Shakespear_Lstm.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "accelerator": "TPU",
    "colab": {
      "name": "Copy of NLP_Week4_Exercise_Shakespeare_Answer.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "zX4Kg8DUTKWO",
        "colab_type": "code",
        "colab": {}
      },
      "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."
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "BOwsuGQQY9OL",
        "colab": {}
      },
      "source": [
        "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
        "from tensorflow.keras.layers import Embedding, LSTM, Dense, Dropout, Bidirectional\n",
        "from tensorflow.keras.preprocessing.text import Tokenizer\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.optimizers import Adam\n",
        "from tensorflow.keras import regularizers\n",
        "import tensorflow.keras.utils as ku \n",
        "import numpy as np "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "PRnDnCW-Z7qv",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 202
        },
        "outputId": "08053e12-c1a5-49f2-90e1-418c7dd7fb64"
      },
      "source": [
        "tokenizer = Tokenizer()\n",
        "!wget --no-check-certificate \\\n",
        "    https://storage.googleapis.com/laurencemoroney-blog.appspot.com/sonnets.txt \\\n",
        "    -O /tmp/sonnets.txt\n",
        "data = open('/tmp/sonnets.txt').read()\n",
        "\n",
        "corpus = data.lower().split(\"\\n\")\n",
        "\n",
        "\n",
        "tokenizer.fit_on_texts(corpus)\n",
        "total_words = len(tokenizer.word_index) + 1\n",
        "\n",
        "# create input sequences using list of tokens\n",
        "input_sequences = []\n",
        "for line in corpus:\n",
        "\ttoken_list = tokenizer.texts_to_sequences([line])[0]\n",
        "\tfor i in range(1, len(token_list)):\n",
        "\t\tn_gram_sequence = token_list[:i+1]\n",
        "\t\tinput_sequences.append(n_gram_sequence)\n",
        "\n",
        "\n",
        "# pad sequences \n",
        "max_sequence_len = max([len(x) for x in input_sequences])\n",
        "input_sequences = np.array(pad_sequences(input_sequences, maxlen=max_sequence_len, padding='pre'))\n",
        "\n",
        "# create predictors and label\n",
        "predictors, label = input_sequences[:,:-1],input_sequences[:,-1]\n",
        "\n",
        "label = ku.to_categorical(label, num_classes=total_words)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2020-06-16 16:23:44--  https://storage.googleapis.com/laurencemoroney-blog.appspot.com/sonnets.txt\n",
            "Resolving storage.googleapis.com (storage.googleapis.com)... 108.177.121.128, 2607:f8b0:4001:c08::80\n",
            "Connecting to storage.googleapis.com (storage.googleapis.com)|108.177.121.128|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 93578 (91K) [text/plain]\n",
            "Saving to: ‘/tmp/sonnets.txt’\n",
            "\n",
            "\r/tmp/sonnets.txt      0%[                    ]       0  --.-KB/s               \r/tmp/sonnets.txt    100%[===================>]  91.38K  --.-KB/s    in 0.002s  \n",
            "\n",
            "2020-06-16 16:23:44 (45.4 MB/s) - ‘/tmp/sonnets.txt’ saved [93578/93578]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "w9vH8Y59ajYL",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 370
        },
        "outputId": "924fd341-5bf6-48b8-8ab0-082be7e9a667"
      },
      "source": [
        "model = Sequential()\n",
        "model.add(Embedding(total_words, 100, input_length=max_sequence_len-1))\n",
        "model.add(Bidirectional(LSTM(150, return_sequences = True)))\n",
        "model.add(Dropout(0.2))\n",
        "model.add(LSTM(100))\n",
        "model.add(Dense(total_words/2, activation='relu', kernel_regularizer=regularizers.l2(0.01)))\n",
        "model.add(Dense(total_words, activation='softmax'))\n",
        "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
        "print(model.summary())\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "embedding (Embedding)        (None, 10, 100)           321100    \n",
            "_________________________________________________________________\n",
            "bidirectional (Bidirectional (None, 10, 300)           301200    \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 10, 300)           0         \n",
            "_________________________________________________________________\n",
            "lstm_1 (LSTM)                (None, 100)               160400    \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 1605)              162105    \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 3211)              5156866   \n",
            "=================================================================\n",
            "Total params: 6,101,671\n",
            "Trainable params: 6,101,671\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "AIg2f1HBxqof",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "6b8e32e9-4799-4d46-e859-d44ada2ce92d"
      },
      "source": [
        " history = model.fit(predictors, label, epochs=100, verbose=1)"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 6.9021 - accuracy: 0.0214\n",
            "Epoch 2/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 6.5019 - accuracy: 0.0220\n",
            "Epoch 3/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 6.4124 - accuracy: 0.0243\n",
            "Epoch 4/100\n",
            "484/484 [==============================] - 55s 114ms/step - loss: 6.2964 - accuracy: 0.0287\n",
            "Epoch 5/100\n",
            "484/484 [==============================] - 55s 113ms/step - loss: 6.1947 - accuracy: 0.0340\n",
            "Epoch 6/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 6.1105 - accuracy: 0.0382\n",
            "Epoch 7/100\n",
            "484/484 [==============================] - 55s 113ms/step - loss: 6.0385 - accuracy: 0.0416\n",
            "Epoch 8/100\n",
            "484/484 [==============================] - 55s 113ms/step - loss: 5.9705 - accuracy: 0.0443\n",
            "Epoch 9/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 5.8872 - accuracy: 0.0483\n",
            "Epoch 10/100\n",
            "484/484 [==============================] - 55s 114ms/step - loss: 5.7948 - accuracy: 0.0540\n",
            "Epoch 11/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 5.7026 - accuracy: 0.0596\n",
            "Epoch 12/100\n",
            "484/484 [==============================] - 55s 113ms/step - loss: 5.6065 - accuracy: 0.0650\n",
            "Epoch 13/100\n",
            "484/484 [==============================] - 54s 113ms/step - loss: 5.5121 - accuracy: 0.0692\n",
            "Epoch 14/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 5.4168 - accuracy: 0.0745\n",
            "Epoch 15/100\n",
            "484/484 [==============================] - 53s 110ms/step - loss: 5.3254 - accuracy: 0.0806\n",
            "Epoch 16/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 5.2329 - accuracy: 0.0848\n",
            "Epoch 17/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 5.1382 - accuracy: 0.0942\n",
            "Epoch 18/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 5.0458 - accuracy: 0.1006\n",
            "Epoch 19/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 4.9440 - accuracy: 0.1053\n",
            "Epoch 20/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 4.8449 - accuracy: 0.1126\n",
            "Epoch 21/100\n",
            "484/484 [==============================] - 53s 110ms/step - loss: 4.7497 - accuracy: 0.1213\n",
            "Epoch 22/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 4.6518 - accuracy: 0.1279\n",
            "Epoch 23/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 4.5608 - accuracy: 0.1379\n",
            "Epoch 24/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 4.4606 - accuracy: 0.1503\n",
            "Epoch 25/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 4.3651 - accuracy: 0.1570\n",
            "Epoch 26/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 4.2676 - accuracy: 0.1679\n",
            "Epoch 27/100\n",
            "484/484 [==============================] - 55s 113ms/step - loss: 4.1729 - accuracy: 0.1788\n",
            "Epoch 28/100\n",
            "484/484 [==============================] - 55s 114ms/step - loss: 4.0767 - accuracy: 0.1936\n",
            "Epoch 29/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 3.9944 - accuracy: 0.2029\n",
            "Epoch 30/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 3.8890 - accuracy: 0.2202\n",
            "Epoch 31/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 3.8049 - accuracy: 0.2337\n",
            "Epoch 32/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 3.7122 - accuracy: 0.2538\n",
            "Epoch 33/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 3.6281 - accuracy: 0.2733\n",
            "Epoch 34/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 3.5362 - accuracy: 0.2868\n",
            "Epoch 35/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 3.4554 - accuracy: 0.3071\n",
            "Epoch 36/100\n",
            "484/484 [==============================] - 55s 113ms/step - loss: 3.3666 - accuracy: 0.3251\n",
            "Epoch 37/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 3.2948 - accuracy: 0.3422\n",
            "Epoch 38/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 3.2161 - accuracy: 0.3608\n",
            "Epoch 39/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 3.1420 - accuracy: 0.3738\n",
            "Epoch 40/100\n",
            "484/484 [==============================] - 53s 110ms/step - loss: 3.0615 - accuracy: 0.3969\n",
            "Epoch 41/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 2.9992 - accuracy: 0.4106\n",
            "Epoch 42/100\n",
            "484/484 [==============================] - 55s 113ms/step - loss: 2.9372 - accuracy: 0.4239\n",
            "Epoch 43/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 2.8660 - accuracy: 0.4361\n",
            "Epoch 44/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 2.8036 - accuracy: 0.4523\n",
            "Epoch 45/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 2.7439 - accuracy: 0.4641\n",
            "Epoch 46/100\n",
            "484/484 [==============================] - 53s 110ms/step - loss: 2.6853 - accuracy: 0.4783\n",
            "Epoch 47/100\n",
            "484/484 [==============================] - 54s 113ms/step - loss: 2.6193 - accuracy: 0.4951\n",
            "Epoch 48/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 2.5695 - accuracy: 0.5012\n",
            "Epoch 49/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 2.5214 - accuracy: 0.5173\n",
            "Epoch 50/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 2.4553 - accuracy: 0.5337\n",
            "Epoch 51/100\n",
            "484/484 [==============================] - 53s 110ms/step - loss: 2.4147 - accuracy: 0.5393\n",
            "Epoch 52/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 2.3655 - accuracy: 0.5556\n",
            "Epoch 53/100\n",
            "484/484 [==============================] - 55s 113ms/step - loss: 2.3078 - accuracy: 0.5669\n",
            "Epoch 54/100\n",
            "484/484 [==============================] - 55s 113ms/step - loss: 2.2723 - accuracy: 0.5746\n",
            "Epoch 55/100\n",
            "484/484 [==============================] - 53s 110ms/step - loss: 2.2264 - accuracy: 0.5823\n",
            "Epoch 56/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 2.1900 - accuracy: 0.5902\n",
            "Epoch 57/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 2.1294 - accuracy: 0.6066\n",
            "Epoch 58/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 2.1016 - accuracy: 0.6068\n",
            "Epoch 59/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 2.0688 - accuracy: 0.6214\n",
            "Epoch 60/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 2.0213 - accuracy: 0.6242\n",
            "Epoch 61/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.9882 - accuracy: 0.6381\n",
            "Epoch 62/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.9678 - accuracy: 0.6391\n",
            "Epoch 63/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.9271 - accuracy: 0.6523\n",
            "Epoch 64/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.8840 - accuracy: 0.6596\n",
            "Epoch 65/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.8491 - accuracy: 0.6683\n",
            "Epoch 66/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.8308 - accuracy: 0.6700\n",
            "Epoch 67/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.7965 - accuracy: 0.6788\n",
            "Epoch 68/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 1.7762 - accuracy: 0.6817\n",
            "Epoch 69/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.7403 - accuracy: 0.6901\n",
            "Epoch 70/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.7075 - accuracy: 0.6987\n",
            "Epoch 71/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.6918 - accuracy: 0.7007\n",
            "Epoch 72/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.6587 - accuracy: 0.7072\n",
            "Epoch 73/100\n",
            "484/484 [==============================] - 58s 120ms/step - loss: 1.6422 - accuracy: 0.7080\n",
            "Epoch 74/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.6203 - accuracy: 0.7146\n",
            "Epoch 75/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 1.5887 - accuracy: 0.7235\n",
            "Epoch 76/100\n",
            "484/484 [==============================] - 55s 113ms/step - loss: 1.5641 - accuracy: 0.7300\n",
            "Epoch 77/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 1.5448 - accuracy: 0.7368\n",
            "Epoch 78/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.5275 - accuracy: 0.7343\n",
            "Epoch 79/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.5123 - accuracy: 0.7327\n",
            "Epoch 80/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.4906 - accuracy: 0.7401\n",
            "Epoch 81/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.4680 - accuracy: 0.7446\n",
            "Epoch 82/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.4454 - accuracy: 0.7507\n",
            "Epoch 83/100\n",
            "484/484 [==============================] - 54s 111ms/step - loss: 1.4361 - accuracy: 0.7511\n",
            "Epoch 84/100\n",
            "484/484 [==============================] - 58s 120ms/step - loss: 1.4197 - accuracy: 0.7504\n",
            "Epoch 85/100\n",
            "484/484 [==============================] - 53s 109ms/step - loss: 1.3991 - accuracy: 0.7617\n",
            "Epoch 86/100\n",
            "484/484 [==============================] - 53s 109ms/step - loss: 1.3865 - accuracy: 0.7585\n",
            "Epoch 87/100\n",
            "484/484 [==============================] - 52s 108ms/step - loss: 1.3700 - accuracy: 0.7630\n",
            "Epoch 88/100\n",
            "484/484 [==============================] - 52s 108ms/step - loss: 1.3552 - accuracy: 0.7648\n",
            "Epoch 89/100\n",
            "484/484 [==============================] - 53s 109ms/step - loss: 1.3281 - accuracy: 0.7738\n",
            "Epoch 90/100\n",
            "484/484 [==============================] - 52s 108ms/step - loss: 1.3257 - accuracy: 0.7734\n",
            "Epoch 91/100\n",
            "484/484 [==============================] - 52s 108ms/step - loss: 1.3214 - accuracy: 0.7716\n",
            "Epoch 92/100\n",
            "484/484 [==============================] - 53s 109ms/step - loss: 1.2991 - accuracy: 0.7755\n",
            "Epoch 93/100\n",
            "484/484 [==============================] - 60s 124ms/step - loss: 1.2726 - accuracy: 0.7817\n",
            "Epoch 94/100\n",
            "484/484 [==============================] - 55s 113ms/step - loss: 1.2614 - accuracy: 0.7859\n",
            "Epoch 95/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 1.2651 - accuracy: 0.7824\n",
            "Epoch 96/100\n",
            "484/484 [==============================] - 60s 124ms/step - loss: 1.2584 - accuracy: 0.7804\n",
            "Epoch 97/100\n",
            "484/484 [==============================] - 55s 113ms/step - loss: 1.2429 - accuracy: 0.7846\n",
            "Epoch 98/100\n",
            "484/484 [==============================] - 54s 112ms/step - loss: 1.2341 - accuracy: 0.7868\n",
            "Epoch 99/100\n",
            "484/484 [==============================] - 55s 113ms/step - loss: 1.2064 - accuracy: 0.7927\n",
            "Epoch 100/100\n",
            "484/484 [==============================] - 55s 113ms/step - loss: 1.1971 - accuracy: 0.7929\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "1fXTEO3GJ282",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 545
        },
        "outputId": "42972426-22d3-44f4-d48e-dee60dd8e310"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "acc = history.history['accuracy']\n",
        "loss = history.history['loss']\n",
        "\n",
        "epochs = range(len(acc))\n",
        "\n",
        "plt.plot(epochs, acc, 'b', label='Training accuracy')\n",
        "plt.title('Training accuracy')\n",
        "\n",
        "plt.figure()\n",
        "\n",
        "plt.plot(epochs, loss, 'b', label='Training Loss')\n",
        "plt.title('Training loss')\n",
        "plt.legend()\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "6Vc6PHgxa6Hm",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 104
        },
        "outputId": "46f51f81-607d-4889-d483-a9765356e4ea"
      },
      "source": [
        "seed_text = \"Help me Obi Wan Kenobi, you're my only hope\"\n",
        "next_words = 100\n",
        "  \n",
        "for _ in range(next_words):\n",
        "\ttoken_list = tokenizer.texts_to_sequences([seed_text])[0]\n",
        "\ttoken_list = pad_sequences([token_list], maxlen=max_sequence_len-1, padding='pre')\n",
        "\tpredicted = model.predict_classes(token_list, verbose=0)\n",
        "\toutput_word = \"\"\n",
        "\tfor word, index in tokenizer.word_index.items():\n",
        "\t\tif index == predicted:\n",
        "\t\t\toutput_word = word\n",
        "\t\t\tbreak\n",
        "\tseed_text += \" \" + output_word\n",
        "print(seed_text)"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From <ipython-input-7-622d307fa19a>:7: Sequential.predict_classes (from tensorflow.python.keras.engine.sequential) is deprecated and will be removed after 2021-01-01.\n",
            "Instructions for updating:\n",
            "Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "Help me Obi Wan Kenobi, you're my only hope his thought more perjured lie to thee of too hell that look you lack of words ride date new ill young well or memory new dear friend eyes will you thine friend not not not love's friend too words that show thee young hell of best so gone out he so write her eyes of eyes false days night will read new allow torn out thee for all memory true ' find thine friend have live do live you live still of every pen hate days lose groan still me being out to trust thee so so much show thee\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}