Transfer_Learning_Semantic_Similarity_with_TF_Hub_Universal_Encoder.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "co7MV6sX7Xto"
   },
   "source": [
    "# [Keras + Universal Sentence Encoder = Transfer Learning for text data](https://www.dlology.com/blog/keras-meets-universal-sentence-encoder-transfer-learning-for-text-data/) Tutorial\n",
    "## Universal Sentence Encoder"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "eAVQGidpL8v5"
   },
   "source": [
    "This notebook illustrates how to access the Universal Sentence Encoder and use it for sentence similarity and sentence classification tasks.\n",
    "\n",
    "The Universal Sentence Encoder makes getting sentence level embeddings as easy as it has historically been to lookup the embeddings for individual words. The sentence embeddings can then be trivially used to compute sentence level meaning similarity as well as to enable better performance on downstream classification tasks using less supervised training data.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "pOTzp8O36CyQ"
   },
   "source": [
    "# Getting Started\n",
    "\n",
    "This section sets up the environment for access to the Universal Sentence Encoder on TF Hub and provides examples of applying the encoder to words, sentences, and paragraphs."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can skip this if you installed the dependencies from `requirements.txt`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "lVjNK8shFKOC"
   },
   "outputs": [],
   "source": [
    "# Install the latest Tensorflow version.\n",
    "!pip3 install --quiet \"tensorflow>=1.7\"\n",
    "# Install TF-Hub.\n",
    "!pip3 install --quiet tensorflow-hub\n",
    "!pip3 install seaborn"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "63Pd3nJnTl-i"
   },
   "source": [
    "More detailed information about installing Tensorflow can be found at [https://www.tensorflow.org/install/](https://www.tensorflow.org/install/)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 8211,
     "status": "ok",
     "timestamp": 1528598360840,
     "user": {
      "displayName": "zhang chengwei",
      "photoUrl": "//lh3.googleusercontent.com/-cNclsvC1ySc/AAAAAAAAAAI/AAAAAAAAAKI/-krROCWVXVQ/s50-c-k-no/photo.jpg",
      "userId": "107720063101388471292"
     },
     "user_tz": -480
    },
    "id": "MSeY-MUQo2Ha",
    "outputId": "c869b494-6d38-4e89-a49a-814e787e4679"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "import tensorflow_hub as hub\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import os\n",
    "import pandas as pd\n",
    "import re\n",
    "import seaborn as sns\n",
    "import keras.layers as layers\n",
    "from keras.models import Model\n",
    "from keras import backend as K\n",
    "np.random.seed(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "zwty8Z6mAkdV"
   },
   "outputs": [],
   "source": [
    "module_url = \"https://tfhub.dev/google/universal-sentence-encoder-large/3\" #@param [\"https://tfhub.dev/google/universal-sentence-encoder/2\", \"https://tfhub.dev/google/universal-sentence-encoder-large/3\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 2055
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 51216,
     "status": "ok",
     "timestamp": 1528598413651,
     "user": {
      "displayName": "zhang chengwei",
      "photoUrl": "//lh3.googleusercontent.com/-cNclsvC1ySc/AAAAAAAAAAI/AAAAAAAAAKI/-krROCWVXVQ/s50-c-k-no/photo.jpg",
      "userId": "107720063101388471292"
     },
     "user_tz": -480
    },
    "id": "Q8F4LNGFqOiq",
    "outputId": "74dd13b8-2106-4f8e-f78b-260634741098"
   },
   "outputs": [],
   "source": [
    "# Import the Universal Sentence Encoder's TF Hub module\n",
    "embed = hub.Module(module_url)\n",
    "\n",
    "# Compute a representation for each message, showing various lengths supported.\n",
    "word = \"Elephant\"\n",
    "sentence = \"I am a sentence for which I would like to get its embedding.\"\n",
    "paragraph = (\n",
    "    \"Universal Sentence Encoder embeddings also support short paragraphs. \"\n",
    "    \"There is no hard limit on how long the paragraph is. Roughly, the longer \"\n",
    "    \"the more 'diluted' the embedding will be.\")\n",
    "messages = [word, sentence, paragraph]\n",
    "\n",
    "# Reduce logging output.\n",
    "tf.logging.set_verbosity(tf.logging.ERROR)\n",
    "\n",
    "with tf.Session() as session:\n",
    "  session.run([tf.global_variables_initializer(), tf.tables_initializer()])\n",
    "  message_embeddings = session.run(embed(messages))\n",
    "\n",
    "  for i, message_embedding in enumerate(np.array(message_embeddings).tolist()):\n",
    "    print(\"Message: {}\".format(messages[i]))\n",
    "    print(\"Embedding size: {}\".format(len(message_embedding)))\n",
    "    message_embedding_snippet = \", \".join(\n",
    "        (str(x) for x in message_embedding[:3]))\n",
    "    print(\"Embedding: [{}, ...]\\n\".format(message_embedding_snippet))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 90
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 5660,
     "status": "ok",
     "timestamp": 1528598421168,
     "user": {
      "displayName": "zhang chengwei",
      "photoUrl": "//lh3.googleusercontent.com/-cNclsvC1ySc/AAAAAAAAAAI/AAAAAAAAAKI/-krROCWVXVQ/s50-c-k-no/photo.jpg",
      "userId": "107720063101388471292"
     },
     "user_tz": -480
    },
    "id": "lNRSk5lukD_u",
    "outputId": "f4b07c61-fbba-4244-ef9e-ab0a7f04708d"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.06587551,  0.02066354, -0.01454356, ...,  0.06447642,\n",
       "         0.01654527, -0.04688655],\n",
       "       [ 0.06909196,  0.01529877,  0.03278331, ...,  0.01220771,\n",
       "         0.03000253, -0.01277521]], dtype=float32)"
      ]
     },
     "execution_count": 4,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Compute a representation for each message, showing various lengths supported.\n",
    "messages = [\"That band rocks!\", \"That song is really cool.\"]\n",
    "\n",
    "with tf.Session() as session:\n",
    "  session.run([tf.global_variables_initializer(), tf.tables_initializer()])\n",
    "  message_embeddings = session.run(embed(messages))\n",
    "message_embeddings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 750,
     "status": "ok",
     "timestamp": 1528598549895,
     "user": {
      "displayName": "zhang chengwei",
      "photoUrl": "//lh3.googleusercontent.com/-cNclsvC1ySc/AAAAAAAAAAI/AAAAAAAAAKI/-krROCWVXVQ/s50-c-k-no/photo.jpg",
      "userId": "107720063101388471292"
     },
     "user_tz": -480
    },
    "id": "O2cYc2WEkSGP",
    "outputId": "627031de-f995-4bf8-fe55-b8837ff33356"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "512"
      ]
     },
     "execution_count": 5,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "embed_size = embed.get_output_info_dict()['default'].get_shape()[1].value\n",
    "embed_size"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can skip this if you have cloned/download the repo since it includes those two files."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 417
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2824,
     "status": "ok",
     "timestamp": 1528710813182,
     "user": {
      "displayName": "Chengwei Zhang",
      "photoUrl": "//lh5.googleusercontent.com/-FK2ckwmh6mM/AAAAAAAAAAI/AAAAAAAAPLw/SX9b1QAzJ5g/s50-c-k-no/photo.jpg",
      "userId": "114808171854651597062"
     },
     "user_tz": -480
    },
    "id": "UgdEV2sQt3DS",
    "outputId": "6ede117c-547e-46a9-dc05-0ede8ee0719b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--2018-06-11 09:53:31--  https://raw.githubusercontent.com/Tony607/Keras-Text-Transfer-Learning/master/train_5500.txt\r\n",
      "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 151.101.0.133, 151.101.64.133, 151.101.128.133, ...\r\n",
      "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|151.101.0.133|:443... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 335860 (328K) [text/plain]\n",
      "Saving to: ‘train_5500.txt’\n",
      "\n",
      "train_5500.txt      100%[===================>] 327.99K  --.-KB/s    in 0.06s   \n",
      "\n",
      "2018-06-11 09:53:31 (5.03 MB/s) - ‘train_5500.txt’ saved [335860/335860]\n",
      "\n",
      "--2018-06-11 09:53:32--  https://raw.githubusercontent.com/Tony607/Keras-Text-Transfer-Learning/master/test_data.txt\n",
      "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 151.101.0.133, 151.101.64.133, 151.101.128.133, ...\n",
      "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|151.101.0.133|:443... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 23354 (23K) [text/plain]\n",
      "Saving to: ‘test_data.txt’\n",
      "\n",
      "test_data.txt       100%[===================>]  22.81K  --.-KB/s    in 0.01s   \n",
      "\n",
      "2018-06-11 09:53:32 (1.57 MB/s) - ‘test_data.txt’ saved [23354/23354]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "!wget https://raw.githubusercontent.com/Tony607/Keras-Text-Transfer-Learning/master/train_5500.txt\n",
    "!wget https://raw.githubusercontent.com/Tony607/Keras-Text-Transfer-Learning/master/test_data.txt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 198
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 719,
     "status": "ok",
     "timestamp": 1528598786205,
     "user": {
      "displayName": "zhang chengwei",
      "photoUrl": "//lh3.googleusercontent.com/-cNclsvC1ySc/AAAAAAAAAAI/AAAAAAAAAKI/-krROCWVXVQ/s50-c-k-no/photo.jpg",
      "userId": "107720063101388471292"
     },
     "user_tz": -480
    },
    "id": "MElqkcDiDKsp",
    "outputId": "2dcf04b0-c631-484c-9137-363fae821197"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>label</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DESC</td>\n",
       "      <td>How did serfdom develop in and then leave Russ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ENTY</td>\n",
       "      <td>What films featured the character Popeye Doyle ?</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>DESC</td>\n",
       "      <td>How can I find a list of celebrities ' real na...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ENTY</td>\n",
       "      <td>What fowl grabs the spotlight after the Chines...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ABBR</td>\n",
       "      <td>What is the full form of .com ?</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  label                                               text\n",
       "0  DESC  How did serfdom develop in and then leave Russ...\n",
       "1  ENTY   What films featured the character Popeye Doyle ?\n",
       "2  DESC  How can I find a list of celebrities ' real na...\n",
       "3  ENTY  What fowl grabs the spotlight after the Chines...\n",
       "4  ABBR                    What is the full form of .com ?"
      ]
     },
     "execution_count": 10,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def get_dataframe(filename):\n",
    "    lines = open(filename, 'r').read().splitlines()\n",
    "    data = []\n",
    "    for i in range(0, len(lines)):\n",
    "        label = lines[i].split(' ')[0]\n",
    "        label = label.split(\":\")[0]\n",
    "        text = ' '.join(lines[i].split(' ')[1:])\n",
    "        text = re.sub('[^A-Za-z0-9 ,\\?\\'\\\"-._\\+\\!/\\`@=;:]+', '', text)\n",
    "        data.append([label, text])\n",
    "\n",
    "    df = pd.DataFrame(data, columns=['label', 'text'])\n",
    "    df.label = df.label.astype('category')\n",
    "    return df\n",
    "\n",
    "df_train = get_dataframe('train_5500.txt')\n",
    "df_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "4TkOWZyCFmZN"
   },
   "outputs": [],
   "source": [
    "df_train = get_dataframe('test_data.txt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 638,
     "status": "ok",
     "timestamp": 1528598800581,
     "user": {
      "displayName": "zhang chengwei",
      "photoUrl": "//lh3.googleusercontent.com/-cNclsvC1ySc/AAAAAAAAAAI/AAAAAAAAAKI/-krROCWVXVQ/s50-c-k-no/photo.jpg",
      "userId": "107720063101388471292"
     },
     "user_tz": -480
    },
    "id": "3HAtd4X5DayF",
    "outputId": "73ada7e8-e7c9-48d7-de4e-97874207a60d"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6"
      ]
     },
     "execution_count": 13,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "category_counts = len(df_train.label.cat.categories)\n",
    "category_counts"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "Sf9A4Xl6J7c6"
   },
   "source": [
    "## Wrap embed module in a Lambda layer\n",
    "Explicitly cast the input as a string"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "PRD3fWgJjOrP"
   },
   "outputs": [],
   "source": [
    "def UniversalEmbedding(x):\n",
    "    return embed(tf.squeeze(tf.cast(x, tf.string)), signature=\"default\", as_dict=True)[\"default\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 290
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 3322,
     "status": "ok",
     "timestamp": 1528601871955,
     "user": {
      "displayName": "zhang chengwei",
      "photoUrl": "//lh3.googleusercontent.com/-cNclsvC1ySc/AAAAAAAAAAI/AAAAAAAAAKI/-krROCWVXVQ/s50-c-k-no/photo.jpg",
      "userId": "107720063101388471292"
     },
     "user_tz": -480
    },
    "id": "t3fllZkVjXKV",
    "outputId": "8827cbe8-2a1e-4240-fff4-cdc661e8e327"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         (None, 1)                 0         \n",
      "_________________________________________________________________\n",
      "lambda_1 (Lambda)            (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 256)               131328    \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 6)                 1542      \n",
      "=================================================================\n",
      "Total params: 132,870\n",
      "Trainable params: 132,870\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "input_text = layers.Input(shape=(1,), dtype=tf.string)\n",
    "embedding = layers.Lambda(UniversalEmbedding, output_shape=(embed_size,))(input_text)\n",
    "dense = layers.Dense(256, activation='relu')(embedding)\n",
    "pred = layers.Dense(category_counts, activation='softmax')(dense)\n",
    "model = Model(inputs=[input_text], outputs=pred)\n",
    "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "5Ube1DvYEJ3q"
   },
   "outputs": [],
   "source": [
    "train_text = df_train['text'].tolist()\n",
    "train_text = np.array(train_text, dtype=object)[:, np.newaxis]\n",
    "\n",
    "train_label = np.asarray(pd.get_dummies(df_train.label), dtype = np.int8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 746,
     "status": "ok",
     "timestamp": 1528601877164,
     "user": {
      "displayName": "zhang chengwei",
      "photoUrl": "//lh3.googleusercontent.com/-cNclsvC1ySc/AAAAAAAAAAI/AAAAAAAAAKI/-krROCWVXVQ/s50-c-k-no/photo.jpg",
      "userId": "107720063101388471292"
     },
     "user_tz": -480
    },
    "id": "WX3s8yIVFWHI",
    "outputId": "5468a7ae-f1c3-4019-8f76-91dab02ec482"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(500, 1)"
      ]
     },
     "execution_count": 17,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_text.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 750,
     "status": "ok",
     "timestamp": 1528601878111,
     "user": {
      "displayName": "zhang chengwei",
      "photoUrl": "//lh3.googleusercontent.com/-cNclsvC1ySc/AAAAAAAAAAI/AAAAAAAAAKI/-krROCWVXVQ/s50-c-k-no/photo.jpg",
      "userId": "107720063101388471292"
     },
     "user_tz": -480
    },
    "id": "9PfsPdG8FZBI",
    "outputId": "b7d95197-8bad-4280-fc43-5a2e2373f1f2"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(500, 6)"
      ]
     },
     "execution_count": 18,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_label.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 72
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 730,
     "status": "ok",
     "timestamp": 1528601878979,
     "user": {
      "displayName": "zhang chengwei",
      "photoUrl": "//lh3.googleusercontent.com/-cNclsvC1ySc/AAAAAAAAAAI/AAAAAAAAAKI/-krROCWVXVQ/s50-c-k-no/photo.jpg",
      "userId": "107720063101388471292"
     },
     "user_tz": -480
    },
    "id": "gPYVmBr2Fbob",
    "outputId": "a6df6922-eae1-4018-9660-f7c6cfdf8d74"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0, 0, 0, 0, 0, 1],\n",
       "       [0, 0, 0, 0, 1, 0],\n",
       "       [0, 0, 0, 1, 0, 0]], dtype=int8)"
      ]
     },
     "execution_count": 19,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_label[:3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "X9MJnOnKI6F0"
   },
   "outputs": [],
   "source": [
    "df_test = get_dataframe('test_data.txt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "QWAjtjdeI9P4"
   },
   "outputs": [],
   "source": [
    "test_text = df_test['text'].tolist()\n",
    "test_text = np.array(test_text, dtype=object)[:, np.newaxis]\n",
    "test_label = np.asarray(pd.get_dummies(df_test.label), dtype = np.int8)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "bqcRy_JWXe0u"
   },
   "source": [
    "## Train Keras model and save weights\n",
    "This only train and save our Keras layers not the embed module' weights."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 419
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 27059,
     "status": "ok",
     "timestamp": 1528602874403,
     "user": {
      "displayName": "zhang chengwei",
      "photoUrl": "//lh3.googleusercontent.com/-cNclsvC1ySc/AAAAAAAAAAI/AAAAAAAAAKI/-krROCWVXVQ/s50-c-k-no/photo.jpg",
      "userId": "107720063101388471292"
     },
     "user_tz": -480
    },
    "id": "_stfC_7VFhS8",
    "outputId": "09262f88-ed9d-4f2a-bda1-7112d0a219b8"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 500 samples, validate on 500 samples\n",
      "Epoch 1/10\n",
      "500/500 [==============================] - 6s 12ms/step - loss: 1.6009 - acc: 0.5500 - val_loss: 1.3178 - val_acc: 0.7460\n",
      "Epoch 2/10\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 1.1025 - acc: 0.7860 - val_loss: 0.8527 - val_acc: 0.8400\n",
      "Epoch 3/10\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.7176 - acc: 0.8820 - val_loss: 0.5661 - val_acc: 0.9040\n",
      "Epoch 4/10\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.4912 - acc: 0.9020 - val_loss: 0.4098 - val_acc: 0.9180\n",
      "Epoch 5/10\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.3742 - acc: 0.9200 - val_loss: 0.3214 - val_acc: 0.9240\n",
      "Epoch 6/10\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2975 - acc: 0.9300 - val_loss: 0.2655 - val_acc: 0.9360\n",
      "Epoch 7/10\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2540 - acc: 0.9340 - val_loss: 0.2231 - val_acc: 0.9420\n",
      "Epoch 8/10\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.2132 - acc: 0.9420 - val_loss: 0.1923 - val_acc: 0.9460\n",
      "Epoch 9/10\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.1856 - acc: 0.9420 - val_loss: 0.1669 - val_acc: 0.9580\n",
      "Epoch 10/10\n",
      "500/500 [==============================] - 2s 4ms/step - loss: 0.1635 - acc: 0.9600 - val_loss: 0.1470 - val_acc: 0.9700\n"
     ]
    }
   ],
   "source": [
    "with tf.Session() as session:\n",
    "  K.set_session(session)\n",
    "  session.run(tf.global_variables_initializer())\n",
    "  session.run(tf.tables_initializer())\n",
    "  history = model.fit(train_text, \n",
    "            train_label,\n",
    "            validation_data=(test_text, test_label),\n",
    "            epochs=10,\n",
    "            batch_size=32)\n",
    "  model.save_weights('./model.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can skip this if running on Windows, it just check the size of the save weights file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1716,
     "status": "ok",
     "timestamp": 1528602876152,
     "user": {
      "displayName": "zhang chengwei",
      "photoUrl": "//lh3.googleusercontent.com/-cNclsvC1ySc/AAAAAAAAAAI/AAAAAAAAAKI/-krROCWVXVQ/s50-c-k-no/photo.jpg",
      "userId": "107720063101388471292"
     },
     "user_tz": -480
    },
    "id": "UW1CiBhnXnxa",
    "outputId": "1c9f90b9-3757-461e-d06e-076e524bd792"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-rw-r--r-- 1 root root 534K Jun 10 03:54 model.h5\r\n"
     ]
    }
   ],
   "source": [
    "!ls -alh | grep model.h5"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "nQux6qLdXabG"
   },
   "source": [
    "## Make predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "fSDxetlfUEiD"
   },
   "outputs": [],
   "source": [
    "new_text = [\"In what year did the titanic sink ?\", \"What is the highest peak in California ?\", \"Who invented the light bulb ?\"]\n",
    "new_text = np.array(new_text, dtype=object)[:, np.newaxis]\n",
    "with tf.Session() as session:\n",
    "  K.set_session(session)\n",
    "  session.run(tf.global_variables_initializer())\n",
    "  session.run(tf.tables_initializer())\n",
    "  model.load_weights('./model.h5')  \n",
    "  predicts = model.predict(new_text, batch_size=32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 126
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 737,
     "status": "ok",
     "timestamp": 1528603169307,
     "user": {
      "displayName": "zhang chengwei",
      "photoUrl": "//lh3.googleusercontent.com/-cNclsvC1ySc/AAAAAAAAAAI/AAAAAAAAAKI/-krROCWVXVQ/s50-c-k-no/photo.jpg",
      "userId": "107720063101388471292"
     },
     "user_tz": -480
    },
    "id": "xaR0d5VPU23Z",
    "outputId": "edeceb9e-d431-42d7-dd5d-dfe4381650a4"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[2.8481445e-04, 1.9496026e-03, 4.7652838e-03, 1.6724154e-03,\n",
       "        5.2437396e-03, 9.8608416e-01],\n",
       "       [1.7583232e-03, 2.1737781e-03, 5.2858055e-02, 5.4584690e-03,\n",
       "        8.8098717e-01, 5.6764200e-02],\n",
       "       [4.1049864e-04, 6.7947753e-04, 5.6051292e-02, 9.3884909e-01,\n",
       "        3.1365198e-03, 8.7315065e-04]], dtype=float32)"
      ]
     },
     "execution_count": 28,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predicts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 751,
     "status": "ok",
     "timestamp": 1528603170166,
     "user": {
      "displayName": "zhang chengwei",
      "photoUrl": "//lh3.googleusercontent.com/-cNclsvC1ySc/AAAAAAAAAAI/AAAAAAAAAKI/-krROCWVXVQ/s50-c-k-no/photo.jpg",
      "userId": "107720063101388471292"
     },
     "user_tz": -480
    },
    "id": "yyDGVtigW57f",
    "outputId": "6261f056-7f3a-4413-c051-3c1d92c4fbc6"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['NUM', 'LOC', 'HUM']"
      ]
     },
     "execution_count": 29,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "categories = df_train.label.cat.categories.tolist()\n",
    "predict_logits = predicts.argmax(axis=1)\n",
    "predict_labels = [categories[logit] for logit in predict_logits]\n",
    "predict_labels"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "BnvjATdy64eR"
   },
   "source": [
    "# Semantic Textual Similarity Task Example\n",
    "\n",
    "The embeddings produced by the Universal Sentence Encoder are approximately normalized. The semantic similarity of two sentences can be trivially computed as the inner product of the encodings."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "h1FFCTKm7ba4"
   },
   "outputs": [],
   "source": [
    "def plot_similarity(labels, features, rotation):\n",
    "  corr = np.inner(features, features)\n",
    "  sns.set(font_scale=1.2)\n",
    "  g = sns.heatmap(\n",
    "      corr,\n",
    "      xticklabels=labels,\n",
    "      yticklabels=labels,\n",
    "      vmin=0,\n",
    "      vmax=1,\n",
    "      cmap=\"YlOrRd\")\n",
    "  g.set_xticklabels(labels, rotation=rotation)\n",
    "  g.set_title(\"Semantic Textual Similarity\")\n",
    "\n",
    "\n",
    "def run_and_plot(session_, input_tensor_, messages_, encoding_tensor):\n",
    "  message_embeddings_ = session_.run(\n",
    "      encoding_tensor, feed_dict={input_tensor_: messages_})\n",
    "  plot_similarity(messages_, message_embeddings_, 90)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "339tuJ5Pwqqv"
   },
   "source": [
    "## Similarity Visualized\n",
    "Here we show the similarity in a heat map. The final graph is a 9x9 matrix where each entry `[i, j]` is colored based on the inner product of the encodings for sentence `i` and `j`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     },
     "base_uri": "https://localhost:8080/",
     "height": 585
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 9551,
     "status": "ok",
     "timestamp": 1528541396848,
     "user": {
      "displayName": "zhang chengwei",
      "photoUrl": "//lh3.googleusercontent.com/-cNclsvC1ySc/AAAAAAAAAAI/AAAAAAAAAKI/-krROCWVXVQ/s50-c-k-no/photo.jpg",
      "userId": "107720063101388471292"
     },
     "user_tz": -480
    },
    "id": "cPMCaxrZwp7t",
    "outputId": "feecd3da-ac01-4ade-c650-fbe18a615aff"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAp8AAAI4CAYAAAAs8YJeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f80cc0fa780>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "messages = [\n",
    "    # Smartphones\n",
    "    \"I like my phone\",\n",
    "    \"My phone is not good.\",\n",
    "    \"Your cellphone looks great.\",\n",
    "\n",
    "    # Weather\n",
    "    \"Will it snow tomorrow?\",\n",
    "    \"Recently a lot of hurricanes have hit the US\",\n",
    "    \"Global warming is real\",\n",
    "\n",
    "    # Food and health\n",
    "    \"An apple a day, keeps the doctors away\",\n",
    "    \"Eating strawberries is healthy\",\n",
    "    \"Is paleo better than keto?\",\n",
    "\n",
    "    # Asking about age\n",
    "    \"How old are you?\",\n",
    "    \"what is your age?\",\n",
    "]\n",
    "\n",
    "similarity_input_placeholder = tf.placeholder(tf.string, shape=(None))\n",
    "similarity_message_encodings = embed(similarity_input_placeholder)\n",
    "with tf.Session() as session:\n",
    "  session.run(tf.global_variables_initializer())\n",
    "  session.run(tf.tables_initializer())\n",
    "  run_and_plot(session, similarity_input_placeholder, messages,\n",
    "               similarity_message_encodings)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "hYhmukbSKpnp"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "default_view": {},
   "name": "Transfer Learning - Semantic Similarity with TF-Hub Universal Encoder",
   "provenance": [
    {
     "file_id": "1UU9zlnuO2lYsTRnjwaX2QPpDXMzmp395",
     "timestamp": 1528633735502
    },
    {
     "file_id": "https://github.com/tensorflow/hub/blob/master/examples/colab/semantic_similarity_with_tf_hub_universal_encoder.ipynb",
     "timestamp": 1528530105070
    }
   ],
   "version": "0.3.2",
   "views": {}
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}