deep_learning/time_series_sequences/TF2Keras_TextClassification_RNN.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"kernelspec":{"name":"python3","display_name":"Python 3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"},"colab":{"name":"TF2Keras_TextClassification_RNN.ipynb","provenance":[],"collapsed_sections":[]},"accelerator":"GPU"},"cells":[{"cell_type":"markdown","metadata":{"id":"SQH5KX-x4IDG","colab_type":"text"},"source":["In this tutorial, you are going to train a text classifier on the IMDB dataset using RNNs."]},{"cell_type":"code","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true,"_kg_hide-output":true,"id":"8GG69T7I4IDK","colab_type":"code","colab":{}},"source":["!pip install -q tf-nightly tfds-nightly\n","#!pip install -q tensorflow_datasets"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","trusted":true,"id":"FI37sneV4IDS","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":51},"outputId":"cb143e24-0bab-46e5-c888-49c7018d2f4b","executionInfo":{"status":"ok","timestamp":1581586260864,"user_tz":-480,"elapsed":5185,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["import tensorflow as tf\n","import tensorflow_datasets as tfds\n","from IPython.display import clear_output\n","import matplotlib.pyplot as plt\n","\n","tfds.disable_progress_bar()\n","\n","print(\"Tensorflow Version: {}\".format(tf.__version__))\n","print(\"GPU {} available.\".format(\"is\" if tf.config.experimental.list_physical_devices(\"GPU\") else \"not\"))"],"execution_count":1,"outputs":[{"output_type":"stream","text":["Tensorflow Version: 2.2.0-dev20200212\n","GPU is available.\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"Uu3-rTtz4IDY","colab_type":"text"},"source":["The function `plot_history` helps you plot the training status."]},{"cell_type":"code","metadata":{"trusted":true,"id":"QYW7SSRv4IDZ","colab_type":"code","colab":{}},"source":["def plot_history(history, metrics):\n","  plt.figure(figsize=(9, 6))\n","  plt.plot(history.history[metrics], label=\"Training\")\n","  plt.plot(history.history[\"val_\" + metrics], label=\"Validation\")\n","  plt.xlabel('Epochs')\n","  plt.ylabel(metrics)\n","  plt.legend()\n","  plt.show()"],"execution_count":0,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"SS4XtP074IDc","colab_type":"text"},"source":["# Input Pipeline"]},{"cell_type":"code","metadata":{"trusted":true,"id":"RYXlm26d4IDf","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":102},"outputId":"ff301ec9-7771-489e-f2f6-03f9893dc43b","executionInfo":{"status":"ok","timestamp":1581586379408,"user_tz":-480,"elapsed":123668,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["dataset, info = tfds.load(name=\"imdb_reviews/subwords8k\", with_info=True, as_supervised=True)\n","train_dataset, test_dataset = dataset['train'], dataset['test']"],"execution_count":3,"outputs":[{"output_type":"stream","text":["\u001b[1mDownloading and preparing dataset imdb_reviews (80.23 MiB) to /root/tensorflow_datasets/imdb_reviews/subwords8k/1.0.0...\u001b[0m\n","Shuffling and writing examples to /root/tensorflow_datasets/imdb_reviews/subwords8k/1.0.0.incomplete7KEKLY/imdb_reviews-train.tfrecord\n","Shuffling and writing examples to /root/tensorflow_datasets/imdb_reviews/subwords8k/1.0.0.incomplete7KEKLY/imdb_reviews-test.tfrecord\n","Shuffling and writing examples to /root/tensorflow_datasets/imdb_reviews/subwords8k/1.0.0.incomplete7KEKLY/imdb_reviews-unsupervised.tfrecord\n","\u001b[1mDataset imdb_reviews downloaded and prepared to /root/tensorflow_datasets/imdb_reviews/subwords8k/1.0.0. Subsequent calls will reuse this data.\u001b[0m\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"zbBgR4Ok4IDi","colab_type":"text"},"source":["The info of the dataset contains the encoder for words."]},{"cell_type":"code","metadata":{"trusted":true,"id":"VTPzqqdE4IDk","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":221},"outputId":"29003afc-b1eb-40d5-e32c-2e0e3541d9b4","executionInfo":{"status":"ok","timestamp":1581586379411,"user_tz":-480,"elapsed":123610,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["encoder = info.features['text'].encoder\n","print('Vocabulary Size: {}'.format(encoder.vocab_size))\n","print('The first few words: \\n{}'.format('\\n'.join(encoder.subwords[:10])))"],"execution_count":4,"outputs":[{"output_type":"stream","text":["Vocabulary Size: 8185\n","The first few words: \n","the_\n",", \n",". \n","a_\n","and_\n","of_\n","to_\n","s_\n","is_\n","br\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"2_O2AK2H4IDo","colab_type":"text"},"source":["The encoder can encode a word into an integer list. This list can also be decoded back to the same word."]},{"cell_type":"code","metadata":{"trusted":true,"id":"sIvSnvnK4IDp","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":51},"outputId":"c45cb9e2-5828-47b4-c56f-46aa1c82dbb1","executionInfo":{"status":"ok","timestamp":1581586379412,"user_tz":-480,"elapsed":123602,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["example_string = 'Hello world, RNN in Tensorflow.'\n","\n","encoded_string = encoder.encode(example_string)\n","print(\"Encoded integer list: {}\".format(encoded_string))\n","\n","decoded_string = encoder.decode(encoded_string)\n","print('Decoded from an integer list: {}'.format(decoded_string))\n","\n","assert decoded_string == example_string"],"execution_count":5,"outputs":[{"output_type":"stream","text":["Encoded integer list: [4025, 222, 562, 2, 8011, 8007, 1978, 11, 6307, 2327, 2934, 7975]\n","Decoded from an integer list: Hello world, RNN in Tensorflow.\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"trusted":true,"id":"lkWiaIpJ4IDt","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":221},"outputId":"399f4f69-d1ff-4ac1-c529-84f2218476ed","executionInfo":{"status":"ok","timestamp":1581586379415,"user_tz":-480,"elapsed":123586,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["for index in encoded_string:\n","  print(\"{} <-> {}\".format(index, encoder.decode([index])))"],"execution_count":6,"outputs":[{"output_type":"stream","text":["4025 <-> Hell\n","222 <-> o \n","562 <-> world\n","2 <-> , \n","8011 <-> R\n","8007 <-> N\n","1978 <-> N \n","11 <-> in \n","6307 <-> Ten\n","2327 <-> sor\n","2934 <-> flow\n","7975 <-> .\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"trusted":true,"id":"muIXF9R74ID6","colab_type":"code","colab":{}},"source":["BUFFER_SIZE = 10000\n","batch_size = 64"],"execution_count":0,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"e5GpHQJi4ID-","colab_type":"text"},"source":["Next, we are going to create batches of these strings. Use the `padded_batch` to pad zeros at the end of the sequences to make the batch of sequences the same length. If the `padded_shapes ` was not set, the default `padded_shape` value is the maximum value o the batch sequence."]},{"cell_type":"code","metadata":{"trusted":true,"id":"eCgjCHdr4IEA","colab_type":"code","colab":{}},"source":["train_dataset = train_dataset.shuffle(BUFFER_SIZE).padded_batch(batch_size, padded_shapes=([-1], []))\n","test_dataset = test_dataset.padded_batch(batch_size, padded_shapes=([-1], []))"],"execution_count":0,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"UZcnh3f84IEG","colab_type":"text"},"source":["Let's take a look at these batch data. The sequence length is variable among batch data."]},{"cell_type":"code","metadata":{"trusted":true,"id":"R9kkb6Pa4IEH","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":85},"outputId":"ba2d8d4f-f644-4da8-ee8e-251c504e6222","executionInfo":{"status":"ok","timestamp":1581586380352,"user_tz":-480,"elapsed":124493,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["for seq, label in train_dataset.take(2):\n","  print(seq.shape, label.shape)\n","  print(\"Label: {}, Data: {}\".format(label[0].numpy(), seq[0].numpy()))"],"execution_count":9,"outputs":[{"output_type":"stream","text":["(64, 1231) (64,)\n","Label: 1, Data: [7180    2    4 ...    0    0    0]\n","(64, 1223) (64,)\n","Label: 1, Data: [  19  651 3966 ...    0    0    0]\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"e9uSAF9p4IEK","colab_type":"text"},"source":["# Create the Model"]},{"cell_type":"markdown","metadata":{"id":"DahEC4HU4IEL","colab_type":"text"},"source":["Let's try the LSTMCell layer first. \n","\n","The `tf.keras.layers.Bidirectional` API provides an easier way to construct a bidirectional mechanism of RNN cells. It can be composed of different RNN cells, like LSTM, GRU or simple RNN cell. The composed cell is a real operating unit, in this case, we use LSTM cells as the unit. When we construct an LSTM layer, five basic parameters are required to be known. \n","\n","First, the `return_sequence` flag is used for returning the last output (default is set to False) or the full sequence (True). If the flag is set to True, the output sequence dimension would become [batch_size, sequence_length, lstm_cell_output_dim] instead of [batch_size, lstm_cell_output_dim] (when it is at default value).\n","\n","The `return_state` flag is used for whether to return the last state (default is set to False). If the flag is set to True, it would return three states sequentially, the sequence output, the final memory state, and the final carry state.\n","\n","![LSTM Cell](https://www.mdpi.com/water/water-11-01387/article_deploy/html/images/water-11-01387-g004-550.jpg)\n","\n","Refer to Xuan-Hien Le, etc. (2019) \n","\n","The `go_backwards` flag (default is set to False) helps you to reverse the input vector. In the previous studies, they showed better performance when the input vector was reversed in some cases.\n","\n","The `stateful` flag (default is set to False) helps you to initialize the cell with a pre-existing state.\n","\n","The `time_major` flag (default is set to False) helps you to train a model on the time major dataset. When the flag is set to True, the shape of batch time would become (timestamps, batch_size, features) instead of (batch_size, timestamps, features) (the default shape)."]},{"cell_type":"code","metadata":{"trusted":true,"id":"08bww-TZ4IEM","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":34},"outputId":"df626867-51d3-4e6e-a883-d9bc6f6eb1c5","executionInfo":{"status":"ok","timestamp":1581586385239,"user_tz":-480,"elapsed":129367,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["x = tf.random.normal([1, 20, 30])  # [Batch, Sequence_Length, Embedded_Vector_Length]\n","lstm = tf.keras.layers.LSTM(64, return_state=True, return_sequences=True)(x)\n","whole_sequence_output, final_memory_state, final_carry_state = lstm\n","whole_sequence_output.shape, final_memory_state.shape, final_carry_state.shape"],"execution_count":10,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(TensorShape([1, 20, 64]), TensorShape([1, 64]), TensorShape([1, 64]))"]},"metadata":{"tags":[]},"execution_count":10}]},{"cell_type":"markdown","metadata":{"id":"eZ25I2EF4IEQ","colab_type":"text"},"source":["Next, let's try the bidirectional layer. The bidirectional layer is a mechanism looking for a long timestamp relationship. There are two core parameters used for building a bidirectional layer.\n","\n","The `layer` parameter is used for the recurrent instance, or say the forward instance.\n","\n","The `backward_layer` is used for another recurrent instance but it is backward (from the end to the start). If no recurrent instance was assigned to the backward layer, it would use the same instance with the one assigned in the `layer` parameter.\n","\n","Because the bidirectional layer is composed of both the forward layer and the backward layer, the output of the bidirectional layer is double than of a recurrent layer."]},{"cell_type":"code","metadata":{"trusted":true,"id":"AOuzlJN74IES","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":102},"outputId":"ae364e4c-e001-49e6-db1e-0c08cc216a89","executionInfo":{"status":"ok","timestamp":1581586385241,"user_tz":-480,"elapsed":129334,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["x = tf.random.normal([1, 20, 30])  # (batch_size, timestamps, features)\n","bi = tf.keras.layers.Bidirectional(\n","  layer=tf.keras.layers.LSTM(64, return_sequences=True, return_state=True)\n",")(x)\n","returned_sequence, layer_memory_state, layer_carry_state, backward_memory_state, backward_carry_state = bi\n","returned_sequence.shape, layer_memory_state.shape, layer_carry_state.shape, backward_memory_state.shape, backward_carry_state.shape"],"execution_count":11,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(TensorShape([1, 20, 128]),\n"," TensorShape([1, 64]),\n"," TensorShape([1, 64]),\n"," TensorShape([1, 64]),\n"," TensorShape([1, 64]))"]},"metadata":{"tags":[]},"execution_count":11}]},{"cell_type":"markdown","metadata":{"id":"Dnj-vIFQ4IEX","colab_type":"text"},"source":["Next, we are going to create a text classifier. It first comes with an `embedding layer` to convert an integer list (represents a sentence) to the embedding vectors. The embedding layer is initialized as random float values. After training the text classification, the embedding layer is also trained and is potential for the dense representation between words.\n","\n","Second, `a bidirectional layer of LSTM cells` takes the output of the embedding layers. This layer consumes the embedding value iteratively, representing passing the value a timestamp by a timestamp. The bidirectional layer is the potential for a long time relationship between words.\n","\n","Finally, we add `two dense layers` to do the classification task."]},{"cell_type":"code","metadata":{"trusted":true,"id":"1N_rh32h4IEZ","colab_type":"code","colab":{}},"source":["def build_model():\n","  def _model(inputs):\n","    embed = tf.keras.layers.Embedding(input_dim=encoder.vocab_size, output_dim=64)(inputs)  # [None, None, 64]\n","    bi_lstm = tf.keras.layers.Bidirectional(\n","      tf.keras.layers.LSTM(units=64)\n","    )(embed)                                                                                # [None, 64*2]\n","    x = tf.keras.layers.Dense(units=64, activation='relu')(bi_lstm)                         # [None, 64]\n","    classifier = tf.keras.layers.Dense(units=1)(x)                                          # [None, 1]\n","    return classifier\n","\n","  inputs = tf.keras.Input(shape=(None, ))\n","  classifier = _model(inputs)\n","  return tf.keras.Model(inputs, classifier)"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"trusted":true,"id":"3H3vBDlu4IEd","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":323},"outputId":"562270b2-df4a-4741-be89-46b66946cc9a","executionInfo":{"status":"ok","timestamp":1581586385558,"user_tz":-480,"elapsed":129532,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["model = build_model()\n","model.summary()"],"execution_count":13,"outputs":[{"output_type":"stream","text":["Model: \"model\"\n","_________________________________________________________________\n","Layer (type)                 Output Shape              Param #   \n","=================================================================\n","input_1 (InputLayer)         [(None, None)]            0         \n","_________________________________________________________________\n","embedding (Embedding)        (None, None, 64)          523840    \n","_________________________________________________________________\n","bidirectional_1 (Bidirection (None, 128)               66048     \n","_________________________________________________________________\n","dense (Dense)                (None, 64)                8256      \n","_________________________________________________________________\n","dense_1 (Dense)              (None, 1)                 65        \n","=================================================================\n","Total params: 598,209\n","Trainable params: 598,209\n","Non-trainable params: 0\n","_________________________________________________________________\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"trusted":true,"id":"WtXFScqZ4IEh","colab_type":"code","colab":{}},"source":["model.compile(loss=tf.keras.losses.BinaryCrossentropy(from_logits=True), \n","              optimizer=tf.keras.optimizers.Adam(1e-4), \n","              metrics=[tf.keras.metrics.BinaryAccuracy()])"],"execution_count":0,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"tN1y7gND4IEl","colab_type":"text"},"source":["# Train the Model"]},{"cell_type":"code","metadata":{"_kg_hide-output":true,"trusted":true,"id":"rRMeyRS64IEn","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":357},"outputId":"6e39f57e-9ce6-4617-88aa-b95510d041a2","executionInfo":{"status":"ok","timestamp":1581586863059,"user_tz":-480,"elapsed":607015,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["history = model.fit(train_dataset, epochs=10, validation_data=test_dataset, validation_steps=20)"],"execution_count":15,"outputs":[{"output_type":"stream","text":["Epoch 1/10\n","391/391 [==============================] - 50s 129ms/step - loss: 0.6456 - binary_accuracy: 0.5595 - val_loss: 0.4751 - val_binary_accuracy: 0.7453\n","Epoch 2/10\n","391/391 [==============================] - 47s 121ms/step - loss: 0.3452 - binary_accuracy: 0.8507 - val_loss: 0.4176 - val_binary_accuracy: 0.8031\n","Epoch 3/10\n","391/391 [==============================] - 47s 121ms/step - loss: 0.2541 - binary_accuracy: 0.9009 - val_loss: 0.3594 - val_binary_accuracy: 0.8617\n","Epoch 4/10\n","391/391 [==============================] - 48s 122ms/step - loss: 0.2169 - binary_accuracy: 0.9172 - val_loss: 0.3358 - val_binary_accuracy: 0.8656\n","Epoch 5/10\n","391/391 [==============================] - 48s 122ms/step - loss: 0.1842 - binary_accuracy: 0.9359 - val_loss: 0.3665 - val_binary_accuracy: 0.8656\n","Epoch 6/10\n","391/391 [==============================] - 48s 122ms/step - loss: 0.1631 - binary_accuracy: 0.9441 - val_loss: 0.3855 - val_binary_accuracy: 0.8578\n","Epoch 7/10\n","391/391 [==============================] - 48s 122ms/step - loss: 0.1486 - binary_accuracy: 0.9491 - val_loss: 0.3989 - val_binary_accuracy: 0.8609\n","Epoch 8/10\n","391/391 [==============================] - 47s 121ms/step - loss: 0.1352 - binary_accuracy: 0.9546 - val_loss: 0.4030 - val_binary_accuracy: 0.8516\n","Epoch 9/10\n","391/391 [==============================] - 47s 121ms/step - loss: 0.1326 - binary_accuracy: 0.9549 - val_loss: 0.5115 - val_binary_accuracy: 0.8523\n","Epoch 10/10\n","391/391 [==============================] - 47s 120ms/step - loss: 0.1255 - binary_accuracy: 0.9573 - val_loss: 0.5034 - val_binary_accuracy: 0.8664\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"trusted":true,"id":"hz4oPR6Q4IEr","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":34},"outputId":"8cf97646-fe55-4222-d07d-e1ea35ae6b60","executionInfo":{"status":"ok","timestamp":1581586880781,"user_tz":-480,"elapsed":624728,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["test_loss, test_acc = model.evaluate(test_dataset)"],"execution_count":16,"outputs":[{"output_type":"stream","text":["    391/Unknown - 18s 46ms/step - loss: 0.4979 - binary_accuracy: 0.8558"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"BFCftWN_4IEu","colab_type":"text"},"source":["The above model does not `mask` the padding applied to the sequences. This causes a skew when applying the model (trained on padded sequences) to the un-padded sequences. More information please refer to https://www.tensorflow.org/guide/keras/masking_and_padding.\n"]},{"cell_type":"code","metadata":{"trusted":true,"id":"kthdBDhi4IEv","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":388},"outputId":"5c003e2c-c9bb-4bce-a979-f6a207bc0ef9","executionInfo":{"status":"ok","timestamp":1581586881494,"user_tz":-480,"elapsed":625432,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["plot_history(history, 'loss')"],"execution_count":17,"outputs":[{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAikAAAFzCAYAAAD7bpkSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nySRFkiRN0t+pJbq21ahk2QAAAABJRU5ErkJggg==\n","text/plain":["<Figure size 648x432 with 1 Axes>"]},"metadata":{"tags":[]}}]},{"cell_type":"code","metadata":{"trusted":true,"id":"x3261g064IEz","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":388},"outputId":"45cfea2e-2554-4080-95e1-9954d77c7da4","executionInfo":{"status":"ok","timestamp":1581586881496,"user_tz":-480,"elapsed":625429,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["plot_history(history, 'binary_accuracy')"],"execution_count":18,"outputs":[{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAjAAAAFzCAYAAAAg1TMwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0\ndHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nOzdeXydZZ338c8v+9ZsbUpp0iWFlrYs\n3dICssOAFVlkEaiiLGJnHJBRxtHqMI7iqDwzjqI8PAgqqIy0IoxaZxAERUFRyQllLUtLe0KTFkib\nk6ZL9vN7/rhP0tM0pTntOTlZvu/X67xyn3s7v6S0+XJd131d5u6IiIiIjCQZ6S5AREREJFEKMCIi\nIjLiKMCIiIjIiKMAIyIiIiOOAoyIiIiMOAowIiIiMuJkpbuAZJowYYJPnz493WWIiIhIktTV1W11\n94r++0dVgJk+fTqhUCjdZYiIiEiSmFn9QPvVhSQiIiIjjgKMiIiIjDgKMCIiIjLijKoxMAPp6uqi\noaGB9vb2dJcyquTl5VFVVUV2dna6SxERkTFo1AeYhoYGxo0bx/Tp0zGzdJczKrg727Zto6Ghgerq\n6nSXIyIiY9Co70Jqb29n/PjxCi9JZGaMHz9erVoiIpI2oz7AAAovKaCfqYiIpNOYCDDptG3bNubP\nn8/8+fOZNGkSlZWVfe87OzsHdY9rrrmG11577V3PueOOO/jJT36SjJJFRESGvVE/Bibdxo8fz3PP\nPQfAl770JYqKivjMZz6z1znujruTkTFwnrz33nsP+DnXX3/9oRcrIiIyQqgFJk3Wr1/P3Llz+fCH\nP8zRRx/Nli1bWL58OTU1NRx99NHccsstfeeefPLJPPfcc3R3d1NaWsqKFSuYN28eJ554Iu+88w4A\nN998M7fddlvf+StWrGDJkiUcddRRPP300wDs2rWLSy65hLlz53LppZdSU1PTF65ERERGkjHVAvPl\nX73M2s2tSb3n3MnF/Ov5Rx/Uta+++io//vGPqampAeDWW2+lvLyc7u5uzjjjDC699FLmzp271zXb\nt2/ntNNO49Zbb+Wmm27innvuYcWKFfvc29155plnWL16NbfccguPPPIIt99+O5MmTeKhhx7i+eef\nZ+HChQdVt4iISLqNqQAz3BxxxBF94QVg5cqV/OAHP6C7u5vNmzezdu3afQJMfn4+73vf+wBYtGgR\nTz311ID3vvjii/vOCYfDAPzxj3/kc5/7HADz5s3j6KMPLniJiMjY4e50dEdp7+qhvSv2tXvPdkd3\nlGnlBUyfUDikdY2pAHOwLSWpUli45w973bp1fPvb3+aZZ56htLSUK6+8csDHlHNycvq2MzMz6e7u\nHvDeubm5BzxHRERGlt4w0dEVpb27p+/rXuGiq4f27j3hoqOrZ7/ho70rSscB7tPZHT1gXZ85ZxY3\nnDlzCH4Ce4ypADOctba2Mm7cOIqLi9myZQuPPvooS5cuTepnnHTSSTzwwAOccsopvPjii6xduzap\n9xcRSYXdnd007+qkeVcnO9q7ibrjDg5EYxuOE40G+9ydqAME50Vjx4NtB+jbHvA+Tmx/7D7uwTlR\nj91/gPvss/9d7hP73O6o7wkKfeEiFjziwkV7XLjo6I4S+4iEZRjkZWcGr6wM8rIzyc3OJC87g7ys\nTMYX5gT7YseC48GxvN7zYl9zs/Zcl5udSVVZ/sH/AR+klAcYM1sKfBvIBL7v7rf2Oz4NuAeoAJqB\nK929IXasB3gxduqb7n5BqutNl4ULFzJ37lxmz57NtGnTOOmkk5L+GZ/85Cf56Ec/yty5c/teJSUl\nSf8cEZH96Yk6Lbs7iezuZNvO2NddnUR29fu6u5PmnZ007+6kvevALQAjRYYF82gZkJVpsTCxJxzk\nxsJFaUHO3kEhLnjkZscFiqwDhYw9+7IybFTN4WV+sFFuMDc3ywReB84GGoBaYJm7r40752fA/7j7\nj8zsTOAad/9I7NhOdy8a7OfV1NR4KBTaa98rr7zCnDlzDv2bGQW6u7vp7u4mLy+PdevWcc4557Bu\n3Tqysg4ux+pnKyLxrSP9XwOFlJa2rv22IBTlZlFWmE15YS7lBbGvhXt/HZeXRWZGEAAsLgyYWRAO\nsNj+YDsjY8++DAP6tvfcIyP2S71vf++1BsRt935WRnDCge8TX9coCg5Dzczq3L2m//5Ut8AsAda7\n+4ZYEauAC4H4vou5wE2x7SeAX6S4pjFr586dnHXWWXR3d+Pu3HXXXQcdXkRk9OltHdkriMS1hMQH\nkwO1jmRmGGUFOYwvzKGsMJs5k4r3DidFuZQX5FBeGLzKCrPJzcoc4u9YRrJU//aqBDbFvW8Aju93\nzvPAxQTdTBcB48xsvLtvA/LMLAR0A7e6u8LNISgtLaWuri7dZYjIEHB3dnZ007K7KwgcSWgdCYJG\nDhPH5XHUYcWML8qhrCBnn1aS8oIcivOz1OogKTUc/vf7M8D/NbOrgSeBRqAndmyauzea2Qzgd2b2\noru/EX+xmS0HlgNMnTp16KoWERkiXT1RIrs7gzCyq5PI7q7YOJKuvvEkkd1dbI+Fld793dGB04ha\nR2Q0SHWAaQSmxL2viu3r4+6bCVpgMLMi4BJ3b4kda4x93WBmvwcWAG/0u/5u4G4IxsCk5LsQEUmC\n/q0ifQGkXyjpCyuxrzs79j8VQk5mBqUF2ZQV5FBakM0RFUWUFWZTWpBDWUHv1z1hpLwwh+I8tY7I\nyJfqAFMLzDSzaoLgcgXwofgTzGwC0OzuUeDzBE8kYWZlwG5374idcxLw7ymuV0RkUDq7o7S0HbhV\nJH5fy+6u/baKABTnZVFWmENpQQ7ji3I4cmJRXziJDyOlBdmUFQb78rMzFUZkTEppgHH3bjO7AXiU\n4DHqe9z9ZTO7BQi5+2rgdODrZuYEXUi9qxLOAe4ysyjBmk23xj+9JCKSbDvau2hsaaMx0sbmlja2\n7uw8uFaRrAzK4lpFZk4s6msRKSvIoWSfUJJNSX42WZlank5ksFI+BsbdHwYe7rfvi3HbDwIPDnDd\n08Cxqa4v1c444wxWrFjBe9/73r59t912G6+99hp33nnngNcUFRWxc+dONm/ezI033siDD+7z4+H0\n00/nG9/4xl5LEfR32223sXz5cgoKCgA499xzuf/++yktLT3E70pk5IlGna07O4KAEhdSGlvaaIgE\nX3e07xtKSvKzg4ChVhGRYWU4DOId1ZYtW8aqVav2CjCrVq3i3//9wL1hkydPHjC8DNZtt93GlVde\n2RdgHn744QNcITJydXZH2bI9CCbxIaWxJQgqm1va6ezZ+5HfcXlZVJbmU1WWz5LqcipL86ksy2dy\naT5VpfmML8olM0NBRGQ4UoBJsUsvvZSbb76Zzs5OcnJyCIfDbN68mQULFnDWWWcRiUTo6uri3/7t\n37jwwgv3ujYcDnPeeefx0ksv0dbWxjXXXMPzzz/P7NmzaWtr6zvvE5/4BLW1tbS1tXHppZfy5S9/\nme985zts3ryZM844gwkTJvDEE08wffp0QqEQEyZM4Jvf/Cb33HMPANdddx2f+tSnCIfDvO997+Pk\nk0/m6aefprKykl/+8pfk5w/9FNEi/bW2dwUtJgMElMZIG007O/Z6BNgMJo7LZXJpPsdUlvDeYyZR\nVRqEk96QUpyXnb5vSEQOydgKML9eAW+9eODzEjHpWHjfrfs9XF5ezpIlS/j1r3/NhRdeyKpVq7js\nssvIz8/n5z//OcXFxWzdupUTTjiBCy64YL/NznfeeScFBQW88sorvPDCCyxcuLDv2Fe/+lXKy8vp\n6enhrLPO4oUXXuDGG2/km9/8Jk888QQTJkzY6151dXXce++9/PWvf8XdOf744znttNMoKytj3bp1\nrFy5ku9973tcdtllPPTQQ1x55ZXJ+VmJ7Edv905DrLUkPpj0hpX+3Ts5mRlMLs2jsiyf04+qCIJJ\nLJxUluYzqSRPj/6KjGJjK8CkSW83Um+A+cEPfoC784UvfIEnn3ySjIwMGhsbefvtt5k0adKA93jy\nySe58cYbATjuuOM47rjj+o498MAD3H333XR3d7NlyxbWrl271/H+/vjHP3LRRRf1rYZ98cUX89RT\nT3HBBRdQXV3N/PnzAVi0aBHhcDhJPwUZyzq6e3hrezuNkbZ9Q0pLG1sG6N4pzsuisqyAqrJ8jq8u\n72s16Q0pEwpzyVD3jsiYNbYCzLu0lKTShRdeyKc//WmeffZZdu/ezaJFi/jhD39IU1MTdXV1ZGdn\nM336dNrb2xO+98aNG/nGN75BbW0tZWVlXH311Qd1n165ubl925mZmXt1VYm8m+6eKM9tamHtltZ9\ngsr+uncqS/M5rqqUpcfkURXXtVNZms84de+IyLsYWwEmTYqKijjjjDO49tprWbZsGQDbt29n4sSJ\nZGdn88QTT1BfX/+u9zj11FO5//77OfPMM3nppZd44YUXAGhtbaWwsJCSkhLefvttfv3rX3P66acD\nMG7cOHbs2LFPF9Ipp5zC1VdfzYoVK3B3fv7zn3Pfffcl/xuXUe/Nbbt5cl0TT77exJ/f2MaO2KPF\nOVkZVJbmM7k0j9OPqqCytKCvu6eqtIBJJXnkZOmRYRE5eAowQ2TZsmVcdNFFrFq1CoAPf/jDnH/+\n+Rx77LHU1NQwe/bsd73+E5/4BNdccw1z5sxhzpw5LFq0CIB58+axYMECZs+ezZQpUzjppJP6rlm+\nfDlLly5l8uTJPPHEE337Fy5cyNVXX82SJUuAYBDvggUL1F0kB7SjvYs/v7GNp9Zt5cl1TdRv2w1A\nZWk+5807nFNnVrBoWhkTitS9IyKpZb6/lbtGoJqaGg+FQnvte+WVV5gzZ06aKhrd9LMd/XqizouN\n23nq9SaeXNfEs2+20BN1CnIyOXHGeE6dVcEpMydQPaFQ856ISEqYWZ277zPpmVpgRGQvm1vaeGpd\nE0+u28qf1m+lZXcXZnDM5BL+9tQZnDqrgoVTy9QFJCJppQAjMsbt7uzmrxua+8ayvNG0C4DDinP5\nmzmHceqsCk46Yjzji3IPcCcRkaGjACMyxkSjztotrcE4ltebqKuP0NkTJTcrg+NnjGfZkqmcOquC\nmROL1C0kIsPWmAgw7q5/iJNsNI2dGgveaW3nqXVbeWpdE39cv5WtOzsBmD1pHFefNJ1TZ1ZQM72M\nvGxN/CYiI8OoDzB5eXls27aN8ePHK8Qkibuzbds28vLy0l2K7Ed7Vw+14ea+VpZX39oBwISiHE4+\ncgKnzqrg5CMnMLFYf4YiMjKN+gBTVVVFQ0MDTU1N6S5lVMnLy6OqqirdZUiMu/P62zv7Bt/+dcM2\nOrqj5GRmUDO9jBXvm80pMycwZ1KxHm8WkVFh1AeY7Oxsqqur012GSNJt29nBH9dv7esaeru1A4Aj\nJxbxoeODcSzHV5dTkDPq/5qLyBikf9lERojO7ih19RGeWtfEU+u28tLm7bhDaUE2Jx05gdNmVnDy\nzAlMLtXq4SIy+inAiAxT7s6Grbtik8ht5S8btrG7s4esDGPh1DJu+ptZnDqrgmMqS8hUt5CIjDEK\nMCLDyPbdXfzpjaBL6MnXt9LYEiymOX18AZcsrOLUWRWcMKNcCx2KyJinACOSRm2dPazZFOEvG5p5\n8vUmXmhoIeowLjeL9xw5nr8/4whOObKCqeML0l2qiMiwogAjMoS2t3VRV9/MMxsjPLNxGy82bqer\nx8kwmDellBvOnMlpsyYwr6qUrExN1S8isj8KMCIp9M6Odmo3RqgNN/PXjc28+lYr7pCdaRxXVcrH\nTp7B8dXlLJxWRkm+uoVERAZLAUYkSdydhkgbz2xs5pmNzdSGm9mwNVhXKD87k4XTSvnUWbNYUl3O\n/Cml5Odo1lsRkYOlACNykKJRZ33Tzr0Cy5bt7QCU5GezeHoZVyyZwuLp5RxTWUK2uoRERJJGAUZk\nkLp7oqzd0rpXYIns7gJg4rhcllSXc3x1OYury5k1cZxmvBURSSEFGJH9aO/q4flNLUFgCTfzbH2E\nXZ09AEwbX8DfzDmMxbHQMrW8QGttiYgMIQUYkZgd7V3U1QcDbp/Z2Mzzm7bT2RMFglWbL15YxZLq\ncpZUl3OYFkEUEUkrBRgZs7bt7IiFlQjPhLexdnMrUYesDOOYyhKuPmk6S6aXUzO9jNKCnHSXKyIi\ncRRgZMxobGmjdmPwOPMzG7fxRlPwhFBuVgYLpgZzsBxfXc6CqaVaAFFEZJjTv9IyKrk7bzTt6usO\nemZjc9+0/OPysqiZVsali6awpLqMYytLycnSE0IiIiNJygOMmS0Fvg1kAt9391v7HZ8G3ANUAM3A\nle7eEDt2FXBz7NR/c/cfpbpeGZl6os4r/Z4Q2rarE4AJRbksqS7j46dUs7i6nNmTirX4oYjICJfS\nAGNmmcAdwNlAA1BrZqvdfW3cad8AfuzuPzKzM4GvAx8xs3LgX4EawIG62LWRVNYsI8cLDS08tW4r\nteFm6sIRdnR0A1BVls9pR1UEjzRPL6d6QqGeEBIRGWVS3QKzBFjv7hsAzGwVcCEQH2DmAjfFtp8A\nfhHbfi/wmLs3x659DFgKrExxzTLMtXf18JX/WctP/vomADMnFnHB/MksiQWWyaX5aa5QRERSLdUB\nphLYFPe+ATi+3znPAxcTdDNdBIwzs/H7ubay/weY2XJgOcDUqVOTVrgMT+vf2cEN96/h1bd2sPzU\nGfzdaUdQXqgnhERExprhMHLxM8BpZrYGOA1oBHoGe7G73+3uNe5eU1FRkaoaJc3cnQdCmzj/9j/x\nzo4O7r1mMV84d47Ci4jIGJXqFphGYErc+6rYvj7uvpmgBQYzKwIucfcWM2sETu937e9TWawMTzs7\nurn55y/yi+c2c+KM8dx2xXxNJCciMsalOsDUAjPNrJoguFwBfCj+BDObADS7exT4PMETSQCPAl8z\ns7LY+3Nix2UMealxOzfc/yxvNu/mprNncf0ZR+oJIhERSW2AcfduM7uBIIxkAve4+8tmdgsQcvfV\nBK0sXzczB54Ero9d22xmXyEIQQC39A7oldHP3fnh02G+/vCrlBfmsPLjJ3D8jPHpLktERIYJc/d0\n15A0NTU1HgqF0l2GHKKW3Z3804Mv8Njatzlr9kS+8cF5lGmsi4jImGRmde5e03+/ZuKVYaU23Mw/\nrFxD084O/uW8uVx70nTN4SIiIvtQgJFhoSfq3Pn79Xzr8XVUleXz0Cfew3FVpekuS0REhikFGEm7\nd1rb+fQDz/Gn9ds4f95kvnbRMYzLy053WSIiMowpwEha/eH1Jv7xgefY2dHN/7nkWC6rmaIuIxER\nOSAFGEmLrp4o//mb1/nuH95g1mFF3P/xE5h12Lh0lyUiIiOEAowMuU3Nu7lx1RrWvNnCsiVT+eJ5\nc8nPyUx3WSIiMoIowMiQeuSlLXz2wRdwh//7oQWcd9zkdJckIiIjkAKMDIn2rh6++r+vcN9f6plX\nVcLtyxYydXxBussSEZERSgFGUm79Ozu54f5n+1aQ/sw5R5GTNRzWERURkZFKAUZS6sG6Bv7lFy+R\nn5PJvVcv5ozZE9NdkoiIjAIKMJISOzu6+eIvXuK/1zRywoxybrt8AZNKtIK0iIgkhwKMJN1Ljdv5\n5Mo11G/bxaf/ZhY3nKkVpEVEJLkUYCRp3J0fPR3maw+/SllhNvd//ARO0ArSIiKSAgowkhQtuzv5\n7IMv8Ju1b3NmbAXpcq0gLSIiKaIAI4csFG7mH1Y9xzs72rn5/XP42MnVWg5ARERSSgFGDlo06tz5\nhzf45mOvU1maz4N/9x7mTdEK0iIiknoKMHJQ3tnRzk0/fZ4/rt/KeccdztcuPpZirSAtIiJDRAFG\nEvbk603cFFtB+taLj+XyxVpBWkREhpYCjAxaV0+Ubz72Onf+XitIi4iMea1bYPOz0PgszDgNqk8d\n0o9XgJFBaYjs5saVa3j2zRaWLZnCF887WitI94pGoXMndLRCxw5ob41tt+7Zbo8di98Xv79zV7q/\ni+TJzoeSSiiugpKqYLtkChRXBtvjDodMdTeKjCi7m4OwsnkNNK4JtndsCY5ZJuSOU4CR4ad3Bemo\nw+3LFnD+vFG0gnRPVyxEbN87fPQFju0DhI/+QWQH4O/+OZYR/AXPLQm+5hVD0SQYPzPYzikERkk3\nXOdO2N4ILW/Cm09D+/a9j1tG8L33hpviWMApqQz2FVdB4QRQt6RIenTshC3P72ld2fwsRMJ7jo8/\nEqafApULYfJCmHQs5Az94rwKMLJf7V09fO3hV/jxn+s5rqqE25ctYNr4wnSXtUd3R/B/Bftr1Rgw\niPRrJeluP/DnZOXFwkdxEDZyx0HhDMgr6be/eE84iQ8qubGAMlZ/IXfsCAJNawNsbwi2tzcE77e8\nAK8+DD0de1+TmbtvuOm/nVecnu9HZDTp7oC3X4oFlTXB162vgUeD4yVTYPICWHR1EFYOnwf5w+Np\nUwUYGdAbTTu54f41vLKlletOruazS2cPrxWk166GX3wi+L/9/bI9AaM3TBRMgLLqPcGiL3jEhZP+\n+7M0Id8hyR0HE2cHr4G4w+5tsH1TLOg07tne3gAb/xA0Vff+g9p335K4VpvKfl1VVVA8GbJyU//9\npVO0J+h+7NwFXbuDvw+97+NfPR1B111ZNZRXQ0F5uiuXdIj2QNOre1pVGp+Ft1+GaFdwvGBC0Koy\n98JY68oCKBq+C/AqwMg+Hqpr4F9++RK5WRncc3UNZ84+LN0l7eEOT98Oj30Rqmpg/odigaOkXwtI\nMeQUQcYwCl0yMLOgy6hwQvAP5kB6uoMQ0xoLNdsb9t5uCEFb877XFU4ceBxO73bRYUPz30g0Ct1t\nsUDRGzL6BY6uAYJH32tnLKD0u7677eDqySuJhZkZQaCJ3y6apL83o4E7NG/Y06qy+dmgW6hrd3A8\ntxgmz4cTrw/+3lUuDP5ejKCWYgUY6bOro5t/ia0gfXx1Od++YpitIN3TDQ9/BuruhbkfgIu+GwwY\nldEvMwtKpwSv/encDa2bg9ab/kGn6TVY/7sgJMTLyApaauIHHMd3VVnm3qHh3Vo59goZ/V79P/dA\nsguDbsecgiCI5xQG4XzcpLhjhbFjBXHbhZAdd03vKzM7aNGKbITmjcEvtsjG4Jfb2l+C9+z57Kw8\nKJseBJreFpvekFM6VQOwhyP34L/9+DErm9fsGX+WlQeTjoOFHw26gSoXQvkRIz6oKsAIAC9v3s4n\n719DeNsu/uGsmdx41szhtYJ0eys8eA2sfxxO/jSc+cUR/5dPkiynACYcGbwG4g5tkVi4aewXdBph\n01/g5c0Q7R78Z2bl7QkJ8cGiYMK+wWJ/4SI+jGQXBK9U/LedVwKHzd13f09X8LNo3hgXcGLbbzyx\ndyuPZQZBb6CWm7LpscHoknK7tgUBJT6w7Hw7OJaRBRPnwtEXBS0rkxfCxDmjMngqwIxx7s59f6nn\n3/73FcoKsvnJdSdw4hHDbAXp7Q1w/+Xwzitw/ndg0VXprkhGIrNg7EdBefDUxECiPbDznT0tN/j+\nWzayC4KWoZEuMzsWQmbse8wddry1J9hEYq03zRvh5Z8HgTBe0WH9Wm7itvPLRlT3xLDRsQM2PxcX\nVtZAS33soMGEmTDjjLgngo4ZMy3T5n6Axz9HkJqaGg+FQukuY8TYvruLzz70PI++/DZnHFXBNz44\nj/FFw2zQ4+bngvDStRsu+xEccWa6KxKRXm2RgVtumjfCjs17n5tbsnd3VHwrzrjD1aIK0NUe90RQ\nLLBsfZ2+aRpKpkLlgj3dQIfPHxNP45lZnbvX9N+f8v99MLOlwLeBTOD77n5rv+NTgR8BpbFzVrj7\nw2Y2HXgFeC126l/c/e9SXe9YUVffzI0r96wgfe1J1WQMpy4jgNcegQevDf6P+SOPDtz8LSLpk18G\nlWXBL9P+utqCuUP6t9xseR5e+dXeXXW9424GarkZDeNu3IOn6KI9wXij3q8tb+49yPbtl/f8XAon\nBj/XYy7Z80RQ4YT0fh/DTEoDjJllAncAZwMNQK2ZrXb3tXGn3Qw84O53mtlc4GFgeuzYG+4+P5U1\njjXRqPPdJ9/gP3/zOpNL84bvCtJ/vQseWRHMObBsVTB4UURGjuz8YOzFxDn7HuvpDsbdRDbu24Kz\n8Q97npSBYOLDkil7WmuKJ8cCQU+/QDBAQDjQ/n2OJXKPaL97vcu5/acA6C+3JHgi6D2f3NO6Ulyp\nLrcDSHULzBJgvbtvADCzVcCFQHyAcaC3DawE6NfuKMn0Tw++wEPPNvD+4w7n68NxBeloDzz6Bfjr\nd+Go98Ml39PAQJHRJjNrT3fSEf2OuQcDUvcKNrGnptb+cu/H5S0jGFickRn3NaPf+8yge2qf8zKD\ngLDPuZlg2YO8R8Z+Pm+QdRRNDAJL+Qx1oR2EVAeYSmBT3PsG4Ph+53wJ+I2ZfRIoBP4m7li1ma0B\nWoGb3f2p/h9gZsuB5QBTp05NXuWjUHtXD796fjNXLJ7C1y8+dvitIN25Cx66Dl57GE64Hs75SvAX\nXUTGDrOgxXXcJJh24r7He7r2hAEZ04bDfwHLgB+6exVwLnCfmWUAW4Cp7r4AuAm438z2Ga3k7ne7\ne42711RUVAxp4SPNi43b6eyJctacw4ZfeGndAve+D15/BM79Biz9msKLiOwrM1vhRYDUt8A0AvEz\nT1XF9sX7GLAUwN3/bGZ5wAR3fwfoiO2vM7M3gFmAHjM6SLXhoOl10bSyNFfSz9svw08uC55oWLYK\nZr033RWJiMgwl+oYWwvMNLNqM8sBrgBW9zvnTeAsADObA+QBTWZWERsEjJnNAGYCG1Jc76gWCkc4\noqKQ8sJhtLbP+sfhB+8NBrpd+4jCi4iIDEpKA4y7dwM3AI8SPBL9gLu/bGa3mNkFsdP+Efi4mT0P\nrASu9mBymlOBF8zsOeBB4O/cfYDFTmQwolEnFG5m8fRhtIhb6N6g5aVsOlz3Wzj8uHRXJCIiI0TK\n54Fx94cJHo2O3/fFuO21wEkDXPcQ8FCq6xsr1r2zk9b2bmqGQ4CJRuHxf4WnvwMzz4FL7wnWeRER\nERmkUTAPtgxG7/iXxdPTPP6lqw3+ezm8shoWXwdL/8/omI5dRESGlH5zjBF19REqxuUytbwgfUXs\nbIKVV0BjHZzz1WAZ9+H2NCNWYoAAACAASURBVJSIiIwICjBjRG24mcXTy9L3+HTTa/CTS4MQc/l9\nMOf89NQhIiKjgh6mHwO2bG+jIdLGomlpGv+y8Un4wdnBQmXX/K/Ci4iIHDIFmDEgFA6WvE/L+Jc1\nP4H7LgpWm73ucahcNPQ1iIjIqKMupDEgFG6mICeTuYcP4bLr7vDEV+HJ/4AZp8MHfwT5w3DRSBER\nGZEUYMaA2nCEBVNLycocoga37g745fXw4s9gwUfgvG8F03+LiIgkibqQRrkd7V28+lYrNUM1/mXX\nNvjxhUF4OeuLcMHtCi8iIpJ0aoEZ5da82ULUGZoZeLe9ETxptL0xmJzumEtS/5kiIjImKcCMcqFw\nMxkG86emePxJ/Z9h1TKwDLjqVzD1+NR+noiIjGnqQhrlasMR5k4upig3hVn1xQfhxxdAwfjgSSOF\nFxERSTEFmFGsqyfKmk2R1I1/cYc//Ac89DGoWgwfewzKZ6Tms0REROKoC2kUW7u5lfauaGrGv3R3\nwv98Cp77CRx3eTBYNys3+Z8jIiIyAAWYUax3AceaZE9g1xaBn34Ewk/B6Z+H0z6nNY1ERGRIKcCM\nYqFwhCnl+RxWnJe8m0bC8JMPQvNGuOgumHdF8u4tIiIySAowo5S7E6pv5tSZFcm7aUMI7r8col3w\nkZ9D9SnJu7eIiEgCNIh3lApv283WnZ3UJGv8y9pfwg/fD7lF8LHHFV5ERCStFGBGqd7xL4e8gKM7\n/Onb8MBVMOk4uO63UDErCRWKiIgcPHUhjVJ14QilBdkcUVF08Dfp6YaHPwN198LcD8BF34Xs/OQV\nKSIicpAUYEap2vpmaqaVkZFxkE8HtbfCz66GN34LJ98EZ/4LZKjBTkREhgcFmFFo284ONjTt4oOL\nphzcDbY3wE8ug6ZX4fzvwKKrklugiIjIIVKAGYVC9RHgIMe/bF4D918BXbvhygfhiDOTXJ2IiMih\nU4AZhULhZnKyMji2qiSxC1/7NTx4bbCm0UcehcPmpqZAERGRQ6RBDaNQbTjCvKoScrMyB3/RX74L\nqz4EFUcFTxopvIiIyDA26ABjZuebmQLPMNfW2cPLm7cPfv6XaA88/Fl45HMw631w9f/CuMNSW6SI\niMghSiSQXA6sM7N/N7PZqSpIDs3zDS109fjgxr907IRVH4Zn7oITrofL74OcwtQXKSIicogGPQbG\n3a80s2JgGfBDM3PgXmClu+9IVYGSmFBsAruFUw8QYFq3wMrL4a0X4dxvwJKPD0F1IiIiyZFQl5C7\ntwIPAquAw4GLgGfN7JMpqE0OQm04wqzDiigtyNn/SW+9BN8/C7auh2WrFF5ERGTESWQMzAVm9nPg\n90A2sMTd3wfMA/7xXa5bamavmdl6M1sxwPGpZvaEma0xsxfM7Ny4Y5+PXfeamb03kW9sLOqJOs/W\nR959/Mv6x+GepeBRuPYRmKUfq4iIjDyJPEZ9CfAtd38yfqe77zazjw10gZllAncAZwMNQK2ZrXb3\ntXGn3Qw84O53mtlc4GFgemz7CuBoYDLwuJnNcveeBGoeU15/ewc7Orr3P/6l9vvBgN2Jc+FDP4WS\nyqEtUEREJEkS6UL6EvBM7xszyzez6QDu/tv9XLMEWO/uG9y9k6Dr6cJ+5zhQHNsuATbHti8EVrl7\nh7tvBNbH7if70Tv+pWZavxaYaA/8egX87z/CkX8D1/5a4UVEREa0RALMz4Bo3Pue2L53Uwlsinvf\nENsX70vAlWbWQND60jueZjDXSpzacIRJxXlUlcUtuNixA1Yug7/eCSf8PSxbCbnj0lekiIhIEiQS\nYLJirSgAxLbfZaTooC0DfujuVcC5wH2JzDdjZsvNLGRmoaampiSUM3KFws0sml6GWWwBx+0NwXiX\n9Y/D+/8Tln4dMhKY3E5ERGSYSiTANJnZBb1vzOxCYOsBrmkE4lcUrIrti/cx4AEAd/8zkAdMGOS1\nuPvd7l7j7jUVFRWD/FZGn8aWNjZvb2fxtNj4l8Y6+N6Z0PImfPgBWHxdegsUERFJokQCzN8BXzCz\nN81sE/A54G8PcE0tMNPMqs0sh2BQ7up+57wJnAVgZnMIAkxT7LwrzCzXzKqBmcSNwZG99Y1/mV4O\na1fDve+HrFz42G+CcS8iIiKjSCIT2b0BnGBmRbH3OwdxTbeZ3QA8CmQC97j7y2Z2CxBy99UEj2B/\nz8w+TTCg92p3d+BlM3sAWAt0A9frCaT9C4UjFOVmMueNH8DvvgxVi+GKlVA0dlulRERk9LIgKwzy\nZLP3EzzWnNe7z91vSUFdB6WmpsZDoVC6y0iL8771W/6p+y5O2/UoHH0xfOD/QXb+gS8UEREZxsys\nzt1r+u8fdAuMmX0XKADOAL4PXIq6dIaF1uZ3+OfIzZyYsRZO/Syc/nnI0LqbIiIyeiXyW+497v5R\nIOLuXwZOBGalpiwZtG1vkHXvOSy013n9Pf8JZ/6zwouIiIx6ifyma4993W1mk4EugvWQJF3CfwrW\nNGpr5iPd/0zV6VenuyIREZEhkchSAr8ys1LgP4BnCQbcfi8lVcmBPXc/rL4Ryqv5nK2go/QwCnIS\n+eMUEREZuQbVAhObWO637t7i7g8B04DZ7v7FlFYn+4pG4bdfgV98AqadSOdVj/KbLQXvvoCjiIjI\nKDOoAOPuUYJFGXvfd7j79pRVJQPraoMHr4GnvgELPwpX/jcvRYyO7uj+F3AUEREZhRIZA/NbM7vE\n+uaplyG142344fth7S/h7K/A+d+BzOy+CewW9V/AUUREZBRLZNDE3wI3Ad1m1g4Y4O5e/O6XySF7\n+2W4/3LYvQ0u/y+Yc17fodpwhOnjC6gYl5vGAkVERIZWIjPxagnjdFj3GPzsGsgtgmt+DZPn9x1y\nd+rqI5w5e2IaCxQRERl6iUxkd+pA+939yeSVI3v5693wyOfgsKNh2U+hpHKvwxu27qJ5V6fGv4iI\nyJiTSBfSP8Vt5wFLgDrgzKRWJNDTDY9+AZ65C446Fy7+XtAC089eCziKiIiMIYl0IZ0f/97MpgC3\nJb2isa69FR68FtY/BifeAGffAhmZA55aG45QXpjDjAmFQ1ykiIhIeh3KzGcNwJxkFSJAy6ZgsG7T\nq3Det6Dm2nc9PRRuZtG0MvRgmIiIjDWJjIG5nWD2XQgev55PMCOvJENDHay8Aro74MoH4Yh375l7\nZ0c74W27+dDxU4eoQBERkeEjkRaYUNx2N7DS3f+U5HrGppd/AT//Wyg6DK76FUycfcBL6sIRQONf\nRERkbEokwDwItLt7D4CZZZpZgbvvTk1pY4A7/PGb8NtbYMrxcMX9UDhhUJeG6iPkZmVwzOSSFBcp\nIiIy/CQ0Ey+QH/c+H3g8ueWMId2d8Iu/D8LLsR+Ej64edHiBYPzL/Cml5GQl8kcoIiIyOiTy2y/P\n3Xf2voltFyS/pDFgdzPcdxE8fz+c/vngMensvMFf3tnNS5tbqdH8LyIiMkYl0oW0y8wWuvuzAGa2\nCGhLTVmj2Nb1cP9lsH0TXPx9OO6DCd/iuTdb6Im6xr+IiMiYlUiA+RTwMzPbTLAO0iTg8pRUNVpt\nfAp+emUwr8tVv4KpJxzUbWrDEcxg4VS1wIiIyNiUyER2tWY2Gzgqtus1d+9KTVmj0Jr/gl99Cspn\nwId+CuXVB32rUH0zRx02jpL87CQWKCIiMnIMegyMmV0PFLr7S+7+ElBkZn+futJGiWgUHv8S/PJ6\nmH4SfOw3hxReunuiPFsfYbG6j0REZAxLZBDvx929pfeNu0eAjye/pFGkczf87Cr447dg0TXw4Qch\nv/SQbvnqWzvY1dmjAbwiIjKmJTIGJtPMzN0dgnlggJzUlDUK7HgLVi6DzWvgnK/CiddDEqb81wKO\nIiIiiQWYR4Cfmtldsfd/G9sn/b31UrCmUVskmJxu9rlJu3VtfYTJJXlUluYf+GQREZFRKpEA8zmC\n0PKJ2PvHgO8nvaKR7vVHg9Wkc4vh2l/D4fOSdmt3JxRu5vjq8Um7p4iIyEiUyFNIUeDO2EsG8te7\n4JEVcNgxwZNGxZOTevuGSBtvt3awWONfRERkjEtkNeqZwNeBuUDftLHuPiMFdY0sPd1BcKn9Hsw+\nDy6+G3IKk/4xoXqNfxEREYHEnkK6l6D1pRs4A/gx8F+pKGpEaW+FlZcH4eU9N8Jl96UkvEAwgd24\nvCxmHTYuJfcXEREZKRIJMPnu/lvA3L3e3b8EvP9AF5nZUjN7zczWm9mKAY5/y8yei71eN7OWuGM9\nccdWJ1Dr0Gh5E+55L2z4PZz/bTjnK5CRusUVQ+FmFk4tIzPj0J9mEhERGckSGcTbYWYZwDozuwFo\nBIre7YLYo9Z3AGcDDUCtma1297W957j7p+PO/ySwIO4Wbe4+P4Eah86mWli1LFhV+sqHYMbpKf24\nlt2dvP72Ti6Yl9xxNSIiIiNRIs0F/0Cw+vSNwCLgSuCqA1yzBFjv7hvcvRNYBVz4LucvA1YmUFN6\nvPQQ/PD9QVfRdY+nPLwA1NVHAI1/ERERgQQCjLvXuvtOd29w92vc/RJ3/0vvcTO7fYDLKoFNce8b\nYvv2YWbTgGrgd3G788wsZGZ/MbMP7Oe65bFzQk1NTYP9dg6OOzz5H8Fj0pMXwHW/g4pZqf3MmFB9\nhOxMY17Voc3kKyIiMhok0oV0ICcd4vVXAA+6e0/cvmnu3mhmM4DfmdmL7v5G/EXufjdwN0BNTY0f\nYg37190Bv/oHeH4lHHc5XHA7ZOWm7OP6C4WbOaayhPyczCH7TBERkeEqdSNOA43AlLj3VbF9A7mC\nft1H7t4Y+7oB+D17j48ZOru2wY8/EISXM/4ZLrprSMNLe1cPz2/argUcRUREYlIdYGqBmWZWbWY5\nBCFln6eJzGw2UAb8OW5fmZnlxrYnELTwrO1/7ZD4zT9DYx1c8gM47bNJWdMoES81bqezJ8qiaZrA\nTkREBJLbhbTPb3V37449sfQokAnc4+4vm9ktQMjde8PMFcCq3oUiY+YAd5lZlCBo3Rr/9NKQeu/X\nYPF1UFWTlo+vDccG8CrAiIiIAInNxHusu7/4Lqd8e6Cd7v4w8HC/fV/s9/5LA1z3NHDsYOtLqYLy\n4JUmoXAzMyoKGV80dN1WIiIiw1kiXUj/z8yeMbO/N7OS/gfd/YfJK0t6RaNO3ZsRFk/T+BcREZFe\niTxGfQrwYYJBuXVmdr+ZnZ2yygSAN5p20rK7ixot4CgiItInoUG87r4OuBn4HHAa8B0ze9XMLk5F\ncRI3/kVPIImIiPQZdIAxs+PM7FvAK8CZwPnuPie2/a0U1TfmhcLNTCjKYfr4gnSXIiIiMmwk8hTS\n7cD3gS+4e1vvTnffbGY3J70yAaC2vpmaaeXYED+6LSIiMpwNqgUmtihjo7vfFx9eern7fUmvTHi7\ntZ1NzW0a/yIiItLPoAJMbHr/KbHJ6GSIhGLjXzQDr4iIyN4S6ULaCPzJzFYDu3p3uvs3k16VAFAb\nbiY/O5O5k4vTXYqIiMiwkkiAeSP2ygDGpaYciReqb2b+lFKyM1O94oOIiMjIMugA4+5fTmUhsred\nHd2s3dzKDWccme5SREREhp1ElhKoAD4LHA3k9e539zNTUNeYt+bNCFHX/C8iIiIDSaRv4ifAq0A1\n8GUgTLDatKRAKBwhw2DB1NJ0lyIiIjLsJBJgxrv7D4Aud/+Du19LMImdpECovpk5hxczLi873aWI\niIgMO4kEmK7Y1y1m9n4zWwCofyMFunqirHmzRY9Pi4iI7EciTyH9W2wV6n8kmJW3GPh0Sqoa417Z\n0sruzh4WTdMEdiIiIgNJ5Cmk/4ltbgfOSE05AvELOCrAiIiIDCTRp5A+DkyPvy42FkaSKBRupqos\nn8NL8tNdioiIyLCUSBfSL4GngMeBntSUI+5OqD7CyUdOSHcpIiIiw1YiAabA3T+XskoEgDebd9O0\no0PdRyIiIu8ikaeQ/sfMzk1ZJQLEjX+ZpieQRERE9ieRAPMPBCGmzcxazWyHmbWmqrCxKhRupjgv\ni5kTi9JdioiIyLCVyFNIWsBxCNSGm6mZXk5GhqW7FBERkWHrgAHGzGa7+6tmtnCg4+7+bPLLGpua\nd3XyRtMuLllUle5SREREhrXBtMDcBCwH/hPwuP0We6/lBJKkrj4Y/6IZeEVERN7dAcfAuPvy2Oa5\nwP8STGTXAqyO7ZMkCYWbycnM4NjKknSXIiIiMqwl8hj1j4BW4Dux9x8CfgxcluyixqracDPHVpWQ\nl52Z7lJERESGtUQCzDHuPjfu/RNmtjbZBY1V7V09vNi4nWtPrk53KSIiIsNeIo9RP2tmJ/S+MbPj\ngVDySxqbnt/UQlePs1jzv4iIiBzQAQOMmb1oZi8Ai4CnzSxsZhuBPwM1g7h+qZm9ZmbrzWzFAMe/\nZWbPxV6vm1lL3LGrzGxd7HVVYt/ayBKKDeDVCtQiIiIHNpgupPMO9uZmlgncAZwNNAC1Zrba3fu6\nntz903HnfxJYENsuB/6VICQ5UBe7NnKw9QxnoXAzMycWUVaYk+5SREREhr0DBhh3rz+E+y8B1rv7\nBgAzWwVcCOxv7MwygtAC8F7gMXdvjl37GLAUWHkI9QxL0WiwgON5x01OdykiIiIjQiJjYA5GJbAp\n7n1DbN8+zGwaUA38LpFrzWy5mYXMLNTU1JSUoofa6+/sYEd7NzXqPhIRERmUVAeYRFwBPOjuPYlc\n5O53u3uNu9dUVFSkqLTU6l3AURPYiYiIDE6qA0wjMCXufVVs30CuYO/uoUSuHdFC4WYmjstlSnl+\nuksREREZEVIdYGqBmWZWbWY5BCFldf+TzGw2UEbwZFOvR4FzzKzMzMqAc2L7Rp1QOMLi6eWYaQFH\nERGRwUhpgHH3buAGguDxCvCAu79sZreY2QVxp14BrHJ3j7u2GfgKQQiqBW7pHdA7mmxuaaOxpY2a\n6Rr/IiIiMliJzMR7UNz9YeDhfvu+2O/9l/Zz7T3APSkrbhjonf+lRhPYiYiIDNpwGsQ7JoXCzRTk\nZDLn8HHpLkVERGTEUIBJs9pwhIVTy8jK1B+FiIjIYOm3Zhq1tnfx6lutGv8iIiKSIAWYNFrzZgvu\nmv9FREQkUQowaRQKN5OZYcyfUpruUkREREYUBZg0qg03M/fwYgpzU/4wmIiIyKiiAJMmnd1RntvU\novEvIiIiB0EBJk1e3ryd9q6oxr+IiIgcBAWYNKnrm8BOLTAiIiKJUoBJk9pwM9PGFzCxOC/dpYiI\niIw4CjBp4O6EwhEtHyAiInKQFGDSYOPWXWzb1akBvCIiIgdJASYNQuFg/MtiBRgREZGDogCTBrXh\nZsoKsjmioijdpYiIiIxICjBpUFcfYdG0csws3aWIiIiMSAowQ2zrzg42bN2l7iMREZFDoAAzxHrH\nv9RoAjsREZGDpgAzxELhZnKyMjimsjjdpYiIiIxYCjBDrLY+wvyqUnKzMtNdioiIyIilADOE2jp7\neLlxu+Z/EREROUQKMEPouU0tdEddCziKiIgcIgWYIRQKN2MGC6eqBUZERORQKMAModr6CLMmjqOk\nIDvdpYiIiIxoCjBDpCfqPFsf0fgXERGRJFCAGSKvvtXKzo5ujX8RERFJAgWYIVJX3zuBnVpgRERE\nDpUCzBCpDUc4vCSPytL8dJciIiIy4inADAF3p3ZjMzXTtYCjiIhIMqQ8wJjZUjN7zczWm9mK/Zxz\nmZmtNbOXzez+uP09ZvZc7LU61bWmSmNLG2+1tlMzTd1HIiIiyZCVypubWSZwB3A20ADUmtlqd18b\nd85M4PPASe4eMbOJcbdoc/f5qaxxKOxZwFEBRkREJBlS3QKzBFjv7hvcvRNYBVzY75yPA3e4ewTA\n3d9JcU1DrjbcTFFuFrMnaQFHERGRZEh1gKkENsW9b4jtizcLmGVmfzKzv5jZ0rhjeWYWiu3/QIpr\nTZm6+ggLp5WRmaHxLyIiIsmQ0i6kQcoCZgKnA1XAk2Z2rLu3ANPcvdHMZgC/M7MX3f2N+IvNbDmw\nHGDq1KlDW/kgbN/dxWtv7+D9xx6e7lJERERGjVS3wDQCU+LeV8X2xWsAVrt7l7tvBF4nCDS4e2Ps\n6wbg98CC/h/g7ne7e42711RUVCT/OzhEz74ZwR1qNIGdiIhI0qQ6wNQCM82s2sxygCuA/k8T/YKg\n9QUzm0DQpbTBzMrMLDdu/0nAWkaY2nAzWRnG/Cml6S5FRERk1EhpF5K7d5vZDcCjQCZwj7u/bGa3\nACF3Xx07do6ZrQV6gH9y921m9h7gLjOLEgStW+OfXhopQuEIR1eWkJ+Tme5SRERERo2Uj4Fx94eB\nh/vt+2LctgM3xV7x5zwNHJvq+lKpo7uH5xpa+OgJ09JdioiIyKiimXhT6KXGVjq7oxr/IiIikmQK\nMCkUCjcDmsBOREQk2RRgUqg2HKF6QiETinLTXYqIiMioogCTItGoU1ffrPWPREREUkABJkU2bN1J\nZHcXizX+RUREJOkUYFKkVgs4ioiIpIwCTIqEwhHGF+ZQPaEw3aWIiIiMOgowKRKqb6ZmehlmWsBR\nREQk2RRgUuCd1nbqt+2mZprGv4iIiKSCAkwKhOo1/kVERCSVFGBSoDbcTF52BkdPLkl3KSIiIqOS\nAkwK1NVHmD+llJws/XhFRERSQb9hk2xXRzcvb27V/C8iIiIppACTZM9taqEn6lrAUUREJIUUYJKs\nNtyMGSyYWpruUkREREYtBZgkC4UjzJ5UTHFedrpLERERGbUUYJKouyfKs29GWKzHp0VERFJKASaJ\nXn1rB7s7ezT+RUREJMUUYJKoNtwMoBYYERGRFFOASaJQOEJlaT6Hl+SnuxQREZFRTQEmSdyd2nCz\nlg8QEREZAgowSbKpuY13dnRo/IuIiMgQUIBJEo1/ERERGToKMEkSqo8wLi+LWRPHpbsUERGRUU8B\nJklC4WZqppWRkWHpLkVERGTUU4BJgsiuTta9s1PjX0RERIaIAkwS1NVHAKiZpvEvIiIiQ0EBJglq\n65vJzjTmTdECjiIiIkMh5QHGzJaa2Wtmtt7MVuznnMvMbK2ZvWxm98ftv8rM1sVeV6W61oNVF45w\nbGUJedmZ6S5FRERkTMhK5c3NLBO4AzgbaABqzWy1u6+NO2cm8HngJHePmNnE2P5y4F+BGsCButi1\nkVTWnKj2rh5eaNjONSdNT3cpIiIiY0aqW2CWAOvdfYO7dwKrgAv7nfNx4I7eYOLu78T2vxd4zN2b\nY8ceA5amuN6Evdi4nc6eqAbwioiIDKFUB5hKYFPc+4bYvnizgFlm9icz+4uZLU3g2rTrncBukQbw\nioiIDJmUdiENUhYwEzgdqAKeNLNjB3uxmS0HlgNMnTo1FfW9q1A4whEVhZQX5gz5Z4uIiIxVqW6B\naQSmxL2viu2L1wCsdvcud98IvE4QaAZzLe5+t7vXuHtNRUVFUos/kGjUCYWbWazuIxERkSGV6gBT\nC8w0s2ozywGuAFb3O+cXBK0vmNkEgi6lDcCjwDlmVmZmZcA5sX3DxvqmnbS2d2v8i4iIyBBLaReS\nu3eb2Q0EwSMTuMfdXzazW4CQu69mT1BZC/QA/+Tu2wDM7CsEIQjgFndvTmW9idICjiIiIumR8jEw\n7v4w8HC/fV+M23bgptir/7X3APekusaDFQpHqBiXy9TygnSXIiIiMqZoJt5DUBtbwNFMCziKiIgM\nJQWYg7RlexsNkTaNfxEREUkDBZiDFAoHEwJr/IuIiMjQU4A5SHX1EQpyMpl7eHG6SxERERlzFGAO\nUm24mQVTS8nK1I9QRERkqOm370HY0d7FK1taWTRN419ERETSQQHmIKx5s4Woa/yLiIhIuijAHIRQ\nuJkMgwVTFWBERETSQQHmIITqI8ydXExR7nBYC1NERGTsUYBJUFdPlDVvtlCj8S8iIiJpowCToLWb\nW2nr6tEK1CIiImmkAJOg3gUcazSAV0REJG0UYBIUCkeYUp7PYcV56S5FRERkzFKASYC7E6pvZrHG\nv4iIiKSVAkwC6rftZuvOTi3gKCIikmYKMAnoHf+iCexERETSSwEmAaFwhNKCbI6oKEp3KSIiImOa\nAkwCauubWTS1jIwMS3cpIiIiY5oCzCBt29nBhqZdGv8iIiIyDCjADFKoPgJo/IuIiMhwoAAzSHX1\nEXKyMji2qiTdpYiIiIx5CjCDVBtuZl5VCblZmekuRUREZMxTgBmEts4eXmrcrvEvIiIiw4QCzCA8\n39BCV49TM03jX0RERIYDBZhBCMUmsFukACMiIjIsKMAMQklBDkuPnkRpQU66SxEREREgK90FjAQf\nOWEaHzlhWrrLEBERkRi1wIiIiMiIowAjIiIiI07KA4yZLTWz18xsvZmtGOD41WbWZGbPxV7XxR3r\nidu/OtW1ioiIyMiQ0jEwZpYJ3AGcDTQAtWa22t3X9jv1p+5+wwC3aHP3+amsUUREREaeVLfALAHW\nu/sGd+8EVgEXpvgzRUREZJRLdYCpBDbFvW+I7evvEjN7wcweNLMpcfvzzCxkZn8xsw8M9AFmtjx2\nTqipqSmJpYuIiMhwNRwG8f4KmO7uxwGPAT+KOzbN3WuADwG3mdkR/S9297vdvcbdayoqKoamYhER\n+f/t3X+oX3Udx/HnqzutqbBGg7Dd1QaNav2w1Qhr0B/aH4Wx/ugPlQqK/knSrYjS+jP6IyKiViOy\npQSO/GOZSMQ01EIobPNH6rYCWUtdk7Zi1iKm03d/nDP9Tje25f3uc473+YAv95zP995zX18+3O99\nfz+fzzlHamraBcw+YHJEZbZve15V/aOqjvS7m4H3Tjy3r/+6B/gNsHqaYSVJ0jhMu4DZDqxMsiLJ\nucAVwHFnEyW5cGJ3HbC7b1+c5NX99hJgLfDixb+SJGkemupZSFV1NMnVwO3ADHBDVe1M8nVgR1Xd\nBqxPsg44CvwT+HT/428DfpTkObpC65snOHtJkiTNQ6mq1hnmzJo1a2rHjh2tY0iSpDmS5L5+Pexx\nhrCIV5Ik6YxYwEiSe9VVLQAABKZJREFUpNF5RU0hJTkA/HVKh18CHJzSsfX/s1+Gxz4ZJvtleOyT\n0/OmqnrJdVJeUQXMNCXZcaI5OLVlvwyPfTJM9svw2Ccvj1NIkiRpdCxgJEnS6FjAnL7rWwfQCdkv\nw2OfDJP9Mjz2ycvgGhhJkjQ6jsBIkqTRsYA5DUk+nOTPSR5Ncl3rPPNdkmVJ7k6yK8nOJBtaZ1In\nyUySB5L8snUWdZK8NsnWJH9KsjvJ+1tnEiT5Yv/+9UiSnyV5TetMY2MBcwpJZoBNwEeAVcCVSVa1\nTTXvHQW+VFWrgIuBz9sng7GB/oasGozvAduq6q3ARdg/zSVZCqwH1lTVO+juFXhF21TjYwFzau8D\nHq2qPVX1NHAz8LHGmea1qtpfVff32/+me0Ne2jaVkswClwGbW2dRJ8ki4IPATwCq6umqOtQ2lXoL\ngIVJFgDnAX9rnGd0LGBObSnw+MT+E/jPcjCSLAdWA/e2TSLgu8BXgOdaB9HzVgAHgBv7qb3NSc5v\nHWq+q6p9wLeBx4D9wFNVdUfbVONjAaPRSnIB8HPgC1X1r9Z55rMkHwX+XlX3tc6i4ywA3gP8sKpW\nA/8BXMfXWJLFdCP5K4A3AOcn+WTbVONjAXNq+4BlE/uzfZsaSnIOXfGypapuaZ1HrAXWJdlLN816\nSZKb2kYS3YjxE1V1bIRyK11Bo7Y+BPylqg5U1TPALcAHGmcaHQuYU9sOrEyyIsm5dAutbmucaV5L\nEro5/d1V9Z3WeQRV9dWqmq2q5XR/I3dVlZ8oG6uqJ4HHk7ylb7oU2NUwkjqPARcnOa9/P7sUF1ef\nsQWtAwxdVR1NcjVwO91K8RuqamfjWPPdWuBTwMNJHuzbvlZVv2qYSRqqa4At/QewPcBnGueZ96rq\n3iRbgfvpzqp8AK/Ke8a8Eq8kSRodp5AkSdLoWMBIkqTRsYCRJEmjYwEjSZJGxwJGkiSNjgWMpLMu\nybNJHpx4zNnVYZMsT/LIXB1P0jB5HRhJLfy3qt7dOoSk8XIERtJgJNmb5FtJHk7yhyRv7tuXJ7kr\nyUNJ7kzyxr799Ul+keSP/ePY5dhnkvw4yc4kdyRZ2H//+iS7+uPc3OhlSpoDFjCSWlj4oimkyyee\ne6qq3gn8gO4O1wDfB35aVe8CtgAb+/aNwG+r6iK6e/wcu0r2SmBTVb0dOAR8vG+/DljdH+dz03px\nkqbPK/FKOuuSHK6qC07Qvhe4pKr29DfsfLKqXpfkIHBhVT3Tt++vqiVJDgCzVXVk4hjLgV9X1cp+\n/1rgnKr6RpJtwGHgVuDWqjo85ZcqaUocgZE0NHWS7TNxZGL7WV5Y73cZsIlutGZ7EtcBSiNlASNp\naC6f+Pr7fvt3dHe5BvgEcE+/fSdwFUCSmSSLTnbQJK8CllXV3cC1wCLgJaNAksbBTx+SWlg4cSdx\ngG1VdexU6sVJHqIbRbmyb7sGuDHJl4EDvHBH5Q3A9Uk+SzfSchWw/yS/cwa4qS9yAmysqkNz9ook\nnVWugZE0GP0amDVVdbB1FknD5hSSJEkaHUdgJEnS6DgCI0mSRscCRpIkjY4FjCRJGh0LGEmSNDoW\nMJIkaXQsYCRJ0uj8DxY5j77BP8ImAAAAAElFTkSuQmCC\n","text/plain":["<Figure size 648x432 with 1 Axes>"]},"metadata":{"tags":[]}}]},{"cell_type":"markdown","metadata":{"id":"hyRuf7eO4IE2","colab_type":"text"},"source":["# Prediction\n","\n","If the prediction is >= 0.5, it is positive else it is negative."]},{"cell_type":"code","metadata":{"trusted":true,"id":"NOtkxHT_4IFB","colab_type":"code","colab":{}},"source":["def padding(vec, size):\n","  if size <= len(vec):\n","    return vec[:size]\n","  zeros = [0] * (size - len(vec))\n","  vec.extend(zeros)\n","  return vec"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"trusted":true,"id":"dEezMROZ4IFL","colab_type":"code","colab":{}},"source":["def predict(used_model, pred_text, pad=False):\n","  encoded_text = encoder.encode(pred_text)\n","\n","  if pad:\n","    encoded_text = padding(encoded_text, size=64)\n","\n","  encoded_text = tf.cast(encoded_text, tf.float32)\n","  predictions = used_model.predict(encoded_text[tf.newaxis, ...])\n","  return predictions"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"trusted":true,"id":"ZDtwzN0j4IFP","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":34},"outputId":"43473b04-6fd8-4a46-9217-29171bd871fb","executionInfo":{"status":"ok","timestamp":1581586881817,"user_tz":-480,"elapsed":625735,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["sample_text = \"The movie is overall nice, however, the story is ridiculous.\"\n","predictions = predict(model, sample_text, pad=True)\n","predictions"],"execution_count":21,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([[-0.47524095]], dtype=float32)"]},"metadata":{"tags":[]},"execution_count":21}]},{"cell_type":"markdown","metadata":{"id":"eGVDcYNd4IFT","colab_type":"text"},"source":["# Stack More LSTM Layers"]},{"cell_type":"markdown","metadata":{"id":"UK5cNNfV4IFU","colab_type":"text"},"source":["The following we are going to stack more LSTM layers to a model. Here, we stack a bidirectional LSTM layer upon the first one. However, a bidirectional LSTM layer would squeeze the timestamp dimension to return the sequence of the final layer. We have to set the `return_sequence` flag to True to keep the sequence alongside the timestamp axis. Now the shape would become [batch_size, timestamps, features]. We then pass these tensor into the next bidirectional LSTM layer."]},{"cell_type":"code","metadata":{"trusted":true,"id":"9mSzwRgG4IFV","colab_type":"code","colab":{}},"source":["def build_stack_model():\n","  def _model(inputs):\n","    embed = tf.keras.layers.Embedding(encoder.vocab_size, output_dim=64)(inputs)  # [None, None, 64]\n","    lstm1 = tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(units=64, return_sequences=True))(embed)  # [None, None, 128]\n","    lstm2 = tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(units=32))(lstm1)  # [None, 64]\n","    x = tf.keras.layers.Dense(64, activation='relu')(lstm2)  # [None, 64]\n","    x = tf.keras.layers.Dropout(0.5)(x)\n","    classifier = tf.keras.layers.Dense(1)(x)  # [None, 1]\n","    return classifier\n","\n","  inputs = tf.keras.Input(shape=(None, ))\n","  classifier = _model(inputs)\n","  return tf.keras.Model(inputs, classifier)"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"trusted":true,"id":"nLzQp9zn4IFc","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":391},"outputId":"7c900de3-7c73-4480-d6e9-c4192b217a0c","executionInfo":{"status":"ok","timestamp":1581586882792,"user_tz":-480,"elapsed":626701,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["stack_model = build_stack_model()\n","stack_model.summary()"],"execution_count":23,"outputs":[{"output_type":"stream","text":["Model: \"model_1\"\n","_________________________________________________________________\n","Layer (type)                 Output Shape              Param #   \n","=================================================================\n","input_2 (InputLayer)         [(None, None)]            0         \n","_________________________________________________________________\n","embedding_1 (Embedding)      (None, None, 64)          523840    \n","_________________________________________________________________\n","bidirectional_2 (Bidirection (None, None, 128)         66048     \n","_________________________________________________________________\n","bidirectional_3 (Bidirection (None, 64)                41216     \n","_________________________________________________________________\n","dense_2 (Dense)              (None, 64)                4160      \n","_________________________________________________________________\n","dropout (Dropout)            (None, 64)                0         \n","_________________________________________________________________\n","dense_3 (Dense)              (None, 1)                 65        \n","=================================================================\n","Total params: 635,329\n","Trainable params: 635,329\n","Non-trainable params: 0\n","_________________________________________________________________\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"trusted":true,"id":"Jj47GvA44IFg","colab_type":"code","colab":{}},"source":["stack_model.compile(loss=tf.keras.losses.BinaryCrossentropy(), \n","              optimizer=tf.keras.optimizers.Adam(1e-4), \n","              metrics=[tf.keras.metrics.BinaryAccuracy()])"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"trusted":true,"id":"F3tcir104IFn","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":357},"outputId":"0ba71c91-bb47-4f7e-ea9a-6af23101c6dc","executionInfo":{"status":"ok","timestamp":1581587681727,"user_tz":-480,"elapsed":1425629,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["history = stack_model.fit(train_dataset, epochs=10, validation_data=test_dataset, validation_steps=20)"],"execution_count":25,"outputs":[{"output_type":"stream","text":["Epoch 1/10\n","391/391 [==============================] - 83s 212ms/step - loss: 1.0341 - binary_accuracy: 0.4999 - val_loss: 0.6838 - val_binary_accuracy: 0.5664\n","Epoch 2/10\n","391/391 [==============================] - 79s 203ms/step - loss: 0.6856 - binary_accuracy: 0.6010 - val_loss: 0.5801 - val_binary_accuracy: 0.7031\n","Epoch 3/10\n","391/391 [==============================] - 79s 203ms/step - loss: 0.5855 - binary_accuracy: 0.7532 - val_loss: 0.7490 - val_binary_accuracy: 0.8023\n","Epoch 4/10\n","391/391 [==============================] - 79s 202ms/step - loss: 0.5135 - binary_accuracy: 0.7888 - val_loss: 0.7528 - val_binary_accuracy: 0.8094\n","Epoch 5/10\n","391/391 [==============================] - 79s 202ms/step - loss: 0.7070 - binary_accuracy: 0.6302 - val_loss: 0.6264 - val_binary_accuracy: 0.6820\n","Epoch 6/10\n","391/391 [==============================] - 80s 205ms/step - loss: 0.5733 - binary_accuracy: 0.7048 - val_loss: 0.5610 - val_binary_accuracy: 0.7320\n","Epoch 7/10\n","391/391 [==============================] - 80s 206ms/step - loss: 0.5137 - binary_accuracy: 0.7762 - val_loss: 0.5887 - val_binary_accuracy: 0.7523\n","Epoch 8/10\n","391/391 [==============================] - 80s 204ms/step - loss: 0.4936 - binary_accuracy: 0.7922 - val_loss: 0.6640 - val_binary_accuracy: 0.7617\n","Epoch 9/10\n","391/391 [==============================] - 80s 204ms/step - loss: 0.4636 - binary_accuracy: 0.8197 - val_loss: 0.5928 - val_binary_accuracy: 0.7750\n","Epoch 10/10\n","391/391 [==============================] - 79s 203ms/step - loss: 0.4272 - binary_accuracy: 0.8306 - val_loss: 0.7972 - val_binary_accuracy: 0.7836\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"trusted":true,"id":"O56X_--i4IFs","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":34},"outputId":"2aa015f5-f533-45f9-d0ac-cf7ac81c7881","executionInfo":{"status":"ok","timestamp":1581587683072,"user_tz":-480,"elapsed":1426967,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["sample_text = \"The movie is overall nice, however, the story is ridiculous.\"\n","predictions = predict(stack_model, sample_text, pad=True)\n","predictions"],"execution_count":26,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([[-0.02119487]], dtype=float32)"]},"metadata":{"tags":[]},"execution_count":26}]},{"cell_type":"code","metadata":{"trusted":true,"id":"ifACdTkx4IFw","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":388},"outputId":"8c6cff05-f76e-4328-fa93-be20df58d0cc","executionInfo":{"status":"ok","timestamp":1581587683074,"user_tz":-480,"elapsed":1426963,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["plot_history(history, 'loss')"],"execution_count":27,"outputs":[{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAikAAAFzCAYAAAD7bpkSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n","text/plain":["<Figure size 648x432 with 1 Axes>"]},"metadata":{"tags":[]}}]},{"cell_type":"code","metadata":{"trusted":true,"id":"08wW_7hR4IF0","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":388},"outputId":"a4a70073-5632-4e70-a341-e95bff18c66c","executionInfo":{"status":"ok","timestamp":1581587683724,"user_tz":-480,"elapsed":1427608,"user":{"displayName":"王DevOps","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mAzLB0C3whTHAdHpq24UrEWqGtbhJElQxTU5_b_4g=s64","userId":"04300517850278510646"}}},"source":["plot_history(history, 'binary_accuracy')"],"execution_count":28,"outputs":[{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAjAAAAFzCAYAAAAg1TMwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n","text/plain":["<Figure size 648x432 with 1 Axes>"]},"metadata":{"tags":[]}}]}]}