\": \"text\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 10,\n        \"samples\": [\n          \"no difference was found between children with gg and ag genotypes who were all breastfed and not low birth weight .\",\n          \"to assess the short-term efficacy and safety of corneal collagen cross-linking ( cxl ) in preventing the progression of keratoconus ( kcn ) .\",\n          \"clinicaltrials.gov id : nct@ .\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"target\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1,\n        \"min\": 0,\n        \"max\": 4,\n        \"num_unique_values\": 5,\n        \"samples\": [\n          0,\n          2,\n          3\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"pred\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 1,\n        \"min\": 0,\n        \"max\": 4,\n        \"num_unique_values\": 4,\n        \"samples\": [\n          3,\n          4,\n          2\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"pred_prob\",\n      \"properties\": {\n        \"dtype\": \"float32\",\n        \"num_unique_values\": 10,\n        \"samples\": [\n          0.9061748385429382,\n          0.9244587421417236,\n          0.9189671874046326\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 169
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "most_wrong.tail() # these are false negeatives"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "an2dEoV9c1YI",
        "outputId": "7787bd98-6eb4-4181-b1dc-8a6b2c29584f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                                    text  target  pred  \\\n",
              "30045  we performed an open-label randomized controll...       2     0   \n",
              "29145  we are starting a two-phase , mixed-method tri...       2     0   \n",
              "13487  we also examined the moderating role of parent...       2     0   \n",
              "24692  in the test group , diode laser was used for v...       2     0   \n",
              "1604   there are no enough studies and research on th...       3     0   \n",
              "\n",
              "       pred_prob  \n",
              "30045   0.249014  \n",
              "29145   0.247008  \n",
              "13487   0.246532  \n",
              "24692   0.237872  \n",
              "1604    0.217423  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-799d9a5a-d7e5-4325-9191-10738c7e3b23\" 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>text</th>\n",
              "      <th>target</th>\n",
              "      <th>pred</th>\n",
              "      <th>pred_prob</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>30045</th>\n",
              "      <td>we performed an open-label randomized controll...</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>0.249014</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29145</th>\n",
              "      <td>we are starting a two-phase , mixed-method tri...</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>0.247008</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13487</th>\n",
              "      <td>we also examined the moderating role of parent...</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>0.246532</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>24692</th>\n",
              "      <td>in the test group , diode laser was used for v...</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>0.237872</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1604</th>\n",
              "      <td>there are no enough studies and research on th...</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0.217423</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-799d9a5a-d7e5-4325-9191-10738c7e3b23')\"\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 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\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",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\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-799d9a5a-d7e5-4325-9191-10738c7e3b23 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-799d9a5a-d7e5-4325-9191-10738c7e3b23');\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",
              "\n",
              "\n",
              "<div id=\"df-27a05b78-35f1-4eae-8f6a-b34856027441\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-27a05b78-35f1-4eae-8f6a-b34856027441')\"\n",
              "            title=\"Suggest charts\"\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",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-27a05b78-35f1-4eae-8f6a-b34856027441 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "summary": "{\n  \"name\": \"most_wrong\",\n  \"rows\": 5,\n  \"fields\": [\n    {\n      \"column\": \"text\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 5,\n        \"samples\": [\n          \"we are starting a two-phase , mixed-method trial to evaluate whether the iphr is scalable across a large number of practices and how its uptake differs for minority and disadvantaged patients .\",\n          \"there are no enough studies and research on the effect of vitamin b@ on the symptoms of pms , therefore , this study was conducted to determine the effect of vitamin b@ on the symptoms of pms in students residing at dormitories of jahrom university of medical sciences in @ .\",\n          \"we also examined the moderating role of parental warmth , as reported in the early family environment , on empathic responding following vasopressin , oxytocin , or placebo administration .\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"target\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0,\n        \"min\": 2,\n        \"max\": 3,\n        \"num_unique_values\": 2,\n        \"samples\": [\n          3,\n          2\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"pred\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 0,\n        \"min\": 0,\n        \"max\": 0,\n        \"num_unique_values\": 1,\n        \"samples\": [\n          0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"pred_prob\",\n      \"properties\": {\n        \"dtype\": \"float32\",\n        \"num_unique_values\": 5,\n        \"samples\": [\n          0.2470082938671112\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 170
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Check the false positives\n",
        "for row in most_wrong[:10].itertuples():\n",
        "  _, text, target, pred, pred_prob = row\n",
        "  print(f\"Target: {target}, Pred: {int(pred)}, Prob: {pred_prob}\")\n",
        "  print(f\"Text:\\n{text}\\n\")\n",
        "  print(\"----\\n\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vCQFHveqdspk",
        "outputId": "3ed9cd66-e9e8-4208-9cab-e14898dbb689"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Target: 4, Pred: 2, Prob: 0.9333879947662354\n",
            "Text:\n",
            "primary outcome measures were symptom self-rating scales ( insomnia severity index -lsb- isi -rsb- and the montgomery sberg depression rating scale -lsb- madrs-s -rsb- ) , assessed before and after treatment with follow-up after @ and @ mo. .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 0, Pred: 3, Prob: 0.9244587421417236\n",
            "Text:\n",
            "to assess the short-term efficacy and safety of corneal collagen cross-linking ( cxl ) in preventing the progression of keratoconus ( kcn ) .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 0, Pred: 3, Prob: 0.9236451387405396\n",
            "Text:\n",
            "to test the hypothesis that the administration of antithrombin concentrate improves disseminated intravascular coagulation ( dic ) , resulting in recovery from dic and better outcomes in patients with sepsis , we conducted a prospective , randomized controlled multicenter trial at @ critical care centers in tertiary care hospitals .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 3, Pred: 2, Prob: 0.9231147766113281\n",
            "Text:\n",
            "the resource consumption , costs , and health-related quality of life were evaluated at baseline , and after @ and @ months by using statutory health insurance information and standardized questionnaires .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 4, Pred: 2, Prob: 0.9209442734718323\n",
            "Text:\n",
            "secondary outcomes included time to tb treatment and mortality .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 1, Pred: 0, Prob: 0.9189671874046326\n",
            "Text:\n",
            "clinicaltrials.gov id : nct@ .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 1, Pred: 2, Prob: 0.9155480265617371\n",
            "Text:\n",
            "a qol questionnaire including a pain scale will assess changes from baseline scores through @ months .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 2, Pred: 0, Prob: 0.9084246754646301\n",
            "Text:\n",
            "( clinicaltrials.gov nct@ ) .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 1, Pred: 4, Prob: 0.9061748385429382\n",
            "Text:\n",
            "no difference was found between children with gg and ag genotypes who were all breastfed and not low birth weight .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 2, Pred: 0, Prob: 0.9052949547767639\n",
            "Text:\n",
            "clinicaltrials.gov : nct@ .\n",
            "\n",
            "----\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Check the false negeatives\n",
        "for row in most_wrong[-10:].itertuples():\n",
        "  _, text, target, pred, pred_prob = row\n",
        "  print(f\"Target: {target}, Pred: {int(pred)}, Prob: {pred_prob}\")\n",
        "  print(f\"Text:\\n{text}\\n\")\n",
        "  print(\"----\\n\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3qzMASemd4cA",
        "outputId": "5e2ed021-1521-4eac-f28b-7d577274ea2c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Target: 3, Pred: 0, Prob: 0.25538209080696106\n",
            "Text:\n",
            "this study aims to compare the complication rates and tissue yield of these two methods in nigerian men .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 4, Pred: 0, Prob: 0.2534535229206085\n",
            "Text:\n",
            "one thousand seven hundred and fifty-eight roadmap patients ( placebo arm : @ ; om arm : @ ) participated in the observational follow up ( ofu ) with an average of @ years .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 4, Pred: 0, Prob: 0.25268369913101196\n",
            "Text:\n",
            "curvature ( or the nonlinear effect ) of experiential avoidance during treatment significantly mediated posttreatment social anxiety symptoms and anhedonic depression in act , but not in cbt , with steeper decline of the acceptance and action questionnaire at the beginning of treatment predicting fewer symptoms in act only .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 2, Pred: 0, Prob: 0.2524600028991699\n",
            "Text:\n",
            "the primary aims are to evaluate the efficacy , safety , and feasibility of a @-week tai chi program for patients with copd .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 4, Pred: 0, Prob: 0.25005635619163513\n",
            "Text:\n",
            "the jump height , as dependent variable , showed a different trend as an effect of the different training protocols , in contrast with the current knowledge .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 2, Pred: 0, Prob: 0.2490142583847046\n",
            "Text:\n",
            "we performed an open-label randomized controlled study to evaluate the role of a nutraceutical containing q@ coenzyme , krill-oil , lipoic acid , resveratrol , vitis vinifera seed oil , vitamin e and selenium in addition to etanercept therapy for patients affected by psoriasis and metabolic syndrome .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 2, Pred: 0, Prob: 0.2470082938671112\n",
            "Text:\n",
            "we are starting a two-phase , mixed-method trial to evaluate whether the iphr is scalable across a large number of practices and how its uptake differs for minority and disadvantaged patients .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 2, Pred: 0, Prob: 0.24653224647045135\n",
            "Text:\n",
            "we also examined the moderating role of parental warmth , as reported in the early family environment , on empathic responding following vasopressin , oxytocin , or placebo administration .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 2, Pred: 0, Prob: 0.23787222802639008\n",
            "Text:\n",
            "in the test group , diode laser was used for vestibular deepening in order to obtain maximum maneuverability of the advanced flap .\n",
            "\n",
            "----\n",
            "\n",
            "Target: 3, Pred: 0, Prob: 0.21742261946201324\n",
            "Text:\n",
            "there are no enough studies and research on the effect of vitamin b@ on the symptoms of pms , therefore , this study was conducted to determine the effect of vitamin b@ on the symptoms of pms in students residing at dormitories of jahrom university of medical sciences in @ .\n",
            "\n",
            "----\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Checking model performance on data from the wild"
      ],
      "metadata": {
        "id": "wLp70ewxg9Kw"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import json\n",
        "# Download and open example abstracts (copy and pasted from PubMed)\n",
        "!wget https://raw.githubusercontent.com/mrdbourke/tensorflow-deep-learning/main/extras/skimlit_example_abstracts.json\n",
        "\n",
        "with open(\"skimlit_example_abstracts.json\", \"r\") as f:\n",
        "  example_abstracts = json.load(f)\n",
        "\n",
        "example_abstracts"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fsusbm9WeGgb",
        "outputId": "585c6d4d-ea5f-45b5-9a65-cebfb1ea8483"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2024-09-02 16:22:57--  https://raw.githubusercontent.com/mrdbourke/tensorflow-deep-learning/main/extras/skimlit_example_abstracts.json\n",
            "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.111.133, 185.199.110.133, 185.199.109.133, ...\n",
            "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.111.133|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 6737 (6.6K) [text/plain]\n",
            "Saving to: ‘skimlit_example_abstracts.json’\n",
            "\n",
            "skimlit_example_abs 100%[===================>]   6.58K  --.-KB/s    in 0s      \n",
            "\n",
            "2024-09-02 16:22:57 (90.4 MB/s) - ‘skimlit_example_abstracts.json’ saved [6737/6737]\n",
            "\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[{'abstract': 'This RCT examined the efficacy of a manualized social intervention for children with HFASDs. Participants were randomly assigned to treatment or wait-list conditions. Treatment included instruction and therapeutic activities targeting social skills, face-emotion recognition, interest expansion, and interpretation of non-literal language. A response-cost program was applied to reduce problem behaviors and foster skills acquisition. Significant treatment effects were found for five of seven primary outcome measures (parent ratings and direct child measures). Secondary measures based on staff ratings (treatment group only) corroborated gains reported by parents. High levels of parent, child and staff satisfaction were reported, along with high levels of treatment fidelity. Standardized effect size estimates were primarily in the medium and large ranges and favored the treatment group.',\n",
              "  'source': 'https://pubmed.ncbi.nlm.nih.gov/20232240/',\n",
              "  'details': 'RCT of a manualized social treatment for high-functioning autism spectrum disorders'},\n",
              " {'abstract': \"Postpartum depression (PPD) is the most prevalent mood disorder associated with childbirth. No single cause of PPD has been identified, however the increased risk of nutritional deficiencies incurred through the high nutritional requirements of pregnancy may play a role in the pathology of depressive symptoms. Three nutritional interventions have drawn particular interest as possible non-invasive and cost-effective prevention and/or treatment strategies for PPD; omega-3 (n-3) long chain polyunsaturated fatty acids (LCPUFA), vitamin D and overall diet. We searched for meta-analyses of randomised controlled trials (RCT's) of nutritional interventions during the perinatal period with PPD as an outcome, and checked for any trials published subsequently to the meta-analyses. Fish oil: Eleven RCT's of prenatal fish oil supplementation RCT's show null and positive effects on PPD symptoms. Vitamin D: no relevant RCT's were identified, however seven observational studies of maternal vitamin D levels with PPD outcomes showed inconsistent associations. Diet: Two Australian RCT's with dietary advice interventions in pregnancy had a positive and null result on PPD. With the exception of fish oil, few RCT's with nutritional interventions during pregnancy assess PPD. Further research is needed to determine whether nutritional intervention strategies during pregnancy can protect against symptoms of PPD. Given the prevalence of PPD and ease of administering PPD measures, we recommend future prenatal nutritional RCT's include PPD as an outcome.\",\n",
              "  'source': 'https://pubmed.ncbi.nlm.nih.gov/28012571/',\n",
              "  'details': 'Formatting removed (can be used to compare model to actual example)'},\n",
              " {'abstract': 'Mental illness, including depression, anxiety and bipolar disorder, accounts for a significant proportion of global disability and poses a substantial social, economic and heath burden. Treatment is presently dominated by pharmacotherapy, such as antidepressants, and psychotherapy, such as cognitive behavioural therapy; however, such treatments avert less than half of the disease burden, suggesting that additional strategies are needed to prevent and treat mental disorders. There are now consistent mechanistic, observational and interventional data to suggest diet quality may be a modifiable risk factor for mental illness. This review provides an overview of the nutritional psychiatry field. It includes a discussion of the neurobiological mechanisms likely modulated by diet, the use of dietary and nutraceutical interventions in mental disorders, and recommendations for further research. Potential biological pathways related to mental disorders include inflammation, oxidative stress, the gut microbiome, epigenetic modifications and neuroplasticity. Consistent epidemiological evidence, particularly for depression, suggests an association between measures of diet quality and mental health, across multiple populations and age groups; these do not appear to be explained by other demographic, lifestyle factors or reverse causality. Our recently published intervention trial provides preliminary clinical evidence that dietary interventions in clinically diagnosed populations are feasible and can provide significant clinical benefit. Furthermore, nutraceuticals including n-3 fatty acids, folate, S-adenosylmethionine, N-acetyl cysteine and probiotics, among others, are promising avenues for future research. Continued research is now required to investigate the efficacy of intervention studies in large cohorts and within clinically relevant populations, particularly in patients with schizophrenia, bipolar and anxiety disorders.',\n",
              "  'source': 'https://pubmed.ncbi.nlm.nih.gov/28942748/',\n",
              "  'details': 'Effect of nutrition on mental health'},\n",
              " {'abstract': \"Hepatitis C virus (HCV) and alcoholic liver disease (ALD), either alone or in combination, count for more than two thirds of all liver diseases in the Western world. There is no safe level of drinking in HCV-infected patients and the most effective goal for these patients is total abstinence. Baclofen, a GABA(B) receptor agonist, represents a promising pharmacotherapy for alcohol dependence (AD). Previously, we performed a randomized clinical trial (RCT), which demonstrated the safety and efficacy of baclofen in patients affected by AD and cirrhosis. The goal of this post-hoc analysis was to explore baclofen's effect in a subgroup of alcohol-dependent HCV-infected cirrhotic patients. Any patient with HCV infection was selected for this analysis. Among the 84 subjects randomized in the main trial, 24 alcohol-dependent cirrhotic patients had a HCV infection; 12 received baclofen 10mg t.i.d. and 12 received placebo for 12-weeks. With respect to the placebo group (3/12, 25.0%), a significantly higher number of patients who achieved and maintained total alcohol abstinence was found in the baclofen group (10/12, 83.3%; p=0.0123). Furthermore, in the baclofen group, compared to placebo, there was a significantly higher increase in albumin values from baseline (p=0.0132) and a trend toward a significant reduction in INR levels from baseline (p=0.0716). In conclusion, baclofen was safe and significantly more effective than placebo in promoting alcohol abstinence, and improving some Liver Function Tests (LFTs) (i.e. albumin, INR) in alcohol-dependent HCV-infected cirrhotic patients. Baclofen may represent a clinically relevant alcohol pharmacotherapy for these patients.\",\n",
              "  'source': 'https://pubmed.ncbi.nlm.nih.gov/22244707/',\n",
              "  'details': 'Baclofen promotes alcohol abstinence in alcohol dependent cirrhotic patients with hepatitis C virus (HCV) infection'}]"
            ]
          },
          "metadata": {},
          "execution_count": 179
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "l83UPnUZgTqz"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}