Text_Classification with Universal_Sentence_Encoder.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"Text_Classification with Universal_Sentence_Encoder.ipynb","version":"0.3.2","provenance":[],"collapsed_sections":[]},"kernelspec":{"name":"python3","display_name":"Python 3"},"accelerator":"GPU"},"cells":[{"metadata":{"id":"wnRD3ysWN7JD","colab_type":"code","colab":{}},"cell_type":"code","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 --quiet seaborn"],"execution_count":0,"outputs":[]},{"metadata":{"id":"lNCU09lsRyN3","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":101},"outputId":"cd3b6621-b0b3-4f3a-a1e7-c62a522b3c03","executionInfo":{"status":"ok","timestamp":1547308158887,"user_tz":300,"elapsed":10763,"user":{"displayName":"CARLOS QUIROS","photoUrl":"https://lh4.googleusercontent.com/-cmLzcUyA-SI/AAAAAAAAAAI/AAAAAAAAAHI/3CEu2CkjTCM/s64/photo.jpg","userId":"07669736790263855207"}}},"cell_type":"code","source":["!pip install contractions\n","!pip install beautifulsoup4"],"execution_count":3,"outputs":[{"output_type":"stream","text":["Collecting contractions\n","  Downloading https://files.pythonhosted.org/packages/52/52/a15f0fb338a462045c7c87a35dbaeda11738c45aa9d2f5c76ac191d6adff/contractions-0.0.17-py2.py3-none-any.whl\n","Installing collected packages: contractions\n","Successfully installed contractions-0.0.17\n","Requirement already satisfied: beautifulsoup4 in /usr/local/lib/python3.6/dist-packages (4.6.3)\n"],"name":"stdout"}]},{"metadata":{"id":"DGTS0s61OBTI","colab_type":"code","colab":{}},"cell_type":"code","source":["import tensorflow as tf\n","import tensorflow_hub as hub\n","import numpy as np\n","import pandas as pd\n","import time\n","import contractions #pip install contractions\n","from bs4 import BeautifulSoup #pip install beautifulsoup4\n","import unicodedata\n","import re"],"execution_count":0,"outputs":[]},{"metadata":{"id":"9HAHmdk8QOJ-","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":121},"outputId":"296fbeaa-b370-4b2d-c938-21681c2f5f60","executionInfo":{"status":"ok","timestamp":1547308196244,"user_tz":300,"elapsed":28986,"user":{"displayName":"CARLOS QUIROS","photoUrl":"https://lh4.googleusercontent.com/-cmLzcUyA-SI/AAAAAAAAAAI/AAAAAAAAAHI/3CEu2CkjTCM/s64/photo.jpg","userId":"07669736790263855207"}}},"cell_type":"code","source":["from google.colab import drive\n","drive.mount('/content/drive/')"],"execution_count":5,"outputs":[{"output_type":"stream","text":["Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3Aietf%3Awg%3Aoauth%3A2.0%3Aoob&scope=email%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdocs.test%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive.photos.readonly%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fpeopleapi.readonly&response_type=code\n","\n","Enter your authorization code:\n","··········\n","Mounted at /content/drive/\n"],"name":"stdout"}]},{"metadata":{"id":"5ZkxTXudOHwB","colab_type":"code","outputId":"d2f29943-42dc-4ef1-cfc6-9753b107c4fa","executionInfo":{"status":"ok","timestamp":1547308215746,"user_tz":300,"elapsed":6212,"user":{"displayName":"CARLOS QUIROS","photoUrl":"https://lh4.googleusercontent.com/-cmLzcUyA-SI/AAAAAAAAAAI/AAAAAAAAAHI/3CEu2CkjTCM/s64/photo.jpg","userId":"07669736790263855207"}},"colab":{"base_uri":"https://localhost:8080/","height":134}},"cell_type":"code","source":["df = pd.read_csv('/content/drive/My Drive/Colab Notebooks/movie_reviews.csv.bz2', compression='bz2')\n","df.info()"],"execution_count":6,"outputs":[{"output_type":"stream","text":["<class 'pandas.core.frame.DataFrame'>\n","RangeIndex: 50000 entries, 0 to 49999\n","Data columns (total 2 columns):\n","review       50000 non-null object\n","sentiment    50000 non-null object\n","dtypes: object(2)\n","memory usage: 781.3+ KB\n"],"name":"stdout"}]},{"metadata":{"id":"bswQ9qvBsyNP","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":195},"outputId":"9fee7a67-f319-479b-a47c-e2641ada41ad","executionInfo":{"status":"ok","timestamp":1547308220101,"user_tz":300,"elapsed":1305,"user":{"displayName":"CARLOS QUIROS","photoUrl":"https://lh4.googleusercontent.com/-cmLzcUyA-SI/AAAAAAAAAAI/AAAAAAAAAHI/3CEu2CkjTCM/s64/photo.jpg","userId":"07669736790263855207"}}},"cell_type":"code","source":["df.head()"],"execution_count":7,"outputs":[{"output_type":"execute_result","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>review</th>\n","      <th>sentiment</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>One of the other reviewers has mentioned that ...</td>\n","      <td>positive</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>A wonderful little production. &lt;br /&gt;&lt;br /&gt;The...</td>\n","      <td>positive</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>I thought this was a wonderful way to spend ti...</td>\n","      <td>positive</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>Basically there's a family where a little boy ...</td>\n","      <td>negative</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>Petter Mattei's \"Love in the Time of Money\" is...</td>\n","      <td>positive</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                                              review sentiment\n","0  One of the other reviewers has mentioned that ...  positive\n","1  A wonderful little production. <br /><br />The...  positive\n","2  I thought this was a wonderful way to spend ti...  positive\n","3  Basically there's a family where a little boy ...  negative\n","4  Petter Mattei's \"Love in the Time of Money\" is...  positive"]},"metadata":{"tags":[]},"execution_count":7}]},{"metadata":{"id":"yro-du_4O6-E","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":195},"outputId":"4da8f731-4da9-4d82-eb92-33766be42625","executionInfo":{"status":"ok","timestamp":1547308236230,"user_tz":300,"elapsed":1315,"user":{"displayName":"CARLOS QUIROS","photoUrl":"https://lh4.googleusercontent.com/-cmLzcUyA-SI/AAAAAAAAAAI/AAAAAAAAAHI/3CEu2CkjTCM/s64/photo.jpg","userId":"07669736790263855207"}}},"cell_type":"code","source":["df['sentiment'] = [1 if sentiment == 'positive' else 0 for sentiment in df['sentiment'].values]\n","df.head()"],"execution_count":8,"outputs":[{"output_type":"execute_result","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>review</th>\n","      <th>sentiment</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>One of the other reviewers has mentioned that ...</td>\n","      <td>1</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>A wonderful little production. &lt;br /&gt;&lt;br /&gt;The...</td>\n","      <td>1</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>I thought this was a wonderful way to spend ti...</td>\n","      <td>1</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>Basically there's a family where a little boy ...</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>Petter Mattei's \"Love in the Time of Money\" is...</td>\n","      <td>1</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                                              review  sentiment\n","0  One of the other reviewers has mentioned that ...          1\n","1  A wonderful little production. <br /><br />The...          1\n","2  I thought this was a wonderful way to spend ti...          1\n","3  Basically there's a family where a little boy ...          0\n","4  Petter Mattei's \"Love in the Time of Money\" is...          1"]},"metadata":{"tags":[]},"execution_count":8}]},{"metadata":{"id":"83VWErpywmTj","colab_type":"text"},"cell_type":"markdown","source":["## Creating Datasets for Training, Testing and Validation:"]},{"metadata":{"id":"f0CflT8eRvh1","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":34},"outputId":"f2f8f12e-0cc0-4dd8-d68a-97130474b456","executionInfo":{"status":"ok","timestamp":1547309566821,"user_tz":300,"elapsed":1323,"user":{"displayName":"CARLOS QUIROS","photoUrl":"https://lh4.googleusercontent.com/-cmLzcUyA-SI/AAAAAAAAAAI/AAAAAAAAAHI/3CEu2CkjTCM/s64/photo.jpg","userId":"07669736790263855207"}}},"cell_type":"code","source":["reviews = df['review'].values\n","labels = df['sentiment'].values\n","\n","train_reviews = reviews[:30000]\n","train_labels = labels[:30000]\n","\n","test_reviews = reviews[30000:40000]\n","test_labels = labels[30000:40000]\n","\n","val_reviews = reviews[45000:]\n","val_labels = labels[45000:]\n","\n","train_reviews.shape, test_reviews.shape, val_reviews.shape,"],"execution_count":23,"outputs":[{"output_type":"execute_result","data":{"text/plain":["((30000,), (10000,), (5000,))"]},"metadata":{"tags":[]},"execution_count":23}]},{"metadata":{"id":"Qb6a2mGWR6rS","colab_type":"code","colab":{}},"cell_type":"code","source":["def strip_html_tags(text):\n","    soup = BeautifulSoup(text, \"html.parser\")\n","    [s.extract() for s in soup(['iframe', 'script'])]\n","    text_ = soup.get_text()\n","    text_ = re.sub(r'[\\r|\\n|\\r\\n]+', '\\n', text_)\n","    return text_\n"],"execution_count":0,"outputs":[]},{"metadata":{"id":"3WX8tHGjR9gn","colab_type":"code","colab":{}},"cell_type":"code","source":["def remove_accented_chars(text):\n","    text = unicodedata.normalize('NFKD', text).encode('ascii', 'ignore').decode('utf-8', 'ignore')\n","    return text"],"execution_count":0,"outputs":[]},{"metadata":{"id":"B9Jp0bFHR_yD","colab_type":"code","colab":{}},"cell_type":"code","source":["def expand_contractions(text):\n","    return contractions.fix(text)"],"execution_count":0,"outputs":[]},{"metadata":{"id":"ACqDkw5MSDHt","colab_type":"code","colab":{}},"cell_type":"code","source":["def remove_special_characters(text, remove_digits=False):\n","    pattern = r'[^a-zA-Z0-9\\s]' if not remove_digits else r'[^a-zA-Z\\s]'\n","    text = re.sub(pattern, '', text)\n","    return text"],"execution_count":0,"outputs":[]},{"metadata":{"id":"_3vNJWzeSBu9","colab_type":"code","colab":{}},"cell_type":"code","source":["def clean_doc(document):\n","    \n","    # strip HTML\n","    doc = strip_html_tags(document)\n","    \n","    # lower case\n","    doc = doc.lower()\n","    \n","    # remove extra newlines (often might be present in really noisy text)\n","    doc = doc.translate(document.maketrans(\"\\n\\t\\r\", \"   \"))\n","    \n","    # remove accented characters\n","    doc = remove_accented_chars(doc)\n","    \n","    # expand contractions    \n","    doc = expand_contractions(doc)\n","               \n","    # remove special characters and\\or digits    \n","    # insert spaces between special characters to isolate them    \n","    special_char_pattern = re.compile(r'([{.(-)!}])')\n","    doc = special_char_pattern.sub(\" \\\\1 \", doc)\n","    doc = remove_special_characters(doc, remove_digits=True)  \n","        \n","    # remove extra whitespace\n","    doc = re.sub(' +', ' ', doc)\n","    doc = doc.strip()\n","    \n","    return doc\n","\n","\n","clean_corpus = np.vectorize(clean_doc)"],"execution_count":0,"outputs":[]},{"metadata":{"id":"BvMUM3VD0BWJ","colab_type":"text"},"cell_type":"markdown","source":["## Preprocessing the Text:"]},{"metadata":{"id":"8RIoUcumSI_b","colab_type":"code","colab":{}},"cell_type":"code","source":["train_reviews = clean_corpus(train_reviews)\n","val_reviews = clean_corpus(val_reviews)\n","test_reviews = clean_corpus(test_reviews)\n"],"execution_count":0,"outputs":[]},{"metadata":{"id":"EOK-v8OF0RPX","colab_type":"text"},"cell_type":"markdown","source":["## Preparing the Training:"]},{"metadata":{"id":"nSBoVJ-NSLoI","colab_type":"code","colab":{}},"cell_type":"code","source":["# Training input on the whole training set with no limit on training epochs.\n","train_input = tf.estimator.inputs.numpy_input_fn({'sentence': train_reviews}, train_labels, batch_size=256, num_epochs=None, shuffle=True)"],"execution_count":0,"outputs":[]},{"metadata":{"id":"gsb2xrCZSO0e","colab_type":"code","colab":{}},"cell_type":"code","source":["# Prediction on the whole training set.\n","predict_train = tf.estimator.inputs.numpy_input_fn({'sentence': train_reviews}, train_labels, shuffle=False)"],"execution_count":0,"outputs":[]},{"metadata":{"id":"lH7eEhxISQ6N","colab_type":"code","colab":{}},"cell_type":"code","source":["# Prediction on the whole validation set.\n","predict_val = tf.estimator.inputs.numpy_input_fn({'sentence': val_reviews}, val_labels, shuffle=False)"],"execution_count":0,"outputs":[]},{"metadata":{"id":"YyCKSST8SS13","colab_type":"code","colab":{}},"cell_type":"code","source":["# Prediction on the test set.\n","predict_test = tf.estimator.inputs.numpy_input_fn({'sentence': test_reviews}, test_labels, shuffle=False)"],"execution_count":0,"outputs":[]},{"metadata":{"id":"hCeW2Jzn0hbk","colab_type":"text"},"cell_type":"markdown","source":["## Loading Embeddings from Universal Sentence Encoder:"]},{"metadata":{"id":"OGB1o_CuSVel","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":101},"outputId":"6682ffb7-1135-4bb0-8d70-c41fe7805e2c","executionInfo":{"status":"ok","timestamp":1547309624162,"user_tz":300,"elapsed":20922,"user":{"displayName":"CARLOS QUIROS","photoUrl":"https://lh4.googleusercontent.com/-cmLzcUyA-SI/AAAAAAAAAAI/AAAAAAAAAHI/3CEu2CkjTCM/s64/photo.jpg","userId":"07669736790263855207"}}},"cell_type":"code","source":["embeddings = hub.text_embedding_column(key='sentence', module_spec=\"https://tfhub.dev/google/universal-sentence-encoder/2\", trainable=False)"],"execution_count":27,"outputs":[{"output_type":"stream","text":["INFO:tensorflow:Using /tmp/tfhub_modules to cache modules.\n","INFO:tensorflow:Downloading TF-Hub Module 'https://tfhub.dev/google/universal-sentence-encoder/2'.\n","INFO:tensorflow:Downloading https://tfhub.dev/google/universal-sentence-encoder/2: 788.00MB\n","INFO:tensorflow:Downloaded https://tfhub.dev/google/universal-sentence-encoder/2, Total size: 993.27MB\n","INFO:tensorflow:Downloaded TF-Hub Module 'https://tfhub.dev/google/universal-sentence-encoder/2'.\n"],"name":"stdout"}]},{"metadata":{"id":"UmInLzON0tlc","colab_type":"text"},"cell_type":"markdown","source":["## Creating the Classifier:"]},{"metadata":{"id":"Q2uqNIJTSX-e","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":188},"outputId":"d324f7ac-aca8-45f2-9c75-386fa6748784","executionInfo":{"status":"ok","timestamp":1547309706706,"user_tz":300,"elapsed":1317,"user":{"displayName":"CARLOS QUIROS","photoUrl":"https://lh4.googleusercontent.com/-cmLzcUyA-SI/AAAAAAAAAAI/AAAAAAAAAHI/3CEu2CkjTCM/s64/photo.jpg","userId":"07669736790263855207"}}},"cell_type":"code","source":["cls = tf.estimator.DNNClassifier(\n","          hidden_units=[512, 128],\n","          feature_columns=[embeddings],\n","          n_classes=2,\n","          activation_fn=tf.nn.relu,\n","          dropout=0.2,\n","          optimizer=tf.train.AdagradOptimizer(learning_rate=0.005))"],"execution_count":28,"outputs":[{"output_type":"stream","text":["INFO:tensorflow:Using default config.\n","WARNING:tensorflow:Using temporary folder as model directory: /tmp/tmp0a7vg04h\n","INFO:tensorflow:Using config: {'_model_dir': '/tmp/tmp0a7vg04h', '_tf_random_seed': None, '_save_summary_steps': 100, '_save_checkpoints_steps': None, '_save_checkpoints_secs': 600, '_session_config': allow_soft_placement: true\n","graph_options {\n","  rewrite_options {\n","    meta_optimizer_iterations: ONE\n","  }\n","}\n",", '_keep_checkpoint_max': 5, '_keep_checkpoint_every_n_hours': 10000, '_log_step_count_steps': 100, '_train_distribute': None, '_device_fn': None, '_protocol': None, '_eval_distribute': None, '_experimental_distribute': None, '_service': None, '_cluster_spec': <tensorflow.python.training.server_lib.ClusterSpec object at 0x7fb48877d240>, '_task_type': 'worker', '_task_id': 0, '_global_id_in_cluster': 0, '_master': '', '_evaluation_master': '', '_is_chief': True, '_num_ps_replicas': 0, '_num_worker_replicas': 1}\n"],"name":"stdout"}]},{"metadata":{"id":"NN1c6BK-02gZ","colab_type":"text"},"cell_type":"markdown","source":["## Training the model:"]},{"metadata":{"id":"aOYVC716Sa7f","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":1146},"outputId":"a92c2683-7f39-4e49-e67e-fafef6227acf","executionInfo":{"status":"ok","timestamp":1547311559899,"user_tz":300,"elapsed":1827410,"user":{"displayName":"CARLOS QUIROS","photoUrl":"https://lh4.googleusercontent.com/-cmLzcUyA-SI/AAAAAAAAAAI/AAAAAAAAAHI/3CEu2CkjTCM/s64/photo.jpg","userId":"07669736790263855207"}}},"cell_type":"code","source":["%%time\n","tf.logging.set_verbosity(tf.logging.ERROR)\n","\n","epochs = 1000\n","step = 100\n","for s in range(0, epochs+1, step):\n","    print()\n","    print('-'*150)\n","    print('Epoch =', s)\n","    ini_time = time.time()\n","    cls.train(input_fn=train_input, steps=step)\n","    total_time = time.time() - ini_time\n","    print('Train Time (s):', total_time)\n","    print('Metrics for Train):', cls.evaluate(input_fn=predict_train))\n","    print('Metrics for Validation):', cls.evaluate(input_fn=predict_val))"],"execution_count":30,"outputs":[{"output_type":"stream","text":["\n","----------------------------------------------------------------------------------------------------\n","Training for step = 0\n","Train Time (s): 85.70678925514221\n","Metrics (Train): {'accuracy': 0.83463335, 'accuracy_baseline': 0.5005, 'auc': 0.9252298, 'auc_precision_recall': 0.925236, 'average_loss': 0.36300963, 'label/mean': 0.5005, 'loss': 46.341656, 'precision': 0.8832139, 'prediction/mean': 0.4462625, 'recall': 0.7716284, 'global_step': 100}\n","Metrics (Validation): {'accuracy': 0.8336, 'accuracy_baseline': 0.50600004, 'auc': 0.9234706, 'auc_precision_recall': 0.9208938, 'average_loss': 0.36624637, 'label/mean': 0.494, 'loss': 45.780796, 'precision': 0.88164026, 'prediction/mean': 0.443118, 'recall': 0.7659919, 'global_step': 100}\n","\n","----------------------------------------------------------------------------------------------------\n","Training for step = 100\n","Train Time (s): 82.87584829330444\n","Metrics (Train): {'accuracy': 0.84506667, 'accuracy_baseline': 0.5005, 'auc': 0.9297993, 'auc_precision_recall': 0.9299148, 'average_loss': 0.34645143, 'label/mean': 0.5005, 'loss': 44.227844, 'precision': 0.8784958, 'prediction/mean': 0.4600715, 'recall': 0.8012654, 'global_step': 200}\n","Metrics (Validation): {'accuracy': 0.8422, 'accuracy_baseline': 0.50600004, 'auc': 0.9274026, 'auc_precision_recall': 0.924976, 'average_loss': 0.3517918, 'label/mean': 0.494, 'loss': 43.973976, 'precision': 0.87843317, 'prediction/mean': 0.45579413, 'recall': 0.78987855, 'global_step': 200}\n","\n","----------------------------------------------------------------------------------------------------\n","Training for step = 200\n","Train Time (s): 82.40753698348999\n","Metrics (Train): {'accuracy': 0.8526667, 'accuracy_baseline': 0.5005, 'auc': 0.9315836, 'auc_precision_recall': 0.9316437, 'average_loss': 0.334904, 'label/mean': 0.5005, 'loss': 42.7537, 'precision': 0.863116, 'prediction/mean': 0.488971, 'recall': 0.83862805, 'global_step': 300}\n","Metrics (Validation): {'accuracy': 0.8508, 'accuracy_baseline': 0.50600004, 'auc': 0.92865837, 'auc_precision_recall': 0.9257389, 'average_loss': 0.3423662, 'label/mean': 0.494, 'loss': 42.795773, 'precision': 0.86157715, 'prediction/mean': 0.48444033, 'recall': 0.83157897, 'global_step': 300}\n","\n","----------------------------------------------------------------------------------------------------\n","Training for step = 300\n","Train Time (s): 83.9992687702179\n","Metrics (Train): {'accuracy': 0.85573334, 'accuracy_baseline': 0.5005, 'auc': 0.9334961, 'auc_precision_recall': 0.93384016, 'average_loss': 0.32997045, 'label/mean': 0.5005, 'loss': 42.123886, 'precision': 0.8516154, 'prediction/mean': 0.50632036, 'recall': 0.86193806, 'global_step': 400}\n","Metrics (Validation): {'accuracy': 0.8508, 'accuracy_baseline': 0.50600004, 'auc': 0.9296134, 'auc_precision_recall': 0.92689806, 'average_loss': 0.33949062, 'label/mean': 0.494, 'loss': 42.43633, 'precision': 0.8489879, 'prediction/mean': 0.5021609, 'recall': 0.8489879, 'global_step': 400}\n","\n","----------------------------------------------------------------------------------------------------\n","Training for step = 400\n","Train Time (s): 83.81255602836609\n","Metrics (Train): {'accuracy': 0.8579, 'accuracy_baseline': 0.5005, 'auc': 0.93499684, 'auc_precision_recall': 0.9350821, 'average_loss': 0.32639915, 'label/mean': 0.5005, 'loss': 41.667976, 'precision': 0.85077643, 'prediction/mean': 0.51076764, 'recall': 0.86839825, 'global_step': 500}\n","Metrics (Validation): {'accuracy': 0.852, 'accuracy_baseline': 0.50600004, 'auc': 0.9310585, 'auc_precision_recall': 0.9283587, 'average_loss': 0.33628622, 'label/mean': 0.494, 'loss': 42.035774, 'precision': 0.8473896, 'prediction/mean': 0.5063163, 'recall': 0.854251, 'global_step': 500}\n","\n","----------------------------------------------------------------------------------------------------\n","Training for step = 500\n","Train Time (s): 82.45590043067932\n","Metrics (Train): {'accuracy': 0.85836667, 'accuracy_baseline': 0.5005, 'auc': 0.93629384, 'auc_precision_recall': 0.9364598, 'average_loss': 0.32741708, 'label/mean': 0.5005, 'loss': 41.797928, 'precision': 0.8371117, 'prediction/mean': 0.53107417, 'recall': 0.8902431, 'global_step': 600}\n","Metrics (Validation): {'accuracy': 0.8526, 'accuracy_baseline': 0.50600004, 'auc': 0.9316056, 'auc_precision_recall': 0.92912173, 'average_loss': 0.33925155, 'label/mean': 0.494, 'loss': 42.406445, 'precision': 0.83161116, 'prediction/mean': 0.52704763, 'recall': 0.8797571, 'global_step': 600}\n","\n","----------------------------------------------------------------------------------------------------\n","Training for step = 600\n","Train Time (s): 83.15762448310852\n","Metrics (Train): {'accuracy': 0.86076665, 'accuracy_baseline': 0.5005, 'auc': 0.93776983, 'auc_precision_recall': 0.9378726, 'average_loss': 0.3192718, 'label/mean': 0.5005, 'loss': 40.758102, 'precision': 0.85539085, 'prediction/mean': 0.50831693, 'recall': 0.8686647, 'global_step': 700}\n","Metrics (Validation): {'accuracy': 0.8568, 'accuracy_baseline': 0.50600004, 'auc': 0.9328485, 'auc_precision_recall': 0.93022805, 'average_loss': 0.33176732, 'label/mean': 0.494, 'loss': 41.470917, 'precision': 0.85506076, 'prediction/mean': 0.50313425, 'recall': 0.85506076, 'global_step': 700}\n","\n","----------------------------------------------------------------------------------------------------\n","Training for step = 700\n","Train Time (s): 83.44287371635437\n","Metrics (Train): {'accuracy': 0.86226666, 'accuracy_baseline': 0.5005, 'auc': 0.9389906, 'auc_precision_recall': 0.938891, 'average_loss': 0.3159721, 'label/mean': 0.5005, 'loss': 40.336864, 'precision': 0.86571676, 'prediction/mean': 0.49626547, 'recall': 0.85787547, 'global_step': 800}\n","Metrics (Validation): {'accuracy': 0.8546, 'accuracy_baseline': 0.50600004, 'auc': 0.9336863, 'auc_precision_recall': 0.9311586, 'average_loss': 0.32983088, 'label/mean': 0.494, 'loss': 41.22886, 'precision': 0.86206895, 'prediction/mean': 0.4909053, 'recall': 0.840081, 'global_step': 800}\n","\n","----------------------------------------------------------------------------------------------------\n","Training for step = 800\n","Train Time (s): 83.20907711982727\n","Metrics (Train): {'accuracy': 0.8637667, 'accuracy_baseline': 0.5005, 'auc': 0.9402503, 'auc_precision_recall': 0.94048727, 'average_loss': 0.31480968, 'label/mean': 0.5005, 'loss': 40.18847, 'precision': 0.8494947, 'prediction/mean': 0.51886964, 'recall': 0.88451546, 'global_step': 900}\n","Metrics (Validation): {'accuracy': 0.8558, 'accuracy_baseline': 0.50600004, 'auc': 0.93341416, 'auc_precision_recall': 0.9311147, 'average_loss': 0.33141056, 'label/mean': 0.494, 'loss': 41.42632, 'precision': 0.84280676, 'prediction/mean': 0.51447934, 'recall': 0.8704454, 'global_step': 900}\n","\n","----------------------------------------------------------------------------------------------------\n","Training for step = 900\n","Train Time (s): 84.39755725860596\n","Metrics (Train): {'accuracy': 0.86653334, 'accuracy_baseline': 0.5005, 'auc': 0.94160867, 'auc_precision_recall': 0.9415913, 'average_loss': 0.31024686, 'label/mean': 0.5005, 'loss': 39.60598, 'precision': 0.8543021, 'prediction/mean': 0.51553917, 'recall': 0.8841159, 'global_step': 1000}\n","Metrics (Validation): {'accuracy': 0.8566, 'accuracy_baseline': 0.50600004, 'auc': 0.9346722, 'auc_precision_recall': 0.93249446, 'average_loss': 0.32807717, 'label/mean': 0.494, 'loss': 41.009647, 'precision': 0.8463058, 'prediction/mean': 0.5110898, 'recall': 0.86720645, 'global_step': 1000}\n","\n","----------------------------------------------------------------------------------------------------\n","Training for step = 1000\n","Train Time (s): 85.02146744728088\n","Metrics (Train): {'accuracy': 0.86553335, 'accuracy_baseline': 0.5005, 'auc': 0.94297737, 'auc_precision_recall': 0.9429562, 'average_loss': 0.3091043, 'label/mean': 0.5005, 'loss': 39.46012, 'precision': 0.8847852, 'prediction/mean': 0.4752403, 'recall': 0.84082586, 'global_step': 1100}\n","Metrics (Validation): {'accuracy': 0.8574, 'accuracy_baseline': 0.50600004, 'auc': 0.93529004, 'auc_precision_recall': 0.9328685, 'average_loss': 0.32854372, 'label/mean': 0.494, 'loss': 41.067966, 'precision': 0.88079756, 'prediction/mean': 0.4697365, 'recall': 0.82267207, 'global_step': 1100}\n"],"name":"stdout"}]},{"metadata":{"id":"0YcREwpf08Wt","colab_type":"text"},"cell_type":"markdown","source":["## Testing the model:"]},{"metadata":{"id":"vLS07YzATlGG","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":202},"outputId":"d28e955f-6d5e-4b9c-c99a-1ee7c21dfc2b","executionInfo":{"status":"ok","timestamp":1547316088559,"user_tz":300,"elapsed":30097,"user":{"displayName":"CARLOS QUIROS","photoUrl":"https://lh4.googleusercontent.com/-cmLzcUyA-SI/AAAAAAAAAAI/AAAAAAAAAHI/3CEu2CkjTCM/s64/photo.jpg","userId":"07669736790263855207"}}},"cell_type":"code","source":["cls.evaluate(input_fn=predict_test)"],"execution_count":32,"outputs":[{"output_type":"execute_result","data":{"text/plain":["{'accuracy': 0.8572,\n"," 'accuracy_baseline': 0.5022,\n"," 'auc': 0.93549216,\n"," 'auc_precision_recall': 0.93376917,\n"," 'average_loss': 0.32665488,\n"," 'global_step': 1100,\n"," 'label/mean': 0.4978,\n"," 'loss': 41.34872,\n"," 'precision': 0.8707185,\n"," 'prediction/mean': 0.47735542,\n"," 'recall': 0.83748496}"]},"metadata":{"tags":[]},"execution_count":32}]},{"metadata":{"id":"s4ObPcC75_Mb","colab_type":"code","colab":{}},"cell_type":"code","source":[""],"execution_count":0,"outputs":[]}]}