Attention_seq2seq.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Attention_seq2seq.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "machine_shape": "hm"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sBNQOOrN1-AK",
        "colab_type": "text"
      },
      "source": [
        "### Attention Based Machine Translation\n",
        "\n",
        "In this note I will build an attention-based LSTM Seq2Seq model for English to Chinese Translation. The details of the Luong Attention model can be found in this paper: [Efective Approaches to Attention-based Neural Machine Translation](https://arxiv.org/pdf/1508.04025.pdf), and dot score is being used for the calculation of alignment weight vector. I will explain the construction of attention layer using visualization.\n",
        "\n",
        "The Chinese-English dataset can be downloaded at http://www.manythings.org/anki/. <br>\n",
        "\n",
        "### Prepare dataset\n",
        "- Add a start of sentence `<sos>` and end of sentence `<eos>` token to each sentence.\n",
        "- Clean the sentences by removing special characters.\n",
        "- Create a word index and reverse word index\n",
        "- Pad each sentence to a maximum length."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6QOOVRhZ5-Rb",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import re\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "import tensorflow.keras as keras\n",
        "from tensorflow.keras.layers import Dense, Embedding, LSTM\n",
        "from tensorflow.keras.preprocessing.text import Tokenizer\n",
        "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
        "from tensorflow.keras.callbacks import EarlyStopping"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ogvg9nRO3fF6",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def preprocess_en_sentence(w):\n",
        "  w = w.lower()\n",
        "\n",
        "  # Add a space between a word and the punctuation following it\n",
        "  # eg: \"he is a boy.\" => \"he is a boy . \" => \"he is a boy .\"\n",
        "  w = re.sub(r\"([?.!,])\", r\" \\1 \", w)\n",
        "  w = re.sub(r'\\s{2,}', \" \", w)\n",
        "\n",
        "  # replacing everything with space except (a-z, A-Z, \".\", \"?\", \"!\", \",\")\n",
        "  w = re.sub(r\"[^a-zA-Z?.!,]+\", \" \", w)\n",
        "  w = w.strip()\n",
        "  return w\n",
        "\n",
        "def preprocess_cn_sentence(w):\n",
        "  # Remove special punctuations\n",
        "  w = re.sub(r'[“”()]', '', w)\n",
        "\n",
        "  # Add empty spaces between words\n",
        "  w = ' '.join([char for char in w])\n",
        "\n",
        "  # Add a start and an end token to the sentence\n",
        "  w = '<sos> ' + w + ' <eos>'\n",
        "  return w"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HDEvHcv355ob",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 72
        },
        "outputId": "49cc10b2-a876-403c-9471-9cf2f0ef61b3"
      },
      "source": [
        "print(preprocess_en_sentence('Hi! How can I help you?'))\n",
        "print()\n",
        "print(preprocess_cn_sentence('你好！我可以帮什么忙吗？'))"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "hi ! how can i help you ?\n",
            "\n",
            "<sos> 你 好 ！ 我 可 以 帮 什 么 忙 吗 ？ <eos>\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lVQc5sBA19IK",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "num_examples = 17000\n",
        "en_sentences = []\n",
        "cn_sentences = []\n",
        "\n",
        "with open('cmn.txt', 'r', encoding='utf-8') as f:\n",
        "  lines = f.read().split('\\n')\n",
        "  for line in lines[:min(len(lines)-1, num_examples)]:\n",
        "      en_sentence, cn_sentence, _ = line.split('\\t')\n",
        "      en_sentences.append(preprocess_en_sentence(en_sentence))\n",
        "      cn_sentences.append(preprocess_cn_sentence(cn_sentence))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5foQrBFW1-vp",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 54
        },
        "outputId": "7e618f19-f908-40c3-af4a-05fa7cafecc4"
      },
      "source": [
        "print(en_sentences[-1])\n",
        "print(cn_sentences[-1])"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "what i want is some peace and quiet .\n",
            "<sos> 我 所 想 要 的 是 一 点 平 和 和 安 静 。 <eos>\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XTN3txQq9W45",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def tokenize(texts):\n",
        "  tokenizer = Tokenizer(filters='')\n",
        "  tokenizer.fit_on_texts(texts)\n",
        "  tensor = tokenizer.texts_to_sequences(texts)\n",
        "  tensor = pad_sequences(tensor, padding='post')\n",
        "  return tensor, tokenizer"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xwZSa4Cu-JMF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "encoder_tensor, encoder_tokenizer = tokenize(en_sentences)\n",
        "decoder_tensor, decoder_tokenizer = tokenize(cn_sentences)\n",
        "\n",
        "# Calculate max_length of the encoder and decoder tensors\n",
        "max_len_encoder = encoder_tensor.shape[1]\n",
        "max_len_decoder = decoder_tensor.shape[1]\n",
        "\n",
        "# Calculate number of unique tokens for the encoder and decoder\n",
        "num_words_encoder = len(encoder_tokenizer.word_index) + 1\n",
        "num_words_decoder = len(decoder_tokenizer.word_index) + 1"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8FOb8Jy_rWBr",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 54
        },
        "outputId": "60ff16ad-138e-4042-d7cc-a327dee0d04a"
      },
      "source": [
        "print(encoder_tensor.shape)\n",
        "print(decoder_tensor.shape)"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(17000, 12)\n",
            "(17000, 25)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sK7PxG-ZAUxm",
        "colab_type": "text"
      },
      "source": [
        "### Encoder and Decoder model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AVvsz-IZAVwU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "hidden_units = 1024\n",
        "embedding_dim = 256"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oIonGWuwb3Gb",
        "colab_type": "text"
      },
      "source": [
        "#### Encoder\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lZtGUotTDhgd",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "encoder_inputs = keras.Input(shape=(max_len_encoder,))\n",
        "encoder_embedding = Embedding(num_words_encoder, \n",
        "                              embedding_dim, \n",
        "                              mask_zero=True,\n",
        "                              input_length=max_len_encoder)(encoder_inputs)\n",
        "encoder_outputs, encoder_state_h, encoder_state_c = LSTM(\n",
        "    hidden_units, return_state=True, \n",
        "    return_sequences=True, unroll=True)(encoder_embedding)\n",
        "encoder_states = [encoder_state_h, encoder_state_c]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "b_XSMF9RPpZW",
        "colab_type": "text"
      },
      "source": [
        "There are two internal states in LSTM, the **hidden state**, usually denoted by **h**, and the **cell state** usually denoted by **c**. The tuple **(c, h)** is the **internal state** of a LSTM. The hidden state h is output for every timestep the LSTM is unrolled. \n",
        "\n",
        "LSTM with `return_sequences=True` returns the hidden state of the LSTM for every timestep in the input to the LSTM. \n",
        "\n",
        "For example, if the input batch is (samples, timesteps, dims), then the call `LSTM(units, return_sequences=True)` will generate output of dimensions (samples, timesteps, units).\n",
        "\n",
        "However, if you also want the final internal cell state of LSTM, use `LSTM(units, return_sequence=True, return_state=True)` instead. <br><br>\n",
        "\n",
        "<img src='https://drive.google.com/uc?id=1uyRKZZ2gl1iSFO65Oq-XY8abTCiXlGNO'></img>\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QuMzxhwubjhA",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 72
        },
        "outputId": "bb34c7fc-d721-4849-a748-9c6072180d11"
      },
      "source": [
        "print('encoder_outputs: ', encoder_outputs)\n",
        "print('encoder_state_h: ', encoder_state_h)\n",
        "print('encoder_state_c: ', encoder_state_c)"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "encoder_outputs:  Tensor(\"lstm/Identity:0\", shape=(None, 12, 1024), dtype=float32)\n",
            "encoder_state_h:  Tensor(\"lstm/Identity_1:0\", shape=(None, 1024), dtype=float32)\n",
            "encoder_state_c:  Tensor(\"lstm/Identity_2:0\", shape=(None, 1024), dtype=float32)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rdGqJlYmb6ZT",
        "colab_type": "text"
      },
      "source": [
        "#### Decoder"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SrXtiWHPMyFw",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "decoder_inputs = keras.Input(shape=(max_len_decoder,))\n",
        "decoder_embedding = Embedding(num_words_decoder, \n",
        "                           embedding_dim, \n",
        "                           input_length=max_len_decoder)(decoder_inputs)\n",
        "decoder_outputs, decoder_state_h, decoder_state_c = LSTM(\n",
        "    hidden_units, return_sequences=True, return_state=True)(\n",
        "        decoder_embedding, initial_state=encoder_states)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_kriXaDpbhK-",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 72
        },
        "outputId": "545a3314-ffaa-4662-cf43-682ecac68bb2"
      },
      "source": [
        "print('decoder_outputs: ', decoder_outputs)\n",
        "print('decoder_state_h: ', decoder_state_h)\n",
        "print('decoder_state_c: ', decoder_state_c)"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "decoder_outputs:  Tensor(\"lstm_1/Identity:0\", shape=(None, 25, 1024), dtype=float32)\n",
            "decoder_state_h:  Tensor(\"lstm_1/Identity_1:0\", shape=(None, 1024), dtype=float32)\n",
            "decoder_state_c:  Tensor(\"lstm_1/Identity_2:0\", shape=(None, 1024), dtype=float32)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Wt5D00WZInWA",
        "colab_type": "text"
      },
      "source": [
        "#### Attention Layer \n",
        "\n",
        "This Attention model is based on the paper: [Efective Approaches to Attention-based Neural Machine Translation](https://arxiv.org/pdf/1508.04025.pdf)."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "60zuszGDRGV4",
        "colab_type": "text"
      },
      "source": [
        "Alignment weight vector $\\alpha_t$ based on the current target hidden state $h_t$ and all sources states $\\bar{h}_s$. <br><br>\n",
        "\n",
        "\n",
        "$$\\alpha_t(s) = align(h_t, \\bar{h_s})=\\frac{exp(score(h_t, \\bar{h_s}))}{\\sum_{s^{'}}exp(score(h_t, \\bar{h_{s^{'}}}))}$$\n",
        "\n",
        "<br>\n",
        "Where\n",
        "\n",
        "$$h_t : \\text{current target hidden state}$$\n",
        "<br>\n",
        "$$\\bar{h}_s :  \\text{source hidden state}$$\n",
        "<br>\n",
        "\n",
        "We applied dot score here:<br>\n",
        "\n",
        "$$score(h_t, \\bar{h_s}) = h_t^T\\bar{h_s}$$\n",
        "\n",
        "<br>\n",
        "\n",
        "Documentation for **Dot** layer: https://keras.io/api/layers/merging_layers/dot/\n",
        "\n",
        "> Layer that computes a dot product between samples in two tensors. \n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z19k1QJ9e8tY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from tensorflow.keras.layers import (Dot, Activation, \n",
        "                                     concatenate, \n",
        "                                     TimeDistributed)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-FZV-YxsHfNH",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "07309776-2f07-4a99-fa17-306c4fcb10b9"
      },
      "source": [
        "# Attention Layer\n",
        "attention = Dot(axes=2)([decoder_outputs, encoder_outputs])\n",
        "attention = Activation('softmax', name='attention')(attention)\n",
        "print('attention', attention)"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "attention Tensor(\"attention/Identity:0\", shape=(None, 25, 12), dtype=float32)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "w-uo2SeCNq-1",
        "colab_type": "text"
      },
      "source": [
        "This process is illustrated below: <br><br>\n",
        "<img src='https://drive.google.com/uc?id=1--xGolyVK-UWkv6S17GM17mMGsbElGHV'></img>\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hP8W4r1ST8T3",
        "colab_type": "text"
      },
      "source": [
        "A **context vector** $c_t$ is then computed as the weighted average of all source states, according to the alignment weight $\\alpha_t$."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cfFeagNOg4Fo",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "2cd582ae-93f8-4276-a90e-0a68a615ed38"
      },
      "source": [
        "context = Dot(axes=[2, 1])([attention, encoder_outputs])\n",
        "print('context', context)"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "context Tensor(\"dot_1/Identity:0\", shape=(None, 25, 1024), dtype=float32)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RyeUXEGpWArG",
        "colab_type": "text"
      },
      "source": [
        "This process is illustrated below: <br><br>\n",
        "<img src='https://drive.google.com/uc?id=1bIGg3_wl3wqYLBGEgx56heKRNSNn-DWc' width=1000></img>\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "spVUqpupUcLI",
        "colab_type": "text"
      },
      "source": [
        "Then given the target hidden state $h_t$ and\n",
        "the source-side context vector $c_t$, we employ a\n",
        "simple concatenation layer to combine the information from both vectors to produce an attentional\n",
        "hidden state:\n",
        "\n",
        "$$[c_t; h_t]$$"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bsSbmvZQHfKd",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "117e5fae-81f1-4c85-855f-b32939507810"
      },
      "source": [
        "decoder_combined_context = concatenate([context, decoder_outputs])\n",
        "print('decoder_combined_context', decoder_combined_context)"
      ],
      "execution_count": 127,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "decoder_combined_context Tensor(\"concatenate_1/Identity:0\", shape=(None, 25, 2048), dtype=float32)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_09uS6GSahOB",
        "colab_type": "text"
      },
      "source": [
        "<img src='https://drive.google.com/uc?id=1ZcyT28CbhhLRAQ7-SCnjOBF85dix9qay' width=1000></img>\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rMR0HdJ3U2ME",
        "colab_type": "text"
      },
      "source": [
        "Then we will produce an **attentional hidden state** $\\tilde{h}_t$ as follows: <br> <br>\n",
        "\n",
        "$$\\tilde{h}_t = tanh(W_c[c_t;h_t])$$\n",
        "\n",
        "<br>\n",
        "\n",
        "This attentional hidden state will then be fed through the **softmax layer** to produce the predictive distribution. <br> <br>\n",
        "\n",
        "$$p(y_t|y_{<t}, x) = softmax(W_s \\tilde{h}_t)$$"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c442MnIjWbu1",
        "colab_type": "text"
      },
      "source": [
        "Documentatin of **TimeDistributed** Layer: https://keras.io/api/layers/recurrent_layers/time_distributed/\n",
        "\n",
        ">This wrapper allows to apply a layer to every temporal slice of an input. The input should be at least 3D, and the dimension of index one will be considered to be the temporal dimension.\n",
        ">\n",
        ">For example, consider a batch of 32 video samples, where each sample is a 128x128 RGB image with channels_last data format, across 10 timesteps. The batch input shape is (32, 10, 128, 128, 3)."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lAZwd0mBHfIM",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "ad0df604-5876-40a4-e65c-cf08fb4bc955"
      },
      "source": [
        "output = TimeDistributed(Dense(hidden_units, activation=\"tanh\"))(decoder_combined_context)\n",
        "output = TimeDistributed(Dense(num_words_decoder, activation=\"softmax\"))(output)\n",
        "print('output', output)"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "output Tensor(\"time_distributed_1/Identity:0\", shape=(None, 25, 3040), dtype=float32)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "p4m1Vpyakcfq",
        "colab_type": "text"
      },
      "source": [
        "<img src='https://drive.google.com/uc?id=1lWueNI_nNByU7_lQqCvroMaHafxE3C8W' width=1000></img>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rdVaOiX3on_K",
        "colab_type": "text"
      },
      "source": [
        "#### The Whole Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "waPhDQTYonXN",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model = keras.Model(inputs=[encoder_inputs, decoder_inputs], outputs=[output])\n",
        "model.compile(optimizer='rmsprop', \n",
        "              loss='categorical_crossentropy',\n",
        "              metrics=['accuracy'])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Mz7EWVHcpuuN",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "0d7a388d-5ea1-42b6-9f79-6bdbfd3385a0"
      },
      "source": [
        "tf.keras.utils.plot_model(model, to_file='my_model.png', show_shapes=True)"
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YcTg2GgYo71X",
        "colab_type": "text"
      },
      "source": [
        "#### Model Training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bOUKyb-ADhjn",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "encoder_input = encoder_tensor.copy()\n",
        "\n",
        "decoder_input = np.zeros_like(decoder_tensor)\n",
        "decoder_input[:, 1:] = decoder_tensor[:,:-1]  # The last one is not included\n",
        "decoder_input[:, 0] = decoder_tokenizer.word_index['<sos>']\n",
        "\n",
        "# decoder_output is the same as decoder_input but offset by one timestep.\n",
        "decoder_output =np.zeros_like(decoder_tensor)\n",
        "decoder_output[:, :-1] = decoder_tensor[:,1:]  # The first one is not included\n",
        "decoder_output = np.eye(num_words_decoder)[decoder_output.astype('int')]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "df0GHk_etmul",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 72
        },
        "outputId": "3a1be287-1fff-4198-b5b0-0fd822dcec23"
      },
      "source": [
        "print(encoder_input.shape)\n",
        "print(decoder_input.shape)\n",
        "print(decoder_output.shape)"
      ],
      "execution_count": 84,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(17000, 12)\n",
            "(17000, 25)\n",
            "(17000, 25, 3040)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lNXhWTjKeTU3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "my_callbacks = [EarlyStopping(patience=3)]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M0LmP0J5tB2f",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "outputId": "a70abe29-9501-497d-a3e2-79eb88c894af"
      },
      "source": [
        "BATCH_SIZE = 16\n",
        "EPOCHS = 10\n",
        "model.fit([encoder_input, decoder_input], \n",
        "           decoder_output,\n",
        "           batch_size=BATCH_SIZE,\n",
        "           epochs=EPOCHS,\n",
        "           callbacks=my_callbacks,\n",
        "           validation_split=0.2)"
      ],
      "execution_count": 86,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n",
            "850/850 [==============================] - 830s 977ms/step - loss: 1.0094 - accuracy: 0.8112 - val_loss: 1.8427 - val_accuracy: 0.6941\n",
            "Epoch 2/10\n",
            "850/850 [==============================] - 830s 976ms/step - loss: 0.8721 - accuracy: 0.8331 - val_loss: 1.8417 - val_accuracy: 0.6998\n",
            "Epoch 3/10\n",
            "850/850 [==============================] - 846s 995ms/step - loss: 0.7285 - accuracy: 0.8585 - val_loss: 1.8453 - val_accuracy: 0.7014\n",
            "Epoch 4/10\n",
            "850/850 [==============================] - 836s 984ms/step - loss: 0.5986 - accuracy: 0.8840 - val_loss: 1.8776 - val_accuracy: 0.7036\n",
            "Epoch 5/10\n",
            "850/850 [==============================] - 840s 988ms/step - loss: 0.4976 - accuracy: 0.9068 - val_loss: 1.8987 - val_accuracy: 0.7046\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7f5046966748>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 86
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fKvA1rIL0tO6",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model.save('seq2seq_attention_model.h5')\n",
        "# from google.colab import files\n",
        "# files.download('seq2seq_attention_model.h5')\n",
        "# from keras.models import load_model\n",
        "# model = load_model('seq2seq_attention_model.h5')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CUR-tC46cxjk",
        "colab_type": "text"
      },
      "source": [
        "<img src='https://drive.google.com/uc?id=1DbDSjuv2a5-XVgXaVg6bRijJ-69Dj-7y' width=800></img>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-CFcsUWbeIm0",
        "colab_type": "text"
      },
      "source": [
        "Note the `decoder input` at time step `t` is the `decoder output` at time step `t-1`."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "13YTe1zGte5m",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def predict(text):\n",
        "  text = preprocess_en_sentence(text)\n",
        "  encoder_input = pad_sequences(encoder_tokenizer.texts_to_sequences([text]), \n",
        "                                maxlen=max_len_encoder, padding='post')\n",
        "  decoder_input = np.zeros(shape=(len(encoder_input), max_len_decoder))\n",
        "  decoder_input[:,0] = decoder_tokenizer.word_index['<sos>']\n",
        "  for i in range(1, max_len_decoder):\n",
        "      output = model.predict([encoder_input, decoder_input]).argmax(axis=2)\n",
        "      decoder_input[:,i] = int(output[:,i-1])\n",
        "  return decoder_input[:, 1:][0]\n",
        "\n",
        "def decode(index_sequence):\n",
        "  text = ''\n",
        "  for index in index_sequence:\n",
        "      if index == 0:\n",
        "         text = text \n",
        "      elif index == decoder_tokenizer.word_index['<eos>']:\n",
        "         break\n",
        "      else:\n",
        "         text += decoder_tokenizer.index_word[index]\n",
        "  return text"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fCG-neDf4kqF",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 72
        },
        "outputId": "c6160249-98f7-444a-a84d-28a18f01aa05"
      },
      "source": [
        "print(decode(predict(text='How is your day?')))\n",
        "print(decode(predict(text='I like you!')))\n",
        "print(decode(predict(text='I want to eat dinner!')))"
      ],
      "execution_count": 114,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "你天怎样？\n",
            "我喜你\n",
            "我想吃饭饭饭？\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6UYuz9mr9Z21",
        "colab_type": "text"
      },
      "source": [
        "### Visualize Attention Layer"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rQczaSB07lFT",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 54
        },
        "outputId": "a0b73fd7-89b6-42e9-ade0-69594fef9f8c"
      },
      "source": [
        "attention_layer = model.get_layer('attention') \n",
        "attention_model = keras.Model(inputs=model.inputs, \n",
        "                              outputs=model.outputs + [attention_layer.output])\n",
        "print(attention_model)\n",
        "print(attention_model.output_shape)"
      ],
      "execution_count": 90,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<tensorflow.python.keras.engine.training.Model object at 0x7f50649bc6d8>\n",
            "[(None, 25, 3040), (None, 25, 12)]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PvHOdFrIG11v",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Use external font for chinese character\n",
        "import matplotlib.font_manager as fm\n",
        "path = '/usr/share/fonts/truetype/chinese.stsong.ttf'\n",
        "fontprop = fm.FontProperties(fname=path, size=14)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P2zFTdmY7zbd",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import matplotlib as mpl\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "def attent_and_generate(text):\n",
        "    text = preprocess_en_sentence(text)\n",
        "    encoder_input = pad_sequences(encoder_tokenizer.texts_to_sequences([text]), \n",
        "                                  maxlen=max_len_encoder, padding='post')\n",
        "    decoder_input = np.zeros(shape=(len(encoder_input), max_len_decoder))\n",
        "    decoder_input[:,0] = decoder_tokenizer.word_index['<sos>']\n",
        "    \n",
        "    for i in range(1, max_len_decoder):\n",
        "        output, attention = attention_model.predict([encoder_input, decoder_input])\n",
        "        decoder_input[:,i] = output.argmax(axis=2)[:,i-1]\n",
        "        attention_density = attention[0]\n",
        "        decoded_output = decode(decoder_input[0][1:])\n",
        "        \n",
        "    return text, attention_density, decoded_output\n",
        "\n",
        "def visualize(text):\n",
        "    text, attention_density, chinese = attent_and_generate(text)\n",
        "    xticklabels = [w for w in text.split()]\n",
        "    yticklabels = [w for w in chinese]\n",
        "\n",
        "    plt.clf()\n",
        "    plt.figure(figsize=(8, 6))\n",
        "    ax = sns.heatmap(attention_density[:len(chinese), :len(text.split())], \n",
        "                     cmap='viridis')\n",
        "    ax.set_xticklabels(xticklabels)\n",
        "    ax.set_yticklabels(yticklabels, fontproperties=fontprop)\n",
        "    ax.invert_yaxis()\n",
        "    plt.show()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sXnzvABK8Oom",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 395
        },
        "outputId": "5862f25c-b8bd-4067-a093-41237a9dcabb"
      },
      "source": [
        "txt = 'How can I help you?'\n",
        "visualize(txt)\n",
        "plt.show()"
      ],
      "execution_count": 126,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x432 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}