Simple_Autoencoder_for_text.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Simple_Autoencoder_for_text.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "VcANdQ6-u8fs",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 72
        },
        "outputId": "77a9afd8-5157-4a67-b770-921355d550e8"
      },
      "source": [
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "\n",
        "import gensim.downloader as api\n",
        "from gensim.models.word2vec import Word2Vec\n",
        "\n",
        "corpus = api.load('text8')\n",
        "w2v = Word2Vec(corpus)\n",
        "w2v.wv.add('<END>', np.random.rand(100))"
      ],
      "execution_count": 75,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/smart_open/smart_open_lib.py:253: UserWarning: This function is deprecated, use smart_open.open instead. See the migration notes for details: https://github.com/RaRe-Technologies/smart_open/blob/master/README.rst#migrating-to-the-new-open-function\n",
            "  'See the migration notes for details: %s' % _MIGRATION_NOTES_URL\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WJ5R53NLvAuu",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class Encoder(tf.keras.layers.Layer):\n",
        "    def __init__(self, intermediate_dim):\n",
        "        super(Encoder, self).__init__()\n",
        "        self.hidden_layer = tf.keras.layers.Dense(units=intermediate_dim, activation=tf.nn.sigmoid)\n",
        "        self.output_layer = tf.keras.layers.Dense(units=intermediate_dim, activation=tf.nn.sigmoid)\n",
        "    \n",
        "    def call(self, input_features):\n",
        "        return self.output_layer(self.hidden_layer(input_features))\n",
        "\n",
        "class Decoder(tf.keras.layers.Layer):\n",
        "    def __init__(self, intermediate_dim, original_dim):\n",
        "        super(Decoder, self).__init__()\n",
        "        self.hidden_layer = tf.keras.layers.Dense(units=intermediate_dim, activation=tf.nn.sigmoid)\n",
        "        self.output_layer = tf.keras.layers.Dense(units=original_dim, activation=tf.nn.sigmoid)\n",
        "  \n",
        "    def call(self, code):\n",
        "        return self.output_layer(self.hidden_layer(code))\n",
        "\n",
        "class Autoencoder(tf.keras.Model):\n",
        "  def __init__(self, intermediate_dim, original_dim):\n",
        "    super(Autoencoder, self).__init__()\n",
        "    self.loss = []\n",
        "    self.encoder = Encoder(intermediate_dim=intermediate_dim)\n",
        "    self.decoder = Decoder(intermediate_dim=intermediate_dim, original_dim=original_dim)\n",
        "\n",
        "  def call(self, input_features):\n",
        "    return self.decoder(self.encoder(input_features))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vV97qVN7vDlh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def loss(preds, real):\n",
        "  return tf.reduce_mean(tf.square(tf.subtract(preds, real)))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DzrGTMLPvFvf",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def train(loss, model, opt, original):\n",
        "  with tf.GradientTape() as tape:\n",
        "    preds = model(original)\n",
        "    reconstruction_error = loss(preds, original)\n",
        "    gradients = tape.gradient(reconstruction_error, model.trainable_variables)\n",
        "    gradient_variables = zip(gradients, model.trainable_variables)\n",
        "    opt.apply_gradients(gradient_variables)\n",
        "  \n",
        "  return reconstruction_error"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "If46P2rjvH6G",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def train_loop(model, opt, loss, dataset, epochs, print_steps):\n",
        "  for epoch in range(epochs):\n",
        "    epoch_loss = 0\n",
        "    for step, batch_features in enumerate(dataset):\n",
        "      loss_values = train(loss, model, opt, batch_features)\n",
        "      epoch_loss += loss_values\n",
        "    model.loss.append(epoch_loss)\n",
        "    if (epoch + 1)%print_steps == 0:\n",
        "      print('Epoch {}/{}. Loss: {}'.format(epoch + 1, epochs, epoch_loss.numpy()))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-t99EgDeVGu_",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 72
        },
        "outputId": "3f4c3a7a-9c55-424e-e90e-f6d7a9e9dc40"
      },
      "source": [
        "input_text = [\"it is a good day to start\",\n",
        "              \"that sport car is dark red and i love it\",\n",
        "              \"everybody likes dark theme when coding\",\n",
        "              \"i like to move it move it\",\n",
        "              \"i am going to code\",\n",
        "              ]\n",
        "\n",
        "input_text = tf.keras.preprocessing.sequence.pad_sequences(\n",
        "            [i.split() for i in input_text],\n",
        "            padding='post', dtype=object, value='<END>')\n",
        "\n",
        "seq_lenght = len(input_text[0])\n",
        "vector_lenght = seq_lenght*100\n",
        "\n",
        "x_train = [[w2v[j] for j in i] for i in input_text]\n",
        "x_train = np.asarray(x_train, dtype=np.float32)\n",
        "x_train = np.reshape(x_train, (x_train.shape[0], vector_lenght))\n",
        "\n",
        "training_dataset = tf.data.Dataset.from_tensor_slices(x_train).batch(5)"
      ],
      "execution_count": 108,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:15: DeprecationWarning: Call to deprecated `__getitem__` (Method will be removed in 4.0.0, use self.wv.__getitem__() instead).\n",
            "  from ipykernel import kernelapp as app\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Q2R1ABwTvbyw",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 367
        },
        "outputId": "f0bc6fae-736b-4a47-961d-028a26ec38d8"
      },
      "source": [
        "model = Autoencoder(intermediate_dim=128, original_dim=vector_lenght)\n",
        "opt = tf.keras.optimizers.Adam(learning_rate=0.01)\n",
        "\n",
        "train_loop(model, opt, loss, training_dataset, 5000, 1000)\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "plt.plot(range(len(model.loss)), model.loss)\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Loss')\n",
        "plt.show()"
      ],
      "execution_count": 109,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1000/5000. Loss: 0.9618300795555115\n",
            "Epoch 2000/5000. Loss: 0.9292380809783936\n",
            "Epoch 3000/5000. Loss: 0.9168202877044678\n",
            "Epoch 4000/5000. Loss: 0.905506432056427\n",
            "Epoch 5000/5000. Loss: 0.8959076404571533\n"
          ],
          "name": "stdout"
        },
        {
          "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": {
        "id": "98dvYmmXvhg2",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 336
        },
        "outputId": "aee1cd19-226f-4480-8733-0877403b53da"
      },
      "source": [
        "for predict in model(x_train):\n",
        "  predict = predict.numpy().reshape(seq_lenght, 100)\n",
        "  result = [w2v.wv.similar_by_vector(i)[0][0] for i in predict]\n",
        "  print()\n",
        "  print('     ',*[i for i in result if not i=='<END>'])\n",
        "  print('_____________________________________________________________________')"
      ],
      "execution_count": 110,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "      it is a good day to start\n",
            "_____________________________________________________________________\n",
            "\n",
            "      that sport car is dark red and i love it\n",
            "_____________________________________________________________________\n",
            "\n",
            "      everyone likes dark theme when coding\n",
            "_____________________________________________________________________\n",
            "\n",
            "      i like to move it move it\n",
            "_____________________________________________________________________\n",
            "\n",
            "      i am going to code\n",
            "_____________________________________________________________________\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/gensim/matutils.py:737: FutureWarning: Conversion of the second argument of issubdtype from `int` to `np.signedinteger` is deprecated. In future, it will be treated as `np.int64 == np.dtype(int).type`.\n",
            "  if np.issubdtype(vec.dtype, np.int):\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7NZFcjPqv6fR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}