08_sentiment_analysis_with_bert.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "08.sentiment-analysis-with-bert.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU",
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "66147342074249e1a117a29c6069ef8f": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_3339b96dcc4942f9b8f7b92aa6712f62",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_35619ccf75044f88b2a4bbd56a500b12",
              "IPY_MODEL_ebb52fd3f69d44b493518e3be68080a5"
            ]
          }
        },
        "3339b96dcc4942f9b8f7b92aa6712f62": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "35619ccf75044f88b2a4bbd56a500b12": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_1f520b7a1dd3432fb259d4d2c517fada",
            "_dom_classes": [],
            "description": "Downloading: 100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 213450,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 213450,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_c551ddc5062049e295986f21b8d590cf"
          }
        },
        "ebb52fd3f69d44b493518e3be68080a5": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_4fe19dede7614e71bfdf0ef28f13ab34",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 213k/213k [00:00&lt;00:00, 752kB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_46778ab955ee45eb8e0bc68f0f212b79"
          }
        },
        "1f520b7a1dd3432fb259d4d2c517fada": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "c551ddc5062049e295986f21b8d590cf": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "4fe19dede7614e71bfdf0ef28f13ab34": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "46778ab955ee45eb8e0bc68f0f212b79": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/aadilganigaie/ML-project/blob/main/08_sentiment_analysis_with_bert.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PGnlRWvkY-2c"
      },
      "source": [
        "# Sentiment Analysis with BERT\n",
        "\n",
        "> TL;DR In this tutorial, you'll learn how to fine-tune BERT for sentiment analysis. You'll do the required text preprocessing (special tokens, padding, and attention masks) and build a Sentiment Classifier using the amazing Transformers library by Hugging Face!\n",
        "\n",
        "- [Read the tutorial](https://www.curiousily.com/posts/sentiment-analysis-with-bert-and-hugging-face-using-pytorch-and-python/)\n",
        "- [Run the notebook in your browser (Google Colab)](https://colab.research.google.com/drive/1PHv-IRLPCtv7oTcIGbsgZHqrB5LPvB7S)\n",
        "- [Read the `Getting Things Done with Pytorch` book](https://github.com/curiousily/Getting-Things-Done-with-Pytorch)\n",
        "\n",
        "You'll learn how to:\n",
        "\n",
        "- Intuitively understand what BERT is\n",
        "- Preprocess text data for BERT and build PyTorch Dataset (tokenization, attention masks, and padding)\n",
        "- Use Transfer Learning to build Sentiment Classifier using the Transformers library by Hugging Face\n",
        "- Evaluate the model on test data\n",
        "- Predict sentiment on raw text\n",
        "\n",
        "Let's get started!"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fH8xHMfdX974",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 441
        },
        "cellView": "form",
        "outputId": "d05f3e10-0e83-4188-fc7d-28dd6dc31448"
      },
      "source": [
        "#@title Watch the video tutorial\n",
        "\n",
        "from IPython.display import YouTubeVideo\n",
        "YouTubeVideo('8N-nM3QW7O0', width=720, height=420)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "\n",
              "        <iframe\n",
              "            width=\"720\"\n",
              "            height=\"420\"\n",
              "            src=\"https://www.youtube.com/embed/8N-nM3QW7O0\"\n",
              "            frameborder=\"0\"\n",
              "            allowfullscreen\n",
              "        ></iframe>\n",
              "        "
            ],
            "text/plain": [
              "<IPython.lib.display.YouTubeVideo at 0x7f4aa4252e48>"
            ],
            "image/jpeg": "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\n"
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NJ6MhJYYBCwu",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f0bc01ed-f802-4440-9a4a-2ff5455b5167"
      },
      "source": [
        "!nvidia-smi"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Wed Feb 17 16:09:30 2021       \n",
            "+-----------------------------------------------------------------------------+\n",
            "| NVIDIA-SMI 460.39       Driver Version: 460.32.03    CUDA Version: 11.2     |\n",
            "|-------------------------------+----------------------+----------------------+\n",
            "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
            "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n",
            "|                               |                      |               MIG M. |\n",
            "|===============================+======================+======================|\n",
            "|   0  Tesla T4            Off  | 00000000:00:04.0 Off |                    0 |\n",
            "| N/A   52C    P8    13W /  70W |      0MiB / 15109MiB |      0%      Default |\n",
            "|                               |                      |                  N/A |\n",
            "+-------------------------------+----------------------+----------------------+\n",
            "                                                                               \n",
            "+-----------------------------------------------------------------------------+\n",
            "| Processes:                                                                  |\n",
            "|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\n",
            "|        ID   ID                                                   Usage      |\n",
            "|=============================================================================|\n",
            "|  No running processes found                                                 |\n",
            "+-----------------------------------------------------------------------------+\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Tbodro8Fpmwr"
      },
      "source": [
        "## What is BERT?\n",
        "\n",
        "BERT (introduced in [this paper](https://arxiv.org/abs/1810.04805)) stands for Bidirectional Encoder Representations from Transformers. If you don't know what most of that means - you've come to the right place! Let's unpack the main ideas:\n",
        "\n",
        "- Bidirectional - to understand the text  you're looking you'll have to look back (at the previous words) and forward (at the next words)\n",
        "- Transformers - The [Attention Is All You Need](https://arxiv.org/abs/1706.03762) paper presented the Transformer model. The Transformer reads entire sequences of tokens at once. In a sense, the model is non-directional, while LSTMs read sequentially (left-to-right or right-to-left). The attention mechanism allows for learning contextual relations between words (e.g. `his` in a sentence refers to Jim).\n",
        "- (Pre-trained) contextualized word embeddings - [The ELMO paper](https://arxiv.org/abs/1802.05365v2) introduced a way to encode words based on their meaning/context. Nails has multiple meanings - fingernails and metal nails.\n",
        "\n",
        "BERT was trained by masking 15% of the tokens with the goal to guess them. An additional objective was to predict the next sentence. Let's look at examples of these tasks:\n",
        "\n",
        "### Masked Language Modeling (Masked LM)\n",
        "\n",
        "The objective of this task is to guess the masked tokens. Let's look at an example, and try to not make it harder than it has to be:\n",
        "\n",
        "That's `[mask]` she `[mask]` -> That's what she said\n",
        "\n",
        "### Next Sentence Prediction (NSP)\n",
        "\n",
        "Given a pair of two sentences, the task is to say whether or not the second follows the first (binary classification). Let's continue with the example:\n",
        "\n",
        "*Input* = `[CLS]` That's `[mask]` she `[mask]`. [SEP] Hahaha, nice! [SEP]\n",
        "\n",
        "*Label* = *IsNext*\n",
        "\n",
        "*Input* = `[CLS]` That's `[mask]` she `[mask]`. [SEP] Dwight, you ignorant `[mask]`! [SEP]\n",
        "\n",
        "*Label* = *NotNext*\n",
        "\n",
        "The training corpus was comprised of two entries: [Toronto Book Corpus](https://arxiv.org/abs/1506.06724) (800M words) and English Wikipedia (2,500M words). While the original Transformer has an encoder (for reading the input) and a decoder (that makes the prediction), BERT uses only the decoder.\n",
        "\n",
        "BERT is simply a pre-trained stack of Transformer Encoders. How many Encoders? We have two versions - with 12 (BERT base) and 24 (BERT Large).\n",
        "\n",
        "### Is This Thing Useful in Practice?\n",
        "\n",
        "The BERT paper was released along with [the source code](https://github.com/google-research/bert) and pre-trained models.\n",
        "\n",
        "The best part is that you can do Transfer Learning (thanks to the ideas from OpenAI Transformer) with BERT for many NLP tasks - Classification, Question Answering, Entity Recognition, etc. You can train with small amounts of data and achieve great performance!"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wmj22-TcZMef"
      },
      "source": [
        "## Setup\n",
        "\n",
        "We'll need [the Transformers library](https://huggingface.co/transformers/) by Hugging Face:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Kj_7Tz0-pK69"
      },
      "source": [
        "!pip install -q -U watermark"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Jjsbi1u3QFEM",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "128a1ba4-0c7c-4e01-a0df-9aefb86dbc03"
      },
      "source": [
        "!pip install -qq transformers"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[K     |████████████████████████████████| 1.8MB 8.4MB/s \n",
            "\u001b[K     |████████████████████████████████| 3.2MB 34.9MB/s \n",
            "\u001b[K     |████████████████████████████████| 890kB 37.0MB/s \n",
            "\u001b[?25h  Building wheel for sacremoses (setup.py) ... \u001b[?25l\u001b[?25hdone\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AJqoaFpVpoM8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "dfe9f16c-e273-44a0-e2ac-8541bbe1653c"
      },
      "source": [
        "%reload_ext watermark\n",
        "%watermark -v -p numpy,pandas,torch,transformers"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "CPython 3.6.9\n",
            "IPython 5.5.0\n",
            "\n",
            "numpy 1.19.5\n",
            "pandas 1.1.5\n",
            "torch 1.7.0+cu101\n",
            "transformers 4.3.2\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "w68CZpOwFoly",
        "cellView": "form",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "50187dd4-7249-4aca-d3b7-0bf9cc3b6fcc"
      },
      "source": [
        "#@title Setup & Config\n",
        "import transformers\n",
        "from transformers import BertModel, BertTokenizer, AdamW, get_linear_schedule_with_warmup\n",
        "import torch\n",
        "\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "from pylab import rcParams\n",
        "import matplotlib.pyplot as plt\n",
        "from matplotlib import rc\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import confusion_matrix, classification_report\n",
        "from collections import defaultdict\n",
        "from textwrap import wrap\n",
        "\n",
        "from torch import nn, optim\n",
        "from torch.utils.data import Dataset, DataLoader\n",
        "import torch.nn.functional as F\n",
        "\n",
        "%matplotlib inline\n",
        "%config InlineBackend.figure_format='retina'\n",
        "\n",
        "sns.set(style='whitegrid', palette='muted', font_scale=1.2)\n",
        "\n",
        "HAPPY_COLORS_PALETTE = [\"#01BEFE\", \"#FFDD00\", \"#FF7D00\", \"#FF006D\", \"#ADFF02\", \"#8F00FF\"]\n",
        "\n",
        "sns.set_palette(sns.color_palette(HAPPY_COLORS_PALETTE))\n",
        "\n",
        "rcParams['figure.figsize'] = 12, 8\n",
        "\n",
        "RANDOM_SEED = 42\n",
        "np.random.seed(RANDOM_SEED)\n",
        "torch.manual_seed(RANDOM_SEED)\n",
        "\n",
        "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
        "device"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "device(type='cuda', index=0)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ufzPdoTtNikq"
      },
      "source": [
        "## Data Exploration\n",
        "\n",
        "We'll load the Google Play app reviews dataset, that we've put together in the previous part:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SgPRhuMzi9ot",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4b8598c0-ac2e-48a4-f8f1-d030e4d6a446"
      },
      "source": [
        "!gdown --id 1S6qMioqPJjyBLpLVz4gmRTnJHnjitnuV\n",
        "!gdown --id 1zdmewp7ayS4js4VtrJEHzAheSW-5NBZv"
      ],
      "execution_count": 68,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading...\n",
            "From: https://drive.google.com/uc?id=1S6qMioqPJjyBLpLVz4gmRTnJHnjitnuV\n",
            "To: /content/apps.csv\n",
            "100% 134k/134k [00:00<00:00, 4.35MB/s]\n",
            "Downloading...\n",
            "From: https://drive.google.com/uc?id=1zdmewp7ayS4js4VtrJEHzAheSW-5NBZv\n",
            "To: /content/reviews.csv\n",
            "7.17MB [00:00, 63.0MB/s]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mUKLyKc7I6Qp",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 734
        },
        "outputId": "d8d0cf18-eb23-4ed9-af88-3ffc805dec39"
      },
      "source": [
        "df = pd.read_csv(\"reviewdata.csv\")\n",
        "df.head()"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>reviewId</th>\n",
              "      <th>userName</th>\n",
              "      <th>userImage</th>\n",
              "      <th>content</th>\n",
              "      <th>score</th>\n",
              "      <th>thumbsUpCount</th>\n",
              "      <th>reviewCreatedVersion</th>\n",
              "      <th>at</th>\n",
              "      <th>replyContent</th>\n",
              "      <th>repliedAt</th>\n",
              "      <th>sortOrder</th>\n",
              "      <th>appId</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>gp:AOqpTOG4l1TpcNizG0OqdK7mFcBjmzzvbKmXlRmsMlx...</td>\n",
              "      <td>Madelyn Albright</td>\n",
              "      <td>https://play-lh.googleusercontent.com/a-/AOh14...</td>\n",
              "      <td>I was happy with the app for a while, but then...</td>\n",
              "      <td>1</td>\n",
              "      <td>3557</td>\n",
              "      <td>1.195.1.1</td>\n",
              "      <td>2021-02-06 23:35:26</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>most_relevant</td>\n",
              "      <td>com.king.candycrushsaga</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>gp:AOqpTOHjZWAhfk4lZUIh9IMyGIpjL8Nb-MHpFUrDktR...</td>\n",
              "      <td>Michael Boucher</td>\n",
              "      <td>https://play-lh.googleusercontent.com/-8uMKfOD...</td>\n",
              "      <td>Another type game where your moves don't dicta...</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>1.195.1.1</td>\n",
              "      <td>2021-02-13 16:22:03</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>most_relevant</td>\n",
              "      <td>com.king.candycrushsaga</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>gp:AOqpTOFvvJ4HLiXFRWzgte5QoXu63XCs_dWwrQlDVYI...</td>\n",
              "      <td>Bridget D</td>\n",
              "      <td>https://play-lh.googleusercontent.com/-5jilF2P...</td>\n",
              "      <td>I used to enjoy this game, but not anymore. Wi...</td>\n",
              "      <td>1</td>\n",
              "      <td>1592</td>\n",
              "      <td>1.195.1.1</td>\n",
              "      <td>2021-02-05 15:44:46</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>most_relevant</td>\n",
              "      <td>com.king.candycrushsaga</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>gp:AOqpTOEu4gcoL3oPP2Zruy345YVz5nFQI34Q2g7iixE...</td>\n",
              "      <td>jatin bhosle</td>\n",
              "      <td>https://play-lh.googleusercontent.com/a-/AOh14...</td>\n",
              "      <td>I'm done. It used to be a good game until last...</td>\n",
              "      <td>1</td>\n",
              "      <td>1956</td>\n",
              "      <td>1.195.1.1</td>\n",
              "      <td>2021-02-08 00:10:59</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>most_relevant</td>\n",
              "      <td>com.king.candycrushsaga</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>gp:AOqpTOHGQ5zPny4IhKbqZ6R_RQr5fYC9PiyHuopkPjh...</td>\n",
              "      <td>Larry Sinclair</td>\n",
              "      <td>https://play-lh.googleusercontent.com/a-/AOh14...</td>\n",
              "      <td>Game is rigged. You can't go any further than ...</td>\n",
              "      <td>1</td>\n",
              "      <td>2620</td>\n",
              "      <td>1.195.1.1</td>\n",
              "      <td>2021-02-11 16:17:44</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>most_relevant</td>\n",
              "      <td>com.king.candycrushsaga</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                            reviewId  ...                    appId\n",
              "0  gp:AOqpTOG4l1TpcNizG0OqdK7mFcBjmzzvbKmXlRmsMlx...  ...  com.king.candycrushsaga\n",
              "1  gp:AOqpTOHjZWAhfk4lZUIh9IMyGIpjL8Nb-MHpFUrDktR...  ...  com.king.candycrushsaga\n",
              "2  gp:AOqpTOFvvJ4HLiXFRWzgte5QoXu63XCs_dWwrQlDVYI...  ...  com.king.candycrushsaga\n",
              "3  gp:AOqpTOEu4gcoL3oPP2Zruy345YVz5nFQI34Q2g7iixE...  ...  com.king.candycrushsaga\n",
              "4  gp:AOqpTOHGQ5zPny4IhKbqZ6R_RQr5fYC9PiyHuopkPjh...  ...  com.king.candycrushsaga\n",
              "\n",
              "[5 rows x 12 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dB2jE6am7Dpo",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d1220acc-8715-46bb-c1b8-ef199466d1ae"
      },
      "source": [
        "df.shape"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(10570, 12)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TWqVNHJbn10l"
      },
      "source": [
        "We have about 16k examples. Let's check for missing values:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VA_wGSLQLKCh",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "bc93ff4c-4300-4167-ddb1-ea6ffc13cd0a"
      },
      "source": [
        "df.info()"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 10570 entries, 0 to 10569\n",
            "Data columns (total 12 columns):\n",
            " #   Column                Non-Null Count  Dtype \n",
            "---  ------                --------------  ----- \n",
            " 0   reviewId              10570 non-null  object\n",
            " 1   userName              10570 non-null  object\n",
            " 2   userImage             10570 non-null  object\n",
            " 3   content               10570 non-null  object\n",
            " 4   score                 10570 non-null  int64 \n",
            " 5   thumbsUpCount         10570 non-null  int64 \n",
            " 6   reviewCreatedVersion  9072 non-null   object\n",
            " 7   at                    10570 non-null  object\n",
            " 8   replyContent          781 non-null    object\n",
            " 9   repliedAt             781 non-null    object\n",
            " 10  sortOrder             10570 non-null  object\n",
            " 11  appId                 10570 non-null  object\n",
            "dtypes: int64(2), object(10)\n",
            "memory usage: 991.1+ KB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H3cL_1qVn_6h"
      },
      "source": [
        "Great, no missing values in the score and review texts! Do we have class imbalance?"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Wwh_rW4Efhs3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 560
        },
        "outputId": "e4036b0f-f044-49cc-ecf9-a470b992b16b"
      },
      "source": [
        "sns.countplot(df.score)\n",
        "plt.xlabel('review score');"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
            "  FutureWarning\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "image/png": {
              "width": 745,
              "height": 489
            }
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nZM0GKviobjM"
      },
      "source": [
        "That's hugely imbalanced, but it's okay. We're going to convert the dataset into negative, neutral and positive sentiment:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ei0xmdi1Chp0"
      },
      "source": [
        "def to_sentiment(rating):\n",
        "  rating = int(rating)\n",
        "  if rating <= 2:\n",
        "    return 0\n",
        "  elif rating == 3:\n",
        "    return 1\n",
        "  else: \n",
        "    return 2\n",
        "\n",
        "df['sentiment'] = df.score.apply(to_sentiment)"
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "V-155O-SFSqE"
      },
      "source": [
        "class_names = ['negative', 'neutral', 'positive']"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "y3tY3ECJDPaz",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 560
        },
        "outputId": "840acec8-a40d-4d1b-f860-965376d0d52b"
      },
      "source": [
        "ax = sns.countplot(df.sentiment)\n",
        "plt.xlabel('review sentiment')\n",
        "ax.set_xticklabels(class_names);"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
            "  FutureWarning\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "image/png": {
              "width": 745,
              "height": 489
            }
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tOssB4CKnAX2"
      },
      "source": [
        "The balance was (mostly) restored."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9aHyGuTFgyPO"
      },
      "source": [
        "## Data Preprocessing\n",
        "\n",
        "You might already know that Machine Learning models don't work with raw text. You need to convert text to numbers (of some sort). BERT requires even more attention (good one, right?). Here are the requirements: \n",
        "\n",
        "- Add special tokens to separate sentences and do classification\n",
        "- Pass sequences of constant length (introduce padding)\n",
        "- Create array of 0s (pad token) and 1s (real token) called *attention mask*\n",
        "\n",
        "The Transformers library provides (you've guessed it) a wide variety of Transformer models (including BERT). It works with TensorFlow and PyTorch! It also includes prebuild tokenizers that do the heavy lifting for us!\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E7Mj-0ne--5t"
      },
      "source": [
        "PRE_TRAINED_MODEL_NAME = 'bert-base-cased'"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fMSr7C-F_sey"
      },
      "source": [
        "> You can use a cased and uncased version of BERT and tokenizer. I've experimented with both. The cased version works better. Intuitively, that makes sense, since \"BAD\" might convey more sentiment than \"bad\"."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NiLb-ltM-ZRz"
      },
      "source": [
        "Let's load a pre-trained [BertTokenizer](https://huggingface.co/transformers/model_doc/bert.html#berttokenizer):"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "H3AfJSZ8NNLF",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 66,
          "referenced_widgets": [
            "66147342074249e1a117a29c6069ef8f",
            "3339b96dcc4942f9b8f7b92aa6712f62",
            "35619ccf75044f88b2a4bbd56a500b12",
            "ebb52fd3f69d44b493518e3be68080a5",
            "1f520b7a1dd3432fb259d4d2c517fada",
            "c551ddc5062049e295986f21b8d590cf",
            "4fe19dede7614e71bfdf0ef28f13ab34",
            "46778ab955ee45eb8e0bc68f0f212b79"
          ]
        },
        "outputId": "000c44f5-4744-41b2-fca3-adb091a0ef59"
      },
      "source": [
        "tokenizer = BertTokenizer.from_pretrained(PRE_TRAINED_MODEL_NAME)"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "66147342074249e1a117a29c6069ef8f",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=213450.0, style=ProgressStyle(descripti…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CfrSbwTQ-wi_"
      },
      "source": [
        "We'll use this text to understand the tokenization process:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HZMitwrqm2eb"
      },
      "source": [
        "sample_txt = 'When was I last outside? I am stuck at home for 2 weeks.'"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yO2qBTVl_KPs"
      },
      "source": [
        "Some basic operations can convert the text to tokens and tokens to unique integers (ids):"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iTFhpHpsoWO7",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "754a2a26-7ba4-4aaf-d30e-eb8542c55c41"
      },
      "source": [
        "tokens = tokenizer.tokenize(sample_txt)\n",
        "token_ids = tokenizer.convert_tokens_to_ids(tokens)\n",
        "\n",
        "print(f' Sentence: {sample_txt}')\n",
        "print(f'   Tokens: {tokens}')\n",
        "print(f'Token IDs: {token_ids}')"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            " Sentence: When was I last outside? I am stuck at home for 2 weeks.\n",
            "   Tokens: ['When', 'was', 'I', 'last', 'outside', '?', 'I', 'am', 'stuck', 'at', 'home', 'for', '2', 'weeks', '.']\n",
            "Token IDs: [1332, 1108, 146, 1314, 1796, 136, 146, 1821, 5342, 1120, 1313, 1111, 123, 2277, 119]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bzbbKLR8lZbu"
      },
      "source": [
        "### Special Tokens\n",
        "\n",
        "`[SEP]` - marker for ending of a sentence\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EXwz47bQvCbc",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "386271d5-c24a-4e1f-9669-bed58d3e8be9"
      },
      "source": [
        "tokenizer.sep_token, tokenizer.sep_token_id"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "('[SEP]', 102)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Mip_eGeXwLFF"
      },
      "source": [
        "`[CLS]` - we must add this token to the start of each sentence, so BERT knows we're doing classification"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_6K4it5HwE6l",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "65180554-8a23-488b-8c7f-d72333fb5274"
      },
      "source": [
        "tokenizer.cls_token, tokenizer.cls_token_id"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "('[CLS]', 101)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Qi6O-yEY09gl"
      },
      "source": [
        "There is also a special token for padding:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Vx7gD5xf1AFK",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e386256c-ccdc-4935-d989-7ccefce88977"
      },
      "source": [
        "tokenizer.pad_token, tokenizer.pad_token_id"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "('[PAD]', 0)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6GWCfijM0TWB"
      },
      "source": [
        "BERT understands tokens that were in the training set. Everything else can be encoded using the `[UNK]` (unknown) token:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4cmfFsbEKQDT",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "1af04897-c27c-46c7-9d30-1264db8c4782"
      },
      "source": [
        "tokenizer.unk_token, tokenizer.unk_token_id"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "('[UNK]', 100)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "W9ap7jdL0LYU"
      },
      "source": [
        "All of that work can be done using the [`encode_plus()`](https://huggingface.co/transformers/main_classes/tokenizer.html#transformers.PreTrainedTokenizer.encode_plus) method:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Vea9edaaxSPO",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f08ff11a-8c66-415a-d252-e79f183dc197"
      },
      "source": [
        "encoding = tokenizer.encode_plus(\n",
        "  sample_txt,\n",
        "  max_length=32,\n",
        "  add_special_tokens=True, # Add '[CLS]' and '[SEP]'\n",
        "  return_token_type_ids=False,\n",
        "  padding='max_length',\n",
        "  return_attention_mask=True,\n",
        "  return_tensors='pt',  # Return PyTorch tensors\n",
        ")\n",
        "\n",
        "encoding.keys()"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys(['input_ids', 'attention_mask'])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sS69c8WvdOED"
      },
      "source": [
        "The token ids are now stored in a Tensor and padded to a length of 32:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YzBmcOla0yQR",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "3d449fc4-a6c8-4d8f-fafb-f43c6b7862d6"
      },
      "source": [
        "print(len(encoding['input_ids'][0]))\n",
        "encoding['input_ids'][0]"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "32\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([ 101, 1332, 1108,  146, 1314, 1796,  136,  146, 1821, 5342, 1120, 1313,\n",
              "        1111,  123, 2277,  119,  102,    0,    0,    0,    0,    0,    0,    0,\n",
              "           0,    0,    0,    0,    0,    0,    0,    0])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "itAyVPsNdyc1"
      },
      "source": [
        "The attention mask has the same length:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Wiv5LLiw03Ox",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e2b1260e-1bb4-4a4f-9af4-bec991831241"
      },
      "source": [
        "print(len(encoding['attention_mask'][0]))\n",
        "encoding['attention_mask']"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "32\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0,\n",
              "         0, 0, 0, 0, 0, 0, 0, 0]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "m1RvhC4jNHHy"
      },
      "source": [
        "We can inverse the tokenization to have a look at the special tokens:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IagGoafKLUwW",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "466a2d04-1084-4406-8d95-8a3821aa5d78"
      },
      "source": [
        "tokenizer.convert_ids_to_tokens(encoding['input_ids'][0])"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['[CLS]',\n",
              " 'When',\n",
              " 'was',\n",
              " 'I',\n",
              " 'last',\n",
              " 'outside',\n",
              " '?',\n",
              " 'I',\n",
              " 'am',\n",
              " 'stuck',\n",
              " 'at',\n",
              " 'home',\n",
              " 'for',\n",
              " '2',\n",
              " 'weeks',\n",
              " '.',\n",
              " '[SEP]',\n",
              " '[PAD]',\n",
              " '[PAD]',\n",
              " '[PAD]',\n",
              " '[PAD]',\n",
              " '[PAD]',\n",
              " '[PAD]',\n",
              " '[PAD]',\n",
              " '[PAD]',\n",
              " '[PAD]',\n",
              " '[PAD]',\n",
              " '[PAD]',\n",
              " '[PAD]',\n",
              " '[PAD]',\n",
              " '[PAD]',\n",
              " '[PAD]']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "waKjYxTDuaWt"
      },
      "source": [
        "### Choosing Sequence Length\n",
        "\n",
        "BERT works with fixed-length sequences. We'll use a simple strategy to choose the max length. Let's store the token length of each review:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BUnE5CT9hbeZ"
      },
      "source": [
        "token_lens = []\n",
        "\n",
        "for txt in df.content:\n",
        "  tokens = tokenizer.encode(txt, max_length=512)\n",
        "  token_lens.append(len(tokens))"
      ],
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tI4goUrHf6da"
      },
      "source": [
        "and plot the distribution:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SzE1j4jxmUtd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 560
        },
        "outputId": "b930b2aa-c769-4a24-a33e-9399e90bbf58"
      },
      "source": [
        "sns.distplot(token_lens)\n",
        "plt.xlim([0, 256]);\n",
        "plt.xlabel('Token count');"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
            "  warnings.warn(msg, FutureWarning)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "image/png": {
              "width": 749,
              "height": 489
            }
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oW6ajl30t6du"
      },
      "source": [
        "Most of the reviews seem to contain less than 128 tokens, but we'll be on the safe side and choose a maximum length of 160."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "t7xSmJtLuoxW"
      },
      "source": [
        "MAX_LEN = 209"
      ],
      "execution_count": 105,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XvvcoU6nurHy"
      },
      "source": [
        "We have all building blocks required to create a PyTorch dataset. Let's do it:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E2BPgRJ7YBK0"
      },
      "source": [
        "class GPReviewDataset(Dataset):\n",
        "\n",
        "  def __init__(self, reviews, targets, tokenizer, max_len):\n",
        "    self.reviews = reviews\n",
        "    self.targets = targets\n",
        "    self.tokenizer = tokenizer\n",
        "    self.max_len = max_len\n",
        "  \n",
        "  def __len__(self):\n",
        "    return len(self.reviews)\n",
        "  \n",
        "  def __getitem__(self, item):\n",
        "    review = str(self.reviews[item])\n",
        "    target = self.targets[item]\n",
        "    max_length = 32\n",
        "    encoding = self.tokenizer.encode_plus(\n",
        "      review,\n",
        "      add_special_tokens=True,\n",
        "      max_length=self.max_len,\n",
        "      return_token_type_ids=False,\n",
        "      padding='max_length',\n",
        "      return_attention_mask=True,\n",
        "      return_tensors='pt',\n",
        "    )\n",
        "\n",
        "    return {\n",
        "      'review_text': review,\n",
        "      'input_ids': encoding['input_ids'].flatten(),\n",
        "      'attention_mask': encoding['attention_mask'].flatten(),\n",
        "      'targets': torch.tensor(target, dtype=torch.long)\n",
        "    }"
      ],
      "execution_count": 106,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 163
        },
        "id": "KJ_G0QqGASES",
        "outputId": "f3046b19-91ae-4a40-b572-bab59e50cf1f"
      },
      "source": [
        "type(input_ids)"
      ],
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-31-def419b8ac7b>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mtype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput_ids\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m: name 'input_ids' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rrnvv_drAR6y"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "x2uwsvCYqDJK"
      },
      "source": [
        "The tokenizer is doing most of the heavy lifting for us. We also return the review texts, so it'll be easier to evaluate the predictions from our model. Let's split the data:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "B-vWzoo81dvO"
      },
      "source": [
        "df_train, df_test = train_test_split(df, test_size=0.1, random_state=RANDOM_SEED)\n",
        "df_val, df_test = train_test_split(df_test, test_size=0.5, random_state=RANDOM_SEED)"
      ],
      "execution_count": 139,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xz3ZOQXVPCwh",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "52b511e8-7f73-4d1b-ff7a-c64e44c0e9aa"
      },
      "source": [
        "df_train.shape, df_val.shape, df_test.shape"
      ],
      "execution_count": 140,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((9513, 13), (528, 13), (529, 13))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 140
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "J4tQ1x-vqNab"
      },
      "source": [
        "We also need to create a couple of data loaders. Here's a helper function to do it:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KEGqcvkuOuTX"
      },
      "source": [
        "def create_data_loader(df, tokenizer, max_len, batch_size):\n",
        "  ds = GPReviewDataset(\n",
        "    reviews=df.content.to_numpy(),\n",
        "    targets=df.sentiment.to_numpy(),\n",
        "    tokenizer=tokenizer,\n",
        "    max_len=max_len\n",
        "  )\n",
        "\n",
        "  return DataLoader(\n",
        "    ds,\n",
        "    batch_size=batch_size,\n",
        "    num_workers=4\n",
        "  )"
      ],
      "execution_count": 141,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vODDxMKsPHqI"
      },
      "source": [
        "BATCH_SIZE = 8\n",
        "\n",
        "train_data_loader = create_data_loader(df_train, tokenizer, MAX_LEN, BATCH_SIZE)\n",
        "val_data_loader = create_data_loader(df_val, tokenizer, MAX_LEN, BATCH_SIZE)\n",
        "test_data_loader = create_data_loader(df_test, tokenizer, MAX_LEN, BATCH_SIZE)"
      ],
      "execution_count": 142,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "A6dlOptwqlhF"
      },
      "source": [
        "Let's have a look at an example batch from our training data loader:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Y93ldSN47FeT",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5b179614-9b3f-4e56-f12d-cc241d43b10c"
      },
      "source": [
        "data = next(iter(train_data_loader))\n",
        "data.keys()"
      ],
      "execution_count": 143,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys(['review_text', 'input_ids', 'attention_mask', 'targets'])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 143
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IdU4YVqb7N8M",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "48421259-cddc-4fef-a51e-0ee698420524"
      },
      "source": [
        "print(data['input_ids'].shape)\n",
        "print(data['attention_mask'].shape)\n",
        "print(data['targets'].shape)"
      ],
      "execution_count": 144,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "torch.Size([8, 209])\n",
            "torch.Size([8, 209])\n",
            "torch.Size([8])\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 163
        },
        "id": "CC1apZ0AAit5",
        "outputId": "57f61096-ed57-403d-b096-e5655f857e64"
      },
      "source": [
        "type(input_ids)"
      ],
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-50-def419b8ac7b>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mtype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput_ids\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m: name 'input_ids' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H63Y-TjyRC7S"
      },
      "source": [
        "## Sentiment Classification with BERT and Hugging Face"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "440Nd31VTHER"
      },
      "source": [
        "There are a lot of helpers that make using BERT easy with the Transformers library. Depending on the task you might want to use [BertForSequenceClassification](https://huggingface.co/transformers/model_doc/bert.html#bertforsequenceclassification), [BertForQuestionAnswering](https://huggingface.co/transformers/model_doc/bert.html#bertforquestionanswering) or something else. \n",
        "\n",
        "But who cares, right? We're *hardcore*! We'll use the basic [BertModel](https://huggingface.co/transformers/model_doc/bert.html#bertmodel) and build our sentiment classifier on top of it. Let's load the model:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0P41FayISNRI"
      },
      "source": [
        "bert_model = BertModel.from_pretrained(PRE_TRAINED_MODEL_NAME)"
      ],
      "execution_count": 145,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aFE7YSbFdY4t"
      },
      "source": [
        "And try to use it on the encoding of our sample text:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "s1aoFxbQSn15"
      },
      "source": [
        "last_hidden_state, pooled_output = bert_model(\n",
        "  input_ids=encoding['input_ids'], \n",
        "  attention_mask=encoding['attention_mask']\n",
        ").to_tuple()"
      ],
      "execution_count": 146,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Hvg-qV4uAxC5",
        "outputId": "045455b5-2bed-4b12-fc9a-8296634568eb"
      },
      "source": [
        "last_hidden_state.shape"
      ],
      "execution_count": 147,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([1, 32, 768])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 147
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mLLu8zmqbaHV"
      },
      "source": [
        "The `last_hidden_state` is a sequence of hidden states of the last layer of the model. Obtaining the `pooled_output` is done by applying the [BertPooler](https://github.com/huggingface/transformers/blob/edf0582c0be87b60f94f41c659ea779876efc7be/src/transformers/modeling_bert.py#L426) on `last_hidden_state`:"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Q4dAot4zbz8k"
      },
      "source": [
        "We have the hidden state for each of our 32 tokens (the length of our example sequence). But why 768? This is the number of hidden units in the feedforward-networks. We can verify that by checking the config:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nsxB7Qy7b5YN",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6cdff2ff-f60f-4d85-da5a-4c6bbbd6d858"
      },
      "source": [
        "bert_model.config.hidden_size"
      ],
      "execution_count": 115,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "768"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 115
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wTKi8-rTd_j4"
      },
      "source": [
        "\n",
        "\n",
        "You can think of the `pooled_output` as a summary of the content, according to BERT. Albeit, you might try and do better. Let's look at the shape of the output:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2jIAtRhaSz9c",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "454ea1a4-3c56-41ea-d524-5173518477f5"
      },
      "source": [
        "pooled_output.shape"
      ],
      "execution_count": 96,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([1, 768])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 96
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0o_NiS3WgOFf"
      },
      "source": [
        "We can use all of this knowledge to create a classifier that uses the BERT model:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "m_mRflxPl32F"
      },
      "source": [
        "class SentimentClassifier(nn.Module):\n",
        "\n",
        "  def __init__(self, n_classes):\n",
        "    super(SentimentClassifier, self).__init__()\n",
        "    self.bert = BertModel.from_pretrained(PRE_TRAINED_MODEL_NAME)\n",
        "    self.drop = nn.Dropout(p=0.3)\n",
        "    self.out = nn.Linear(self.bert.config.hidden_size, n_classes)\n",
        "  \n",
        "  def forward(self, input_ids, attention_mask):\n",
        "    _, pooled_output = self.bert(\n",
        "      input_ids=input_ids,\n",
        "      attention_mask=attention_mask\n",
        "    ).to_tuple()\n",
        "    output = self.drop(pooled_output)\n",
        "    return self.out(output)"
      ],
      "execution_count": 148,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UJg8m3NQJahc"
      },
      "source": [
        "Our classifier delegates most of the heavy lifting to the BertModel. We use a dropout layer for some regularization and a fully-connected layer for our output. Note that we're returning the raw output of the last layer since that is required for the cross-entropy loss function in PyTorch to work.\n",
        "\n",
        "This should work like any other PyTorch model. Let's create an instance and move it to the GPU:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GbWQwFLU4DBW",
        "outputId": "8ac19e2a-0ed8-42ae-a0ad-aeef28b68b9b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 137
        }
      },
      "source": [
        "  torch.cuda.memory_summary(device=None, abbreviated=False)"
      ],
      "execution_count": 138,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'|===========================================================================|\\n|                  PyTorch CUDA memory summary, device ID 0                 |\\n|---------------------------------------------------------------------------|\\n|            CUDA OOMs: 4            |        cudaMalloc retries: 5         |\\n|===========================================================================|\\n|        Metric         | Cur Usage  | Peak Usage | Tot Alloc  | Tot Freed  |\\n|---------------------------------------------------------------------------|\\n| Allocated memory      |   13551 MB |   13644 MB |  323102 MB |  309550 MB |\\n|       from large pool |   13542 MB |   13634 MB |  322915 MB |  309372 MB |\\n|       from small pool |       9 MB |       9 MB |     186 MB |     177 MB |\\n|---------------------------------------------------------------------------|\\n| Active memory         |   13551 MB |   13644 MB |  323102 MB |  309550 MB |\\n|       from large pool |   13542 MB |   13634 MB |  322915 MB |  309372 MB |\\n|       from small pool |       9 MB |       9 MB |     186 MB |     177 MB |\\n|---------------------------------------------------------------------------|\\n| GPU reserved memory   |   14080 MB |   14162 MB |   15732 MB |    1652 MB |\\n|       from large pool |   14070 MB |   14150 MB |   15720 MB |    1650 MB |\\n|       from small pool |      10 MB |      12 MB |      12 MB |       2 MB |\\n|---------------------------------------------------------------------------|\\n| Non-releasable memory |  541016 KB |     893 MB |  208022 MB |  207493 MB |\\n|       from large pool |  540076 KB |     891 MB |  207728 MB |  207201 MB |\\n|       from small pool |     940 KB |       2 MB |     293 MB |     292 MB |\\n|---------------------------------------------------------------------------|\\n| Allocations           |    3454    |    3479    |   65500    |   62046    |\\n|       from large pool |    1559    |    1569    |   32593    |   31034    |\\n|       from small pool |    1895    |    1913    |   32907    |   31012    |\\n|---------------------------------------------------------------------------|\\n| Active allocs         |    3454    |    3479    |   65500    |   62046    |\\n|       from large pool |    1559    |    1569    |   32593    |   31034    |\\n|       from small pool |    1895    |    1913    |   32907    |   31012    |\\n|---------------------------------------------------------------------------|\\n| GPU reserved segments |     592    |     599    |     658    |      66    |\\n|       from large pool |     587    |     594    |     652    |      65    |\\n|       from small pool |       5    |       6    |       6    |       1    |\\n|---------------------------------------------------------------------------|\\n| Non-releasable allocs |     392    |     420    |   33669    |   33277    |\\n|       from large pool |     319    |     347    |   20253    |   19934    |\\n|       from small pool |      73    |      75    |   13416    |   13343    |\\n|===========================================================================|\\n'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 138
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "i0yQnuSFsjDp",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 340
        },
        "outputId": "abbff27a-4099-453a-cfe0-43add5e1981f"
      },
      "source": [
        "model = SentimentClassifier(len(class_names))\n",
        "model = model.to(device)"
      ],
      "execution_count": 149,
      "outputs": [
        {
          "output_type": "error",
          "ename": "RuntimeError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mRuntimeError\u001b[0m                              Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-149-83098f2c468c>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0mmodel\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mSentimentClassifier\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclass_names\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mmodel\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mto\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdevice\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36mto\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m    610\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mto\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdevice\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mis_floating_point\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnon_blocking\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    611\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 612\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_apply\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconvert\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    613\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    614\u001b[0m     def register_backward_hook(\n",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_apply\u001b[0;34m(self, fn)\u001b[0m\n\u001b[1;32m    357\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_apply\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    358\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mmodule\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mchildren\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 359\u001b[0;31m             \u001b[0mmodule\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_apply\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    360\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    361\u001b[0m         \u001b[0;32mdef\u001b[0m \u001b[0mcompute_should_use_set_data\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtensor\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtensor_applied\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_apply\u001b[0;34m(self, fn)\u001b[0m\n\u001b[1;32m    357\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_apply\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    358\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mmodule\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mchildren\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 359\u001b[0;31m             \u001b[0mmodule\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_apply\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    360\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    361\u001b[0m         \u001b[0;32mdef\u001b[0m \u001b[0mcompute_should_use_set_data\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtensor\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtensor_applied\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_apply\u001b[0;34m(self, fn)\u001b[0m\n\u001b[1;32m    357\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_apply\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    358\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mmodule\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mchildren\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 359\u001b[0;31m             \u001b[0mmodule\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_apply\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    360\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    361\u001b[0m         \u001b[0;32mdef\u001b[0m \u001b[0mcompute_should_use_set_data\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtensor\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtensor_applied\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_apply\u001b[0;34m(self, fn)\u001b[0m\n\u001b[1;32m    379\u001b[0m                 \u001b[0;31m# `with torch.no_grad():`\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    380\u001b[0m                 \u001b[0;32mwith\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mno_grad\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 381\u001b[0;31m                     \u001b[0mparam_applied\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mparam\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    382\u001b[0m                 \u001b[0mshould_use_set_data\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcompute_should_use_set_data\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mparam\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mparam_applied\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    383\u001b[0m                 \u001b[0;32mif\u001b[0m \u001b[0mshould_use_set_data\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36mconvert\u001b[0;34m(t)\u001b[0m\n\u001b[1;32m    608\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mconvert_to_format\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdim\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m4\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    609\u001b[0m                 \u001b[0;32mreturn\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mto\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdevice\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mis_floating_point\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnon_blocking\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmemory_format\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mconvert_to_format\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 610\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mto\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdevice\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mis_floating_point\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnon_blocking\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    611\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    612\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_apply\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconvert\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mRuntimeError\u001b[0m: CUDA out of memory. Tried to allocate 86.00 MiB (GPU 0; 14.76 GiB total capacity; 13.23 GiB already allocated; 87.75 MiB free; 13.75 GiB reserved in total by PyTorch)"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VCPCFDLlKIQd"
      },
      "source": [
        "We'll move the example batch of our training data to the GPU:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mz7p__CqdaMO",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d183fe32-0cf1-42c7-aa6f-740063fbdc8b"
      },
      "source": [
        "input_ids = data['input_ids'].to(device)\n",
        "attention_mask = data['attention_mask'].to(device)\n",
        "\n",
        "print(input_ids.shape) # batch size x seq length\n",
        "print(attention_mask.shape) # batch size x seq length"
      ],
      "execution_count": 118,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "torch.Size([16, 209])\n",
            "torch.Size([16, 209])\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oEZTFFt4DbiN",
        "outputId": "c7c4abdb-4caf-4877-9299-39ec32c754db"
      },
      "source": [
        "type(input_ids)"
      ],
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Tensor"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 61
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Hr1EgkEtKOIB"
      },
      "source": [
        "To get the predicted probabilities from our trained model, we'll apply the softmax function to the outputs:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2rTCj46Zamry",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e0cc8774-666b-4354-c925-766a5def3156"
      },
      "source": [
        "F.softmax(model(input_ids, attention_mask), dim=1)"
      ],
      "execution_count": 119,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[0.6215, 0.1849, 0.1936],\n",
              "        [0.6676, 0.1809, 0.1515],\n",
              "        [0.6772, 0.2088, 0.1140],\n",
              "        [0.5668, 0.1237, 0.3096],\n",
              "        [0.4701, 0.3390, 0.1909],\n",
              "        [0.4209, 0.4560, 0.1230],\n",
              "        [0.5589, 0.2836, 0.1575],\n",
              "        [0.5501, 0.2461, 0.2038],\n",
              "        [0.4254, 0.3482, 0.2264],\n",
              "        [0.6456, 0.0981, 0.2564],\n",
              "        [0.5474, 0.2145, 0.2381],\n",
              "        [0.5172, 0.2427, 0.2401],\n",
              "        [0.6003, 0.2163, 0.1834],\n",
              "        [0.4043, 0.3583, 0.2374],\n",
              "        [0.6919, 0.1512, 0.1569],\n",
              "        [0.5742, 0.1457, 0.2801]], device='cuda:0', grad_fn=<SoftmaxBackward>)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 119
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "g9xikRdtRN1N"
      },
      "source": [
        "### Training"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "76g7FV85H-T8"
      },
      "source": [
        "To reproduce the training procedure from the BERT paper, we'll use the [AdamW](https://huggingface.co/transformers/main_classes/optimizer_schedules.html#adamw) optimizer provided by Hugging Face. It corrects weight decay, so it's similar to the original paper. We'll also use a linear scheduler with no warmup steps:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5v-ArJ2fCCcU"
      },
      "source": [
        "EPOCHS = 12\n",
        "\n",
        "optimizer = AdamW(model.parameters(), lr=5e-5, correct_bias=False)\n",
        "total_steps = len(train_data_loader) * EPOCHS\n",
        "\n",
        "scheduler = get_linear_schedule_with_warmup(\n",
        "  optimizer,\n",
        "  num_warmup_steps=0,\n",
        "  num_training_steps=total_steps\n",
        ")\n",
        "\n",
        "loss_fn = nn.CrossEntropyLoss().to(device)"
      ],
      "execution_count": 120,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "A8522g7JIu5J"
      },
      "source": [
        "How do we come up with all hyperparameters? The BERT authors have some recommendations for fine-tuning:\n",
        "\n",
        "- Batch size: 16, 32\n",
        "- Learning rate (Adam): 5e-5, 3e-5, 2e-5\n",
        "- Number of epochs: 2, 3, 4\n",
        "\n",
        "We're going to ignore the number of epochs recommendation but stick with the rest. Note that increasing the batch size reduces the training time significantly, but gives you lower accuracy.\n",
        "\n",
        "Let's continue with writing a helper function for training our model for one epoch:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bzl9UhuNx1_Q"
      },
      "source": [
        "def train_epoch(\n",
        "  model, \n",
        "  data_loader, \n",
        "  loss_fn, \n",
        "  optimizer, \n",
        "  device, \n",
        "  scheduler, \n",
        "  n_examples\n",
        "):\n",
        "  model = model.train()\n",
        "\n",
        "  losses = []\n",
        "  correct_predictions = 0\n",
        "  \n",
        "  for d in data_loader:\n",
        "    input_ids = d[\"input_ids\"].to(device)\n",
        "    attention_mask = d[\"attention_mask\"].to(device)\n",
        "    targets = d[\"targets\"].to(device)\n",
        "\n",
        "    outputs = model(\n",
        "      input_ids=input_ids,\n",
        "      attention_mask=attention_mask\n",
        "    )\n",
        "\n",
        "    _, preds = torch.max(outputs, dim=1)\n",
        "    loss = loss_fn(outputs, targets)\n",
        "\n",
        "    correct_predictions += torch.sum(preds == targets)\n",
        "    losses.append(loss.item())\n",
        "\n",
        "    loss.backward()\n",
        "    nn.utils.clip_grad_norm_(model.parameters(), max_norm=1.0)\n",
        "    optimizer.step()\n",
        "    scheduler.step()\n",
        "    optimizer.zero_grad()\n",
        "\n",
        "  return correct_predictions.double() / n_examples, np.mean(losses)"
      ],
      "execution_count": 121,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "E4PniYIte0fr"
      },
      "source": [
        "Training the model should look familiar, except for two things. The scheduler gets called every time a batch is fed to the model. We're avoiding exploding gradients by clipping the gradients of the model using [clip_grad_norm_](https://pytorch.org/docs/stable/nn.html#clip-grad-norm).\n",
        "\n",
        "Let's write another one that helps us evaluate the model on a given data loader:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CXeRorVGIKre"
      },
      "source": [
        "def eval_model(model, data_loader, loss_fn, device, n_examples):\n",
        "  model = model.eval()\n",
        "\n",
        "  losses = []\n",
        "  correct_predictions = 0\n",
        "\n",
        "  with torch.no_grad():\n",
        "    for d in data_loader:\n",
        "      input_ids = d[\"input_ids\"].to(device)\n",
        "      attention_mask = d[\"attention_mask\"].to(device)\n",
        "      targets = d[\"targets\"].to(device)\n",
        "\n",
        "      outputs = model(\n",
        "        input_ids=input_ids,\n",
        "        attention_mask=attention_mask\n",
        "      )\n",
        "      _, preds = torch.max(outputs, dim=1)\n",
        "\n",
        "      loss = loss_fn(outputs, targets)\n",
        "\n",
        "      correct_predictions += torch.sum(preds == targets)\n",
        "      losses.append(loss.item())\n",
        "\n",
        "  return correct_predictions.double() / n_examples, np.mean(losses)"
      ],
      "execution_count": 122,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "a_rdSDBHhhCh"
      },
      "source": [
        "Using those two, we can write our training loop. We'll also store the training history:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1zhHoFNsxufs",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 510
        },
        "outputId": "51c7d27e-6505-4eca-ae7c-81ccdad0e5e7"
      },
      "source": [
        "%%time\n",
        "\n",
        "history = defaultdict(list)\n",
        "best_accuracy = 0\n",
        "\n",
        "for epoch in range(EPOCHS):\n",
        "\n",
        "  print(f'Epoch {epoch + 1}/{EPOCHS}')\n",
        "  print('-' * 10)\n",
        "\n",
        "  train_acc, train_loss = train_epoch(\n",
        "    model,\n",
        "    train_data_loader,    \n",
        "    loss_fn, \n",
        "    optimizer, \n",
        "    device, \n",
        "    scheduler, \n",
        "    len(df_train)\n",
        "  )\n",
        "\n",
        "  print(f'Train loss {train_loss} accuracy {train_acc}')\n",
        "\n",
        "  val_acc, val_loss = eval_model(\n",
        "    model,\n",
        "    val_data_loader,\n",
        "    loss_fn, \n",
        "    device, \n",
        "    len(df_val)\n",
        "  )\n",
        "\n",
        "  print(f'Val   loss {val_loss} accuracy {val_acc}')\n",
        "  print()\n",
        "\n",
        "  history['train_acc'].append(train_acc)\n",
        "  history['train_loss'].append(train_loss)\n",
        "  history['val_acc'].append(val_acc)\n",
        "  history['val_loss'].append(val_loss)\n",
        "\n",
        "  if val_acc > best_accuracy:\n",
        "    torch.save(model.state_dict(), 'best_model_state.bin')\n",
        "    best_accuracy = val_acc"
      ],
      "execution_count": 126,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/12\n",
            "----------\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "error",
          "ename": "RuntimeError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mRuntimeError\u001b[0m                              Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-126-ec27ef2a7e99>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mget_ipython\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun_cell_magic\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'time'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m''\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"\\nhistory = defaultdict(list)\\nbest_accuracy = 0\\n\\nfor epoch in range(EPOCHS):\\n\\n  print(f'Epoch {epoch + 1}/{EPOCHS}')\\n  print('-' * 10)\\n\\n  train_acc, train_loss = train_epoch(\\n    model,\\n    train_data_loader,    \\n    loss_fn, \\n    optimizer, \\n    device, \\n    scheduler, \\n    len(df_train)\\n  )\\n\\n  print(f'Train loss {train_loss} accuracy {train_acc}')\\n\\n  val_acc, val_loss = eval_model(\\n    model,\\n    val_data_loader,\\n    loss_fn, \\n    device, \\n    len(df_val)\\n  )\\n\\n  print(f'Val   loss {val_loss} accuracy {val_acc}')\\n  print()\\n\\n  history['train_acc'].append(train_acc)\\n  history['train_loss'].append(train_loss)\\n  history['val_acc'].append(val_acc)\\n  history['val_loss'].append(val_loss)\\n\\n  if val_acc > best_accuracy:\\n    torch.save(model.state_dict(), 'best_model_state.bin')\\n    best_accuracy = val_acc\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/IPython/core/interactiveshell.py\u001b[0m in \u001b[0;36mrun_cell_magic\u001b[0;34m(self, magic_name, line, cell)\u001b[0m\n\u001b[1;32m   2115\u001b[0m             \u001b[0mmagic_arg_s\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvar_expand\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mline\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstack_depth\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2116\u001b[0m             \u001b[0;32mwith\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbuiltin_trap\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2117\u001b[0;31m                 \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmagic_arg_s\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcell\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2118\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0mresult\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2119\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m<decorator-gen-60>\u001b[0m in \u001b[0;36mtime\u001b[0;34m(self, line, cell, local_ns)\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/IPython/core/magic.py\u001b[0m in \u001b[0;36m<lambda>\u001b[0;34m(f, *a, **k)\u001b[0m\n\u001b[1;32m    186\u001b[0m     \u001b[0;31m# but it's overkill for just that one bit of state.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    187\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mmagic_deco\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0marg\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 188\u001b[0;31m         \u001b[0mcall\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mlambda\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mk\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mk\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    189\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    190\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mcallable\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0marg\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/IPython/core/magics/execution.py\u001b[0m in \u001b[0;36mtime\u001b[0;34m(self, line, cell, local_ns)\u001b[0m\n\u001b[1;32m   1191\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1192\u001b[0m             \u001b[0mst\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mclock2\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1193\u001b[0;31m             \u001b[0mexec\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcode\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mglob\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlocal_ns\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1194\u001b[0m             \u001b[0mend\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mclock2\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1195\u001b[0m             \u001b[0mout\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m<timed exec>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n",
            "\u001b[0;32m<ipython-input-121-46992c3ae043>\u001b[0m in \u001b[0;36mtrain_epoch\u001b[0;34m(model, data_loader, loss_fn, optimizer, device, scheduler, n_examples)\u001b[0m\n\u001b[1;32m     20\u001b[0m     outputs = model(\n\u001b[1;32m     21\u001b[0m       \u001b[0minput_ids\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0minput_ids\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 22\u001b[0;31m       \u001b[0mattention_mask\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mattention_mask\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     23\u001b[0m     )\n\u001b[1;32m     24\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_call_impl\u001b[0;34m(self, *input, **kwargs)\u001b[0m\n\u001b[1;32m    725\u001b[0m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_slow_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    726\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 727\u001b[0;31m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    728\u001b[0m         for hook in itertools.chain(\n\u001b[1;32m    729\u001b[0m                 \u001b[0m_global_forward_hooks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m<ipython-input-116-c7722bd2dde2>\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, input_ids, attention_mask)\u001b[0m\n\u001b[1;32m     10\u001b[0m     _, pooled_output = self.bert(\n\u001b[1;32m     11\u001b[0m       \u001b[0minput_ids\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0minput_ids\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 12\u001b[0;31m       \u001b[0mattention_mask\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mattention_mask\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     13\u001b[0m     ).to_tuple()\n\u001b[1;32m     14\u001b[0m     \u001b[0moutput\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdrop\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpooled_output\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_call_impl\u001b[0;34m(self, *input, **kwargs)\u001b[0m\n\u001b[1;32m    725\u001b[0m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_slow_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    726\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 727\u001b[0;31m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    728\u001b[0m         for hook in itertools.chain(\n\u001b[1;32m    729\u001b[0m                 \u001b[0m_global_forward_hooks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/transformers/models/bert/modeling_bert.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, input_ids, attention_mask, token_type_ids, position_ids, head_mask, inputs_embeds, encoder_hidden_states, encoder_attention_mask, past_key_values, use_cache, output_attentions, output_hidden_states, return_dict)\u001b[0m\n\u001b[1;32m    962\u001b[0m             \u001b[0mtoken_type_ids\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mtoken_type_ids\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    963\u001b[0m             \u001b[0minputs_embeds\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0minputs_embeds\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 964\u001b[0;31m             \u001b[0mpast_key_values_length\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mpast_key_values_length\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    965\u001b[0m         )\n\u001b[1;32m    966\u001b[0m         encoder_outputs = self.encoder(\n",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_call_impl\u001b[0;34m(self, *input, **kwargs)\u001b[0m\n\u001b[1;32m    725\u001b[0m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_slow_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    726\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 727\u001b[0;31m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    728\u001b[0m         for hook in itertools.chain(\n\u001b[1;32m    729\u001b[0m                 \u001b[0m_global_forward_hooks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/transformers/models/bert/modeling_bert.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, input_ids, token_type_ids, position_ids, inputs_embeds, past_key_values_length)\u001b[0m\n\u001b[1;32m    206\u001b[0m             \u001b[0membeddings\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0mposition_embeddings\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    207\u001b[0m         \u001b[0membeddings\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mLayerNorm\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0membeddings\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 208\u001b[0;31m         \u001b[0membeddings\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdropout\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0membeddings\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    209\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0membeddings\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    210\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_call_impl\u001b[0;34m(self, *input, **kwargs)\u001b[0m\n\u001b[1;32m    725\u001b[0m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_slow_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    726\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 727\u001b[0;31m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    728\u001b[0m         for hook in itertools.chain(\n\u001b[1;32m    729\u001b[0m                 \u001b[0m_global_forward_hooks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/dropout.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, input)\u001b[0m\n\u001b[1;32m     56\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     57\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mTensor\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m->\u001b[0m \u001b[0mTensor\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 58\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mF\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdropout\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtraining\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minplace\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     59\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     60\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/functional.py\u001b[0m in \u001b[0;36mdropout\u001b[0;34m(input, p, training, inplace)\u001b[0m\n\u001b[1;32m    981\u001b[0m     return (_VF.dropout_(input, p, training)\n\u001b[1;32m    982\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0minplace\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 983\u001b[0;31m             else _VF.dropout(input, p, training))\n\u001b[0m\u001b[1;32m    984\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    985\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mRuntimeError\u001b[0m: CUDA out of memory. Tried to allocate 20.00 MiB (GPU 0; 14.76 GiB total capacity; 13.32 GiB already allocated; 5.75 MiB free; 13.83 GiB reserved in total by PyTorch)"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4r8-5zWsiVur"
      },
      "source": [
        "Note that we're storing the state of the best model, indicated by the highest validation accuracy."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wLQf52c7fbzr"
      },
      "source": [
        "Whoo, this took some time! We can look at the training vs validation accuracy:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-FWG7kBm372V",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 523
        },
        "outputId": "cc97304f-348e-4d74-ca02-5af121c2ad99"
      },
      "source": [
        "plt.plot(history['train_acc'], label='train accuracy')\n",
        "plt.plot(history['val_acc'], label='validation accuracy')\n",
        "\n",
        "plt.title('Training history')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend()\n",
        "plt.ylim([0, 1]);"
      ],
      "execution_count": 166,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "image/png": {
              "width": 732,
              "height": 506
            }
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZsHqkLAuf8pv"
      },
      "source": [
        "The training accuracy starts to approach 100% after 10 epochs or so. You might try to fine-tune the parameters a bit more, but this will be good enough for us.\n",
        "\n",
        "Don't want to wait? Uncomment the next cell to download my pre-trained model:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zoGUH8VZ-pPQ"
      },
      "source": [
        "# !gdown --id 1V8itWtowCYnb2Bc9KlK9SxGff9WwmogA\n",
        "\n",
        "# model = SentimentClassifier(len(class_names))\n",
        "# model.load_state_dict(torch.load('best_model_state.bin'))\n",
        "# model = model.to(device)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "U3HZb3NWFtFf"
      },
      "source": [
        "## Evaluation\n",
        "\n",
        "So how good is our model on predicting sentiment? Let's start by calculating the accuracy on the test data:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jS3gJ_qBEljD",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "06706905-1f7c-4373-a3fe-5ca30896910d"
      },
      "source": [
        "test_acc, _ = eval_model(\n",
        "  model,\n",
        "  test_data_loader,\n",
        "  loss_fn,\n",
        "  device,\n",
        "  len(df_test)\n",
        ")\n",
        "\n",
        "test_acc.item()"
      ],
      "execution_count": 167,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py:2155: FutureWarning: The `pad_to_max_length` argument is deprecated and will be removed in a future version, use `padding=True` or `padding='longest'` to pad to the longest sequence in the batch, or use `padding='max_length'` to pad to a max length. In this case, you can give a specific length with `max_length` (e.g. `max_length=45`) or leave max_length to None to pad to the maximal input size of the model (e.g. 512 for Bert).\n",
            "  FutureWarning,\n",
            "/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py:2155: FutureWarning: The `pad_to_max_length` argument is deprecated and will be removed in a future version, use `padding=True` or `padding='longest'` to pad to the longest sequence in the batch, or use `padding='max_length'` to pad to a max length. In this case, you can give a specific length with `max_length` (e.g. `max_length=45`) or leave max_length to None to pad to the maximal input size of the model (e.g. 512 for Bert).\n",
            "  FutureWarning,\n",
            "/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py:2155: FutureWarning: The `pad_to_max_length` argument is deprecated and will be removed in a future version, use `padding=True` or `padding='longest'` to pad to the longest sequence in the batch, or use `padding='max_length'` to pad to a max length. In this case, you can give a specific length with `max_length` (e.g. `max_length=45`) or leave max_length to None to pad to the maximal input size of the model (e.g. 512 for Bert).\n",
            "  FutureWarning,\n",
            "/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py:2155: FutureWarning: The `pad_to_max_length` argument is deprecated and will be removed in a future version, use `padding=True` or `padding='longest'` to pad to the longest sequence in the batch, or use `padding='max_length'` to pad to a max length. In this case, you can give a specific length with `max_length` (e.g. `max_length=45`) or leave max_length to None to pad to the maximal input size of the model (e.g. 512 for Bert).\n",
            "  FutureWarning,\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.8553299492385786"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 167
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mdQ7-ylCj8Gd"
      },
      "source": [
        "The accuracy is about 1% lower on the test set. Our model seems to generalize well.\n",
        "\n",
        "We'll define a helper function to get the predictions from our model:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EgR6MuNS8jr_"
      },
      "source": [
        "def get_predictions(model, data_loader):\n",
        "  model = model.eval()\n",
        "  \n",
        "  review_texts = []\n",
        "  predictions = []\n",
        "  prediction_probs = []\n",
        "  real_values = []\n",
        "\n",
        "  with torch.no_grad():\n",
        "    for d in data_loader:\n",
        "\n",
        "      texts = d[\"review_text\"]\n",
        "      input_ids = d[\"input_ids\"].to(device)\n",
        "      attention_mask = d[\"attention_mask\"].to(device)\n",
        "      targets = d[\"targets\"].to(device)\n",
        "\n",
        "      outputs = model(\n",
        "        input_ids=input_ids,\n",
        "        attention_mask=attention_mask\n",
        "      )\n",
        "      _, preds = torch.max(outputs, dim=1)\n",
        "\n",
        "      probs = F.softmax(outputs, dim=1)\n",
        "\n",
        "      review_texts.extend(texts)\n",
        "      predictions.extend(preds)\n",
        "      prediction_probs.extend(probs)\n",
        "      real_values.extend(targets)\n",
        "\n",
        "  predictions = torch.stack(predictions).cpu()\n",
        "  prediction_probs = torch.stack(prediction_probs).cpu()\n",
        "  real_values = torch.stack(real_values).cpu()\n",
        "  return review_texts, predictions, prediction_probs, real_values"
      ],
      "execution_count": 168,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dkbnBTI7kd_y"
      },
      "source": [
        "This is similar to the evaluation function, except that we're storing the text of the reviews and the predicted probabilities (by applying the softmax on the model outputs):"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zHdPZr60-0c_",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "347e3a35-c26b-4135-b2e2-7c3ff71f58d4"
      },
      "source": [
        "y_review_texts, y_pred, y_pred_probs, y_test = get_predictions(\n",
        "  model,\n",
        "  test_data_loader\n",
        ")"
      ],
      "execution_count": 169,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py:2155: FutureWarning: The `pad_to_max_length` argument is deprecated and will be removed in a future version, use `padding=True` or `padding='longest'` to pad to the longest sequence in the batch, or use `padding='max_length'` to pad to a max length. In this case, you can give a specific length with `max_length` (e.g. `max_length=45`) or leave max_length to None to pad to the maximal input size of the model (e.g. 512 for Bert).\n",
            "  FutureWarning,\n",
            "/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py:2155: FutureWarning: The `pad_to_max_length` argument is deprecated and will be removed in a future version, use `padding=True` or `padding='longest'` to pad to the longest sequence in the batch, or use `padding='max_length'` to pad to a max length. In this case, you can give a specific length with `max_length` (e.g. `max_length=45`) or leave max_length to None to pad to the maximal input size of the model (e.g. 512 for Bert).\n",
            "  FutureWarning,\n",
            "/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py:2155: FutureWarning: The `pad_to_max_length` argument is deprecated and will be removed in a future version, use `padding=True` or `padding='longest'` to pad to the longest sequence in the batch, or use `padding='max_length'` to pad to a max length. In this case, you can give a specific length with `max_length` (e.g. `max_length=45`) or leave max_length to None to pad to the maximal input size of the model (e.g. 512 for Bert).\n",
            "  FutureWarning,\n",
            "/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py:2155: FutureWarning: The `pad_to_max_length` argument is deprecated and will be removed in a future version, use `padding=True` or `padding='longest'` to pad to the longest sequence in the batch, or use `padding='max_length'` to pad to a max length. In this case, you can give a specific length with `max_length` (e.g. `max_length=45`) or leave max_length to None to pad to the maximal input size of the model (e.g. 512 for Bert).\n",
            "  FutureWarning,\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gVwoVij2lC7F"
      },
      "source": [
        "Let's have a look at the classification report"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "L8a9_8-ND3Is",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b55a3511-db89-449c-a2fc-d1b1825ab18c"
      },
      "source": [
        "print(classification_report(y_test, y_pred, target_names=class_names))"
      ],
      "execution_count": 170,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "    negative       0.89      0.84      0.87       245\n",
            "     neutral       0.77      0.85      0.81       254\n",
            "    positive       0.91      0.87      0.89       289\n",
            "\n",
            "    accuracy                           0.86       788\n",
            "   macro avg       0.86      0.85      0.86       788\n",
            "weighted avg       0.86      0.86      0.86       788\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rFAekw3mmWUi"
      },
      "source": [
        "Looks like it is really hard to classify neutral (3 stars) reviews. And I can tell you from experience, looking at many reviews, those are hard to classify.\n",
        "\n",
        "We'll continue with the confusion matrix:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6d1qxsc__DTh",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 537
        },
        "outputId": "06813562-1c27-4e98-95b8-57545f1f2551"
      },
      "source": [
        "def show_confusion_matrix(confusion_matrix):\n",
        "  hmap = sns.heatmap(confusion_matrix, annot=True, fmt=\"d\", cmap=\"Blues\")\n",
        "  hmap.yaxis.set_ticklabels(hmap.yaxis.get_ticklabels(), rotation=0, ha='right')\n",
        "  hmap.xaxis.set_ticklabels(hmap.xaxis.get_ticklabels(), rotation=30, ha='right')\n",
        "  plt.ylabel('True sentiment')\n",
        "  plt.xlabel('Predicted sentiment');\n",
        "\n",
        "cm = confusion_matrix(y_test, y_pred)\n",
        "df_cm = pd.DataFrame(cm, index=class_names, columns=class_names)\n",
        "show_confusion_matrix(df_cm)"
      ],
      "execution_count": 171,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x576 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "image/png": {
              "width": 725,
              "height": 520
            }
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wx0U7oNsnZ3A"
      },
      "source": [
        "This confirms that our model is having difficulty classifying neutral reviews. It mistakes those for negative and positive at a roughly equal frequency.\n",
        "\n",
        "That's a good overview of the performance of our model. But let's have a look at an example from our test data:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iANBiY3sLo-K"
      },
      "source": [
        "idx = 2\n",
        "\n",
        "review_text = y_review_texts[idx]\n",
        "true_sentiment = y_test[idx]\n",
        "pred_df = pd.DataFrame({\n",
        "  'class_names': class_names,\n",
        "  'values': y_pred_probs[idx]\n",
        "})"
      ],
      "execution_count": 172,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-8D0rb1yfnv4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "cdb8abf4-7bae-4543-d770-671c03bbb4ae"
      },
      "source": [
        "print(\"\\n\".join(wrap(review_text)))\n",
        "print()\n",
        "print(f'True sentiment: {class_names[true_sentiment]}')"
      ],
      "execution_count": 173,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "I used to use Habitica, and I must say this is a great step up. I'd\n",
            "like to see more social features, such as sharing tasks - only one\n",
            "person has to perform said task for it to be checked off, but only\n",
            "giving that person the experience and gold. Otherwise, the price for\n",
            "subscription is too steep, thus resulting in a sub-perfect score. I\n",
            "could easily justify $0.99/month or eternal subscription for $15. If\n",
            "that price could be met, as well as fine tuning, this would be easily\n",
            "worth 5 stars.\n",
            "\n",
            "True sentiment: neutral\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "f7hj_IZFnn2X"
      },
      "source": [
        "Now we can look at the confidence of each sentiment of our model:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qj4d8lZyMkhf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 506
        },
        "outputId": "f5209992-b157-4a9f-bbb0-730746d1cce8"
      },
      "source": [
        "sns.barplot(x='values', y='class_names', data=pred_df, orient='h')\n",
        "plt.ylabel('sentiment')\n",
        "plt.xlabel('probability')\n",
        "plt.xlim([0, 1]);"
      ],
      "execution_count": 174,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "image/png": {
              "width": 779,
              "height": 489
            }
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7WL5pDmvFyaU"
      },
      "source": [
        "### Predicting on Raw Text\n",
        "\n",
        "Let's use our model to predict the sentiment of some raw text:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QEPi7zQRsDhH"
      },
      "source": [
        "review_text = \" Best app eve, it is really good!!!\""
      ],
      "execution_count": 183,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GaN4RnqMnxYw"
      },
      "source": [
        "We have to use the tokenizer to encode the text:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zA5Or4D2sLc9",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "41b2b16d-7e9f-49c2-ea9e-ee5192166226"
      },
      "source": [
        "encoded_review = tokenizer.encode_plus(\n",
        "  review_text,\n",
        "  max_length=MAX_LEN,\n",
        "  add_special_tokens=True,\n",
        "  return_token_type_ids=False,\n",
        "  pad_to_max_length=True,\n",
        "  return_attention_mask=True,\n",
        "  return_tensors='pt',\n",
        ")"
      ],
      "execution_count": 184,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py:2155: FutureWarning: The `pad_to_max_length` argument is deprecated and will be removed in a future version, use `padding=True` or `padding='longest'` to pad to the longest sequence in the batch, or use `padding='max_length'` to pad to a max length. In this case, you can give a specific length with `max_length` (e.g. `max_length=45`) or leave max_length to None to pad to the maximal input size of the model (e.g. 512 for Bert).\n",
            "  FutureWarning,\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "et8xlDrKpH60"
      },
      "source": [
        "Let's get the predictions from our model:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Qr_t3rUksumr",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "77f17661-7a61-43f5-a430-bba58171b9ae"
      },
      "source": [
        "input_ids = encoded_review['input_ids'].to(device)\n",
        "attention_mask = encoded_review['attention_mask'].to(device)\n",
        "\n",
        "output = model(input_ids, attention_mask)\n",
        "_, prediction = torch.max(output, dim=1)\n",
        "\n",
        "print(f'Review text: {review_text}')\n",
        "print(f'Sentiment  : {class_names[prediction]}')"
      ],
      "execution_count": 185,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Review text:  Best app eve, it is really good!!!\n",
            "Sentiment  : neutral\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PVhwzq7bpPRl"
      },
      "source": [
        "## Summary\n",
        "\n",
        "Nice job! You learned how to use BERT for sentiment analysis. You built a custom classifier using the Hugging Face library and trained it on our app reviews dataset!\n",
        "\n",
        "- [Read the tutorial](https://www.curiousily.com/posts/sentiment-analysis-with-bert-and-hugging-face-using-pytorch-and-python/)\n",
        "- [Run the notebook in your browser (Google Colab)](https://colab.research.google.com/drive/1PHv-IRLPCtv7oTcIGbsgZHqrB5LPvB7S)\n",
        "- [Read the `Getting Things Done with Pytorch` book](https://github.com/curiousily/Getting-Things-Done-with-Pytorch)\n",
        "\n",
        "You learned how to:\n",
        "\n",
        "- Intuitively understand what BERT is\n",
        "- Preprocess text data for BERT and build PyTorch Dataset (tokenization, attention masks, and padding)\n",
        "- Use Transfer Learning to build Sentiment Classifier using the Transformers library by Hugging Face\n",
        "- Evaluate the model on test data\n",
        "- Predict sentiment on raw text\n",
        "\n",
        "Next, we'll learn how to deploy our trained model behind a REST API and build a simple web app to access it."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Wf39tauBa2V2"
      },
      "source": [
        "## References\n",
        "\n",
        "- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://arxiv.org/abs/1810.04805)\n",
        "- [L11 Language Models - Alec Radford (OpenAI)](https://www.youtube.com/watch?v=BnpB3GrpsfM)\n",
        "- [The Illustrated BERT, ELMo, and co.](https://jalammar.github.io/illustrated-bert/)\n",
        "- [BERT Fine-Tuning Tutorial with PyTorch](https://mccormickml.com/2019/07/22/BERT-fine-tuning/)\n",
        "- [How to Fine-Tune BERT for Text Classification?](https://arxiv.org/pdf/1905.05583.pdf)\n",
        "- [Huggingface Transformers](https://huggingface.co/transformers/)\n",
        "- [BERT Explained: State of the art language model for NLP](https://towardsdatascience.com/bert-explained-state-of-the-art-language-model-for-nlp-f8b21a9b6270)"
      ]
    }
  ]
}