detect_fake_text_kerasnlp_tf_torch_jax_train.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aQmkJQfmWsyg"
      },
      "source": [
        "# LLM - Detect AI Generated Text by Daniel Fang and Tony Lyu\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
        "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
        "id": "1_KJ8osSWsyj",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "f9cdc998-1a47-4e77-e54a-fbbe5d661cc8"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m950.8/950.8 kB\u001b[0m \u001b[31m9.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m584.5/584.5 kB\u001b[0m \u001b[31m7.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m5.2/5.2 MB\u001b[0m \u001b[31m30.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m475.2/475.2 MB\u001b[0m \u001b[31m2.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m5.5/5.5 MB\u001b[0m \u001b[31m100.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m442.0/442.0 kB\u001b[0m \u001b[31m39.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.7/1.7 MB\u001b[0m \u001b[31m89.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Installing build dependencies ... \u001b[?25l\u001b[?25hdone\n",
            "  Getting requirements to build wheel ... \u001b[?25l\u001b[?25hdone\n",
            "  Preparing metadata (pyproject.toml) ... \u001b[?25l\u001b[?25hdone\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m2.1/2.1 MB\u001b[0m \u001b[31m15.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m190.6/190.6 kB\u001b[0m \u001b[31m17.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m252.8/252.8 kB\u001b[0m \u001b[31m21.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m236.0/236.0 kB\u001b[0m \u001b[31m20.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m2.1/2.1 MB\u001b[0m \u001b[31m37.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m62.7/62.7 kB\u001b[0m \u001b[31m9.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m78.7/78.7 kB\u001b[0m \u001b[31m11.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m512.2/512.2 kB\u001b[0m \u001b[31m36.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m453.8/453.8 kB\u001b[0m \u001b[31m36.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m54.5/54.5 kB\u001b[0m \u001b[31m9.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m3.8/3.8 MB\u001b[0m \u001b[31m63.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m3.3/3.3 MB\u001b[0m \u001b[31m66.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m468.9/468.9 kB\u001b[0m \u001b[31m51.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Building wheel for wandb-addons (pyproject.toml) ... \u001b[?25l\u001b[?25hdone\n",
            "  Building wheel for pathtools (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "Collecting TensorFlow==2.10.1\n",
            "  Downloading tensorflow-2.10.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (578.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m578.1/578.1 MB\u001b[0m \u001b[31m3.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: absl-py>=1.0.0 in /usr/local/lib/python3.10/dist-packages (from TensorFlow==2.10.1) (1.4.0)\n",
            "Requirement already satisfied: astunparse>=1.6.0 in /usr/local/lib/python3.10/dist-packages (from TensorFlow==2.10.1) (1.6.3)\n",
            "Requirement already satisfied: flatbuffers>=2.0 in /usr/local/lib/python3.10/dist-packages (from TensorFlow==2.10.1) (23.5.26)\n",
            "Collecting gast<=0.4.0,>=0.2.1 (from TensorFlow==2.10.1)\n",
            "  Downloading gast-0.4.0-py3-none-any.whl (9.8 kB)\n",
            "Requirement already satisfied: google-pasta>=0.1.1 in /usr/local/lib/python3.10/dist-packages (from TensorFlow==2.10.1) (0.2.0)\n",
            "Requirement already satisfied: grpcio<2.0,>=1.24.3 in /usr/local/lib/python3.10/dist-packages (from TensorFlow==2.10.1) (1.59.3)\n",
            "Requirement already satisfied: h5py>=2.9.0 in /usr/local/lib/python3.10/dist-packages (from TensorFlow==2.10.1) (3.9.0)\n",
            "Collecting keras<2.11,>=2.10.0 (from TensorFlow==2.10.1)\n",
            "  Downloading keras-2.10.0-py2.py3-none-any.whl (1.7 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.7/1.7 MB\u001b[0m \u001b[31m85.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting keras-preprocessing>=1.1.1 (from TensorFlow==2.10.1)\n",
            "  Downloading Keras_Preprocessing-1.1.2-py2.py3-none-any.whl (42 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m42.6/42.6 kB\u001b[0m \u001b[31m6.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: libclang>=13.0.0 in /usr/local/lib/python3.10/dist-packages (from TensorFlow==2.10.1) (16.0.6)\n",
            "Requirement already satisfied: numpy>=1.20 in /usr/local/lib/python3.10/dist-packages (from TensorFlow==2.10.1) (1.23.5)\n",
            "Requirement already satisfied: opt-einsum>=2.3.2 in /usr/local/lib/python3.10/dist-packages (from TensorFlow==2.10.1) (3.3.0)\n",
            "Requirement already satisfied: packaging in /usr/local/lib/python3.10/dist-packages (from TensorFlow==2.10.1) (23.2)\n",
            "Collecting protobuf<3.20,>=3.9.2 (from TensorFlow==2.10.1)\n",
            "  Downloading protobuf-3.19.6-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.1/1.1 MB\u001b[0m \u001b[31m52.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: setuptools in /usr/local/lib/python3.10/dist-packages (from TensorFlow==2.10.1) (67.7.2)\n",
            "Requirement already satisfied: six>=1.12.0 in /usr/local/lib/python3.10/dist-packages (from TensorFlow==2.10.1) (1.16.0)\n",
            "Collecting tensorboard<2.11,>=2.10 (from TensorFlow==2.10.1)\n",
            "  Downloading tensorboard-2.10.1-py3-none-any.whl (5.9 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m5.9/5.9 MB\u001b[0m \u001b[31m110.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: tensorflow-io-gcs-filesystem>=0.23.1 in /usr/local/lib/python3.10/dist-packages (from TensorFlow==2.10.1) (0.34.0)\n",
            "Collecting tensorflow-estimator<2.11,>=2.10.0 (from TensorFlow==2.10.1)\n",
            "  Downloading tensorflow_estimator-2.10.0-py2.py3-none-any.whl (438 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m438.7/438.7 kB\u001b[0m \u001b[31m49.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: termcolor>=1.1.0 in /usr/local/lib/python3.10/dist-packages (from TensorFlow==2.10.1) (2.3.0)\n",
            "Requirement already satisfied: typing-extensions>=3.6.6 in /usr/local/lib/python3.10/dist-packages (from TensorFlow==2.10.1) (4.5.0)\n",
            "Requirement already satisfied: wrapt>=1.11.0 in /usr/local/lib/python3.10/dist-packages (from TensorFlow==2.10.1) (1.14.1)\n",
            "Requirement already satisfied: wheel<1.0,>=0.23.0 in /usr/local/lib/python3.10/dist-packages (from astunparse>=1.6.0->TensorFlow==2.10.1) (0.42.0)\n",
            "Requirement already satisfied: google-auth<3,>=1.6.3 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.11,>=2.10->TensorFlow==2.10.1) (2.17.3)\n",
            "Collecting google-auth-oauthlib<0.5,>=0.4.1 (from tensorboard<2.11,>=2.10->TensorFlow==2.10.1)\n",
            "  Downloading google_auth_oauthlib-0.4.6-py2.py3-none-any.whl (18 kB)\n",
            "Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.11,>=2.10->TensorFlow==2.10.1) (3.5.1)\n",
            "Requirement already satisfied: requests<3,>=2.21.0 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.11,>=2.10->TensorFlow==2.10.1) (2.31.0)\n",
            "Collecting tensorboard-data-server<0.7.0,>=0.6.0 (from tensorboard<2.11,>=2.10->TensorFlow==2.10.1)\n",
            "  Downloading tensorboard_data_server-0.6.1-py3-none-manylinux2010_x86_64.whl (4.9 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m4.9/4.9 MB\u001b[0m \u001b[31m111.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting tensorboard-plugin-wit>=1.6.0 (from tensorboard<2.11,>=2.10->TensorFlow==2.10.1)\n",
            "  Downloading tensorboard_plugin_wit-1.8.1-py3-none-any.whl (781 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m781.3/781.3 kB\u001b[0m \u001b[31m71.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: werkzeug>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.11,>=2.10->TensorFlow==2.10.1) (3.0.1)\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<2.11,>=2.10->TensorFlow==2.10.1) (5.3.2)\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<2.11,>=2.10->TensorFlow==2.10.1) (0.3.0)\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<2.11,>=2.10->TensorFlow==2.10.1) (4.9)\n",
            "Requirement already satisfied: requests-oauthlib>=0.7.0 in /usr/local/lib/python3.10/dist-packages (from google-auth-oauthlib<0.5,>=0.4.1->tensorboard<2.11,>=2.10->TensorFlow==2.10.1) (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<2.11,>=2.10->TensorFlow==2.10.1) (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<2.11,>=2.10->TensorFlow==2.10.1) (3.6)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.21.0->tensorboard<2.11,>=2.10->TensorFlow==2.10.1) (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<2.11,>=2.10->TensorFlow==2.10.1) (2023.11.17)\n",
            "Requirement already satisfied: MarkupSafe>=2.1.1 in /usr/local/lib/python3.10/dist-packages (from werkzeug>=1.0.1->tensorboard<2.11,>=2.10->TensorFlow==2.10.1) (2.1.3)\n",
            "Requirement already satisfied: pyasn1<0.6.0,>=0.4.6 in /usr/local/lib/python3.10/dist-packages (from pyasn1-modules>=0.2.1->google-auth<3,>=1.6.3->tensorboard<2.11,>=2.10->TensorFlow==2.10.1) (0.5.1)\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<0.5,>=0.4.1->tensorboard<2.11,>=2.10->TensorFlow==2.10.1) (3.2.2)\n",
            "Installing collected packages: tensorboard-plugin-wit, keras, tensorflow-estimator, tensorboard-data-server, protobuf, keras-preprocessing, gast, google-auth-oauthlib, tensorboard, TensorFlow\n",
            "  Attempting uninstall: keras\n",
            "    Found existing installation: keras 2.15.0\n",
            "    Uninstalling keras-2.15.0:\n",
            "      Successfully uninstalled keras-2.15.0\n",
            "  Attempting uninstall: tensorflow-estimator\n",
            "    Found existing installation: tensorflow-estimator 2.15.0\n",
            "    Uninstalling tensorflow-estimator-2.15.0:\n",
            "      Successfully uninstalled tensorflow-estimator-2.15.0\n",
            "  Attempting uninstall: tensorboard-data-server\n",
            "    Found existing installation: tensorboard-data-server 0.7.2\n",
            "    Uninstalling tensorboard-data-server-0.7.2:\n",
            "      Successfully uninstalled tensorboard-data-server-0.7.2\n",
            "  Attempting uninstall: protobuf\n",
            "    Found existing installation: protobuf 3.20.3\n",
            "    Uninstalling protobuf-3.20.3:\n",
            "      Successfully uninstalled protobuf-3.20.3\n",
            "  Attempting uninstall: gast\n",
            "    Found existing installation: gast 0.5.4\n",
            "    Uninstalling gast-0.5.4:\n",
            "      Successfully uninstalled gast-0.5.4\n",
            "  Attempting uninstall: google-auth-oauthlib\n",
            "    Found existing installation: google-auth-oauthlib 1.0.0\n",
            "    Uninstalling google-auth-oauthlib-1.0.0:\n",
            "      Successfully uninstalled google-auth-oauthlib-1.0.0\n",
            "  Attempting uninstall: tensorboard\n",
            "    Found existing installation: tensorboard 2.15.1\n",
            "    Uninstalling tensorboard-2.15.1:\n",
            "      Successfully uninstalled tensorboard-2.15.1\n",
            "  Attempting uninstall: TensorFlow\n",
            "    Found existing installation: tensorflow 2.15.0\n",
            "    Uninstalling tensorflow-2.15.0:\n",
            "      Successfully uninstalled tensorflow-2.15.0\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",
            "tensorflow-datasets 4.9.3 requires protobuf>=3.20, but you have protobuf 3.19.6 which is incompatible.\n",
            "tensorflow-metadata 1.14.0 requires protobuf<4.21,>=3.20.3, but you have protobuf 3.19.6 which is incompatible.\n",
            "tensorflow-text 2.15.0 requires tensorflow<2.16,>=2.15.0; platform_machine != \"arm64\" or platform_system != \"Darwin\", but you have tensorflow 2.10.1 which is incompatible.\u001b[0m\u001b[31m\n",
            "\u001b[0mSuccessfully installed TensorFlow-2.10.1 gast-0.4.0 google-auth-oauthlib-0.4.6 keras-2.10.0 keras-preprocessing-1.1.2 protobuf-3.19.6 tensorboard-2.10.1 tensorboard-data-server-0.6.1 tensorboard-plugin-wit-1.8.1 tensorflow-estimator-2.10.0\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.colab-display-data+json": {
              "pip_warning": {
                "packages": [
                  "google"
                ]
              }
            }
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting tensorflow-text==2.10.0\n",
            "  Downloading tensorflow_text-2.10.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (5.9 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m5.9/5.9 MB\u001b[0m \u001b[31m24.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: tensorflow-hub>=0.8.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow-text==2.10.0) (0.15.0)\n",
            "Requirement already satisfied: tensorflow<2.11,>=2.10.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow-text==2.10.0) (2.10.1)\n",
            "Requirement already satisfied: absl-py>=1.0.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (1.4.0)\n",
            "Requirement already satisfied: astunparse>=1.6.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (1.6.3)\n",
            "Requirement already satisfied: flatbuffers>=2.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (23.5.26)\n",
            "Requirement already satisfied: gast<=0.4.0,>=0.2.1 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (0.4.0)\n",
            "Requirement already satisfied: google-pasta>=0.1.1 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (0.2.0)\n",
            "Requirement already satisfied: grpcio<2.0,>=1.24.3 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (1.59.3)\n",
            "Requirement already satisfied: h5py>=2.9.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (3.9.0)\n",
            "Requirement already satisfied: keras<2.11,>=2.10.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (2.10.0)\n",
            "Requirement already satisfied: keras-preprocessing>=1.1.1 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (1.1.2)\n",
            "Requirement already satisfied: libclang>=13.0.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (16.0.6)\n",
            "Requirement already satisfied: numpy>=1.20 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (1.23.5)\n",
            "Requirement already satisfied: opt-einsum>=2.3.2 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (3.3.0)\n",
            "Requirement already satisfied: packaging in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (23.2)\n",
            "Requirement already satisfied: protobuf<3.20,>=3.9.2 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (3.19.6)\n",
            "Requirement already satisfied: setuptools in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (67.7.2)\n",
            "Requirement already satisfied: six>=1.12.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (1.16.0)\n",
            "Requirement already satisfied: tensorboard<2.11,>=2.10 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (2.10.1)\n",
            "Requirement already satisfied: tensorflow-io-gcs-filesystem>=0.23.1 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (0.34.0)\n",
            "Requirement already satisfied: tensorflow-estimator<2.11,>=2.10.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (2.10.0)\n",
            "Requirement already satisfied: termcolor>=1.1.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (2.3.0)\n",
            "Requirement already satisfied: typing-extensions>=3.6.6 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (4.5.0)\n",
            "Requirement already satisfied: wrapt>=1.11.0 in /usr/local/lib/python3.10/dist-packages (from tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (1.14.1)\n",
            "Requirement already satisfied: wheel<1.0,>=0.23.0 in /usr/local/lib/python3.10/dist-packages (from astunparse>=1.6.0->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (0.42.0)\n",
            "Requirement already satisfied: google-auth<3,>=1.6.3 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (2.17.3)\n",
            "Requirement already satisfied: google-auth-oauthlib<0.5,>=0.4.1 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (0.4.6)\n",
            "Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (3.5.1)\n",
            "Requirement already satisfied: requests<3,>=2.21.0 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (2.31.0)\n",
            "Requirement already satisfied: tensorboard-data-server<0.7.0,>=0.6.0 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (0.6.1)\n",
            "Requirement already satisfied: tensorboard-plugin-wit>=1.6.0 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (1.8.1)\n",
            "Requirement already satisfied: werkzeug>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from tensorboard<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (3.0.1)\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<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (5.3.2)\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<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (0.3.0)\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<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (4.9)\n",
            "Requirement already satisfied: requests-oauthlib>=0.7.0 in /usr/local/lib/python3.10/dist-packages (from google-auth-oauthlib<0.5,>=0.4.1->tensorboard<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (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<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (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<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (3.6)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2.21.0->tensorboard<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (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<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (2023.11.17)\n",
            "Requirement already satisfied: MarkupSafe>=2.1.1 in /usr/local/lib/python3.10/dist-packages (from werkzeug>=1.0.1->tensorboard<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (2.1.3)\n",
            "Requirement already satisfied: pyasn1<0.6.0,>=0.4.6 in /usr/local/lib/python3.10/dist-packages (from pyasn1-modules>=0.2.1->google-auth<3,>=1.6.3->tensorboard<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (0.5.1)\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<0.5,>=0.4.1->tensorboard<2.11,>=2.10->tensorflow<2.11,>=2.10.0->tensorflow-text==2.10.0) (3.2.2)\n",
            "Installing collected packages: tensorflow-text\n",
            "  Attempting uninstall: tensorflow-text\n",
            "    Found existing installation: tensorflow-text 2.15.0\n",
            "    Uninstalling tensorflow-text-2.15.0:\n",
            "      Successfully uninstalled tensorflow-text-2.15.0\n",
            "Successfully installed tensorflow-text-2.10.0\n"
          ]
        }
      ],
      "source": [
        "!pip install -q keras-core --upgrade\n",
        "!pip install -q keras-nlp --upgrade\n",
        "!pip install --upgrade -q wandb git+https://github.com/soumik12345/wandb-addons\n",
        "!pip install TensorFlow==2.10.1\n",
        "!pip install tensorflow-text==2.10.0"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "-zAUhEHMtcY-"
      },
      "outputs": [],
      "source": [
        "#!pip install matplotlib"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "_kg_hide-output": true,
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dEVMQuiLWsyl",
        "outputId": "fb35e746-b7b1-40c6-d67b-91fe84c02375"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Using PyTorch backend.\n"
          ]
        }
      ],
      "source": [
        "import os\n",
        "os.environ[\"KERAS_BACKEND\"] = \"torch\"\n",
        "os.environ['WANDB_NOTEBOOK_NAME'] = 'detect_fake_text_kerasnlp_tf_torch_jax_train.ipynb'\n",
        "\n",
        "import keras_nlp\n",
        "import keras_core as keras\n",
        "import keras_core.backend as K\n",
        "\n",
        "\n",
        "import torch\n",
        "import tensorflow as tf\n",
        "\n",
        "\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "import matplotlib as mpl\n",
        "\n",
        "cmap = mpl.colormaps.get_cmap('coolwarm')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GeP9g0XfWsym"
      },
      "source": [
        "## Library Version"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "_kg_hide-input": true,
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FE874pN_Wsym",
        "outputId": "9e43b2b9-231b-4187-e3b5-e484c1d06ce4"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "TensorFlow: 2.10.1\n",
            "Keras: 0.1.7\n",
            "KerasNLP: 0.6.3\n"
          ]
        }
      ],
      "source": [
        "print(\"TensorFlow:\", tf.__version__)\n",
        "print(\"Keras:\", keras.__version__)\n",
        "print(\"KerasNLP:\", keras_nlp.__version__)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XmOk6A0tWsym"
      },
      "source": [
        "# Configuration"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "_kg_hide-input": true,
        "id": "o0oV0o4nWsyn"
      },
      "outputs": [],
      "source": [
        "class CFG:\n",
        "    verbose = 0\n",
        "\n",
        "    wandb = True\n",
        "    competition = 'llm-detect-ai-generated-text'\n",
        "    _wandb_kernel = 'awsaf49'\n",
        "    comment = 'DebertaV3-MaxSeq_200-ext_s-torch'\n",
        "\n",
        "    preset = \"deberta_v3_base_en\"\n",
        "    sequence_length = 200\n",
        "\n",
        "    device = 'TPU'\n",
        "\n",
        "    seed = 42  # Random seed for reproducibility\n",
        "\n",
        "    num_folds = 4  # Total folds\n",
        "    selected_folds = [0, 1, 2]  # Folds to train on\n",
        "\n",
        "    epochs = 3\n",
        "    batch_size = 16\n",
        "    drop_remainder = True\n",
        "    cache = True\n",
        "    replicas = 3\n",
        "\n",
        "    scheduler = 'cosine'\n",
        "\n",
        "    class_names = [\"real\", \"fake\"]\n",
        "    num_classes = len(class_names)\n",
        "    class_labels = list(range(num_classes))\n",
        "    label2name = dict(zip(class_labels, class_names))\n",
        "    name2label = {v: k for k, v in label2name.items()}"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "hdRSmaKWWsyo"
      },
      "outputs": [],
      "source": [
        "keras.utils.set_random_seed(CFG.seed)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jL0DG_NGWsyo"
      },
      "source": [
        "# Hardware Setup\n",
        "Following codes automatically detects hardware (TPU or GPU)."
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def get_device():\n",
        "    \"Detect and intializes GPU/TPU automatically\"\n",
        "    try:\n",
        "        # Connect to TPU, this doesn't work on our colab pro\n",
        "        resolver = tf.distribute.cluster_resolver.TPUClusterResolver(tpu='')\n",
        "        tf.config.experimental_connect_to_cluster(resolver)\n",
        "        tf.tpu.experimental.initialize_tpu_system(resolver)\n",
        "        print(\"All devices: \", tf.config.list_logical_devices('TPU'))\n",
        "        # Set TPU strategy\n",
        "        strategy = tf.distribute.TPUStrategy(resolver)\n",
        "        print(f'> Running on TPU', resolver.master(), end=' | ')\n",
        "        print('Num of TPUs: ', strategy.num_replicas_in_sync)\n",
        "        device=CFG.device\n",
        "    except:\n",
        "        # If TPU is not available, detect GPUs\n",
        "        gpus = tf.config.list_logical_devices('GPU')\n",
        "        ngpu = len(gpus)\n",
        "         # Check number of GPUs\n",
        "        if ngpu:\n",
        "            strategy = tf.distribute.MirroredStrategy(gpus)\n",
        "            print(\"> Running on GPU\", end=' | ')\n",
        "            print(\"Num of GPUs: \", ngpu)\n",
        "            device='GPU'\n",
        "        else:\n",
        "            print(\"> Running on CPU\")\n",
        "            strategy = tf.distribute.get_strategy()\n",
        "            device='CPU'\n",
        "    return strategy, device"
      ],
      "metadata": {
        "id": "q8zfpCHRyRYj"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "strategy, CFG.device = get_device()\n",
        "CFG.replicas = strategy.num_replicas_in_sync"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fvIsqEh_y3gJ",
        "outputId": "837d8459-6383-4b61-e698-436f72b80a50"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "> Running on GPU | Num of GPUs:  1\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sVISNajXWsyp"
      },
      "source": [
        "# Dataset Path"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive"
      ],
      "metadata": {
        "id": "tW_ymHvP0FI7"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "aA7fK1q-ZOnh"
      },
      "outputs": [],
      "source": [
        "base_path = '/content/'\n",
        "\n",
        "train_prompts_path = base_path + 'train_prompts.csv'\n",
        "test_essays_path = base_path + 'test_essays.csv'\n",
        "train_essays_path = base_path + 'train_essays.csv'\n",
        "sample_submission_path = base_path + 'sample_submission.csv'\n",
        "\n",
        "train_prompts_df = pd.read_csv(train_prompts_path)\n",
        "test_essays_df = pd.read_csv(test_essays_path)\n",
        "train_essays_df = pd.read_csv(train_essays_path)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jhpEr_rmWsyq"
      },
      "source": [
        "# Meta Data\n",
        "* `{test|train}_essays.csv`\n",
        "    * `id` - A unique identifier for each essay.\n",
        "    * `prompt_id` - Identifies the prompt the essay was written in response to.\n",
        "    * `text` - The essay text itself.\n",
        "    * `generated` - Whether the essay was written by a student (0) or generated by an LLM (1). This field is the target and is not present in test_essays.csv."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WYGb7V2aWsyq"
      },
      "source": [
        "## Train Data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "UUv1wnALtcZA",
        "outputId": "a54210a1-287f-413a-cc3d-1fed61cd0fde"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "         id  prompt_id                                               text  \\\n",
              "0  0059830c          0  Cars. Cars have been around since they became ...   \n",
              "1  005db917          0  Transportation is a large necessity in most co...   \n",
              "2  008f63e3          0  \"America's love affair with it's vehicles seem...   \n",
              "3  00940276          0  How often do you ride in a car? Do you drive a...   \n",
              "4  00c39458          0  Cars are a wonderful thing. They are perhaps o...   \n",
              "\n",
              "   generated  \n",
              "0          0  \n",
              "1          0  \n",
              "2          0  \n",
              "3          0  \n",
              "4          0  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-b5eca019-77f7-40c4-8ef2-0987fee7fd8d\" 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>id</th>\n",
              "      <th>prompt_id</th>\n",
              "      <th>text</th>\n",
              "      <th>generated</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0059830c</td>\n",
              "      <td>0</td>\n",
              "      <td>Cars. Cars have been around since they became ...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>005db917</td>\n",
              "      <td>0</td>\n",
              "      <td>Transportation is a large necessity in most co...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>008f63e3</td>\n",
              "      <td>0</td>\n",
              "      <td>\"America's love affair with it's vehicles seem...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>00940276</td>\n",
              "      <td>0</td>\n",
              "      <td>How often do you ride in a car? Do you drive a...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>00c39458</td>\n",
              "      <td>0</td>\n",
              "      <td>Cars are a wonderful thing. They are perhaps o...</td>\n",
              "      <td>0</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-b5eca019-77f7-40c4-8ef2-0987fee7fd8d')\"\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-b5eca019-77f7-40c4-8ef2-0987fee7fd8d 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-b5eca019-77f7-40c4-8ef2-0987fee7fd8d');\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-c64e4f94-5437-4a6c-b9df-59f2331ad5ca\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-c64e4f94-5437-4a6c-b9df-59f2331ad5ca')\"\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-c64e4f94-5437-4a6c-b9df-59f2331ad5ca button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ],
      "source": [
        "train_essays_df.head()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "_kg_hide-input": true,
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 555
        },
        "id": "4FsuvBKDWsyq",
        "outputId": "6f04f7d1-7df9-4b4b-b5ab-912d5cdd2d13"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "# Train Data: 1,378\n",
            "# Sample:\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "         id  prompt_id                                               text  \\\n",
              "0  0059830c          0  Cars. Cars have been around since they became ...   \n",
              "1  005db917          0  Transportation is a large necessity in most co...   \n",
              "\n",
              "   generated  label  name  \n",
              "0          0      0  real  \n",
              "1          0      0  real  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-e06bc01a-f8af-4f5c-ae31-51ca650cbc53\" 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>id</th>\n",
              "      <th>prompt_id</th>\n",
              "      <th>text</th>\n",
              "      <th>generated</th>\n",
              "      <th>label</th>\n",
              "      <th>name</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0059830c</td>\n",
              "      <td>0</td>\n",
              "      <td>Cars. Cars have been around since they became ...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>real</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>005db917</td>\n",
              "      <td>0</td>\n",
              "      <td>Transportation is a large necessity in most co...</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>real</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-e06bc01a-f8af-4f5c-ae31-51ca650cbc53')\"\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-e06bc01a-f8af-4f5c-ae31-51ca650cbc53 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-e06bc01a-f8af-4f5c-ae31-51ca650cbc53');\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-ed7ebd70-71ce-4916-800e-19dc90a8cd49\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-ed7ebd70-71ce-4916-800e-19dc90a8cd49')\"\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-ed7ebd70-71ce-4916-800e-19dc90a8cd49 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 800x400 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "train_essays_df['label'] = train_essays_df['generated'].copy()\n",
        "\n",
        "train_essays_df['name'] = train_essays_df['generated'].map(CFG.label2name) #change later, switch numeric labels to class cateogories\n",
        "\n",
        "print(\"# Train Data: {:,}\".format(len(train_essays_df)))\n",
        "print(\"# Sample:\")\n",
        "display(train_essays_df.head(2))\n",
        "\n",
        "plt.figure(figsize=(8, 4))\n",
        "train_essays_df['name'].value_counts().plot.bar()\n",
        "plt.xlabel(\"Class\")\n",
        "plt.ylabel(\"Count\")\n",
        "plt.title(\"Class distribution for Train Data\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "This is too little data, and we need more!"
      ],
      "metadata": {
        "id": "IAB2fE_R3N0r"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SUTBwtXhWsyr"
      },
      "source": [
        "## External Datasets\n",
        "\n",
        "We'll be utilizing the following external datasets:\n",
        "\n",
        "* [Augmented data for LLM - Detect AI Generated Text](https://www.kaggle.com/datasets/jdragonxherrera/augmented-data-for-llm-detect-ai-generated-text?select=final_train.csv) @jdragonxherrera\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "_kg_hide-input": true,
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 461
        },
        "id": "m9hpSllbWsyr",
        "outputId": "d632fe5f-32ef-4a03-fa8c-25900669eaa6"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "# External Data: 433,564\n",
            "# Sample:\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 800x400 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "file_names = ['final_train.csv', 'final_train2.csv']\n",
        "\n",
        "ext_df = pd.concat([pd.read_csv(base_path + file) for file in file_names])\n",
        "ext_df['name'] = ext_df.label.map(CFG.label2name)\n",
        "\n",
        "print(\"# External Data: {:,}\".format(len(ext_df)))\n",
        "print(\"# Sample:\")\n",
        "\n",
        "plt.figure(figsize=(8, 4))\n",
        "ext_df.name.value_counts().plot.bar(color=[cmap(0.0), cmap(0.65)])\n",
        "plt.xlabel(\"Class\")\n",
        "plt.ylabel(\"Count\")\n",
        "plt.title(\"Answer distribution for External Data\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "LssoMUiGtcZA",
        "outputId": "26c0b7dc-3831-4d3e-8356-b9e7aa240d6b"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                                text  label  name\n",
              "0  We should keep the Electoral College for a num...      0  real\n",
              "1  More and more money is spent on building theat...      1  fake\n",
              "2  Limiting car usage can actually be effective b...      0  real\n",
              "3  Dear Mrs. Smith,\\n\\nI am writing to you today ...      1  fake\n",
              "4  Dear Principal,\\n\\nAfter school or during scho...      0  real"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-e712c8c4-5df3-443d-8aa5-105e8e0e696e\" 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>label</th>\n",
              "      <th>name</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>We should keep the Electoral College for a num...</td>\n",
              "      <td>0</td>\n",
              "      <td>real</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>More and more money is spent on building theat...</td>\n",
              "      <td>1</td>\n",
              "      <td>fake</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Limiting car usage can actually be effective b...</td>\n",
              "      <td>0</td>\n",
              "      <td>real</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Dear Mrs. Smith,\\n\\nI am writing to you today ...</td>\n",
              "      <td>1</td>\n",
              "      <td>fake</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Dear Principal,\\n\\nAfter school or during scho...</td>\n",
              "      <td>0</td>\n",
              "      <td>real</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-e712c8c4-5df3-443d-8aa5-105e8e0e696e')\"\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-e712c8c4-5df3-443d-8aa5-105e8e0e696e 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-e712c8c4-5df3-443d-8aa5-105e8e0e696e');\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-6c0477fc-adb7-45db-a4d0-04c7f9657d6f\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-6c0477fc-adb7-45db-a4d0-04c7f9657d6f')\"\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-6c0477fc-adb7-45db-a4d0-04c7f9657d6f button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ],
      "source": [
        "ext_df.head()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cjGvHOzwtcZA",
        "outputId": "421c87a3-cff9-433e-dc3a-47f3d09ac087"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "155565"
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ],
      "source": [
        "len(ext_df[ext_df['label'] == 1])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UiC3pl1utcZA"
      },
      "source": [
        "## Some More Balancing and Shrinking"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "gu4VuQAJtcZA"
      },
      "outputs": [],
      "source": [
        "ext_df_final = pd.concat([ext_df[ext_df['label'] == 0].sample(20000), ext_df[ext_df['label'] == 1].sample(20000)]).reset_index(drop=True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6zkOym9ltcZA",
        "outputId": "7015a9a6-8bf6-4705-c7ec-1dcd0c13a625"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "20000"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ],
      "source": [
        "len(ext_df_final[ext_df_final['label'] == 0])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 461
        },
        "id": "VjLL65eqtcZA",
        "outputId": "53e5b332-41ce-4c2f-e369-436b07d2b6ee"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "# External Data: 40,000\n",
            "# Sample:\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 800x400 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "print(\"# External Data: {:,}\".format(len(ext_df_final)))\n",
        "print(\"# Sample:\")\n",
        "\n",
        "plt.figure(figsize=(8, 4))\n",
        "ext_df_final.name.value_counts().plot.bar(color=[cmap(0.0), cmap(0.65)])\n",
        "plt.xlabel(\"Class\")\n",
        "plt.ylabel(\"Count\")\n",
        "plt.title(\"Answer distribution for External Data\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tjUdiq-hWsys"
      },
      "source": [
        "## Combine External and Train Data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "GDVaM7DbWsys",
        "outputId": "8c3f4293-2de7-49b3-a670-03c9148d909c"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                                text  label  name   id  \\\n",
              "0  It is necessary for a student to keep their mi...      0  real  NaN   \n",
              "1  THC FACS, or Facial Action Coding System, can ...      0  real  NaN   \n",
              "2  I\\n\\nWould disagree for adding a one hour and ...      0  real  NaN   \n",
              "3  Have you ever wondered what it is like to have...      0  real  NaN   \n",
              "4  Senator of Florida, YM writing you today to ad...      0  real  NaN   \n",
              "\n",
              "   prompt_id  generated  \n",
              "0        NaN        NaN  \n",
              "1        NaN        NaN  \n",
              "2        NaN        NaN  \n",
              "3        NaN        NaN  \n",
              "4        NaN        NaN  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-2cc4e4e3-dfd8-4da7-84ba-fd82e6ed2595\" 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>label</th>\n",
              "      <th>name</th>\n",
              "      <th>id</th>\n",
              "      <th>prompt_id</th>\n",
              "      <th>generated</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>It is necessary for a student to keep their mi...</td>\n",
              "      <td>0</td>\n",
              "      <td>real</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>THC FACS, or Facial Action Coding System, can ...</td>\n",
              "      <td>0</td>\n",
              "      <td>real</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>I\\n\\nWould disagree for adding a one hour and ...</td>\n",
              "      <td>0</td>\n",
              "      <td>real</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Have you ever wondered what it is like to have...</td>\n",
              "      <td>0</td>\n",
              "      <td>real</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Senator of Florida, YM writing you today to ad...</td>\n",
              "      <td>0</td>\n",
              "      <td>real</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</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-2cc4e4e3-dfd8-4da7-84ba-fd82e6ed2595')\"\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-2cc4e4e3-dfd8-4da7-84ba-fd82e6ed2595 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-2cc4e4e3-dfd8-4da7-84ba-fd82e6ed2595');\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-6bb42453-eb5e-405a-a806-e0cdfabb7788\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-6bb42453-eb5e-405a-a806-e0cdfabb7788')\"\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-6bb42453-eb5e-405a-a806-e0cdfabb7788 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ],
      "source": [
        "df = pd.concat([ext_df_final, train_essays_df], axis=0).reset_index(drop=True)\n",
        "\n",
        "df.head()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IPCz98tDWsys"
      },
      "source": [
        "# Data Split\n",
        "\n",
        "In the code snippet provided below, we will divide the existing **train** data into folds using a stratification of `label` column."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "Ot21eFH1tcZB",
        "outputId": "cc5aeb78-abff-4daa-d319-f64af3473e8a"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                                text  label  name   id  \\\n",
              "0  It is necessary for a student to keep their mi...      0  real  NaN   \n",
              "1  THC FACS, or Facial Action Coding System, can ...      0  real  NaN   \n",
              "2  I\\n\\nWould disagree for adding a one hour and ...      0  real  NaN   \n",
              "3  Have you ever wondered what it is like to have...      0  real  NaN   \n",
              "4  Senator of Florida, YM writing you today to ad...      0  real  NaN   \n",
              "\n",
              "   prompt_id  generated  \n",
              "0        NaN        NaN  \n",
              "1        NaN        NaN  \n",
              "2        NaN        NaN  \n",
              "3        NaN        NaN  \n",
              "4        NaN        NaN  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-c2cb7ff3-fe89-4847-aa7c-847ae418779a\" 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>label</th>\n",
              "      <th>name</th>\n",
              "      <th>id</th>\n",
              "      <th>prompt_id</th>\n",
              "      <th>generated</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>It is necessary for a student to keep their mi...</td>\n",
              "      <td>0</td>\n",
              "      <td>real</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>THC FACS, or Facial Action Coding System, can ...</td>\n",
              "      <td>0</td>\n",
              "      <td>real</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>I\\n\\nWould disagree for adding a one hour and ...</td>\n",
              "      <td>0</td>\n",
              "      <td>real</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Have you ever wondered what it is like to have...</td>\n",
              "      <td>0</td>\n",
              "      <td>real</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Senator of Florida, YM writing you today to ad...</td>\n",
              "      <td>0</td>\n",
              "      <td>real</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</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-c2cb7ff3-fe89-4847-aa7c-847ae418779a')\"\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-c2cb7ff3-fe89-4847-aa7c-847ae418779a 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-c2cb7ff3-fe89-4847-aa7c-847ae418779a');\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-140810b1-6ed4-43d3-8ecf-7214de8b03f6\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-140810b1-6ed4-43d3-8ecf-7214de8b03f6')\"\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-140810b1-6ed4-43d3-8ecf-7214de8b03f6 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ],
      "source": [
        "df.head()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "_kg_hide-input": true,
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XBRoiXTxWsyt",
        "outputId": "58b4888d-858e-4ef8-df19-24255d3b13f6"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "fold  name\n",
              "0     fake    5001\n",
              "      real    5344\n",
              "1     fake    5001\n",
              "      real    5344\n",
              "2     fake    5000\n",
              "      real    5344\n",
              "3     fake    5001\n",
              "      real    5343\n",
              "dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ],
      "source": [
        "from sklearn.model_selection import StratifiedKFold\n",
        "\n",
        "skf = StratifiedKFold(n_splits=CFG.num_folds, shuffle=True, random_state=CFG.seed)\n",
        "\n",
        "df = df.reset_index(drop=True)\n",
        "\n",
        "df['stratify'] = df.label.astype(str)\n",
        "\n",
        "df[\"fold\"] = -1\n",
        "\n",
        "# Assign folds using StratifiedKFold\n",
        "for fold, (train_idx, val_idx) in enumerate(skf.split(df, df['stratify'])):\n",
        "    df.loc[val_idx, 'fold'] = fold\n",
        "\n",
        "df.groupby([\"fold\", \"name\"]).size()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CBWkOoT8Wsyt"
      },
      "source": [
        "# Preprocessing\n",
        "\n",
        "**What it does:** The preprocessor takes input strings and transforms them into a dictionary (`token_ids`, `padding_mask`) containing preprocessed tensors. This process starts with tokenization, where input strings are converted into sequences of token IDs.\n",
        "\n",
        "**Why it's important:** Initially, raw text data is complex and challenging for modeling due to its high dimensionality. By converting text into a compact set of tokens, we simplify the data. Many models rely on special tokens and additional tensors to understand input. These tokens help divide input and identify padding, among other tasks. Making all sequences the same length through padding boosts computational efficiency, making subsequent steps smoother."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "19XUzvQSWsyt"
      },
      "outputs": [],
      "source": [
        "preprocessor = keras_nlp.models.DebertaV3Preprocessor.from_preset(\n",
        "    preset=CFG.preset, # Name of the model\n",
        "    sequence_length=CFG.sequence_length,\n",
        ")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "frIrYLHLWsyu"
      },
      "source": [
        "Now, let's examine what the output shape of the preprocessing layer looks like. The output shape of the layer can be represented as the following:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BMj6VvXlWsyu",
        "outputId": "25790320-5658-4170-d5fa-4ea40ba95127"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "token_ids : torch.Size([200])\n",
            "padding_mask : torch.Size([200])\n"
          ]
        }
      ],
      "source": [
        "inp = preprocessor(df.text.iloc[0])\n",
        "\n",
        "for k, v in inp.items():\n",
        "    print(k, \":\", v.shape)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "jPbQdZBhWsyu"
      },
      "outputs": [],
      "source": [
        "# We'll use the `preprocessing_fn` function to transform each text option using the `dataset.map(preprocessing_fn)` method.\n",
        "def preprocess_fn(text, label=None):\n",
        "    text = preprocessor(text)  # Preprocess text\n",
        "    return (text, label) if label is not None else text"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NbD_hcJYWsyu"
      },
      "source": [
        "# DataLoader\n",
        "\n",
        "The code below sets up a robust data flow pipeline using `tf.data.Dataset` for data processing. Notable aspects of `tf.data` include its ability to simplify pipeline construction and represent components in sequences.\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "_kg_hide-input": true,
        "id": "mgnqy5r0Wsyu"
      },
      "outputs": [],
      "source": [
        "def build_dataset(texts, labels=None, batch_size=32,\n",
        "                  cache=False, drop_remainder=True,\n",
        "                  repeat=False, shuffle=1024):\n",
        "    AUTO = tf.data.AUTOTUNE\n",
        "    slices = (texts,) if labels is None else (texts, labels)\n",
        "    ds = tf.data.Dataset.from_tensor_slices(slices)\n",
        "    ds = ds.cache() if cache else ds\n",
        "    ds = ds.map(preprocess_fn, num_parallel_calls=AUTO)\n",
        "    ds = ds.repeat() if repeat else ds\n",
        "    opt = tf.data.Options()\n",
        "    if shuffle:\n",
        "        ds = ds.shuffle(shuffle, seed=CFG.seed)\n",
        "        opt.experimental_deterministic = False\n",
        "    ds = ds.with_options(opt)\n",
        "    ds = ds.batch(batch_size, drop_remainder=drop_remainder)\n",
        "    ds = ds.prefetch(AUTO)\n",
        "    return ds"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0WjXJjsmWsyv"
      },
      "source": [
        "## Fetch Train/Valid Dataset\n",
        "\n",
        "The function below generates the training and validation datasets for a given fold."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "_kg_hide-input": false,
        "id": "IWZYyRtRWsyv"
      },
      "outputs": [],
      "source": [
        "def get_datasets(fold):\n",
        "    train_df = df[df.fold!=fold].sample(frac=1)\n",
        "\n",
        "    train_texts = train_df.text.tolist()\n",
        "    train_labels = train_df.label.tolist()\n",
        "\n",
        "    # Build training dataset\n",
        "    train_ds = build_dataset(train_texts, train_labels,\n",
        "                             batch_size=CFG.batch_size*CFG.replicas, cache=CFG.cache,\n",
        "                             shuffle=True, drop_remainder=True, repeat=True)\n",
        "\n",
        "    valid_df = df[df.fold==fold].sample(frac=1)  # Get validation fold data\n",
        "    valid_texts = valid_df.text.tolist()\n",
        "    valid_labels = valid_df.label.tolist()\n",
        "\n",
        "    # Build validation dataset\n",
        "    valid_ds = build_dataset(valid_texts, valid_labels,\n",
        "                             batch_size=min(CFG.batch_size*CFG.replicas, len(valid_df)), cache=CFG.cache,\n",
        "                             shuffle=False, drop_remainder=True, repeat=False)\n",
        "\n",
        "    return (train_ds, train_df), (valid_ds, valid_df)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SojWcpmNWsyv"
      },
      "source": [
        "# WandB\n",
        "\n",
        "To monitor the training of my text-based model, I'll make use of **Weights & Biases**. Weights & Biases (W&B) is an MLOps platform that offers experiment tracking, dataset versioning, and model management functionalities, aiding in efficient model development."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        },
        "id": "mKnREALmWsyv",
        "outputId": "756c08a7-ad8a-4044-c2cb-bbb70f00abd2"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: \u001b[33mWARNING\u001b[0m WANDB_NOTEBOOK_NAME should be a path to a notebook file, couldn't find detect_fake_text_kerasnlp_tf_torch_jax_train.ipynb.\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ],
            "application/javascript": [
              "\n",
              "        window._wandbApiKey = new Promise((resolve, reject) => {\n",
              "            function loadScript(url) {\n",
              "            return new Promise(function(resolve, reject) {\n",
              "                let newScript = document.createElement(\"script\");\n",
              "                newScript.onerror = reject;\n",
              "                newScript.onload = resolve;\n",
              "                document.body.appendChild(newScript);\n",
              "                newScript.src = url;\n",
              "            });\n",
              "            }\n",
              "            loadScript(\"https://cdn.jsdelivr.net/npm/postmate/build/postmate.min.js\").then(() => {\n",
              "            const iframe = document.createElement('iframe')\n",
              "            iframe.style.cssText = \"width:0;height:0;border:none\"\n",
              "            document.body.appendChild(iframe)\n",
              "            const handshake = new Postmate({\n",
              "                container: iframe,\n",
              "                url: 'https://wandb.ai/authorize'\n",
              "            });\n",
              "            const timeout = setTimeout(() => reject(\"Couldn't auto authenticate\"), 5000)\n",
              "            handshake.then(function(child) {\n",
              "                child.on('authorize', data => {\n",
              "                    clearTimeout(timeout)\n",
              "                    resolve(data)\n",
              "                });\n",
              "            });\n",
              "            })\n",
              "        });\n",
              "    "
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: Appending key for api.wandb.ai to your netrc file: /root/.netrc\n"
          ]
        }
      ],
      "source": [
        "import wandb\n",
        "import wandb_addons\n",
        "\n",
        "try:\n",
        "    from kaggle_secrets import UserSecretsClient\n",
        "    user_secrets = UserSecretsClient()\n",
        "    api_key = user_secrets.get_secret(\"WANDB\")\n",
        "    wandb.login(key=api_key)\n",
        "    anonymous = 'allow'\n",
        "except:\n",
        "    anonymous = 'must'\n",
        "    wandb.login(anonymous=anonymous, relogin=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QM7lKFX_Wsyw"
      },
      "source": [
        "## Logger\n",
        "\n",
        "The following code cell contains code to log data to WandB. It is noteworthy that the newly released callbacks offer more flexibility in terms of customization, and they are more compact compared to the classic `WandbCallback`, making it easier to use. Here's a brief introduction to them:\n",
        "\n",
        "* **WandbModelCheckpoint**: This callback saves the model or weights using `tf.keras.callbacks.ModelCheckpoint`. Hence, we can harness the power of the official TensorFlow callback to log even `tf.keras.Model` subclass model in TPU.\n",
        "* **WandbMetricsLogger**: This callback simply logs all the metrics and losses.\n",
        "* **WandbEvalCallback**: This one is even more special. We can use it to log the model's prediction after a certain epoch/frequency. We can use it to save segmentation masks, bounding boxes, GradCAM within epochs to check intermediate results and so on.\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "_kg_hide-input": true,
        "id": "sNy_hEYlWsyw"
      },
      "outputs": [],
      "source": [
        "\n",
        "def wandb_init(fold):\n",
        "    config = {k: v for k, v in dict(vars(CFG)).items() if '__' not in k}\n",
        "    config.update({\"fold\": int(fold)})  # Add fold to config\n",
        "    run = wandb.init(project=\"llm-fake-text\",\n",
        "                     name=f\"fold-{fold}|max_seq-{CFG.sequence_length}|model-{CFG.preset}\",\n",
        "                     config=config,\n",
        "                     group=CFG.comment,\n",
        "                     save_code=True,\n",
        "                     anonymous=\"allow\")\n",
        "    return run\n",
        "\n",
        "def log_wandb():\n",
        "    wandb.log({'best_auc': best_auc, 'best_loss': best_loss, 'best_epoch': best_epoch})\n",
        "\n",
        "# Fetch W&B callbacks\n",
        "def get_wb_callbacks(fold):\n",
        "#     wb_ckpt = WandbModelCheckpoint(f'fold{fold}.keras',\n",
        "#                                    monitor='val_auc',\n",
        "#                                    save_best_only=True,\n",
        "#                                    save_weights_only=False,\n",
        "#                                    mode='max')\n",
        "    wb_metr = wandb.keras.WandbMetricsLogger()\n",
        "    return [wb_metr]  # Return WandB callbacks\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rnaeu1QVWsyw"
      },
      "source": [
        "# ⚓ | LR Schedule\n",
        "\n",
        "Implementing a learning rate scheduler is crucial for transfer learning. The learning rate initiates at `lr_start` and gradually tapers down to `lr_min` using various techniques, including:\n",
        "- `step`: Lowering the learning rate in step-wise manner resembling stairs.\n",
        "- `cos`: Utilizing a cosine curve to gradually reduce the learning rate.\n",
        "- `exp`: Exponentially decreasing the learning rate.\n",
        "\n",
        "**Importance:** A well-structured learning rate schedule is essential for efficient model training, ensuring optimal convergence and avoiding issues such as overshooting or stagnation. Initially, a smaller learning rate is often used to avoid large updates that could destroy the pre-learned features. Over time, adjusting the learning rate can help in finding an optimal set of weights for the new task.\n",
        "A learning rate scheduler can be particularly useful when fine-tuning the last few layers of a large neural network while keeping the earlier layers fixed."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "_kg_hide-input": true,
        "id": "e5YmgzV2Wsyw"
      },
      "outputs": [],
      "source": [
        "import math\n",
        "\n",
        "def get_lr_callback(batch_size=8, mode='cos', epochs=10, plot=False):\n",
        "    lr_start, lr_max, lr_min = 0.6e-6, 0.5e-6 * batch_size, 0.3e-6\n",
        "    lr_ramp_ep, lr_sus_ep, lr_decay = 1, 0, 0.75\n",
        "\n",
        "    def lrfn(epoch):  # Learning rate update function\n",
        "        if epoch < lr_ramp_ep: lr = (lr_max - lr_start) / lr_ramp_ep * epoch + lr_start\n",
        "        elif epoch < lr_ramp_ep + lr_sus_ep: lr = lr_max\n",
        "        elif mode == 'exp': lr = (lr_max - lr_min) * lr_decay**(epoch - lr_ramp_ep - lr_sus_ep) + lr_min\n",
        "        elif mode == 'step': lr = lr_max * lr_decay**((epoch - lr_ramp_ep - lr_sus_ep) // 2)\n",
        "        elif mode == 'cos':\n",
        "            decay_total_epochs, decay_epoch_index = epochs - lr_ramp_ep - lr_sus_ep + 3, epoch - lr_ramp_ep - lr_sus_ep\n",
        "            phase = math.pi * decay_epoch_index / decay_total_epochs\n",
        "            lr = (lr_max - lr_min) * 0.5 * (1 + math.cos(phase)) + lr_min\n",
        "        return lr\n",
        "\n",
        "    if plot:\n",
        "        plt.figure(figsize=(10, 5))\n",
        "        plt.plot(np.arange(epochs), [lrfn(epoch) for epoch in np.arange(epochs)], marker='o')\n",
        "        plt.xlabel('epoch'); plt.ylabel('lr')\n",
        "        plt.title('LR Scheduler')\n",
        "        plt.show()\n",
        "\n",
        "    return keras.callbacks.LearningRateScheduler(lrfn, verbose=False)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 487
        },
        "id": "VBghNgATWsyx",
        "outputId": "a77b839a-c216-4d5d-aadc-eff010c9f7db"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x500 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "_=get_lr_callback(CFG.batch_size*CFG.replicas, plot=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZT2gGN8DWsyx"
      },
      "source": [
        "# Callbacks\n",
        "\n",
        "The function below will gather all the training callbacks, such as `lr_scheduler`, `model_checkpoint`, `wandb_logger`, and etc."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "_kg_hide-input": true,
        "id": "AAos-iuDWsyx"
      },
      "outputs": [],
      "source": [
        "def get_callbacks(fold):\n",
        "    callbacks = []\n",
        "    lr_cb = get_lr_callback(CFG.batch_size*CFG.replicas)\n",
        "    ckpt_cb = keras.callbacks.ModelCheckpoint(f'fold{fold}.keras',\n",
        "                                              monitor='val_auc',\n",
        "                                              save_best_only=True,\n",
        "                                              save_weights_only=False,\n",
        "                                              mode='max')\n",
        "    callbacks.extend([lr_cb, ckpt_cb])\n",
        "\n",
        "    if CFG.wandb:\n",
        "        wb_cbs = get_wb_callbacks(fold)  # Get WandB callbacks\n",
        "        callbacks.extend(wb_cbs)\n",
        "\n",
        "    return callbacks  # Return the list of callbacks"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XS4LAuC1Wsyx"
      },
      "source": [
        "# Modeling\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H-zUqr6oWsyx"
      },
      "source": [
        "## KerasNLP Classifier\n",
        "\n",
        "The `KerasNLP` library provides comprehensive, ready-to-use implementations of popular NLP model architectures. It features a variety of pre-trained models including `Bert`, `Roberta`, `DebertaV3`, and more. In this notebook, we'll showcase the usage of `DebertaV3`.\n",
        "\n",
        "Our approach involves using `keras_nlp.models.XXClassifier` to process each text and generatie logits. These logits are passed through a `softmax` function to produce the final output."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "JotE9T4ZWsyy"
      },
      "outputs": [],
      "source": [
        "def build_model():\n",
        "    # Create a DebertaV3Classifier model\n",
        "    classifier = keras_nlp.models.DebertaV3Classifier.from_preset(\n",
        "        CFG.preset,\n",
        "        preprocessor=None,\n",
        "        num_classes=1 # one output per one option, for five options total 5 outputs\n",
        "    )\n",
        "    inputs = classifier.input\n",
        "    logits = classifier(inputs)\n",
        "\n",
        "    outputs = keras.layers.Activation(\"sigmoid\")(logits)\n",
        "    model = keras.Model(inputs, outputs)\n",
        "\n",
        "    model.compile(\n",
        "        optimizer=keras.optimizers.AdamW(3e-6),\n",
        "        loss=keras.losses.BinaryCrossentropy(label_smoothing=0.01),\n",
        "        metrics=[\n",
        "            keras.metrics.AUC(name=\"auc\"),\n",
        "        ],\n",
        "        jit_compile=True\n",
        "    )\n",
        "    return model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "V7SRP6X3Wsyy",
        "outputId": "ac472d30-0823-483e-fefd-4f9400378698"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/keras_core/src/trainers/trainer.py:166: UserWarning: `jit_compile` is not yet enabled for the PyTorch backend. Proceeding with `jit_compile=False`.\n",
            "  warnings.warn(\n"
          ]
        }
      ],
      "source": [
        "model = build_model()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xdQ5Wu4BWsyy"
      },
      "source": [
        "### Model Summary"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "_kg_hide-input": true,
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 273
        },
        "id": "s8c8ZR4JWsyy",
        "outputId": "caebfe0f-108c-4ddf-dc39-c4986b9ec6dc"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1mModel: \"functional_1\"\u001b[0m\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"functional_1\"</span>\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
              "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)             \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape          \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m   Param #\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConnected to              \u001b[0m\u001b[1m \u001b[0m┃\n",
              "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
              "│ padding_mask (\u001b[38;5;33mInputLayer\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;45mNone\u001b[0m)           │          \u001b[38;5;34m0\u001b[0m │ -                          │\n",
              "├───────────────────────────┼────────────────────────┼────────────┼────────────────────────────┤\n",
              "│ token_ids (\u001b[38;5;33mInputLayer\u001b[0m)    │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;45mNone\u001b[0m)           │          \u001b[38;5;34m0\u001b[0m │ -                          │\n",
              "├───────────────────────────┼────────────────────────┼────────────┼────────────────────────────┤\n",
              "│ deberta_v3_classifier     │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m)              │ \u001b[38;5;34m184,422,9…\u001b[0m │ padding_mask[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m],        │\n",
              "│ (\u001b[38;5;33mDebertaV3Classifier\u001b[0m)     │                        │            │ token_ids[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]            │\n",
              "├───────────────────────────┼────────────────────────┼────────────┼────────────────────────────┤\n",
              "│ activation (\u001b[38;5;33mActivation\u001b[0m)   │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m)              │          \u001b[38;5;34m0\u001b[0m │ deberta_v3_classifier[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m…\u001b[0m │\n",
              "└───────────────────────────┴────────────────────────┴────────────┴────────────────────────────┘\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
              "┃<span style=\"font-weight: bold\"> Layer (type)              </span>┃<span style=\"font-weight: bold\"> Output Shape           </span>┃<span style=\"font-weight: bold\">    Param # </span>┃<span style=\"font-weight: bold\"> Connected to               </span>┃\n",
              "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
              "│ padding_mask (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>) │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>)           │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ -                          │\n",
              "├───────────────────────────┼────────────────────────┼────────────┼────────────────────────────┤\n",
              "│ token_ids (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)    │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>)           │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ -                          │\n",
              "├───────────────────────────┼────────────────────────┼────────────┼────────────────────────────┤\n",
              "│ deberta_v3_classifier     │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)              │ <span style=\"color: #00af00; text-decoration-color: #00af00\">184,422,9…</span> │ padding_mask[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>],        │\n",
              "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">DebertaV3Classifier</span>)     │                        │            │ token_ids[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]            │\n",
              "├───────────────────────────┼────────────────────────┼────────────┼────────────────────────────┤\n",
              "│ activation (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Activation</span>)   │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)              │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ deberta_v3_classifier[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">…</span> │\n",
              "└───────────────────────────┴────────────────────────┴────────────┴────────────────────────────┘\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m184,422,913\u001b[0m (703.52 MB)\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">184,422,913</span> (703.52 MB)\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m184,422,913\u001b[0m (703.52 MB)\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">184,422,913</span> (703.52 MB)\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        }
      ],
      "source": [
        "model.summary()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9TDdmiGTWsyy"
      },
      "source": [
        "### Model Plot"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "_kg_hide-input": true,
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 939
        },
        "id": "MwYOCQrgWsyz",
        "outputId": "2573d0a0-3135-46c6-f69a-83697acd613b"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 33
        }
      ],
      "source": [
        "keras.utils.plot_model(model, show_shapes=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zKxEkfq-Wsyz"
      },
      "source": [
        "# Training"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "_kg_hide-input": true,
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "SceMRBAsWsyz",
        "outputId": "7534827f-540b-4d35-ed9e-7852cfa33be4"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: Currently logged in as: \u001b[33mdaniel-zhijin-fang\u001b[0m (\u001b[33maitext\u001b[0m). Use \u001b[1m`wandb login --relogin`\u001b[0m to force relogin\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "wandb version 0.16.0 is available!  To upgrade, please run:\n",
              " $ pip install wandb --upgrade"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Tracking run with wandb version 0.15.12"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Run data is saved locally in <code>/content/wandb/run-20231203_220340-zm86gb43</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Syncing run <strong><a href='https://wandb.ai/aitext/llm-fake-text/runs/zm86gb43' target=\"_blank\">fold-0|max_seq-200|model-deberta_v3_base_en</a></strong> to <a href='https://wandb.ai/aitext/llm-fake-text' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View project at <a href='https://wandb.ai/aitext/llm-fake-text' target=\"_blank\">https://wandb.ai/aitext/llm-fake-text</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run at <a href='https://wandb.ai/aitext/llm-fake-text/runs/zm86gb43' target=\"_blank\">https://wandb.ai/aitext/llm-fake-text/runs/zm86gb43</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "##################################################\n",
            "\tFold: 1 | Model: deberta_v3_base_en\n",
            "\tBatch Size: 16 | Scheduler: cosine\n",
            "\tNum Train: 31033 | Num Valid: 10345\n",
            "##################################################\n",
            "Epoch 1/3\n",
            "\u001b[1m1939/1939\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 372ms/step - auc: 0.8873 - loss: 0.3814"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\u001b[34m\u001b[1mwandb\u001b[0m: \u001b[32m\u001b[41mERROR\u001b[0m Unable to log learning rate.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r\u001b[1m1939/1939\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m822s\u001b[0m 424ms/step - auc: 0.8873 - loss: 0.3813 - val_auc: 0.9978 - val_loss: 0.0830 - learning_rate: 6.0000e-07\n",
            "Epoch 2/3\n",
            "\u001b[1m1939/1939\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m829s\u001b[0m 428ms/step - auc: 0.9978 - loss: 0.0697 - val_auc: 0.9991 - val_loss: 0.0511 - learning_rate: 8.0000e-06\n",
            "Epoch 3/3\n",
            "\u001b[1m1939/1939\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m816s\u001b[0m 421ms/step - auc: 0.9992 - loss: 0.0430 - val_auc: 0.9990 - val_loss: 0.0484 - learning_rate: 7.8688e-06\n",
            "\n",
            "================= FOLD 0 RESULTS =================\n",
            ">>>> BEST Loss  : 0.051\n",
            ">>>> BEST AUC   : 0.999\n",
            ">>>> BEST Epoch : 1\n",
            "==================================================\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<style>\n",
              "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
              "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
              "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
              "    </style>\n",
              "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>best_auc</td><td>▁</td></tr><tr><td>best_epoch</td><td>▁</td></tr><tr><td>best_loss</td><td>▁</td></tr><tr><td>epoch/auc</td><td>▁██</td></tr><tr><td>epoch/epoch</td><td>▁▅█</td></tr><tr><td>epoch/learning_rate</td><td>▁██</td></tr><tr><td>epoch/loss</td><td>█▂▁</td></tr><tr><td>epoch/val_auc</td><td>▁█▇</td></tr><tr><td>epoch/val_loss</td><td>█▂▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>best_auc</td><td>0.99911</td></tr><tr><td>best_epoch</td><td>1</td></tr><tr><td>best_loss</td><td>0.05112</td></tr><tr><td>epoch/auc</td><td>0.99917</td></tr><tr><td>epoch/epoch</td><td>2</td></tr><tr><td>epoch/learning_rate</td><td>1e-05</td></tr><tr><td>epoch/loss</td><td>0.0445</td></tr><tr><td>epoch/val_auc</td><td>0.99899</td></tr><tr><td>epoch/val_loss</td><td>0.04839</td></tr></table><br/></div></div>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run <strong style=\"color:#cdcd00\">fold-0|max_seq-200|model-deberta_v3_base_en</strong> at: <a href='https://wandb.ai/aitext/llm-fake-text/runs/zm86gb43' target=\"_blank\">https://wandb.ai/aitext/llm-fake-text/runs/zm86gb43</a><br/> View job at <a href='https://wandb.ai/aitext/llm-fake-text/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjEyMDQzODIwMQ==/version_details/v6' target=\"_blank\">https://wandb.ai/aitext/llm-fake-text/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjEyMDQzODIwMQ==/version_details/v6</a><br/>Synced 5 W&B file(s), 0 media file(s), 4 artifact file(s) and 1 other file(s)"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Find logs at: <code>./wandb/run-20231203_220340-zm86gb43/logs</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "wandb version 0.16.0 is available!  To upgrade, please run:\n",
              " $ pip install wandb --upgrade"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Tracking run with wandb version 0.15.12"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Run data is saved locally in <code>/content/wandb/run-20231203_224517-x5l694hs</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Syncing run <strong><a href='https://wandb.ai/aitext/llm-fake-text/runs/x5l694hs' target=\"_blank\">fold-1|max_seq-200|model-deberta_v3_base_en</a></strong> to <a href='https://wandb.ai/aitext/llm-fake-text' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View project at <a href='https://wandb.ai/aitext/llm-fake-text' target=\"_blank\">https://wandb.ai/aitext/llm-fake-text</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run at <a href='https://wandb.ai/aitext/llm-fake-text/runs/x5l694hs' target=\"_blank\">https://wandb.ai/aitext/llm-fake-text/runs/x5l694hs</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "##################################################\n",
            "\tFold: 2 | Model: deberta_v3_base_en\n",
            "\tBatch Size: 16 | Scheduler: cosine\n",
            "\tNum Train: 31033 | Num Valid: 10345\n",
            "##################################################\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/keras_core/src/trainers/trainer.py:166: UserWarning: `jit_compile` is not yet enabled for the PyTorch backend. Proceeding with `jit_compile=False`.\n",
            "  warnings.warn(\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/3\n",
            "\u001b[1m1939/1939\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m828s\u001b[0m 427ms/step - auc: 0.8984 - loss: 0.3940 - val_auc: 0.9984 - val_loss: 0.0893 - learning_rate: 6.0000e-07\n",
            "Epoch 2/3\n",
            "\u001b[1m1939/1939\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m830s\u001b[0m 428ms/step - auc: 0.9979 - loss: 0.0661 - val_auc: 0.9994 - val_loss: 0.0499 - learning_rate: 8.0000e-06\n",
            "Epoch 3/3\n",
            "\u001b[1m1939/1939\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m829s\u001b[0m 428ms/step - auc: 0.9994 - loss: 0.0415 - val_auc: 0.9998 - val_loss: 0.0436 - learning_rate: 7.8688e-06\n",
            "\n",
            "================= FOLD 1 RESULTS =================\n",
            ">>>> BEST Loss  : 0.044\n",
            ">>>> BEST AUC   : 1.000\n",
            ">>>> BEST Epoch : 2\n",
            "==================================================\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<style>\n",
              "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
              "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
              "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
              "    </style>\n",
              "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>best_auc</td><td>▁</td></tr><tr><td>best_epoch</td><td>▁</td></tr><tr><td>best_loss</td><td>▁</td></tr><tr><td>epoch/auc</td><td>▁██</td></tr><tr><td>epoch/epoch</td><td>▁▅█</td></tr><tr><td>epoch/learning_rate</td><td>▁██</td></tr><tr><td>epoch/loss</td><td>█▂▁</td></tr><tr><td>epoch/val_auc</td><td>▁▆█</td></tr><tr><td>epoch/val_loss</td><td>█▂▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>best_auc</td><td>0.99982</td></tr><tr><td>best_epoch</td><td>2</td></tr><tr><td>best_loss</td><td>0.04365</td></tr><tr><td>epoch/auc</td><td>0.99941</td></tr><tr><td>epoch/epoch</td><td>2</td></tr><tr><td>epoch/learning_rate</td><td>1e-05</td></tr><tr><td>epoch/loss</td><td>0.042</td></tr><tr><td>epoch/val_auc</td><td>0.99982</td></tr><tr><td>epoch/val_loss</td><td>0.04365</td></tr></table><br/></div></div>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run <strong style=\"color:#cdcd00\">fold-1|max_seq-200|model-deberta_v3_base_en</strong> at: <a href='https://wandb.ai/aitext/llm-fake-text/runs/x5l694hs' target=\"_blank\">https://wandb.ai/aitext/llm-fake-text/runs/x5l694hs</a><br/> View job at <a href='https://wandb.ai/aitext/llm-fake-text/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjEyMDQzODIwMQ==/version_details/v7' target=\"_blank\">https://wandb.ai/aitext/llm-fake-text/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjEyMDQzODIwMQ==/version_details/v7</a><br/>Synced 5 W&B file(s), 0 media file(s), 4 artifact file(s) and 1 other file(s)"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Find logs at: <code>./wandb/run-20231203_224517-x5l694hs/logs</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "wandb version 0.16.0 is available!  To upgrade, please run:\n",
              " $ pip install wandb --upgrade"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Tracking run with wandb version 0.15.12"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Run data is saved locally in <code>/content/wandb/run-20231203_232706-xlvt9zgf</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Syncing run <strong><a href='https://wandb.ai/aitext/llm-fake-text/runs/xlvt9zgf' target=\"_blank\">fold-2|max_seq-200|model-deberta_v3_base_en</a></strong> to <a href='https://wandb.ai/aitext/llm-fake-text' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View project at <a href='https://wandb.ai/aitext/llm-fake-text' target=\"_blank\">https://wandb.ai/aitext/llm-fake-text</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run at <a href='https://wandb.ai/aitext/llm-fake-text/runs/xlvt9zgf' target=\"_blank\">https://wandb.ai/aitext/llm-fake-text/runs/xlvt9zgf</a>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "##################################################\n",
            "\tFold: 3 | Model: deberta_v3_base_en\n",
            "\tBatch Size: 16 | Scheduler: cosine\n",
            "\tNum Train: 31034 | Num Valid: 10344\n",
            "##################################################\n",
            "Epoch 1/3\n",
            "\u001b[1m1939/1939\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m826s\u001b[0m 426ms/step - auc: 0.8462 - loss: 0.4372 - val_auc: 0.9983 - val_loss: 0.0884 - learning_rate: 6.0000e-07\n",
            "Epoch 2/3\n",
            "\u001b[1m1939/1939\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m833s\u001b[0m 430ms/step - auc: 0.9980 - loss: 0.0657 - val_auc: 0.9987 - val_loss: 0.0719 - learning_rate: 8.0000e-06\n",
            "Epoch 3/3\n",
            "\u001b[1m1939/1939\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m834s\u001b[0m 430ms/step - auc: 0.9996 - loss: 0.0430 - val_auc: 0.9993 - val_loss: 0.0563 - learning_rate: 7.8688e-06\n",
            "\n",
            "================= FOLD 2 RESULTS =================\n",
            ">>>> BEST Loss  : 0.056\n",
            ">>>> BEST AUC   : 0.999\n",
            ">>>> BEST Epoch : 2\n",
            "==================================================\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Waiting for W&B process to finish... <strong style=\"color:green\">(success).</strong>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<style>\n",
              "    table.wandb td:nth-child(1) { padding: 0 10px; text-align: left ; width: auto;} td:nth-child(2) {text-align: left ; width: 100%}\n",
              "    .wandb-row { display: flex; flex-direction: row; flex-wrap: wrap; justify-content: flex-start; width: 100% }\n",
              "    .wandb-col { display: flex; flex-direction: column; flex-basis: 100%; flex: 1; padding: 10px; }\n",
              "    </style>\n",
              "<div class=\"wandb-row\"><div class=\"wandb-col\"><h3>Run history:</h3><br/><table class=\"wandb\"><tr><td>best_auc</td><td>▁</td></tr><tr><td>best_epoch</td><td>▁</td></tr><tr><td>best_loss</td><td>▁</td></tr><tr><td>epoch/auc</td><td>▁██</td></tr><tr><td>epoch/epoch</td><td>▁▅█</td></tr><tr><td>epoch/learning_rate</td><td>▁██</td></tr><tr><td>epoch/loss</td><td>█▂▁</td></tr><tr><td>epoch/val_auc</td><td>▁▄█</td></tr><tr><td>epoch/val_loss</td><td>█▄▁</td></tr></table><br/></div><div class=\"wandb-col\"><h3>Run summary:</h3><br/><table class=\"wandb\"><tr><td>best_auc</td><td>0.99934</td></tr><tr><td>best_epoch</td><td>2</td></tr><tr><td>best_loss</td><td>0.05626</td></tr><tr><td>epoch/auc</td><td>0.9996</td></tr><tr><td>epoch/epoch</td><td>2</td></tr><tr><td>epoch/learning_rate</td><td>1e-05</td></tr><tr><td>epoch/loss</td><td>0.04264</td></tr><tr><td>epoch/val_auc</td><td>0.99934</td></tr><tr><td>epoch/val_loss</td><td>0.05626</td></tr></table><br/></div></div>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              " View run <strong style=\"color:#cdcd00\">fold-2|max_seq-200|model-deberta_v3_base_en</strong> at: <a href='https://wandb.ai/aitext/llm-fake-text/runs/xlvt9zgf' target=\"_blank\">https://wandb.ai/aitext/llm-fake-text/runs/xlvt9zgf</a><br/> View job at <a href='https://wandb.ai/aitext/llm-fake-text/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjEyMDQzODIwMQ==/version_details/v8' target=\"_blank\">https://wandb.ai/aitext/llm-fake-text/jobs/QXJ0aWZhY3RDb2xsZWN0aW9uOjEyMDQzODIwMQ==/version_details/v8</a><br/>Synced 5 W&B file(s), 0 media file(s), 4 artifact file(s) and 1 other file(s)"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "Find logs at: <code>./wandb/run-20231203_232706-xlvt9zgf/logs</code>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "\n",
            "\n"
          ]
        }
      ],
      "source": [
        "for fold in CFG.selected_folds:\n",
        "    if CFG.wandb:\n",
        "        run = wandb_init(fold)\n",
        "\n",
        "    (train_ds, train_df), (valid_ds, valid_df) = get_datasets(fold)\n",
        "    callbacks = get_callbacks(fold)\n",
        "\n",
        "    print('#' * 50)\n",
        "    print(f'\\tFold: {fold + 1} | Model: {CFG.preset}\\n\\tBatch Size: {CFG.batch_size * CFG.replicas} | Scheduler: {CFG.scheduler}')\n",
        "    print(f'\\tNum Train: {len(train_df)} | Num Valid: {len(valid_df)}')\n",
        "    print('#' * 50)\n",
        "\n",
        "    K.clear_session()\n",
        "    with strategy.scope():\n",
        "        model = build_model()\n",
        "\n",
        "    history = model.fit(\n",
        "        train_ds,\n",
        "        epochs=CFG.epochs,\n",
        "        validation_data=valid_ds,\n",
        "        callbacks=callbacks,\n",
        "        steps_per_epoch=int(len(train_df) / CFG.batch_size / CFG.replicas),\n",
        "    )\n",
        "\n",
        "    best_epoch = np.argmax(model.history.history['val_auc'])\n",
        "    best_auc = model.history.history['val_auc'][best_epoch]\n",
        "    best_loss = model.history.history['val_loss'][best_epoch]\n",
        "\n",
        "    print(f'\\n{\"=\" * 17} FOLD {fold} RESULTS {\"=\" * 17}')\n",
        "    print(f'>>>> BEST Loss  : {best_loss:.3f}\\n>>>> BEST AUC   : {best_auc:.3f}\\n>>>> BEST Epoch : {best_epoch}')\n",
        "    print('=' * 50)\n",
        "\n",
        "    if CFG.wandb:\n",
        "        log_wandb()\n",
        "        wandb.run.finish()\n",
        "    print(\"\\n\\n\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lt2d-o5RWsyz"
      },
      "source": [
        "## Training Log\n",
        "### Check all the training logs in **WandB** dashboard.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "itFZcMg4Wsyz"
      },
      "source": [
        "# 🧪 | Prediction"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "BcfOxeP_Wsyz",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "7c45970c-fa50-4164-99ce-d1c895a49766"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m646/646\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m93s\u001b[0m 144ms/step\n"
          ]
        }
      ],
      "source": [
        "# Make predictions using the trained model on last validation data\n",
        "predictions = model.predict(\n",
        "    valid_ds,\n",
        "    batch_size=min(CFG.batch_size * CFG.replicas * 2, len(valid_df)), # max batch size = valid size\n",
        "    verbose=1\n",
        ")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "_kg_hide-input": true,
        "id": "noYJoh55Wsyz",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6477f240-84c4-4a1b-98df-6ce01fbf5db1"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "# Predictions\n",
            "\n",
            "Text 1:\n",
            "Have you ever wanted to be apart of NASA or a scientist? Scientist get to explore many parts of the  .... eryone would not be satisfied with my agreement, I now have proof that it was not created by aliens.\n",
            "\n",
            "True: real\n",
            "\n",
            "Predicted: real\n",
            "\n",
            "------------------------------------------------------------------------------------------ \n",
            "\n",
            "Text 2:\n",
            "Every dry, people make decisions about their lives and futures, sometimes without seeking expert gui .... d cons. Each individual must weigh the merits and demerits and arrive rt the best possible solution.\n",
            "\n",
            "True: fake\n",
            "\n",
            "Predicted: fake\n",
            "\n",
            "------------------------------------------------------------------------------------------ \n",
            "\n",
            "Text 3:\n",
            "I strongly disagree with the principal's decision to make extracurricular activities mandatory. Whil .... happy BND engaged in their extracurricular activities, rather then feeling obligated to participate.\n",
            "\n",
            "True: fake\n",
            "\n",
            "Predicted: fake\n",
            "\n",
            "------------------------------------------------------------------------------------------ \n",
            "\n",
            "Text 4:\n",
            "There is a lot controversy regarding the Electoral College system of voting in the United States. So .... and maintain our democracy as members of our wonderful country. Remove the Electoral College system.\n",
            "\n",
            "True: real\n",
            "\n",
            "Predicted: real\n",
            "\n",
            "------------------------------------------------------------------------------------------ \n",
            "\n",
            "Text 5:\n",
            "Eitle: the Importance of Failure in the Path to Success\n",
            "\n",
            "Introduction:\n",
            "\n",
            "Winston Churchill, the forme ....  Statement is true:\n",
            "\n",
            "Churchill's statement is true because failure allows us to learn and grow. When\n",
            "\n",
            "True: fake\n",
            "\n",
            "Predicted: fake\n",
            "\n",
            "------------------------------------------------------------------------------------------ \n",
            "\n"
          ]
        }
      ],
      "source": [
        "pred_answers = (predictions > 0.5).astype(int).squeeze()\n",
        "true_answers = valid_df.label.values\n",
        "\n",
        "# Check 5 Predictions\n",
        "print(\"# Predictions\\n\")\n",
        "for i in range(5):\n",
        "    row = valid_df.iloc[i]\n",
        "    text  = row.text\n",
        "    pred_answer = CFG.label2name[pred_answers[i]]\n",
        "    true_answer = CFG.label2name[true_answers[i]]\n",
        "    print(f\"Text {i+1}:\\n{text[:100]} .... {text[-100:]}\\n\")\n",
        "    print(f\"True: {true_answer}\\n\")\n",
        "    print(f\"Predicted: {pred_answer}\\n\")\n",
        "    print(\"-\"*90, \"\\n\")"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "provenance": [],
      "machine_shape": "hm",
      "gpuType": "V100"
    },
    "kaggle": {
      "accelerator": "nvidiaTeslaT4",
      "dataSources": [
        {
          "databundleVersionId": 6888007,
          "sourceId": 61542,
          "sourceType": "competition"
        },
        {
          "datasetId": 3936750,
          "sourceId": 6847931,
          "sourceType": "datasetVersion"
        },
        {
          "datasetId": 3946973,
          "sourceId": 6867914,
          "sourceType": "datasetVersion"
        },
        {
          "datasetId": 3942644,
          "sourceId": 6890527,
          "sourceType": "datasetVersion"
        }
      ],
      "isGpuEnabled": true,
      "isInternetEnabled": true,
      "language": "python",
      "sourceType": "notebook"
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.9.18"
    },
    "accelerator": "GPU"
  },
  "nbformat": 4,
  "nbformat_minor": 0
}