l09c05_nlp_tweaking_the_model.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "punL79CN7Ox6"
      },
      "source": [
        "##### Copyright 2020 The TensorFlow Authors."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "cellView": "form",
        "id": "_ckMIh7O7s6D"
      },
      "outputs": [],
      "source": [
        "#@title Licensed under the Apache License, Version 2.0 (the \"License\");\n",
        "# you may not use this file except in compliance with the License.\n",
        "# You may obtain a copy of the License at\n",
        "#\n",
        "# https://www.apache.org/licenses/LICENSE-2.0\n",
        "#\n",
        "# Unless required by applicable law or agreed to in writing, software\n",
        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
        "# See the License for the specific language governing permissions and\n",
        "# limitations under the License."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QrxSyyyhygUR"
      },
      "source": [
        "# Tweaking the Model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "S5Uhzt6vVIB2"
      },
      "source": [
        "<table class=\"tfo-notebook-buttons\" align=\"left\">\n",
        "  <td>\n",
        "    <a target=\"_blank\" href=\"https://colab.research.google.com/github/tensorflow/examples/blob/master/courses/udacity_intro_to_tensorflow_for_deep_learning/l09c05_nlp_tweaking_the_model.ipynb\"><img src=\"https://www.tensorflow.org/images/colab_logo_32px.png\" />Run in Google Colab</a>\n",
        "  </td>\n",
        "  <td>\n",
        "    <a target=\"_blank\" href=\"https://github.com/tensorflow/examples/blob/master/courses/udacity_intro_to_tensorflow_for_deep_learning/l09c05_nlp_tweaking_the_model.ipynb\"><img src=\"https://www.tensorflow.org/images/GitHub-Mark-32px.png\" />View source on GitHub</a>\n",
        "  </td>\n",
        "</table>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xiWacy71Cu54"
      },
      "source": [
        "In this colab, you'll investigate how various tweaks to data processing and the model itself can impact results. At the end, you'll once again be able to visualize how the network sees the related sentiment of each word in the dataset."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hY-fjvwfy2P9"
      },
      "source": [
        "## Import TensorFlow and related functions"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "id": "drsUfVVXyxJl"
      },
      "outputs": [],
      "source": [
        "import tensorflow as tf\n",
        "\n",
        "from tensorflow.keras.preprocessing.text import Tokenizer\n",
        "from tensorflow.keras.preprocessing.sequence import pad_sequences"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZIf1N46jy6Ed"
      },
      "source": [
        "## Get the dataset\n",
        "\n",
        "We'll once again use the dataset containing Amazon and Yelp reviews. This dataset was originally extracted from [here](https://www.kaggle.com/marklvl/sentiment-labelled-sentences-data-set)."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "id": "m83g42sJzGO0",
        "outputId": "31cf9c14-53a2-418e-f0f6-a8a205d48ae8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2021-12-07 21:59:03--  https://drive.google.com/uc?id=13ySLC_ue6Umt9RJYSeM2t-V0kCv-4C-P\n",
            "Resolving drive.google.com (drive.google.com)... 74.125.140.138, 74.125.140.102, 74.125.140.139, ...\n",
            "Connecting to drive.google.com (drive.google.com)|74.125.140.138|:443... connected.\n",
            "HTTP request sent, awaiting response... 302 Moved Temporarily\n",
            "Location: https://doc-08-ak-docs.googleusercontent.com/docs/securesc/ha0ro937gcuc7l7deffksulhg5h7mbp1/nv1dnmpiiqt7j9ju8t1irfhb4r6gnv1l/1638914325000/11118900490791463723/*/13ySLC_ue6Umt9RJYSeM2t-V0kCv-4C-P [following]\n",
            "Warning: wildcards not supported in HTTP.\n",
            "--2021-12-07 21:59:03--  https://doc-08-ak-docs.googleusercontent.com/docs/securesc/ha0ro937gcuc7l7deffksulhg5h7mbp1/nv1dnmpiiqt7j9ju8t1irfhb4r6gnv1l/1638914325000/11118900490791463723/*/13ySLC_ue6Umt9RJYSeM2t-V0kCv-4C-P\n",
            "Resolving doc-08-ak-docs.googleusercontent.com (doc-08-ak-docs.googleusercontent.com)... 64.233.166.132, 2a00:1450:400c:c09::84\n",
            "Connecting to doc-08-ak-docs.googleusercontent.com (doc-08-ak-docs.googleusercontent.com)|64.233.166.132|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 127831 (125K) [text/csv]\n",
            "Saving to: ‘/tmp/sentiment.csv’\n",
            "\n",
            "/tmp/sentiment.csv  100%[===================>] 124.83K  --.-KB/s    in 0.001s  \n",
            "\n",
            "2021-12-07 21:59:04 (88.7 MB/s) - ‘/tmp/sentiment.csv’ saved [127831/127831]\n",
            "\n"
          ]
        }
      ],
      "source": [
        "!wget --no-check-certificate \\\n",
        "    https://drive.google.com/uc?id=13ySLC_ue6Umt9RJYSeM2t-V0kCv-4C-P \\\n",
        "    -O /tmp/sentiment.csv"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "y4e6GG2CzJUq"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "\n",
        "dataset = pd.read_csv('/tmp/sentiment.csv')\n",
        "\n",
        "sentences = dataset['text'].tolist()\n",
        "labels = dataset['sentiment'].tolist()\n",
        "\n",
        "# Separate out the sentences and labels into training and test sets\n",
        "training_size = int(len(sentences) * 0.8)\n",
        "\n",
        "training_sentences = sentences[0:training_size]\n",
        "testing_sentences = sentences[training_size:]\n",
        "training_labels = labels[0:training_size]\n",
        "testing_labels = labels[training_size:]\n",
        "\n",
        "# Make labels into numpy arrays for use with the network later\n",
        "training_labels_final = np.array(training_labels)\n",
        "testing_labels_final = np.array(testing_labels)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "drDkTFMuzW6N"
      },
      "source": [
        "## Tokenize the dataset (with tweaks!)\n",
        "\n",
        "Now, we'll tokenize the dataset, but we can make some changes to this from before. Previously, we used: \n",
        "```\n",
        "vocab_size = 1000\n",
        "embedding_dim = 16\n",
        "max_length = 100\n",
        "trunc_type='post'\n",
        "padding_type='post'\n",
        "```\n",
        "\n",
        "How might changing the `vocab_size`, `embedding_dim` or `max_length` affect how the model performs?"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "hjPUJFhQzuee"
      },
      "outputs": [],
      "source": [
        "vocab_size = 500\n",
        "embedding_dim = 16\n",
        "max_length = 50\n",
        "trunc_type='post'\n",
        "padding_type='post'\n",
        "oov_tok = \"<OOV>\"\n",
        "\n",
        "tokenizer = Tokenizer(num_words = vocab_size, oov_token=oov_tok)\n",
        "tokenizer.fit_on_texts(training_sentences)\n",
        "word_index = tokenizer.word_index\n",
        "training_sequences = tokenizer.texts_to_sequences(training_sentences)\n",
        "training_padded = pad_sequences(training_sequences, maxlen=max_length, padding=padding_type, truncating=trunc_type)\n",
        "\n",
        "testing_sequences = tokenizer.texts_to_sequences(testing_sentences)\n",
        "testing_padded = pad_sequences(testing_sequences, maxlen=max_length, padding=padding_type, truncating=trunc_type)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FwFjO1kg0UUK"
      },
      "source": [
        "## Train a Sentiment Model (with tweaks!)\n",
        "\n",
        "We'll use a slightly different model here, using `GlobalAveragePooling1D` instead of `Flatten()`."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "ectP92fl0dFO",
        "outputId": "2d0c8cb3-bbc9-4713-d69a-61f7cc4dffe2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " embedding (Embedding)       (None, 50, 16)            8000      \n",
            "                                                                 \n",
            " global_average_pooling1d (G  (None, 16)               0         \n",
            " lobalAveragePooling1D)                                          \n",
            "                                                                 \n",
            " dense (Dense)               (None, 6)                 102       \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 1)                 7         \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 8,109\n",
            "Trainable params: 8,109\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "model = tf.keras.Sequential([\n",
        "    tf.keras.layers.Embedding(vocab_size, embedding_dim, input_length=max_length),\n",
        "    tf.keras.layers.GlobalAveragePooling1D(),\n",
        "    tf.keras.layers.Dense(6, activation='relu'),\n",
        "    tf.keras.layers.Dense(1, activation='sigmoid')\n",
        "])\n",
        "model.compile(loss='binary_crossentropy',optimizer='adam',metrics=['accuracy'])\n",
        "model.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "id": "7TQIaGjs073w",
        "outputId": "05c98a5f-cd8a-4bca-efa3-53840254e569",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/30\n",
            "50/50 [==============================] - 3s 11ms/step - loss: 0.6932 - accuracy: 0.5066 - val_loss: 0.6944 - val_accuracy: 0.4110\n",
            "Epoch 2/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.6927 - accuracy: 0.5223 - val_loss: 0.6953 - val_accuracy: 0.4110\n",
            "Epoch 3/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.6924 - accuracy: 0.5223 - val_loss: 0.6956 - val_accuracy: 0.4110\n",
            "Epoch 4/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.6920 - accuracy: 0.5223 - val_loss: 0.6968 - val_accuracy: 0.4110\n",
            "Epoch 5/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.6885 - accuracy: 0.5298 - val_loss: 0.6927 - val_accuracy: 0.4511\n",
            "Epoch 6/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.6826 - accuracy: 0.5361 - val_loss: 0.6893 - val_accuracy: 0.4912\n",
            "Epoch 7/30\n",
            "50/50 [==============================] - 0s 6ms/step - loss: 0.6742 - accuracy: 0.6083 - val_loss: 0.6875 - val_accuracy: 0.4787\n",
            "Epoch 8/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.6631 - accuracy: 0.6246 - val_loss: 0.6750 - val_accuracy: 0.5489\n",
            "Epoch 9/30\n",
            "50/50 [==============================] - 0s 6ms/step - loss: 0.6490 - accuracy: 0.6717 - val_loss: 0.6603 - val_accuracy: 0.6366\n",
            "Epoch 10/30\n",
            "50/50 [==============================] - 0s 6ms/step - loss: 0.6312 - accuracy: 0.7081 - val_loss: 0.6455 - val_accuracy: 0.6842\n",
            "Epoch 11/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.6113 - accuracy: 0.7489 - val_loss: 0.6394 - val_accuracy: 0.6241\n",
            "Epoch 12/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.5889 - accuracy: 0.7659 - val_loss: 0.6158 - val_accuracy: 0.7343\n",
            "Epoch 13/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.5661 - accuracy: 0.7960 - val_loss: 0.6006 - val_accuracy: 0.7393\n",
            "Epoch 14/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.5412 - accuracy: 0.8129 - val_loss: 0.5920 - val_accuracy: 0.6942\n",
            "Epoch 15/30\n",
            "50/50 [==============================] - 0s 6ms/step - loss: 0.5177 - accuracy: 0.8198 - val_loss: 0.5680 - val_accuracy: 0.7619\n",
            "Epoch 16/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.4951 - accuracy: 0.8330 - val_loss: 0.5668 - val_accuracy: 0.7093\n",
            "Epoch 17/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.4743 - accuracy: 0.8336 - val_loss: 0.5655 - val_accuracy: 0.6967\n",
            "Epoch 18/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.4541 - accuracy: 0.8443 - val_loss: 0.5320 - val_accuracy: 0.7644\n",
            "Epoch 19/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.4348 - accuracy: 0.8475 - val_loss: 0.5264 - val_accuracy: 0.7644\n",
            "Epoch 20/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.4176 - accuracy: 0.8537 - val_loss: 0.5250 - val_accuracy: 0.7393\n",
            "Epoch 21/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.4024 - accuracy: 0.8588 - val_loss: 0.5167 - val_accuracy: 0.7419\n",
            "Epoch 22/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.3866 - accuracy: 0.8619 - val_loss: 0.5398 - val_accuracy: 0.7068\n",
            "Epoch 23/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.3748 - accuracy: 0.8638 - val_loss: 0.5107 - val_accuracy: 0.7444\n",
            "Epoch 24/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.3623 - accuracy: 0.8638 - val_loss: 0.5150 - val_accuracy: 0.7318\n",
            "Epoch 25/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.3502 - accuracy: 0.8713 - val_loss: 0.4967 - val_accuracy: 0.7619\n",
            "Epoch 26/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.3398 - accuracy: 0.8757 - val_loss: 0.4918 - val_accuracy: 0.7619\n",
            "Epoch 27/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.3314 - accuracy: 0.8732 - val_loss: 0.4937 - val_accuracy: 0.7594\n",
            "Epoch 28/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.3205 - accuracy: 0.8807 - val_loss: 0.4893 - val_accuracy: 0.7644\n",
            "Epoch 29/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.3139 - accuracy: 0.8801 - val_loss: 0.4865 - val_accuracy: 0.7719\n",
            "Epoch 30/30\n",
            "50/50 [==============================] - 0s 7ms/step - loss: 0.3045 - accuracy: 0.8858 - val_loss: 0.5013 - val_accuracy: 0.7343\n"
          ]
        }
      ],
      "source": [
        "num_epochs = 30\n",
        "history = model.fit(training_padded, training_labels_final, epochs=num_epochs, validation_data=(testing_padded, testing_labels_final))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "alAlYort7gWV"
      },
      "source": [
        "## Visualize the training graph\n",
        "\n",
        "You can use the code below to visualize the training and validation accuracy while you try out different tweaks to the hyperparameters and model."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "id": "o9l5vBeU71vH",
        "outputId": "f9f9c40e-c034-4f77-ab55-6f1a0d3966cf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 541
        }
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "\n",
        "def plot_graphs(history, string):\n",
        "  plt.plot(history.history[string])\n",
        "  plt.plot(history.history['val_'+string])\n",
        "  plt.xlabel(\"Epochs\")\n",
        "  plt.ylabel(string)\n",
        "  plt.legend([string, 'val_'+string])\n",
        "  plt.show()\n",
        "  \n",
        "plot_graphs(history, \"accuracy\")\n",
        "plot_graphs(history, \"loss\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SZzXE-pT8K57"
      },
      "source": [
        "## Get files for visualizing the network\n",
        "\n",
        "The code below will download two files for visualizing how your network \"sees\" the sentiment related to each word. Head to http://projector.tensorflow.org/ and load these files, then click the checkbox to \"sphereize\" the data.\n",
        "\n",
        "Note: You may run into errors with the projection if your `vocab_size` earlier was larger than the actual number of words in the vocabulary, in which case you'll need to decrease this variable and re-train in order to visualize."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "id": "2Ex4o7Lc8Njl",
        "outputId": "4ceabf38-0180-4bb2-b07a-e5eace68de7b",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(500, 16)\n"
          ]
        }
      ],
      "source": [
        "# First get the weights of the embedding layer\n",
        "e = model.layers[0]\n",
        "weights = e.get_weights()[0]\n",
        "print(weights.shape) # shape: (vocab_size, embedding_dim)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "id": "bUL1zk5p8WIV"
      },
      "outputs": [],
      "source": [
        "import io\n",
        "\n",
        "# Create the reverse word index\n",
        "reverse_word_index = dict([(value, key) for (key, value) in word_index.items()])\n",
        "\n",
        "# Write out the embedding vectors and metadata\n",
        "out_v = io.open('vecs.tsv', 'w', encoding='utf-8')\n",
        "out_m = io.open('meta.tsv', 'w', encoding='utf-8')\n",
        "for word_num in range(1, vocab_size):\n",
        "  word = reverse_word_index[word_num]\n",
        "  embeddings = weights[word_num]\n",
        "  out_m.write(word + \"\\n\")\n",
        "  out_v.write('\\t'.join([str(x) for x in embeddings]) + \"\\n\")\n",
        "out_v.close()\n",
        "out_m.close()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "id": "lqyV8QYnD46U",
        "outputId": "0b671b4d-4825-467d-fa64-c8226f93ae70",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 17
        }
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "\n",
              "    async function download(id, filename, size) {\n",
              "      if (!google.colab.kernel.accessAllowed) {\n",
              "        return;\n",
              "      }\n",
              "      const div = document.createElement('div');\n",
              "      const label = document.createElement('label');\n",
              "      label.textContent = `Downloading \"${filename}\": `;\n",
              "      div.appendChild(label);\n",
              "      const progress = document.createElement('progress');\n",
              "      progress.max = size;\n",
              "      div.appendChild(progress);\n",
              "      document.body.appendChild(div);\n",
              "\n",
              "      const buffers = [];\n",
              "      let downloaded = 0;\n",
              "\n",
              "      const channel = await google.colab.kernel.comms.open(id);\n",
              "      // Send a message to notify the kernel that we're ready.\n",
              "      channel.send({})\n",
              "\n",
              "      for await (const message of channel.messages) {\n",
              "        // Send a message to notify the kernel that we're ready.\n",
              "        channel.send({})\n",
              "        if (message.buffers) {\n",
              "          for (const buffer of message.buffers) {\n",
              "            buffers.push(buffer);\n",
              "            downloaded += buffer.byteLength;\n",
              "            progress.value = downloaded;\n",
              "          }\n",
              "        }\n",
              "      }\n",
              "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
              "      const a = document.createElement('a');\n",
              "      a.href = window.URL.createObjectURL(blob);\n",
              "      a.download = filename;\n",
              "      div.appendChild(a);\n",
              "      a.click();\n",
              "      div.remove();\n",
              "    }\n",
              "  "
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "download(\"download_b8345366-b694-4b43-b156-0cee4bc934e8\", \"vecs.tsv\", 90863)"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "\n",
              "    async function download(id, filename, size) {\n",
              "      if (!google.colab.kernel.accessAllowed) {\n",
              "        return;\n",
              "      }\n",
              "      const div = document.createElement('div');\n",
              "      const label = document.createElement('label');\n",
              "      label.textContent = `Downloading \"${filename}\": `;\n",
              "      div.appendChild(label);\n",
              "      const progress = document.createElement('progress');\n",
              "      progress.max = size;\n",
              "      div.appendChild(progress);\n",
              "      document.body.appendChild(div);\n",
              "\n",
              "      const buffers = [];\n",
              "      let downloaded = 0;\n",
              "\n",
              "      const channel = await google.colab.kernel.comms.open(id);\n",
              "      // Send a message to notify the kernel that we're ready.\n",
              "      channel.send({})\n",
              "\n",
              "      for await (const message of channel.messages) {\n",
              "        // Send a message to notify the kernel that we're ready.\n",
              "        channel.send({})\n",
              "        if (message.buffers) {\n",
              "          for (const buffer of message.buffers) {\n",
              "            buffers.push(buffer);\n",
              "            downloaded += buffer.byteLength;\n",
              "            progress.value = downloaded;\n",
              "          }\n",
              "        }\n",
              "      }\n",
              "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
              "      const a = document.createElement('a');\n",
              "      a.href = window.URL.createObjectURL(blob);\n",
              "      a.download = filename;\n",
              "      div.appendChild(a);\n",
              "      a.click();\n",
              "      div.remove();\n",
              "    }\n",
              "  "
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "download(\"download_ce212c32-cbae-4393-9e3f-4bfd5e6410bc\", \"meta.tsv\", 3025)"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {}
        }
      ],
      "source": [
        "# Download the files\n",
        "try:\n",
        "  from google.colab import files\n",
        "except ImportError:\n",
        "  pass\n",
        "else:\n",
        "  files.download('vecs.tsv')\n",
        "  files.download('meta.tsv')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XUXAlNNk59gG"
      },
      "source": [
        "## Predicting Sentiment in New Reviews\n",
        "\n",
        "Below, we've again included some example new reviews you can test your results on."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "id": "JbFTTcaK6Dan",
        "outputId": "0a7fb3f7-88b2-4145-c601-6bba7b844e5e",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "['I love this phone', 'I hate spaghetti', 'Everything was cold', 'Everything was hot exactly as I wanted', 'Everything was green', 'the host seated us immediately', 'they gave us free chocolate cake', 'not sure about the wilted flowers on the table', 'only works when I stand on tippy toes', 'does not work when I stand on my head']\n",
            "\n",
            "HOT OFF THE PRESS! HERE ARE SOME NEWLY MINTED, ABSOLUTELY GENUINE REVIEWS!\n",
            "\n",
            "I love this phone\n",
            "[0.87587047]\n",
            "\n",
            "\n",
            "I hate spaghetti\n",
            "[0.26441818]\n",
            "\n",
            "\n",
            "Everything was cold\n",
            "[0.60760665]\n",
            "\n",
            "\n",
            "Everything was hot exactly as I wanted\n",
            "[0.46314725]\n",
            "\n",
            "\n",
            "Everything was green\n",
            "[0.60760665]\n",
            "\n",
            "\n",
            "the host seated us immediately\n",
            "[0.5670158]\n",
            "\n",
            "\n",
            "they gave us free chocolate cake\n",
            "[0.78867424]\n",
            "\n",
            "\n",
            "not sure about the wilted flowers on the table\n",
            "[0.15370889]\n",
            "\n",
            "\n",
            "only works when I stand on tippy toes\n",
            "[0.77417576]\n",
            "\n",
            "\n",
            "does not work when I stand on my head\n",
            "[0.04670684]\n",
            "\n",
            "\n"
          ]
        }
      ],
      "source": [
        "# Use the model to predict a review   \n",
        "fake_reviews = ['I love this phone', 'I hate spaghetti', \n",
        "                'Everything was cold',\n",
        "                'Everything was hot exactly as I wanted', \n",
        "                'Everything was green', \n",
        "                'the host seated us immediately',\n",
        "                'they gave us free chocolate cake', \n",
        "                'not sure about the wilted flowers on the table',\n",
        "                'only works when I stand on tippy toes', \n",
        "                'does not work when I stand on my head']\n",
        "\n",
        "print(fake_reviews) \n",
        "\n",
        "# Create the sequences\n",
        "padding_type='post'\n",
        "sample_sequences = tokenizer.texts_to_sequences(fake_reviews)\n",
        "fakes_padded = pad_sequences(sample_sequences, padding=padding_type, maxlen=max_length)           \n",
        "\n",
        "print('\\nHOT OFF THE PRESS! HERE ARE SOME NEWLY MINTED, ABSOLUTELY GENUINE REVIEWS!\\n')              \n",
        "\n",
        "classes = model.predict(fakes_padded)\n",
        "\n",
        "# The closer the class is to 1, the more positive the review is deemed to be\n",
        "for x in range(len(fake_reviews)):\n",
        "  print(fake_reviews[x])\n",
        "  print(classes[x])\n",
        "  print('\\n')\n",
        "\n",
        "# Try adding reviews of your own\n",
        "# Add some negative words (such as \"not\") to the good reviews and see what happens\n",
        "# For example:\n",
        "# they gave us free chocolate cake and did not charge us"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "collapsed_sections": [],
      "name": "l09c05_nlp_tweaking_the_model.ipynb",
      "toc_visible": true,
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}