analyticsindiamagazine

analyticsindiamagazine

analyticsindiamagazine

analyticsindiamagazine

#The base image - The container is built on top of this image --# reference: https://hub.docker.com/_/ubuntu/
FROM ubuntu:18.04

# Adds metadata to the image as a key value pair
LABEL version="1.0"

# Set environment variables
ENV LANG=C.UTF-8 LC_ALL=C.UTF-8

# Create empty directory to attach volume

analyticsindiamagazine

model = tf.keras.models.Sequential([
    tf.keras.layers.InputLayer(input_shape = 26),                       
    tf.keras.layers.Dense(100, activation=tf.nn.relu),
    tf.keras.layers.Dropout(0.2),
    tf.keras.layers.Dense(1),
  ])
model.compile(
      optimizer=’adam’,
      loss='mean_squared_logarithmic_error',
      metrics=['RootMeanSquaredError'],

analyticsindiamagazine

#The hyperparameters & their values to be tested are stored in aspecial type called HParam
HP_NUM_UNITS = hp.HParam('num_units', hp.Discrete([50, 100, 200]))
HP_DROPOUT = hp.HParam('dropout', hp.RealInterval(0.1, 0.3))
HP_OPTIMIZER = hp.HParam('optimizer', hp.Discrete(['adam', 'sgd','nadam']))

#Settinf the Metric to RMSE
METRIC_RMSE = 'RootMeanSquaredError'

#Clear any logs from previous runs
!rm -rf ./logs/ 

analyticsindiamagazine

#A function that trains and validates the model and returns the rmse
def train_test_model(hparams):
  #Keras sequential model with Hyperparameters passed from the argument
  model = tf.keras.models.Sequential([
    tf.keras.layers.InputLayer(input_shape = 26),                       
    tf.keras.layers.Dense(hparams[HP_NUM_UNITS], activation=tf.nn.relu,  kernel_initializer = 'uniform'),
    tf.keras.layers.Dropout(hparams[HP_DROPOUT]),
    tf.keras.layers.Dense(1),
  ])

analyticsindiamagazine

#A function to log the training process
def run(run_dir, hparams):
  with tf.summary.create_file_writer(run_dir).as_default():
    hp.hparams(hparams)  
    rmse = train_test_model(hparams)
    tf.summary.scalar(METRIC_RMSE, rmse, step=10)

analyticsindiamagazine

# Training the model for each combination of the hyperparameters.

x_train = X_train
x_test, y_test = X_val , y_val

#A unique number for each training session
session_num = 0

#Nested for loop training with all possible  combinathon of hyperparameters
for num_units in HP_NUM_UNITS.domain.values:

analyticsindiamagazine

train_InputExamples = train.apply(lambda x: bert.run_classifier.InputExample(guid=None,
                                                                   text_a = x[DATA_COLUMN], 
                                                                   text_b = None, 
                                                                   label = x[LABEL_COLUMN]), axis = 1)
val_InputExamples = val.apply(lambda x: bert.run_classifier.InputExample(guid=None, 
                                                                   text_a = x[DATA_COLUMN], 
                                                                   text_b = None, 
                                                                   label = x[LABEL_COLUMN]), axis = 1)