How to include normalization of features in Keras regression model?

问题

I have a data for a regression task. The independent features(X_train) are scaled with a standard scaler. Built a Keras sequential model adding hidden layers. Compiled the model. Then fitting the model with model.fit(X_train_scaled, y_train ) Then I saved the model in a .hdf5 file.

Now how to include the scaling part inside the saved model, so that the same scaling parameters can be applied to unseen test data.

#imported all the libraries for training and evaluating the model
X_train, X_test, y_train, y_test = train_test_split( X, y, test_size=0.3, random_state=42)
sc = StandardScaler()
X_train_scaled = sc.fit_transform(X_train)
X_test_scaled= sc.transform (X_test)



def build_model():
    model = keras.Sequential([layers.Dense(64, activation=tf.nn.relu,input_shape=[len(train_dataset.keys())]),
    layers.Dense(64, activation=tf.nn.relu),
    layers.Dense(1)
    ])

    optimizer = tf.keras.optimizers.RMSprop(0.001)

    model.compile(loss='mean_squared_error',
                optimizer=optimizer,
                metrics=['mean_absolute_error', 'mean_squared_error'])
    return model
model = build_model()
EPOCHS=1000
history = model.fit(X_train_scaled, y_train, epochs=EPOCHS,
                    validation_split = 0.2, verbose=0)

loss, mae, mse = model.evaluate(X_test_scaled, y_test, verbose=0)

回答1:

The standard and efficient way, as per my understanding is, to use Tensorflow Transform. It doesn't essentially mean that we should use entire TFX Pipeline if we have to use TF Transform. TF Transform can be used as a Standalone as well.

Tensorflow Transform creates a Beam Transormation Graph, which injects these Transformations as Constants in Tensorflow Graph. As these transformations are represented as Constants in the Graph, they will be consistent across Training and Serving. Advantages of that consistency across Training and Serving are

1. Eliminates Training-Serving Skew
2. Eliminates the need for having code in the Serving System, which improves the latency.

Sample Code for TF Transform is mentioned below:

Code for Importing all the Dependencies:

try:
  import tensorflow_transform as tft
  import apache_beam as beam
except ImportError:
  print('Installing TensorFlow Transform.  This will take a minute, ignore the warnings')
  !pip install -q tensorflow_transform
  print('Installing Apache Beam.  This will take a minute, ignore the warnings')
  !pip install -q apache_beam
  import tensorflow_transform as tft
  import apache_beam as beam

import tensorflow as tf
import tensorflow_transform.beam as tft_beam
from tensorflow_transform.tf_metadata import dataset_metadata
from tensorflow_transform.tf_metadata import dataset_schema

Below mentioned is the Pre-Processing function where we mention all the Transformations:

def preprocessing_fn(inputs):
  """Preprocess input columns into transformed columns."""
  # Since we are modifying some features and leaving others unchanged, we
  # start by setting `outputs` to a copy of `inputs.
  outputs = inputs.copy()

  # Scale numeric columns to have range [0, 1].
  for key in NUMERIC_FEATURE_KEYS:
    outputs[key] = tft.scale_to_0_1(outputs[key])

  for key in OPTIONAL_NUMERIC_FEATURE_KEYS:
    # This is a SparseTensor because it is optional. Here we fill in a default
    # value when it is missing.
    dense = tf.sparse_to_dense(outputs[key].indices,
                               [outputs[key].dense_shape[0], 1],
                               outputs[key].values, default_value=0.)
    # Reshaping from a batch of vectors of size 1 to a batch to scalars.
    dense = tf.squeeze(dense, axis=1)
    outputs[key] = tft.scale_to_0_1(dense)

  return outputs

In addition to

tft.scale_to_0_1

You can also use other APIs for Normalization, like

tft.scale_by_min_max, tft.scale_to_z_score

You can refer below mentioned link for the detailed information and for the Tutorial of TF Transform.

https://www.tensorflow.org/tfx/transform/get_started

https://www.tensorflow.org/tfx/tutorials/transform/census

来源：https://stackoverflow.com/questions/55430596/how-to-include-normalization-of-features-in-keras-regression-model