analyticsindiamagazine

#Importing tensorflow 2
try:
  %tensorflow_version 2.x  #gpu
except Exception:
  pass
import tensorflow as tf

#Installing the What-If Tool widget 
try:
  !pip install --upgrade witwidget

analyticsindiamagazine

# Creates a tf feature spec from the dataframe and columns specified.
def create_feature_spec(df, columns=None):
    feature_spec = {}
    if columns == None:
        columns = df.columns.values.tolist()
    for f in columns:
        if df[f].dtype is np.dtype(np.int64):
            feature_spec[f] = tf.compat.v1.FixedLenFeature(shape=(), dtype=tf.int64)
        elif df[f].dtype is np.dtype(np.float64):
            feature_spec[f] = tf.compat.v1.FixedLenFeature(shape=(), dtype=tf.float32)

analyticsindiamagazine

# Specify the dataframe
df = train

# Setting the categorical feature to predict or classify (Target)
label_column = 'Delivery_Time'

#List the uique classes in the target column
classes = list(df[label_column].unique())

#Encode the categories in the target column using the unique classes

analyticsindiamagazine

# Convert dataset to tf.Example protos
examples = df_to_examples(df)

#Create feature spec for the data
feature_spec = create_feature_spec(df, features_and_labels)

#Creating input features from training set 
train_inpf = functools.partial(tfexamples_input_fn, examples, feature_spec, label_column)

analyticsindiamagazine

# LINEAR CLASSIFIER

# Set the number of training steps
num_steps = 1500

# A logger for logging the training data
tf.compat.v1.logging.set_verbosity(tf.compat.v1.logging.DEBUG)

# Defining the classifier
classifier = tf.estimator.LinearClassifier(

analyticsindiamagazine

#Setting the number of datapoints to choose
num_datapoints = len(val)  

# Preprocessing The validation data
make_label_column_numeric(val, label_column, classes)
test_examples = df_to_examples(val[0:num_datapoints])

analyticsindiamagazine

# Importing The WIT essentials
from witwidget.notebook.visualization import WitConfigBuilder
from witwidget.notebook.visualization import WitWidget

#Setting the WIT tool box height in pixels
tool_height_in_px = 600 

# Configuring the WIT tool with the test examples and the trained classifier
config_builder = WitConfigBuilder(test_examples[0:num_datapoints]).set_estimator_and_feature_spec(
    classifier, feature_spec).set_compare_estimator_and_feature_spec(

analyticsindiamagazine

analyticsindiamagazine

def get_masks(tokens, max_seq_length):
    """Mask for padding"""
    if len(tokens)>max_seq_length:
        #Cutting down the excess length
        tokens = tokens[0:max_seq_length]
        return [1]*len(tokens)
    else :
      return [1]*len(tokens) + [0] * (max_seq_length - len(tokens))

def get_segments(tokens, max_seq_length):

analyticsindiamagazine

input_word_ids = tf.keras.layers.Input(shape=(max_seq_length,), dtype=tf.int32,
                                       name="input_word_ids")
input_mask = tf.keras.layers.Input(shape=(max_seq_length,), dtype=tf.int32,
                                   name="input_mask")
segment_ids = tf.keras.layers.Input(shape=(max_seq_length,), dtype=tf.int32,
                                    name="segment_ids")
bert_inputs = [input_word_ids, input_mask, segment_ids]

pooled_output, _ = bert_layer(bert_inputs)