NNLM_using_tf_keras.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "NNLM_using_tf_keras.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": []
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.1"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "s89hB9f_ogWR",
        "colab_type": "text"
      },
      "source": [
        "## Using the [`nnlm`](https://tfhub.dev/google/tf2-preview/nnlm-en-dim128/1) embeddings to classify IMDB reviews\n",
        "\n",
        "The motivation behind this notebook comes from the this tutorial [Deep Transfer Learning for Natural Language Processing — Text Classification with Universal Embeddings](https://towardsdatascience.com/deep-transfer-learning-for-natural-language-processing-text-classification-with-universal-1a2c69e5baa9) written by [Diapnjan Sarkar](https://in.linkedin.com/in/dipanzan). Specifically from the following lines:\n",
        "\n",
        "> Note: This tutorial is built using TensorFlow entirely given that they provide an easy access to the sentence encoders. However I’m not a big fan of their old APIs and I’m looking for someone to assist me on re-implementing the code using the tf.keras APIs instead of tf.estimator. Do reach out to me if you are interested in contributing and we can even feature your work on the same! (contact links in my profile and in the footer)\n",
        "\n",
        "I am a newbie in NLP and I have always found Dipanjan's tutorial to be extremely informative and well-written. He not only shows you how to deal with a problem statement with the best set of tools but also discusses the corresponding literature in a very broad manner. It is been quite a while that I am exploring the newly released `tensorflow-2.0.0alpha0` (and I am kind of loving it). So, I decided to refactor Dipanjan's code to `tf.keras` as wanted by him. "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FZbXpJoyqIzQ",
        "colab_type": "text"
      },
      "source": [
        "I had to use the CPU installation of `tensorflow-2.0.0alpha0` because the GPU installation is not stable yet to be compatible with [TensorFlow hub](https://tfhub.dev/google/tf2-preview) modules.\n",
        "\n",
        "In the above-mentioned tutorial, Dipanjan presented a hands-on demo on the two very popular sentence embeddings - \n",
        "- [Universal Sentence Encoder](https://arxiv.org/abs/1803.11175)\n",
        "- Neural-Net Language Models based on the seminal paper by Bengio et al. - [A Neural Probabilistic Language Model](http://www.jmlr.org/papers/volume3/bengio03a/bengio03a.pdf)\n",
        "\n",
        "**This notebook uses the `nnlm` model.**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-DtPB3s_sDAz",
        "colab_type": "text"
      },
      "source": [
        "### Library imports"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "MvwdaMOiTxHo",
        "colab": {}
      },
      "source": [
        "!pip install tensorflow==2.0.0-alpha0 \n",
        "!pip install tensorflow-hub\n",
        "!pip install contractions\n",
        "!pip install beautifulsoup4"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "k9faaqF6YqT4",
        "colab": {}
      },
      "source": [
        "import tensorflow as tf\n",
        "import tensorflow_hub as tf_hub\n",
        "import pandas as pd\n",
        "import numpy as np"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "QIwV82Kyxc0O",
        "outputId": "d7589e78-f086-4361-84de-2b86a800ade0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        }
      },
      "source": [
        "tf.test.is_gpu_available()"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "False"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jMqElM-hsGQ4",
        "colab_type": "text"
      },
      "source": [
        "### Data loading and previewing"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "QYUSjjxoZCiu",
        "outputId": "66c1a88d-e328-4032-b897-750b60aafce6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        }
      },
      "source": [
        "url = 'https://github.com/dipanjanS/data_science_for_all/blob/master/tds_deep_transfer_learning_nlp_classification/movie_reviews.csv.bz2?raw=true'\n",
        "dataset = pd.read_csv(url, compression='bz2')\n",
        "dataset.head()"
      ],
      "execution_count": 4,
      "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>review</th>\n",
              "      <th>sentiment</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>One of the other reviewers has mentioned that ...</td>\n",
              "      <td>positive</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>A wonderful little production. &lt;br /&gt;&lt;br /&gt;The...</td>\n",
              "      <td>positive</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>I thought this was a wonderful way to spend ti...</td>\n",
              "      <td>positive</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Basically there's a family where a little boy ...</td>\n",
              "      <td>negative</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Petter Mattei's \"Love in the Time of Money\" is...</td>\n",
              "      <td>positive</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                              review sentiment\n",
              "0  One of the other reviewers has mentioned that ...  positive\n",
              "1  A wonderful little production. <br /><br />The...  positive\n",
              "2  I thought this was a wonderful way to spend ti...  positive\n",
              "3  Basically there's a family where a little boy ...  negative\n",
              "4  Petter Mattei's \"Love in the Time of Money\" is...  positive"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "E7zulWXAZgMd",
        "outputId": "6797bbb2-829f-4bfc-fb59-e2845d92fbc4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 150
        }
      },
      "source": [
        "dataset.info()"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 50000 entries, 0 to 49999\n",
            "Data columns (total 2 columns):\n",
            "review       50000 non-null object\n",
            "sentiment    50000 non-null object\n",
            "dtypes: object(2)\n",
            "memory usage: 781.3+ KB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hGJEBk9AsJ95",
        "colab_type": "text"
      },
      "source": [
        "### Numericalization of the sentiments to ease the computations"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "3o5vhfZ9Ziut",
        "outputId": "bb42cc59-f179-4169-cc05-8db982601f17",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        }
      },
      "source": [
        "dataset['sentiment'] = dataset.sentiment.map({\"positive\": 1, \"negative\": 0})\n",
        "dataset.head()"
      ],
      "execution_count": 5,
      "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>review</th>\n",
              "      <th>sentiment</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>One of the other reviewers has mentioned that ...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>A wonderful little production. &lt;br /&gt;&lt;br /&gt;The...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>I thought this was a wonderful way to spend ti...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Basically there's a family where a little boy ...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Petter Mattei's \"Love in the Time of Money\" is...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                              review  sentiment\n",
              "0  One of the other reviewers has mentioned that ...          1\n",
              "1  A wonderful little production. <br /><br />The...          1\n",
              "2  I thought this was a wonderful way to spend ti...          1\n",
              "3  Basically there's a family where a little boy ...          0\n",
              "4  Petter Mattei's \"Love in the Time of Money\" is...          1"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bY3dkXfZsTUj",
        "colab_type": "text"
      },
      "source": [
        "### Data splits"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "BNHJkMfBZyNR",
        "outputId": "654a0908-a67f-4dc8-cd13-e33b6af6f0a2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        }
      },
      "source": [
        "reviews = dataset['review'].values\n",
        "sentiments = dataset['sentiment'].values\n",
        "\n",
        "train_reviews = reviews[:30000]\n",
        "train_sentiments = sentiments[:30000]\n",
        "\n",
        "val_reviews = reviews[30000:35000]\n",
        "val_sentiments = sentiments[30000:35000]\n",
        "\n",
        "test_reviews = reviews[35000:]\n",
        "test_sentiments = sentiments[35000:]\n",
        "\n",
        "train_reviews.shape, val_reviews.shape, test_reviews.shape"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((30000,), (5000,), (15000,))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "08Xx63zKsWvR",
        "colab_type": "text"
      },
      "source": [
        "### Little helper functions to preprocess the reviews\n",
        "\n",
        "These functions are all written by Dipanjan and I have copy-pasted them (yes, no shame). "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "5dwfmMjUaF_B",
        "colab": {}
      },
      "source": [
        "import contractions\n",
        "from bs4 import BeautifulSoup\n",
        "import unicodedata\n",
        "import re\n",
        "\n",
        "def strip_html_tags(text):\n",
        "    soup = BeautifulSoup(text, \"html.parser\")\n",
        "    [s.extract() for s in soup(['iframe', 'script'])]\n",
        "    stripped_text = soup.get_text()\n",
        "    stripped_text = re.sub(r'[\\r|\\n|\\r\\n]+', '\\n', stripped_text)\n",
        "    return stripped_text\n",
        "\n",
        "def remove_accented_chars(text):\n",
        "    text = unicodedata.normalize('NFKD', text).encode('ascii', 'ignore').decode('utf-8', 'ignore')\n",
        "    return text\n",
        "\n",
        "def expand_contractions(text):\n",
        "    return contractions.fix(text)\n",
        "\n",
        "def remove_special_characters(text, remove_digits=False):\n",
        "    pattern = r'[^a-zA-Z0-9\\s]' if not remove_digits else r'[^a-zA-Z\\s]'\n",
        "    text = re.sub(pattern, '', text)\n",
        "    return text\n",
        "\n",
        "def pre_process_document(document):\n",
        "    # strip HTML\n",
        "    document = strip_html_tags(document)\n",
        "    # lower case\n",
        "    document = document.lower()\n",
        "    # remove extra newlines (often might be present in really noisy text)\n",
        "    document = document.translate(document.maketrans(\"\\n\\t\\r\", \"   \"))\n",
        "    # remove accented characters\n",
        "    document = remove_accented_chars(document)\n",
        "    # expand contractions    \n",
        "    document = expand_contractions(document)  \n",
        "    # remove special characters and\\or digits    \n",
        "    # insert spaces between special characters to isolate them    \n",
        "    special_char_pattern = re.compile(r'([{.(-)!}])')\n",
        "    document = special_char_pattern.sub(\" \\\\1 \", document)\n",
        "    document = remove_special_characters(document, remove_digits=True)  \n",
        "    # remove extra whitespace\n",
        "    document = re.sub(' +', ' ', document)\n",
        "    document = document.strip()\n",
        "    \n",
        "    return document\n",
        "\n",
        "\n",
        "pre_process_corpus = np.vectorize(pre_process_document)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "O3y9H3IXaT8j",
        "colab": {}
      },
      "source": [
        "train_reviews = pre_process_corpus(train_reviews)\n",
        "val_reviews = pre_process_corpus(val_reviews)\n",
        "test_reviews = pre_process_corpus(test_reviews)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "E0Lz46misp0j",
        "colab_type": "text"
      },
      "source": [
        "### Preparing `numpy` arrays of the reviews and the labels"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "kt9nq6vXaXAb",
        "colab": {}
      },
      "source": [
        "X_train = np.asarray(train_reviews)\n",
        "X_valid = np.asarray(val_reviews)\n",
        "\n",
        "y_train = np.array(train_sentiments)\n",
        "y_valid = np.array(val_sentiments)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1WplmUuTsxWy",
        "colab_type": "text"
      },
      "source": [
        "### Get the embedding model from TensorFlow hub and see a few embeddings \n",
        "\n",
        "**Caution**: 128D ahead."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "J8vnCxcv7-Sc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model = \"https://tfhub.dev/google/tf2-preview/nnlm-en-dim128/1\"\n",
        "hub_layer = tf_hub.KerasLayer(model, output_shape=[128], input_shape=[], \n",
        "                           dtype=tf.string, trainable=True)\n",
        "hub_layer(train_reviews[:3])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3JSPB7Fks_wf",
        "colab_type": "text"
      },
      "source": [
        "### Constructing a sequential `keras` model\n",
        "\n",
        "The architecture I used is a bit different from the one Dipanjan used. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h_2Eqrq_8Srq",
        "colab_type": "code",
        "outputId": "c10136ba-e51d-4a91-db03-41d5f2144dc4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        }
      },
      "source": [
        "model = tf.keras.Sequential()\n",
        "model.add(hub_layer)\n",
        "model.add(tf.keras.layers.Dense(16, activation='relu'))\n",
        "model.add(tf.keras.layers.Dense(1, activation='sigmoid'))\n",
        "\n",
        "model.summary()"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "keras_layer (KerasLayer)     (None, 128)               124642688 \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 16)                2064      \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 1)                 17        \n",
            "=================================================================\n",
            "Total params: 124,644,769\n",
            "Trainable params: 124,644,769\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "p0L4TY28ti0U",
        "colab_type": "text"
      },
      "source": [
        "The model does not contain any **Dropout** layers. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NCeDKdy89dN1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model.compile(optimizer='adam',\n",
        "              loss='binary_crossentropy',\n",
        "              metrics=['accuracy'])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "vYC2ifCFmnt9",
        "outputId": "29d89ea0-2bb6-4bc1-946f-6ba6e4c5aedb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 492
        }
      },
      "source": [
        "history = model.fit(X_train, y_train, validation_data=(X_valid, y_valid),\n",
        "                        epochs=12, batch_size=256)"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 30000 samples, validate on 5000 samples\n",
            "Epoch 1/12\n",
            "30000/30000 [==============================] - 177s 6ms/sample - loss: 0.4459 - accuracy: 0.8132 - val_loss: 0.2909 - val_accuracy: 0.8766\n",
            "Epoch 2/12\n",
            "30000/30000 [==============================] - 174s 6ms/sample - loss: 0.2084 - accuracy: 0.9224 - val_loss: 0.2641 - val_accuracy: 0.8908\n",
            "Epoch 3/12\n",
            "30000/30000 [==============================] - 173s 6ms/sample - loss: 0.1298 - accuracy: 0.9566 - val_loss: 0.2630 - val_accuracy: 0.8964\n",
            "Epoch 4/12\n",
            "30000/30000 [==============================] - 172s 6ms/sample - loss: 0.0808 - accuracy: 0.9766 - val_loss: 0.2967 - val_accuracy: 0.8910\n",
            "Epoch 5/12\n",
            "30000/30000 [==============================] - 174s 6ms/sample - loss: 0.0485 - accuracy: 0.9884 - val_loss: 0.3345 - val_accuracy: 0.8898\n",
            "Epoch 6/12\n",
            "30000/30000 [==============================] - 172s 6ms/sample - loss: 0.0283 - accuracy: 0.9950 - val_loss: 0.3734 - val_accuracy: 0.8878\n",
            "Epoch 7/12\n",
            "30000/30000 [==============================] - 172s 6ms/sample - loss: 0.0164 - accuracy: 0.9979 - val_loss: 0.4127 - val_accuracy: 0.8842\n",
            "Epoch 8/12\n",
            "30000/30000 [==============================] - 173s 6ms/sample - loss: 0.0100 - accuracy: 0.9994 - val_loss: 0.4502 - val_accuracy: 0.8800\n",
            "Epoch 9/12\n",
            "30000/30000 [==============================] - 172s 6ms/sample - loss: 0.0061 - accuracy: 0.9996 - val_loss: 0.4893 - val_accuracy: 0.8778\n",
            "Epoch 10/12\n",
            "30000/30000 [==============================] - 173s 6ms/sample - loss: 0.0036 - accuracy: 0.9998 - val_loss: 0.5297 - val_accuracy: 0.8768\n",
            "Epoch 11/12\n",
            "30000/30000 [==============================] - 173s 6ms/sample - loss: 0.0022 - accuracy: 0.9999 - val_loss: 0.5634 - val_accuracy: 0.8762\n",
            "Epoch 12/12\n",
            "30000/30000 [==============================] - 174s 6ms/sample - loss: 0.0013 - accuracy: 1.0000 - val_loss: 0.5946 - val_accuracy: 0.8762\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "K2PS0G1htpY-",
        "colab_type": "text"
      },
      "source": [
        "Absence of the **Dropout** layers is quite evident. The mode clearly overfits. "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CVJwVhu6twUg",
        "colab_type": "text"
      },
      "source": [
        "### Performance visualization of the baseline model (fine-tuned, though)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "v2t8nY9NRT4K",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        },
        "outputId": "46ba9b86-90bd-4f47-e4fc-ffc8b6a80954"
      },
      "source": [
        "history_dict = history.history\n",
        "history_dict.keys()"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys(['loss', 'accuracy', 'val_loss', 'val_accuracy'])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7NKGRf6wZZ3r",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 517
        },
        "outputId": "1e8e2e80-3884-48e7-ec5d-964dd7c391ae"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.style.use('ggplot')\n",
        "\n",
        "acc = history_dict['accuracy']\n",
        "val_acc = history_dict['val_accuracy']\n",
        "loss = history_dict['loss']\n",
        "val_loss = history_dict['val_loss']\n",
        "\n",
        "epochs = range(1, len(acc) + 1)\n",
        "\n",
        "plt.figure(figsize=(10,8))\n",
        "plt.plot(epochs, loss, label='Training loss')\n",
        "plt.plot(epochs, val_loss, label='Validation loss')\n",
        "plt.plot(epochs, acc, label='Training accuracy')\n",
        "plt.plot(epochs, val_acc, label='Validation accuracy')\n",
        "\n",
        "plt.title('Training and validation accuracy and loss')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Accuracy/Loss')\n",
        "plt.legend()\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmcAAAH0CAYAAAB4qIphAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 720x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TAjfL9I4t5pa",
        "colab_type": "text"
      },
      "source": [
        "### Accuracy on the test set"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "X16RpgBlZ8at",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 55
        },
        "outputId": "c4d5b05b-21bb-4884-a6e0-9b3ef7282028"
      },
      "source": [
        "model.evaluate(np.array(test_reviews), np.array(test_sentiments))"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "15000/15000 [==============================] - 82s 5ms/sample - loss: 0.5725 - accuracy: 0.8793\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.572544362214456, 0.8793333]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Jk9KsXfTt8RP",
        "colab_type": "text"
      },
      "source": [
        "### Classification report"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-HVPtbzjanb3",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 188
        },
        "outputId": "3d041fc0-fe0a-4d8a-9606-f894194c8a60"
      },
      "source": [
        "from sklearn.metrics import classification_report\n",
        "\n",
        "y_pred = model.predict_classes(np.array(test_reviews))\n",
        "\n",
        "print(classification_report(np.array(test_sentiments), y_pred))"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.88      0.87      0.88      7490\n",
            "           1       0.87      0.89      0.88      7510\n",
            "\n",
            "   micro avg       0.88      0.88      0.88     15000\n",
            "   macro avg       0.88      0.88      0.88     15000\n",
            "weighted avg       0.88      0.88      0.88     15000\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DVC2Kd8auAUW",
        "colab_type": "text"
      },
      "source": [
        "### Initial model + Dropout layers + EarlyStopping"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yXh9sxZ4bI3Y",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 359
        },
        "outputId": "d5023276-fad4-4d36-b17c-5564bb749909"
      },
      "source": [
        "model = tf.keras.Sequential()\n",
        "model.add(hub_layer)\n",
        "model.add(tf.keras.layers.Dropout(0.3))\n",
        "model.add(tf.keras.layers.Dense(16, activation='relu'))\n",
        "model.add(tf.keras.layers.Dropout(0.3))\n",
        "model.add(tf.keras.layers.Dense(1, activation='sigmoid'))\n",
        "\n",
        "model.summary()"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "keras_layer (KerasLayer)     (None, 128)               124642688 \n",
            "_________________________________________________________________\n",
            "dropout_2 (Dropout)          (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense_4 (Dense)              (None, 16)                2064      \n",
            "_________________________________________________________________\n",
            "dropout_3 (Dropout)          (None, 16)                0         \n",
            "_________________________________________________________________\n",
            "dense_5 (Dense)              (None, 1)                 17        \n",
            "=================================================================\n",
            "Total params: 124,644,769\n",
            "Trainable params: 124,644,769\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FZlKsdf9bsZc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model.compile(optimizer='adam',\n",
        "              loss='binary_crossentropy',\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "es_cb = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=5)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "s-N0JOFRi9K5",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 264
        },
        "outputId": "182b7dd6-c5e5-481e-f97c-37ca795d098d"
      },
      "source": [
        "history = model.fit(X_train, y_train, validation_data=(X_valid, y_valid),\n",
        "                        epochs=12, batch_size=256, callbacks=[es_cb])"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 30000 samples, validate on 5000 samples\n",
            "Epoch 1/12\n",
            "30000/30000 [==============================] - 171s 6ms/sample - loss: 0.1887 - accuracy: 0.9472 - val_loss: 0.3106 - val_accuracy: 0.8850\n",
            "Epoch 2/12\n",
            "30000/30000 [==============================] - 171s 6ms/sample - loss: 0.0354 - accuracy: 0.9955 - val_loss: 0.3925 - val_accuracy: 0.8820\n",
            "Epoch 3/12\n",
            "30000/30000 [==============================] - 171s 6ms/sample - loss: 0.0177 - accuracy: 0.9981 - val_loss: 0.4588 - val_accuracy: 0.8802\n",
            "Epoch 4/12\n",
            "30000/30000 [==============================] - 171s 6ms/sample - loss: 0.0101 - accuracy: 0.9993 - val_loss: 0.5355 - val_accuracy: 0.8786\n",
            "Epoch 5/12\n",
            "30000/30000 [==============================] - 171s 6ms/sample - loss: 0.0071 - accuracy: 0.9992 - val_loss: 0.5745 - val_accuracy: 0.8774\n",
            "Epoch 6/12\n",
            "30000/30000 [==============================] - 171s 6ms/sample - loss: 0.0052 - accuracy: 0.9997 - val_loss: 0.6162 - val_accuracy: 0.8782\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zHhLXZMtuNjH",
        "colab_type": "text"
      },
      "source": [
        "The model still overfits. More sophisticated parameter initialization would be helpful. "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aTLiTpJgubVV",
        "colab_type": "text"
      },
      "source": [
        "### Performance visualization of Model-2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "smmQPlXvbzvp",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 517
        },
        "outputId": "9bd95239-45ec-4ad0-d312-f9ab14d9b569"
      },
      "source": [
        "acc = history_dict['accuracy']\n",
        "val_acc = history_dict['val_accuracy']\n",
        "loss = history_dict['loss']\n",
        "val_loss = history_dict['val_loss']\n",
        "\n",
        "epochs = range(1, len(acc) + 1)\n",
        "\n",
        "plt.figure(figsize=(10,8))\n",
        "plt.plot(epochs, loss, label='Training loss')\n",
        "plt.plot(epochs, val_loss, label='Validation loss')\n",
        "plt.plot(epochs, acc, label='Training accuracy')\n",
        "plt.plot(epochs, val_acc, label='Validation accuracy')\n",
        "\n",
        "plt.title('Training and validation accuracy and loss')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Accuracy/Loss')\n",
        "plt.legend()\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmcAAAH0CAYAAAB4qIphAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 720x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GGrC9uZ4ukGn",
        "colab_type": "text"
      },
      "source": [
        "### Accuracy has increased a bit, although"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9d3XQPERn06Q",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 55
        },
        "outputId": "2054f774-ed56-4a40-a8b6-4385abdf6500"
      },
      "source": [
        "model.evaluate(np.array(test_reviews), np.array(test_sentiments))"
      ],
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "15000/15000 [==============================] - 81s 5ms/sample - loss: 0.5889 - accuracy: 0.8803\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.5888620198649664, 0.8803333]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 29
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bbewV2h4unxA",
        "colab_type": "text"
      },
      "source": [
        "### Classification report of Model-2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u0f91y81n_SE",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 188
        },
        "outputId": "22b719ee-715a-4562-9dc5-260d6a0053e6"
      },
      "source": [
        "from sklearn.metrics import classification_report\n",
        "\n",
        "y_pred = model.predict_classes(np.array(test_reviews))\n",
        "\n",
        "print(classification_report(np.array(test_sentiments), y_pred))"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.88      0.88      0.88      7490\n",
            "           1       0.88      0.88      0.88      7510\n",
            "\n",
            "   micro avg       0.88      0.88      0.88     15000\n",
            "   macro avg       0.88      0.88      0.88     15000\n",
            "weighted avg       0.88      0.88      0.88     15000\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8egqJsyUurki",
        "colab_type": "text"
      },
      "source": [
        "Here are the details of the model' evaluation which is taken from the original tutorial: \n",
        "\n",
        "![](https://cdn-images-1.medium.com/max/1200/1*MMpvIAOh0Ac3-0GKAhV_gw.png)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SCsIo--evmvV",
        "colab_type": "text"
      },
      "source": [
        "The results are more or less very similar (yes, I did not print the AUC score). "
      ]
    }
  ]
}