Evaluation Mode

We'll cover the following

Chapter Goals:

  • Set up the regression function’s evaluation code

A. Evaluating the model

When evaluating the model, we use mean absolute error as the metric. This is because our goal is to get the model's sales predictions as close to the actual labels as possible, which is equivalent to minimizing the mean absolute error between predictions and labels.

Since we use the same evaluation metric as the loss function, this makes the evaluation code extremely easy. We just need to return an EstimatorSpec containing the model's loss on the evaluation set.

Press + to interact
mode = tf.estimator.ModeKeys.TRAIN
estimator_spec = tf.estimator.EstimatorSpec(
    mode, loss=loss, train_op=train_op)

Time to Code!

All code for this chapter goes in the regression_fn function.

The code for this chapter focuses on model evaluation. Since we evaluate the model using mean absolute error, which is the same as the loss function, we don't need to do anything other than return the EstimatorSpec.

Outside the if block from the previous chapter, create another if block. This one should check if mode is equal to tf.estimator.ModeKeys.EVAL.

Inside the if block, return tf.estimator.EstimatorSpec initialized with mode as the required argument and loss as the loss keyword argument.

Press + to interact
class SalesModel(object):
  def __init__(self, hidden_layers):
    self.hidden_layers = hidden_layers
  def regression_fn(self, features, labels, mode, params):
    feature_columns = create_feature_columns()
    inputs = tf.compat.v1.feature_column.input_layer(features, feature_columns)
    batch_predictions = self.model_layers(inputs) 
    predictions = tf.squeeze(batch_predictions)
    if labels is not None:
      loss = tf.compat.v1.losses.absolute_difference(labels, predictions)
    if mode == tf.estimator.ModeKeys.TRAIN:
      global_step = tf.compat.v1.train.get_or_create_global_step()
      adam = tf.compat.v1.train.AdamOptimizer()
      train_op = adam.minimize(
        loss, global_step=global_step)
      return tf.estimator.EstimatorSpec(mode, loss=loss, train_op=train_op)
    # CODE HERE
    
    
    
  def model_layers(self, inputs):
    layer = inputs
    for num_nodes in self.hidden_layers:
      layer = tf.keras.layers.Dense(num_nodes,
        activation=tf.nn.relu)(layer)
    batch_predictions = tf.keras.layers.Dense(1)(layer)
    return batch_predictions

Get hands-on with 1300+ tech skills courses.