Preprocessing data with TensorFlow Transform

The Feature Engineering Component of TensorFlow Extended (TFX)

This example colab notebook provides a somewhat more advanced example of how TensorFlow Transform (tf.Transform) can be used to preprocess data using exactly the same code for both training a model and serving inferences in production.

TensorFlow Transform is a library for preprocessing input data for TensorFlow, including creating features that require a full pass over the training dataset. For example, using TensorFlow Transform you could:

  • Normalize an input value by using the mean and standard deviation
  • Convert strings to integers by generating a vocabulary over all of the input values
  • Convert floats to integers by assigning them to buckets, based on the observed data distribution

TensorFlow has built-in support for manipulations on a single example or a batch of examples. tf.Transform extends these capabilities to support full passes over the entire training dataset.

The output of tf.Transform is exported as a TensorFlow graph which you can use for both training and serving. Using the same graph for both training and serving can prevent skew, since the same transformations are applied in both stages.

What we're doing in this example

In this example we'll be processing a widely used dataset containing census data, and training a model to do classification. Along the way we'll be transforming the data using tf.Transform.

Install TensorFlow Transform

pip install tensorflow-transform
# This cell is only necessary because packages were installed while python was
# running. It avoids the need to restart the runtime when running in Colab.
import pkg_resources
import importlib

importlib.reload(pkg_resources)
<module 'pkg_resources' from '/tmpfs/src/tf_docs_env/lib/python3.9/site-packages/pkg_resources/__init__.py'>

Imports and globals

First import the stuff we need.

import math
import os
import pprint

import pandas as pd
import matplotlib.pyplot as plt

import tensorflow as tf
print('TF: {}'.format(tf.__version__))

import apache_beam as beam
print('Beam: {}'.format(beam.__version__))

import tensorflow_transform as tft
import tensorflow_transform.beam as tft_beam
print('Transform: {}'.format(tft.__version__))

from tfx_bsl.public import tfxio
from tfx_bsl.coders.example_coder import RecordBatchToExamples
TF: 2.8.2
Beam: 2.39.0
Transform: 1.8.0

Next download the data files:

!wget https://storage.googleapis.com/artifacts.tfx-oss-public.appspot.com/datasets/census/adult.data
!wget https://storage.googleapis.com/artifacts.tfx-oss-public.appspot.com/datasets/census/adult.test

train_path = './adult.data'
test_path = './adult.test'
--2022-06-01 09:13:39--  https://storage.googleapis.com/artifacts.tfx-oss-public.appspot.com/datasets/census/adult.data
Resolving storage.googleapis.com (storage.googleapis.com)... 142.251.6.128, 142.250.1.128, 172.217.214.128, ...
Connecting to storage.googleapis.com (storage.googleapis.com)|142.251.6.128|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 3974305 (3.8M) [application/octet-stream]
Saving to: ‘adult.data’

adult.data          100%[===================>]   3.79M  --.-KB/s    in 0.02s   

2022-06-01 09:13:40 (251 MB/s) - ‘adult.data’ saved [3974305/3974305]

--2022-06-01 09:13:40--  https://storage.googleapis.com/artifacts.tfx-oss-public.appspot.com/datasets/census/adult.test
Resolving storage.googleapis.com (storage.googleapis.com)... 142.251.6.128, 142.250.1.128, 172.217.214.128, ...
Connecting to storage.googleapis.com (storage.googleapis.com)|142.251.6.128|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 2003153 (1.9M) [application/octet-stream]
Saving to: ‘adult.test’

adult.test          100%[===================>]   1.91M  --.-KB/s    in 0.01s   

2022-06-01 09:13:40 (152 MB/s) - ‘adult.test’ saved [2003153/2003153]

Name our columns

We'll create some handy lists for referencing the columns in our dataset.

CATEGORICAL_FEATURE_KEYS = [
    'workclass',
    'education',
    'marital-status',
    'occupation',
    'relationship',
    'race',
    'sex',
    'native-country',
]

NUMERIC_FEATURE_KEYS = [
    'age',
    'capital-gain',
    'capital-loss',
    'hours-per-week',
    'education-num'
]

ORDERED_CSV_COLUMNS = [
    'age', 'workclass', 'fnlwgt', 'education', 'education-num',
    'marital-status', 'occupation', 'relationship', 'race', 'sex',
    'capital-gain', 'capital-loss', 'hours-per-week', 'native-country', 'label'
]

LABEL_KEY = 'label'

Here's a quick preview of the data:

pandas_train = pd.read_csv(train_path, header=None, names=ORDERED_CSV_COLUMNS)

pandas_train.head(5)
one_row = dict(pandas_train.loc[0])
COLUMN_DEFAULTS = [
  '' if isinstance(v, str) else 0.0
  for v in  dict(pandas_train.loc[1]).values()]

The test data has 1 header line that needs to be skipped, and a trailing "." at the end of each line.

pandas_test = pd.read_csv(test_path, header=1, names=ORDERED_CSV_COLUMNS)

pandas_test.head(5)
testing = os.getenv("WEB_TEST_BROWSER", False)
if testing:
  pandas_train = pandas_train.loc[:1]
  pandas_test = pandas_test.loc[:1]

Define our features and schema

Let's define a schema based on what types the columns are in our input. Among other things this will help with importing them correctly.

RAW_DATA_FEATURE_SPEC = dict(
    [(name, tf.io.FixedLenFeature([], tf.string))
     for name in CATEGORICAL_FEATURE_KEYS] +
    [(name, tf.io.FixedLenFeature([], tf.float32))
     for name in NUMERIC_FEATURE_KEYS] + 
    [(LABEL_KEY, tf.io.FixedLenFeature([], tf.string))]
)

SCHEMA = tft.tf_metadata.dataset_metadata.DatasetMetadata(
    tft.tf_metadata.schema_utils.schema_from_feature_spec(RAW_DATA_FEATURE_SPEC)).schema

[Optional] Encode and decode tf.train.Example protos

This tutorial needs to convert examples from the dataset to and from tf.train.Example protos in a few places.

The hidden encode_example function below converts a dictionary of features forom the dataset to a tf.train.Example.

Now you can convert dataset examples into Example protos:

tf_example = encode_example(pandas_train.loc[0])
tf_example.features.feature['age']
float_list {
  value: 39.0
}
serialized_example_batch = tf.constant([
  encode_example(pandas_train.loc[i]).SerializeToString()
  for i in range(3)
])

serialized_example_batch
<tf.Tensor: shape=(3,), dtype=string, numpy=
array([b'\n\xf9\x02\n\x1a\n\tworkclass\x12\r\n\x0b\n\tState-gov\n\x18\n\x0ccapital-gain\x12\x08\x12\x06\n\x04\x00\xe0\x07E\n#\n\x0emarital-status\x12\x11\n\x0f\n\rNever-married\n\x1a\n\x0ehours-per-week\x12\x08\x12\x06\n\x04\x00\x00 B\n\x1e\n\noccupation\x12\x10\n\x0e\n\x0cAdm-clerical\n\x0f\n\x03sex\x12\x08\n\x06\n\x04Male\n\x11\n\x04race\x12\t\n\x07\n\x05White\n\x0f\n\x03age\x12\x08\x12\x06\n\x04\x00\x00\x1cB\n\x1a\n\teducation\x12\r\n\x0b\n\tBachelors\n#\n\x0enative-country\x12\x11\n\x0f\n\rUnited-States\n!\n\x0crelationship\x12\x11\n\x0f\n\rNot-in-family\n\x19\n\reducation-num\x12\x08\x12\x06\n\x04\x00\x00PA\n\x12\n\x05label\x12\t\n\x07\n\x05<=50K\n\x18\n\x0ccapital-loss\x12\x08\x12\x06\n\x04\x00\x00\x00\x00',
       b'\n\x82\x03\n\x12\n\x05label\x12\t\n\x07\n\x05<=50K\n!\n\tworkclass\x12\x14\n\x12\n\x10Self-emp-not-inc\n\x18\n\x0ccapital-loss\x12\x08\x12\x06\n\x04\x00\x00\x00\x00\n\x19\n\reducation-num\x12\x08\x12\x06\n\x04\x00\x00PA\n(\n\x0emarital-status\x12\x16\n\x14\n\x12Married-civ-spouse\n\x18\n\x0ccapital-gain\x12\x08\x12\x06\n\x04\x00\x00\x00\x00\n\x0f\n\x03sex\x12\x08\n\x06\n\x04Male\n\x0f\n\x03age\x12\x08\x12\x06\n\x04\x00\x00HB\n#\n\x0enative-country\x12\x11\n\x0f\n\rUnited-States\n\x1b\n\x0crelationship\x12\x0b\n\t\n\x07Husband\n\x1a\n\teducation\x12\r\n\x0b\n\tBachelors\n\x11\n\x04race\x12\t\n\x07\n\x05White\n\x1a\n\x0ehours-per-week\x12\x08\x12\x06\n\x04\x00\x00PA\n!\n\noccupation\x12\x13\n\x11\n\x0fExec-managerial',
       b'\n\xf5\x02\n\x0f\n\x03sex\x12\x08\n\x06\n\x04Male\n#\n\x0enative-country\x12\x11\n\x0f\n\rUnited-States\n!\n\x0crelationship\x12\x11\n\x0f\n\rNot-in-family\n\x18\n\x0ccapital-loss\x12\x08\x12\x06\n\x04\x00\x00\x00\x00\n\x11\n\x04race\x12\t\n\x07\n\x05White\n\x1e\n\x0emarital-status\x12\x0c\n\n\n\x08Divorced\n#\n\noccupation\x12\x15\n\x13\n\x11Handlers-cleaners\n\x0f\n\x03age\x12\x08\x12\x06\n\x04\x00\x00\x18B\n\x18\n\x0ccapital-gain\x12\x08\x12\x06\n\x04\x00\x00\x00\x00\n\x18\n\teducation\x12\x0b\n\t\n\x07HS-grad\n\x12\n\x05label\x12\t\n\x07\n\x05<=50K\n\x19\n\reducation-num\x12\x08\x12\x06\n\x04\x00\x00\x10A\n\x18\n\tworkclass\x12\x0b\n\t\n\x07Private\n\x1a\n\x0ehours-per-week\x12\x08\x12\x06\n\x04\x00\x00 B'],
      dtype=object)>

You can also convert batches of serialized Example protos back into a dictionary of tensors:

decoded_tensors = tf.io.parse_example(
    serialized_example_batch,
    features=RAW_DATA_FEATURE_SPEC
)

In some cases the label will not be passed in, so the encode function is written so that the label is optional:

features_dict = dict(pandas_train.loc[0])
features_dict.pop(LABEL_KEY)

LABEL_KEY in features_dict
False

When creating an Example proto it will simply not contain the label key.

no_label_example = encode_example(features_dict)

LABEL_KEY in no_label_example.features.feature.keys()
False

Setting hyperparameters and basic housekeeping

Constants and hyperparameters used for training.

NUM_OOV_BUCKETS = 1

EPOCH_SPLITS = 10
TRAIN_NUM_EPOCHS = 2*EPOCH_SPLITS
NUM_TRAIN_INSTANCES = len(pandas_train)
NUM_TEST_INSTANCES = len(pandas_test)

BATCH_SIZE = 128

STEPS_PER_TRAIN_EPOCH = tf.math.ceil(NUM_TRAIN_INSTANCES/BATCH_SIZE/EPOCH_SPLITS)
EVALUATION_STEPS = tf.math.ceil(NUM_TEST_INSTANCES/BATCH_SIZE)

# Names of temp files
TRANSFORMED_TRAIN_DATA_FILEBASE = 'train_transformed'
TRANSFORMED_TEST_DATA_FILEBASE = 'test_transformed'
EXPORTED_MODEL_DIR = 'exported_model_dir'
if testing:
  TRAIN_NUM_EPOCHS = 1

Preprocessing with tf.Transform

Create a tf.Transform preprocessing_fn

The preprocessing function is the most important concept of tf.Transform. A preprocessing function is where the transformation of the dataset really happens. It accepts and returns a dictionary of tensors, where a tensor means a Tensor or SparseTensor. There are two main groups of API calls that typically form the heart of a preprocessing function:

  1. TensorFlow Ops: Any function that accepts and returns tensors, which usually means TensorFlow ops. These add TensorFlow operations to the graph that transforms raw data into transformed data one feature vector at a time. These will run for every example, during both training and serving.
  2. Tensorflow Transform Analyzers/Mappers: Any of the analyzers/mappers provided by tf.Transform. These also accept and return tensors, and typically contain a combination of Tensorflow ops and Beam computation, but unlike TensorFlow ops they only run in the Beam pipeline during analysis requiring a full pass over the entire training dataset. The Beam computation runs only once, (prior to training, during analysis), and typically make a full pass over the entire training dataset. They create tf.constant tensors, which are added to your graph. For example, tft.min computes the minimum of a tensor over the training dataset.

Here is a preprocessing_fn for this dataset. It does several things:

  1. Using tft.scale_to_0_1, it scales the numeric features to the [0,1] range.
  2. Using tft.compute_and_apply_vocabulary, it computes a vocabulary for each of the categorical features, and returns the integer IDs for each input as an tf.int64. This applies both to string and integer categorical-inputs.
  3. It applies some manual transformations to the data using standard TensorFlow operations. Here these operations are applied to the label but could transform the features as well. The TensorFlow operations do several things:
    • They build a lookup table for the label (the tf.init_scope ensures that the table is only created the first time the function is called).
    • They normalize the text of the label.
    • They convert the label to a one-hot.
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(inputs[key])

  # For all categorical columns except the label column, we generate a
  # vocabulary but do not modify the feature.  This vocabulary is instead
  # used in the trainer, by means of a feature column, to convert the feature
  # from a string to an integer id.
  for key in CATEGORICAL_FEATURE_KEYS:
    outputs[key] = tft.compute_and_apply_vocabulary(
        tf.strings.strip(inputs[key]),
        num_oov_buckets=NUM_OOV_BUCKETS,
        vocab_filename=key)

  # For the label column we provide the mapping from string to index.
  table_keys = ['>50K', '<=50K']
  with tf.init_scope():
    initializer = tf.lookup.KeyValueTensorInitializer(
        keys=table_keys,
        values=tf.cast(tf.range(len(table_keys)), tf.int64),
        key_dtype=tf.string,
        value_dtype=tf.int64)
    table = tf.lookup.StaticHashTable(initializer, default_value=-1)

  # Remove trailing periods for test data when the data is read with tf.data.
  # label_str  = tf.sparse.to_dense(inputs[LABEL_KEY])
  label_str = inputs[LABEL_KEY]
  label_str = tf.strings.regex_replace(label_str, r'\.$', '')
  label_str = tf.strings.strip(label_str)
  data_labels = table.lookup(label_str)
  transformed_label = tf.one_hot(
      indices=data_labels, depth=len(table_keys), on_value=1.0, off_value=0.0)
  outputs[LABEL_KEY] = tf.reshape(transformed_label, [-1, len(table_keys)])

  return outputs

Syntax

You're almost ready to put everything together and use Apache Beam to run it.

Apache Beam uses a special syntax to define and invoke transforms. For example, in this line:

result = pass_this | 'name this step' >> to_this_call

The method to_this_call is being invoked and passed the object called pass_this, and this operation will be referred to as name this step in a stack trace. The result of the call to to_this_call is returned in result. You will often see stages of a pipeline chained together like this:

result = apache_beam.Pipeline() | 'first step' >> do_this_first() | 'second step' >> do_this_last()

and since that started with a new pipeline, you can continue like this:

next_result = result | 'doing more stuff' >> another_function()

Transform the data

Now we're ready to start transforming our data in an Apache Beam pipeline.

  1. Read in the data using the tfxio.CsvTFXIO CSV reader (to process lines of text in a pipeline use tfxio.BeamRecordCsvTFXIO instead).
  2. Analyse and transform the data using the preprocessing_fn defined above.
  3. Write out the result as a TFRecord of Example protos, which we will use for training a model later
def transform_data(train_data_file, test_data_file, working_dir):
  """Transform the data and write out as a TFRecord of Example protos.

  Read in the data using the CSV reader, and transform it using a
  preprocessing pipeline that scales numeric data and converts categorical data
  from strings to int64 values indices, by creating a vocabulary for each
  category.

  Args:
    train_data_file: File containing training data
    test_data_file: File containing test data
    working_dir: Directory to write transformed data and metadata to
  """

  # The "with" block will create a pipeline, and run that pipeline at the exit
  # of the block.
  with beam.Pipeline() as pipeline:
    with tft_beam.Context(temp_dir=tempfile.mkdtemp()):
      # Create a TFXIO to read the census data with the schema. To do this we
      # need to list all columns in order since the schema doesn't specify the
      # order of columns in the csv.
      # We first read CSV files and use BeamRecordCsvTFXIO whose .BeamSource()
      # accepts a PCollection[bytes] because we need to patch the records first
      # (see "FixCommasTrainData" below). Otherwise, tfxio.CsvTFXIO can be used
      # to both read the CSV files and parse them to TFT inputs:
      # csv_tfxio = tfxio.CsvTFXIO(...)
      # raw_data = (pipeline | 'ToRecordBatches' >> csv_tfxio.BeamSource())
      train_csv_tfxio = tfxio.CsvTFXIO(
          file_pattern=train_data_file,
          telemetry_descriptors=[],
          column_names=ORDERED_CSV_COLUMNS,
          schema=SCHEMA)

      # Read in raw data and convert using CSV TFXIO.
      raw_data = (
          pipeline |
          'ReadTrainCsv' >> train_csv_tfxio.BeamSource())

      # Combine data and schema into a dataset tuple.  Note that we already used
      # the schema to read the CSV data, but we also need it to interpret
      # raw_data.
      cfg = train_csv_tfxio.TensorAdapterConfig()
      raw_dataset = (raw_data, cfg)

      # The TFXIO output format is chosen for improved performance.
      transformed_dataset, transform_fn = (
          raw_dataset | tft_beam.AnalyzeAndTransformDataset(
              preprocessing_fn, output_record_batches=True))

      # Transformed metadata is not necessary for encoding.
      transformed_data, _ = transformed_dataset

      # Extract transformed RecordBatches, encode and write them to the given
      # directory.
      # TODO(b/223384488): Switch to `RecordBatchToExamplesEncoder`.
      _ = (
          transformed_data
          | 'EncodeTrainData' >>
          beam.FlatMapTuple(lambda batch, _: RecordBatchToExamples(batch))
          | 'WriteTrainData' >> beam.io.WriteToTFRecord(
              os.path.join(working_dir, TRANSFORMED_TRAIN_DATA_FILEBASE)))

      # Now apply transform function to test data.  In this case we remove the
      # trailing period at the end of each line, and also ignore the header line
      # that is present in the test data file.
      test_csv_tfxio = tfxio.CsvTFXIO(
          file_pattern=test_data_file,
          skip_header_lines=1,
          telemetry_descriptors=[],
          column_names=ORDERED_CSV_COLUMNS,
          schema=SCHEMA)
      raw_test_data = (
          pipeline
          | 'ReadTestCsv' >> test_csv_tfxio.BeamSource())

      raw_test_dataset = (raw_test_data, test_csv_tfxio.TensorAdapterConfig())

      # The TFXIO output format is chosen for improved performance.
      transformed_test_dataset = (
          (raw_test_dataset, transform_fn)
          | tft_beam.TransformDataset(output_record_batches=True))

      # Transformed metadata is not necessary for encoding.
      transformed_test_data, _ = transformed_test_dataset

      # Extract transformed RecordBatches, encode and write them to the given
      # directory.
      _ = (
          transformed_test_data
          | 'EncodeTestData' >>
          beam.FlatMapTuple(lambda batch, _: RecordBatchToExamples(batch))
          | 'WriteTestData' >> beam.io.WriteToTFRecord(
              os.path.join(working_dir, TRANSFORMED_TEST_DATA_FILEBASE)))

      # Will write a SavedModel and metadata to working_dir, which can then
      # be read by the tft.TFTransformOutput class.
      _ = (
          transform_fn
          | 'WriteTransformFn' >> tft_beam.WriteTransformFn(working_dir))

Run the pipeline:

import tempfile
import pathlib

output_dir = os.path.join(tempfile.mkdtemp(), 'keras')


transform_data(train_path, test_path, output_dir)
WARNING:apache_beam.runners.interactive.interactive_environment:Dependencies required for Interactive Beam PCollection visualization are not available, please use: `pip install apache-beam[interactive]` to install necessary dependencies to enable all data visualization features.
WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.9/site-packages/tensorflow_transform/tf_utils.py:326: Tensor.experimental_ref (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use ref() instead.
WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.9/site-packages/tensorflow_transform/tf_utils.py:326: Tensor.experimental_ref (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use ref() instead.
WARNING:root:Make sure that locally built Python SDK docker image has Python 3.9 interpreter.
2022-06-01 09:13:50.864347: W tensorflow/python/util/util.cc:368] Sets are not currently considered sequences, but this may change in the future, so consider avoiding using them.
INFO:tensorflow:Assets written to: /tmpfs/tmp/tmp__tuyp6f/tftransform_tmp/c2c22fd1a2a94acf84732ff009f34d08/assets
INFO:tensorflow:Assets written to: /tmpfs/tmp/tmp__tuyp6f/tftransform_tmp/c2c22fd1a2a94acf84732ff009f34d08/assets
INFO:tensorflow:struct2tensor is not available.
INFO:tensorflow:struct2tensor is not available.
INFO:tensorflow:tensorflow_decision_forests is not available.
INFO:tensorflow:tensorflow_decision_forests is not available.
INFO:tensorflow:tensorflow_text is not available.
INFO:tensorflow:tensorflow_text is not available.
INFO:tensorflow:Assets written to: /tmpfs/tmp/tmp__tuyp6f/tftransform_tmp/3faf274195414d2fb6d2e981b9e0f7b8/assets
INFO:tensorflow:Assets written to: /tmpfs/tmp/tmp__tuyp6f/tftransform_tmp/3faf274195414d2fb6d2e981b9e0f7b8/assets
INFO:tensorflow:struct2tensor is not available.
INFO:tensorflow:struct2tensor is not available.
INFO:tensorflow:tensorflow_decision_forests is not available.
INFO:tensorflow:tensorflow_decision_forests is not available.
INFO:tensorflow:tensorflow_text is not available.
INFO:tensorflow:tensorflow_text is not available.
WARNING:apache_beam.io.tfrecordio:Couldn't find python-snappy so the implementation of _TFRecordUtil._masked_crc32c is not as fast as it could be.
INFO:tensorflow:struct2tensor is not available.
INFO:tensorflow:struct2tensor is not available.
INFO:tensorflow:tensorflow_decision_forests is not available.
INFO:tensorflow:tensorflow_decision_forests is not available.
INFO:tensorflow:tensorflow_text is not available.
INFO:tensorflow:tensorflow_text is not available.

Wrap up the output directory as a tft.TFTransformOutput:

tf_transform_output = tft.TFTransformOutput(output_dir)
tf_transform_output.transformed_feature_spec()
{'age': FixedLenFeature(shape=[], dtype=tf.float32, default_value=None),
 'capital-gain': FixedLenFeature(shape=[], dtype=tf.float32, default_value=None),
 'capital-loss': FixedLenFeature(shape=[], dtype=tf.float32, default_value=None),
 'education': FixedLenFeature(shape=[], dtype=tf.int64, default_value=None),
 'education-num': FixedLenFeature(shape=[], dtype=tf.float32, default_value=None),
 'hours-per-week': FixedLenFeature(shape=[], dtype=tf.float32, default_value=None),
 'label': FixedLenFeature(shape=[2], dtype=tf.float32, default_value=None),
 'marital-status': FixedLenFeature(shape=[], dtype=tf.int64, default_value=None),
 'native-country': FixedLenFeature(shape=[], dtype=tf.int64, default_value=None),
 'occupation': FixedLenFeature(shape=[], dtype=tf.int64, default_value=None),
 'race': FixedLenFeature(shape=[], dtype=tf.int64, default_value=None),
 'relationship': FixedLenFeature(shape=[], dtype=tf.int64, default_value=None),
 'sex': FixedLenFeature(shape=[], dtype=tf.int64, default_value=None),
 'workclass': FixedLenFeature(shape=[], dtype=tf.int64, default_value=None)}

If you look in the directory you'll see it contains three things:

  1. The train_transformed and test_transformed data files
  2. The transform_fn directory (a tf.saved_model)
  3. The transformed_metadata

The followning sections show how to use these artifacts to train a model.

ls -l {output_dir}
total 15704
-rw-rw-r-- 1 kbuilder kbuilder  5356449 Jun  1 09:14 test_transformed-00000-of-00001
-rw-rw-r-- 1 kbuilder kbuilder 10712569 Jun  1 09:14 train_transformed-00000-of-00001
drwxr-xr-x 4 kbuilder kbuilder     4096 Jun  1 09:13 transform_fn
drwxr-xr-x 2 kbuilder kbuilder     4096 Jun  1 09:13 transformed_metadata

Using our preprocessed data to train a model using tf.keras

To show how tf.Transform enables us to use the same code for both training and serving, and thus prevent skew, we're going to train a model. To train our model and prepare our trained model for production we need to create input functions. The main difference between our training input function and our serving input function is that training data contains the labels, and production data does not. The arguments and returns are also somewhat different.

Create an input function for training

Running the pipeline in the previous section created TFRecord files containing the the transformed data.

The following code uses tf.data.experimental.make_batched_features_dataset and tft.TFTransformOutput.transformed_feature_spec to read these data files as a tf.data.Dataset:

def _make_training_input_fn(tf_transform_output, train_file_pattern,
                            batch_size):
  """An input function reading from transformed data, converting to model input.

  Args:
    tf_transform_output: Wrapper around output of tf.Transform.
    transformed_examples: Base filename of examples.
    batch_size: Batch size.

  Returns:
    The input data for training or eval, in the form of k.
  """
  def input_fn():
    return tf.data.experimental.make_batched_features_dataset(
        file_pattern=train_file_pattern,
        batch_size=batch_size,
        features=tf_transform_output.transformed_feature_spec(),
        reader=tf.data.TFRecordDataset,
        label_key=LABEL_KEY,
        shuffle=True)

  return input_fn
train_file_pattern = pathlib.Path(output_dir)/f'{TRANSFORMED_TRAIN_DATA_FILEBASE}*'

input_fn = _make_training_input_fn(
    tf_transform_output=tf_transform_output,
    train_file_pattern = str(train_file_pattern),
    batch_size = 10
)

Below you can see a transformed sample of the data. Note how the numeric columns like education-num and hourd-per-week are converted to floats with a range of [0,1], and the string columns have been converted to IDs:

for example, label in input_fn().take(1):
  break

pd.DataFrame(example)
label
<tf.Tensor: shape=(10, 2), dtype=float32, numpy=
array([[0., 1.],
       [0., 1.],
       [0., 1.],
       [1., 0.],
       [1., 0.],
       [0., 1.],
       [0., 1.],
       [0., 1.],
       [0., 1.],
       [1., 0.]], dtype=float32)>

Train, Evaluate the model

Build the model

def build_keras_model(working_dir):
  inputs = build_keras_inputs(working_dir)

  encoded_inputs = encode_inputs(inputs)

  stacked_inputs = tf.concat(tf.nest.flatten(encoded_inputs), axis=1)
  output = tf.keras.layers.Dense(100, activation='relu')(stacked_inputs)
  output = tf.keras.layers.Dense(50, activation='relu')(output)
  output = tf.keras.layers.Dense(2)(output)
  model = tf.keras.Model(inputs=inputs, outputs=output)

  return model
def build_keras_inputs(working_dir):
  tf_transform_output = tft.TFTransformOutput(working_dir)

  feature_spec = tf_transform_output.transformed_feature_spec().copy()
  feature_spec.pop(LABEL_KEY)

  # Build the `keras.Input` objects.
  inputs = {}
  for key, spec in feature_spec.items():
    if isinstance(spec, tf.io.VarLenFeature):
      inputs[key] = tf.keras.layers.Input(
          shape=[None], name=key, dtype=spec.dtype, sparse=True)
    elif isinstance(spec, tf.io.FixedLenFeature):
      inputs[key] = tf.keras.layers.Input(
          shape=spec.shape, name=key, dtype=spec.dtype)
    else:
      raise ValueError('Spec type is not supported: ', key, spec)

  return inputs
def encode_inputs(inputs):
  encoded_inputs = {}
  for key in inputs:
    feature = tf.expand_dims(inputs[key], -1)
    if key in CATEGORICAL_FEATURE_KEYS:
      num_buckets = tf_transform_output.num_buckets_for_transformed_feature(key)
      encoding_layer = (
          tf.keras.layers.CategoryEncoding(
              num_tokens=num_buckets, output_mode='binary', sparse=False))
      encoded_inputs[key] = encoding_layer(feature)
    else:
      encoded_inputs[key] = feature

  return encoded_inputs
model = build_keras_model(output_dir)

tf.keras.utils.plot_model(model,rankdir='LR', show_shapes=True)
You must install pydot (`pip install pydot`) and install graphviz (see instructions at https://graphviz.gitlab.io/download/) for plot_model/model_to_dot to work.

Build the datasets

def get_dataset(working_dir, filebase):
  tf_transform_output = tft.TFTransformOutput(working_dir)

  data_path_pattern = os.path.join(
      working_dir,
      filebase + '*')

  input_fn = _make_training_input_fn(
      tf_transform_output,
      data_path_pattern,
      batch_size=BATCH_SIZE)

  dataset = input_fn()

  return dataset

Train and evaluate the model:

def train_and_evaluate(
    model,
    working_dir):
  """Train the model on training data and evaluate on test data.

  Args:
    working_dir: The location of the Transform output.
    num_train_instances: Number of instances in train set
    num_test_instances: Number of instances in test set

  Returns:
    The results from the estimator's 'evaluate' method
  """
  train_dataset = get_dataset(working_dir, TRANSFORMED_TRAIN_DATA_FILEBASE)
  validation_dataset = get_dataset(working_dir, TRANSFORMED_TEST_DATA_FILEBASE)

  model = build_keras_model(working_dir)

  history = train_model(model, train_dataset, validation_dataset)

  metric_values = model.evaluate(validation_dataset,
                                 steps=EVALUATION_STEPS,
                                 return_dict=True)
  return model, history, metric_values
def train_model(model, train_dataset, validation_dataset):
  model.compile(optimizer='adam',
                loss=tf.losses.CategoricalCrossentropy(from_logits=True),
                metrics=['accuracy'])

  history = model.fit(train_dataset, validation_data=validation_dataset,
      epochs=TRAIN_NUM_EPOCHS,
      steps_per_epoch=STEPS_PER_TRAIN_EPOCH,
      validation_steps=EVALUATION_STEPS)
  return history
model, history, metric_values = train_and_evaluate(model, output_dir)
Epoch 1/20
26/26 [==============================] - 2s 39ms/step - loss: 0.5241 - accuracy: 0.7536 - val_loss: 0.4254 - val_accuracy: 0.7982
Epoch 2/20
26/26 [==============================] - 1s 27ms/step - loss: 0.3884 - accuracy: 0.8251 - val_loss: 0.3712 - val_accuracy: 0.8301
Epoch 3/20
26/26 [==============================] - 1s 27ms/step - loss: 0.3566 - accuracy: 0.8338 - val_loss: 0.3576 - val_accuracy: 0.8328
Epoch 4/20
26/26 [==============================] - 1s 27ms/step - loss: 0.3654 - accuracy: 0.8272 - val_loss: 0.3538 - val_accuracy: 0.8359
Epoch 5/20
26/26 [==============================] - 1s 27ms/step - loss: 0.3605 - accuracy: 0.8281 - val_loss: 0.3492 - val_accuracy: 0.8371
Epoch 6/20
26/26 [==============================] - 1s 27ms/step - loss: 0.3399 - accuracy: 0.8401 - val_loss: 0.3458 - val_accuracy: 0.8393
Epoch 7/20
26/26 [==============================] - 1s 27ms/step - loss: 0.3309 - accuracy: 0.8459 - val_loss: 0.3435 - val_accuracy: 0.8389
Epoch 8/20
26/26 [==============================] - 1s 27ms/step - loss: 0.3472 - accuracy: 0.8326 - val_loss: 0.3454 - val_accuracy: 0.8362
Epoch 9/20
26/26 [==============================] - 1s 27ms/step - loss: 0.3525 - accuracy: 0.8359 - val_loss: 0.3436 - val_accuracy: 0.8383
Epoch 10/20
26/26 [==============================] - 1s 27ms/step - loss: 0.3617 - accuracy: 0.8320 - val_loss: 0.3395 - val_accuracy: 0.8422
Epoch 11/20
26/26 [==============================] - 1s 27ms/step - loss: 0.3408 - accuracy: 0.8407 - val_loss: 0.3393 - val_accuracy: 0.8422
Epoch 12/20
26/26 [==============================] - 1s 27ms/step - loss: 0.3439 - accuracy: 0.8350 - val_loss: 0.3353 - val_accuracy: 0.8420
Epoch 13/20
26/26 [==============================] - 1s 27ms/step - loss: 0.3418 - accuracy: 0.8347 - val_loss: 0.3347 - val_accuracy: 0.8425
Epoch 14/20
26/26 [==============================] - 1s 27ms/step - loss: 0.3327 - accuracy: 0.8483 - val_loss: 0.3347 - val_accuracy: 0.8426
Epoch 15/20
26/26 [==============================] - 1s 27ms/step - loss: 0.3177 - accuracy: 0.8507 - val_loss: 0.3317 - val_accuracy: 0.8450
Epoch 16/20
26/26 [==============================] - 1s 27ms/step - loss: 0.3198 - accuracy: 0.8480 - val_loss: 0.3338 - val_accuracy: 0.8449
Epoch 17/20
26/26 [==============================] - 1s 26ms/step - loss: 0.3316 - accuracy: 0.8434 - val_loss: 0.3320 - val_accuracy: 0.8442
Epoch 18/20
26/26 [==============================] - 1s 27ms/step - loss: 0.3194 - accuracy: 0.8573 - val_loss: 0.3321 - val_accuracy: 0.8446
Epoch 19/20
26/26 [==============================] - 1s 26ms/step - loss: 0.3385 - accuracy: 0.8389 - val_loss: 0.3321 - val_accuracy: 0.8440
Epoch 20/20
26/26 [==============================] - 1s 26ms/step - loss: 0.3359 - accuracy: 0.8410 - val_loss: 0.3306 - val_accuracy: 0.8443
128/128 [==============================] - 1s 4ms/step - loss: 0.3300 - accuracy: 0.8447
plt.plot(history.history['loss'], label='Train')
plt.plot(history.history['val_loss'], label='Eval')
plt.ylim(0,max(plt.ylim()))
plt.legend()
plt.title('Loss');

png

Transform new data

In the previous section the training process used the hard-copies of the transformed data that were generated by tft_beam.AnalyzeAndTransformDataset in the transform_dataset function.

For operating on new data you'll need to load final version of the preprocessing_fn that was saved by tft_beam.WriteTransformFn.

The TFTransformOutput.transform_features_layer method loads the preprocessing_fn SavedModel from the output directory.

Here's a function to load new, unprocessed batches from a source file:

def read_csv(file_name, batch_size):
  return tf.data.experimental.make_csv_dataset(
        file_pattern=file_name,
        batch_size=batch_size,
        column_names=ORDERED_CSV_COLUMNS,
        column_defaults=COLUMN_DEFAULTS,
        prefetch_buffer_size=0,
        ignore_errors=True)
for ex in read_csv(test_path, batch_size=5):
  break

pd.DataFrame(ex)

Load the tft.TransformFeaturesLayer to transform this data with the preprocessing_fn:

ex2 = ex.copy()
ex2.pop('fnlwgt')

tft_layer = tf_transform_output.transform_features_layer()
t_ex = tft_layer(ex2)

label = t_ex.pop(LABEL_KEY)
pd.DataFrame(t_ex)
INFO:tensorflow:struct2tensor is not available.
INFO:tensorflow:struct2tensor is not available.
INFO:tensorflow:tensorflow_decision_forests is not available.
INFO:tensorflow:tensorflow_decision_forests is not available.
INFO:tensorflow:tensorflow_text is not available.
INFO:tensorflow:tensorflow_text is not available.

The tft_layer is smart enough to still execute the transformation if only a subset of features are passed in. For example, if you only pass in two features, you'll get just the transformed versions of those features back:

ex2 = pd.DataFrame(ex)[['education', 'hours-per-week']]
ex2
pd.DataFrame(tft_layer(dict(ex2)))

Here's a more robust version that drops features that are not in the feature-spec, and returns a (features, label) pair if the label is in the provided features:

class Transform(tf.Module):
  def __init__(self, working_dir):
    self.working_dir = working_dir
    self.tf_transform_output = tft.TFTransformOutput(working_dir)
    self.tft_layer = tf_transform_output.transform_features_layer()

  @tf.function
  def __call__(self, features):
    raw_features = {}

    for key, val in features.items():
      # Skip unused keys
      if key not in RAW_DATA_FEATURE_SPEC:
        continue

      raw_features[key] = val

    # Apply the `preprocessing_fn`.
    transformed_features = tft_layer(raw_features)

    if LABEL_KEY in transformed_features:
      # Pop the label and return a (features, labels) pair.
      data_labels = transformed_features.pop(LABEL_KEY)
      return (transformed_features, data_labels)
    else:
      return transformed_features
transform = Transform(output_dir)
t_ex, t_label = transform(ex)
pd.DataFrame(t_ex)

Now you can use Dataset.map to apply that transformation, on the fly to new data:

model.evaluate(
    read_csv(test_path, batch_size=5).map(transform),
    steps=EVALUATION_STEPS,
    return_dict=True
)
128/128 [==============================] - 1s 6ms/step - loss: 0.3150 - accuracy: 0.8547
{'loss': 0.3149792551994324, 'accuracy': 0.854687511920929}

Export the model

So you have a trained model, and a method to apply the preporcessing_fn to new data. Assemble them into a new model that accepts serialized tf.train.Example protos as input.

class ServingModel(tf.Module):
  def __init__(self, model, working_dir):
    self.model = model
    self.working_dir = working_dir
    self.transform = Transform(working_dir)

  @tf.function(input_signature=[tf.TensorSpec(shape=[None], dtype=tf.string)])
  def __call__(self, serialized_tf_examples):
    # parse the tf.train.Example
    feature_spec = RAW_DATA_FEATURE_SPEC.copy()
    feature_spec.pop(LABEL_KEY)
    parsed_features = tf.io.parse_example(serialized_tf_examples, feature_spec)
    # Apply the `preprocessing_fn`
    transformed_features = self.transform(parsed_features)
    # Run the model
    outputs = self.model(transformed_features)
    # Format the output
    classes_names = tf.constant([['0', '1']])
    classes = tf.tile(classes_names, [tf.shape(outputs)[0], 1])
    return {'classes': classes, 'scores': outputs}

  def export(self, output_dir):
    # Increment the directory number. This is required in order to make this
    # model servable with model_server.
    save_model_dir = pathlib.Path(output_dir)/'model'
    number_dirs = [int(p.name) for p in save_model_dir.glob('*')
                  if p.name.isdigit()]
    id = max([0] + number_dirs)+1
    save_model_dir = save_model_dir/str(id)

    # Set the signature to make it visible for serving.
    concrete_serving_fn = self.__call__.get_concrete_function()
    signatures = {'serving_default': concrete_serving_fn}

    # Export the model.
    tf.saved_model.save(
        self,
        str(save_model_dir),
        signatures=signatures)

    return save_model_dir

Build the model and test-run it on the batch of serialized examples:

serving_model = ServingModel(model, output_dir)

serving_model(serialized_example_batch)
{'classes': <tf.Tensor: shape=(3, 2), dtype=string, numpy=
 array([[b'0', b'1'],
        [b'0', b'1'],
        [b'0', b'1']], dtype=object)>,
 'scores': <tf.Tensor: shape=(3, 2), dtype=float32, numpy=
 array([[-1.3098445 ,  1.1097956 ],
        [-0.44916642,  0.15945318],
        [-2.1779492 ,  1.6706921 ]], dtype=float32)>}

Export the model as a SavedModel:

saved_model_dir = serving_model.export(output_dir)
saved_model_dir
INFO:tensorflow:Assets written to: /tmpfs/tmp/tmp48qrwzrx/keras/model/1/assets
INFO:tensorflow:Assets written to: /tmpfs/tmp/tmp48qrwzrx/keras/model/1/assets
PosixPath('/tmpfs/tmp/tmp48qrwzrx/keras/model/1')

Reload the the model and test it on the same batch of examples:

reloaded = tf.saved_model.load(str(saved_model_dir))
run_model = reloaded.signatures['serving_default']
run_model(serialized_example_batch)
{'scores': <tf.Tensor: shape=(3, 2), dtype=float32, numpy=
 array([[-1.3098445 ,  1.1097956 ],
        [-0.44916642,  0.15945318],
        [-2.1779492 ,  1.6706921 ]], dtype=float32)>,
 'classes': <tf.Tensor: shape=(3, 2), dtype=string, numpy=
 array([[b'0', b'1'],
        [b'0', b'1'],
        [b'0', b'1']], dtype=object)>}

What we did

In this example we used tf.Transform to preprocess a dataset of census data, and train a model with the cleaned and transformed data. We also created an input function that we could use when we deploy our trained model in a production environment to perform inference. By using the same code for both training and inference we avoid any issues with data skew. Along the way we learned about creating an Apache Beam transform to perform the transformation that we needed for cleaning the data. We also saw how to use this transformed data to train a model using tf.keras. This is just a small piece of what TensorFlow Transform can do! We encourage you to dive into tf.Transform and discover what it can do for you.

[Optional] Using our preprocessed data to train a model using tf.estimator

Create an input function for training

def _make_training_input_fn(tf_transform_output, transformed_examples,
                            batch_size):
  """Creates an input function reading from transformed data.

  Args:
    tf_transform_output: Wrapper around output of tf.Transform.
    transformed_examples: Base filename of examples.
    batch_size: Batch size.

  Returns:
    The input function for training or eval.
  """
  def input_fn():
    """Input function for training and eval."""
    dataset = tf.data.experimental.make_batched_features_dataset(
        file_pattern=transformed_examples,
        batch_size=batch_size,
        features=tf_transform_output.transformed_feature_spec(),
        reader=tf.data.TFRecordDataset,
        shuffle=True)

    transformed_features = tf.compat.v1.data.make_one_shot_iterator(
        dataset).get_next()

    # Extract features and label from the transformed tensors.
    transformed_labels = tf.where(
        tf.equal(transformed_features.pop(LABEL_KEY), 1))

    return transformed_features, transformed_labels[:,1]

  return input_fn

Create an input function for serving

Let's create an input function that we could use in production, and prepare our trained model for serving.

def _make_serving_input_fn(tf_transform_output):
  """Creates an input function reading from raw data.

  Args:
    tf_transform_output: Wrapper around output of tf.Transform.

  Returns:
    The serving input function.
  """
  raw_feature_spec = RAW_DATA_FEATURE_SPEC.copy()
  # Remove label since it is not available during serving.
  raw_feature_spec.pop(LABEL_KEY)

  def serving_input_fn():
    """Input function for serving."""
    # Get raw features by generating the basic serving input_fn and calling it.
    # Here we generate an input_fn that expects a parsed Example proto to be fed
    # to the model at serving time.  See also
    # tf.estimator.export.build_raw_serving_input_receiver_fn.
    raw_input_fn = tf.estimator.export.build_parsing_serving_input_receiver_fn(
        raw_feature_spec, default_batch_size=None)
    serving_input_receiver = raw_input_fn()

    # Apply the transform function that was used to generate the materialized
    # data.
    raw_features = serving_input_receiver.features
    transformed_features = tf_transform_output.transform_raw_features(
        raw_features)

    return tf.estimator.export.ServingInputReceiver(
        transformed_features, serving_input_receiver.receiver_tensors)

  return serving_input_fn

Wrap our input data in FeatureColumns

Our model will expect our data in TensorFlow FeatureColumns.

def get_feature_columns(tf_transform_output):
  """Returns the FeatureColumns for the model.

  Args:
    tf_transform_output: A `TFTransformOutput` object.

  Returns:
    A list of FeatureColumns.
  """
  # Wrap scalars as real valued columns.
  real_valued_columns = [tf.feature_column.numeric_column(key, shape=())
                         for key in NUMERIC_FEATURE_KEYS]

  # Wrap categorical columns.
  one_hot_columns = [
      tf.feature_column.indicator_column(
          tf.feature_column.categorical_column_with_identity(
              key=key,
              num_buckets=(NUM_OOV_BUCKETS +
                  tf_transform_output.vocabulary_size_by_name(
                      vocab_filename=key))))
      for key in CATEGORICAL_FEATURE_KEYS]

  return real_valued_columns + one_hot_columns

Train, Evaluate, and Export our model

def train_and_evaluate(working_dir, num_train_instances=NUM_TRAIN_INSTANCES,
                       num_test_instances=NUM_TEST_INSTANCES):
  """Train the model on training data and evaluate on test data.

  Args:
    working_dir: Directory to read transformed data and metadata from and to
        write exported model to.
    num_train_instances: Number of instances in train set
    num_test_instances: Number of instances in test set

  Returns:
    The results from the estimator's 'evaluate' method
  """
  tf_transform_output = tft.TFTransformOutput(working_dir)

  run_config = tf.estimator.RunConfig()

  estimator = tf.estimator.LinearClassifier(
      feature_columns=get_feature_columns(tf_transform_output),
      config=run_config,
      loss_reduction=tf.losses.Reduction.SUM)

  # Fit the model using the default optimizer.
  train_input_fn = _make_training_input_fn(
      tf_transform_output,
      os.path.join(working_dir, TRANSFORMED_TRAIN_DATA_FILEBASE + '*'),
      batch_size=BATCH_SIZE)
  estimator.train(
      input_fn=train_input_fn,
      max_steps=TRAIN_NUM_EPOCHS * num_train_instances / BATCH_SIZE)

  # Evaluate model on test dataset.
  eval_input_fn = _make_training_input_fn(
      tf_transform_output,
      os.path.join(working_dir, TRANSFORMED_TEST_DATA_FILEBASE + '*'),
      batch_size=1)

  # Export the model.
  serving_input_fn = _make_serving_input_fn(tf_transform_output)
  exported_model_dir = os.path.join(working_dir, EXPORTED_MODEL_DIR)
  estimator.export_saved_model(exported_model_dir, serving_input_fn)

  return estimator.evaluate(input_fn=eval_input_fn, steps=num_test_instances)

Put it all together

We've created all the stuff we need to preprocess our census data, train a model, and prepare it for serving. So far we've just been getting things ready. It's time to start running!

import tempfile
temp = temp = os.path.join(tempfile.mkdtemp(),'estimator')

transform_data(train_path, test_path, temp)
results = train_and_evaluate(temp)
WARNING:root:Make sure that locally built Python SDK docker image has Python 3.9 interpreter.
INFO:tensorflow:Assets written to: /tmpfs/tmp/tmpj48_kjis/tftransform_tmp/e81ba6c0678c42218346c74f0d2fcd25/assets
INFO:tensorflow:Assets written to: /tmpfs/tmp/tmpj48_kjis/tftransform_tmp/e81ba6c0678c42218346c74f0d2fcd25/assets
INFO:tensorflow:struct2tensor is not available.
INFO:tensorflow:struct2tensor is not available.
INFO:tensorflow:tensorflow_decision_forests is not available.
INFO:tensorflow:tensorflow_decision_forests is not available.
INFO:tensorflow:tensorflow_text is not available.
INFO:tensorflow:tensorflow_text is not available.
INFO:tensorflow:Assets written to: /tmpfs/tmp/tmpj48_kjis/tftransform_tmp/9c4bf75f24a24783945a343f36db82d4/assets
INFO:tensorflow:Assets written to: /tmpfs/tmp/tmpj48_kjis/tftransform_tmp/9c4bf75f24a24783945a343f36db82d4/assets
INFO:tensorflow:struct2tensor is not available.
INFO:tensorflow:struct2tensor is not available.
INFO:tensorflow:tensorflow_decision_forests is not available.
INFO:tensorflow:tensorflow_decision_forests is not available.
INFO:tensorflow:tensorflow_text is not available.
INFO:tensorflow:tensorflow_text is not available.
INFO:tensorflow:struct2tensor is not available.
INFO:tensorflow:struct2tensor is not available.
INFO:tensorflow:tensorflow_decision_forests is not available.
INFO:tensorflow:tensorflow_decision_forests is not available.
INFO:tensorflow:tensorflow_text is not available.
INFO:tensorflow:tensorflow_text is not available.
WARNING:tensorflow:Using temporary folder as model directory: /tmpfs/tmp/tmpos6nccqo
WARNING:tensorflow:Using temporary folder as model directory: /tmpfs/tmp/tmpos6nccqo
INFO:tensorflow:Using config: {'_model_dir': '/tmpfs/tmp/tmpos6nccqo', '_tf_random_seed': None, '_save_summary_steps': 100, '_save_checkpoints_steps': None, '_save_checkpoints_secs': 600, '_session_config': allow_soft_placement: true
graph_options {
  rewrite_options {
    meta_optimizer_iterations: ONE
  }
}
, '_keep_checkpoint_max': 5, '_keep_checkpoint_every_n_hours': 10000, '_log_step_count_steps': 100, '_train_distribute': None, '_device_fn': None, '_protocol': None, '_eval_distribute': None, '_experimental_distribute': None, '_experimental_max_worker_delay_secs': None, '_session_creation_timeout_secs': 7200, '_checkpoint_save_graph_def': True, '_service': None, '_cluster_spec': ClusterSpec({}), '_task_type': 'worker', '_task_id': 0, '_global_id_in_cluster': 0, '_master': '', '_evaluation_master': '', '_is_chief': True, '_num_ps_replicas': 0, '_num_worker_replicas': 1}
INFO:tensorflow:Using config: {'_model_dir': '/tmpfs/tmp/tmpos6nccqo', '_tf_random_seed': None, '_save_summary_steps': 100, '_save_checkpoints_steps': None, '_save_checkpoints_secs': 600, '_session_config': allow_soft_placement: true
graph_options {
  rewrite_options {
    meta_optimizer_iterations: ONE
  }
}
, '_keep_checkpoint_max': 5, '_keep_checkpoint_every_n_hours': 10000, '_log_step_count_steps': 100, '_train_distribute': None, '_device_fn': None, '_protocol': None, '_eval_distribute': None, '_experimental_distribute': None, '_experimental_max_worker_delay_secs': None, '_session_creation_timeout_secs': 7200, '_checkpoint_save_graph_def': True, '_service': None, '_cluster_spec': ClusterSpec({}), '_task_type': 'worker', '_task_id': 0, '_global_id_in_cluster': 0, '_master': '', '_evaluation_master': '', '_is_chief': True, '_num_ps_replicas': 0, '_num_worker_replicas': 1}
WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.9/site-packages/tensorflow/python/training/training_util.py:396: Variable.initialized_value (from tensorflow.python.ops.variables) is deprecated and will be removed in a future version.
Instructions for updating:
Use Variable.read_value. Variables in 2.X are initialized automatically both in eager and graph (inside tf.defun) contexts.
WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.9/site-packages/tensorflow/python/training/training_util.py:396: Variable.initialized_value (from tensorflow.python.ops.variables) is deprecated and will be removed in a future version.
Instructions for updating:
Use Variable.read_value. Variables in 2.X are initialized automatically both in eager and graph (inside tf.defun) contexts.
INFO:tensorflow:Calling model_fn.
INFO:tensorflow:Calling model_fn.
/tmpfs/src/tf_docs_env/lib/python3.9/site-packages/tensorflow_estimator/python/estimator/canned/linear.py:1468: UserWarning: `layer.add_variable` is deprecated and will be removed in a future version. Please use `layer.add_weight` method instead.
  self.bias = self.add_variable(
WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.9/site-packages/keras/optimizer_v2/ftrl.py:148: calling Constant.__init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version.
Instructions for updating:
Call initializer instance with the dtype argument instead of passing it to the constructor
WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.9/site-packages/keras/optimizer_v2/ftrl.py:148: calling Constant.__init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version.
Instructions for updating:
Call initializer instance with the dtype argument instead of passing it to the constructor
INFO:tensorflow:Done calling model_fn.
INFO:tensorflow:Done calling model_fn.
INFO:tensorflow:Create CheckpointSaverHook.
INFO:tensorflow:Create CheckpointSaverHook.
INFO:tensorflow:Graph was finalized.
INFO:tensorflow:Graph was finalized.
INFO:tensorflow:Running local_init_op.
INFO:tensorflow:Running local_init_op.
INFO:tensorflow:Done running local_init_op.
INFO:tensorflow:Done running local_init_op.
INFO:tensorflow:Calling checkpoint listeners before saving checkpoint 0...
INFO:tensorflow:Calling checkpoint listeners before saving checkpoint 0...
INFO:tensorflow:Saving checkpoints for 0 into /tmpfs/tmp/tmpos6nccqo/model.ckpt.
INFO:tensorflow:Saving checkpoints for 0 into /tmpfs/tmp/tmpos6nccqo/model.ckpt.
INFO:tensorflow:Calling checkpoint listeners after saving checkpoint 0...
INFO:tensorflow:Calling checkpoint listeners after saving checkpoint 0...
INFO:tensorflow:loss = 88.72284, step = 0
INFO:tensorflow:loss = 88.72284, step = 0
INFO:tensorflow:global_step/sec: 188.678
INFO:tensorflow:global_step/sec: 188.678
INFO:tensorflow:loss = 45.22135, step = 100 (0.531 sec)
INFO:tensorflow:loss = 45.22135, step = 100 (0.531 sec)
INFO:tensorflow:global_step/sec: 256.761
INFO:tensorflow:global_step/sec: 256.761
INFO:tensorflow:loss = 31.924831, step = 200 (0.389 sec)
INFO:tensorflow:loss = 31.924831, step = 200 (0.389 sec)
INFO:tensorflow:global_step/sec: 250.463
INFO:tensorflow:global_step/sec: 250.463
INFO:tensorflow:loss = 46.127586, step = 300 (0.400 sec)
INFO:tensorflow:loss = 46.127586, step = 300 (0.400 sec)
INFO:tensorflow:global_step/sec: 254.799
INFO:tensorflow:global_step/sec: 254.799
INFO:tensorflow:loss = 40.06756, step = 400 (0.392 sec)
INFO:tensorflow:loss = 40.06756, step = 400 (0.392 sec)
INFO:tensorflow:global_step/sec: 259.988
INFO:tensorflow:global_step/sec: 259.988
INFO:tensorflow:loss = 44.59692, step = 500 (0.385 sec)
INFO:tensorflow:loss = 44.59692, step = 500 (0.385 sec)
INFO:tensorflow:global_step/sec: 252.579
INFO:tensorflow:global_step/sec: 252.579
INFO:tensorflow:loss = 37.153473, step = 600 (0.396 sec)
INFO:tensorflow:loss = 37.153473, step = 600 (0.396 sec)
INFO:tensorflow:global_step/sec: 256.924
INFO:tensorflow:global_step/sec: 256.924
INFO:tensorflow:loss = 47.779545, step = 700 (0.389 sec)
INFO:tensorflow:loss = 47.779545, step = 700 (0.389 sec)
INFO:tensorflow:global_step/sec: 263.156
INFO:tensorflow:global_step/sec: 263.156
INFO:tensorflow:loss = 38.260384, step = 800 (0.380 sec)
INFO:tensorflow:loss = 38.260384, step = 800 (0.380 sec)
INFO:tensorflow:global_step/sec: 258.758
INFO:tensorflow:global_step/sec: 258.758
INFO:tensorflow:loss = 38.531883, step = 900 (0.386 sec)
INFO:tensorflow:loss = 38.531883, step = 900 (0.386 sec)
INFO:tensorflow:global_step/sec: 256.946
INFO:tensorflow:global_step/sec: 256.946
INFO:tensorflow:loss = 38.950302, step = 1000 (0.389 sec)
INFO:tensorflow:loss = 38.950302, step = 1000 (0.389 sec)
INFO:tensorflow:global_step/sec: 256.414
INFO:tensorflow:global_step/sec: 256.414
INFO:tensorflow:loss = 41.970627, step = 1100 (0.390 sec)
INFO:tensorflow:loss = 41.970627, step = 1100 (0.390 sec)
INFO:tensorflow:global_step/sec: 255.951
INFO:tensorflow:global_step/sec: 255.951
INFO:tensorflow:loss = 37.365116, step = 1200 (0.391 sec)
INFO:tensorflow:loss = 37.365116, step = 1200 (0.391 sec)
INFO:tensorflow:global_step/sec: 260.234
INFO:tensorflow:global_step/sec: 260.234
INFO:tensorflow:loss = 39.425877, step = 1300 (0.384 sec)
INFO:tensorflow:loss = 39.425877, step = 1300 (0.384 sec)
INFO:tensorflow:global_step/sec: 256.037
INFO:tensorflow:global_step/sec: 256.037
INFO:tensorflow:loss = 37.087532, step = 1400 (0.390 sec)
INFO:tensorflow:loss = 37.087532, step = 1400 (0.390 sec)
INFO:tensorflow:global_step/sec: 258.279
INFO:tensorflow:global_step/sec: 258.279
INFO:tensorflow:loss = 35.798325, step = 1500 (0.387 sec)
INFO:tensorflow:loss = 35.798325, step = 1500 (0.387 sec)
INFO:tensorflow:global_step/sec: 260.315
INFO:tensorflow:global_step/sec: 260.315
INFO:tensorflow:loss = 43.254173, step = 1600 (0.384 sec)
INFO:tensorflow:loss = 43.254173, step = 1600 (0.384 sec)
INFO:tensorflow:global_step/sec: 260.515
INFO:tensorflow:global_step/sec: 260.515
INFO:tensorflow:loss = 46.526375, step = 1700 (0.384 sec)
INFO:tensorflow:loss = 46.526375, step = 1700 (0.384 sec)
INFO:tensorflow:global_step/sec: 259.538
INFO:tensorflow:global_step/sec: 259.538
INFO:tensorflow:loss = 54.708443, step = 1800 (0.385 sec)
INFO:tensorflow:loss = 54.708443, step = 1800 (0.385 sec)
INFO:tensorflow:global_step/sec: 255.042
INFO:tensorflow:global_step/sec: 255.042
INFO:tensorflow:loss = 48.13112, step = 1900 (0.392 sec)
INFO:tensorflow:loss = 48.13112, step = 1900 (0.392 sec)
INFO:tensorflow:global_step/sec: 258.149
INFO:tensorflow:global_step/sec: 258.149
INFO:tensorflow:loss = 36.763084, step = 2000 (0.387 sec)
INFO:tensorflow:loss = 36.763084, step = 2000 (0.387 sec)
INFO:tensorflow:global_step/sec: 254.948
INFO:tensorflow:global_step/sec: 254.948
INFO:tensorflow:loss = 46.21298, step = 2100 (0.392 sec)
INFO:tensorflow:loss = 46.21298, step = 2100 (0.392 sec)
INFO:tensorflow:global_step/sec: 258.605
INFO:tensorflow:global_step/sec: 258.605
INFO:tensorflow:loss = 38.09401, step = 2200 (0.387 sec)
INFO:tensorflow:loss = 38.09401, step = 2200 (0.387 sec)
INFO:tensorflow:global_step/sec: 258.731
INFO:tensorflow:global_step/sec: 258.731
INFO:tensorflow:loss = 34.317585, step = 2300 (0.387 sec)
INFO:tensorflow:loss = 34.317585, step = 2300 (0.387 sec)
INFO:tensorflow:global_step/sec: 256.2
INFO:tensorflow:global_step/sec: 256.2
INFO:tensorflow:loss = 34.32302, step = 2400 (0.390 sec)
INFO:tensorflow:loss = 34.32302, step = 2400 (0.390 sec)
INFO:tensorflow:global_step/sec: 251.282
INFO:tensorflow:global_step/sec: 251.282
INFO:tensorflow:loss = 35.036278, step = 2500 (0.398 sec)
INFO:tensorflow:loss = 35.036278, step = 2500 (0.398 sec)
INFO:tensorflow:global_step/sec: 255.532
INFO:tensorflow:global_step/sec: 255.532
INFO:tensorflow:loss = 37.84765, step = 2600 (0.391 sec)
INFO:tensorflow:loss = 37.84765, step = 2600 (0.391 sec)
INFO:tensorflow:global_step/sec: 261.708
INFO:tensorflow:global_step/sec: 261.708
INFO:tensorflow:loss = 42.411026, step = 2700 (0.382 sec)
INFO:tensorflow:loss = 42.411026, step = 2700 (0.382 sec)
INFO:tensorflow:global_step/sec: 256.98
INFO:tensorflow:global_step/sec: 256.98
INFO:tensorflow:loss = 42.035652, step = 2800 (0.389 sec)
INFO:tensorflow:loss = 42.035652, step = 2800 (0.389 sec)
INFO:tensorflow:global_step/sec: 262.37
INFO:tensorflow:global_step/sec: 262.37
INFO:tensorflow:loss = 44.098034, step = 2900 (0.381 sec)
INFO:tensorflow:loss = 44.098034, step = 2900 (0.381 sec)
INFO:tensorflow:global_step/sec: 258.461
INFO:tensorflow:global_step/sec: 258.461
INFO:tensorflow:loss = 44.148544, step = 3000 (0.387 sec)
INFO:tensorflow:loss = 44.148544, step = 3000 (0.387 sec)
INFO:tensorflow:global_step/sec: 259.328
INFO:tensorflow:global_step/sec: 259.328
INFO:tensorflow:loss = 29.088625, step = 3100 (0.386 sec)
INFO:tensorflow:loss = 29.088625, step = 3100 (0.386 sec)
INFO:tensorflow:global_step/sec: 267.598
INFO:tensorflow:global_step/sec: 267.598
INFO:tensorflow:loss = 32.90155, step = 3200 (0.374 sec)
INFO:tensorflow:loss = 32.90155, step = 3200 (0.374 sec)
INFO:tensorflow:global_step/sec: 264.989
INFO:tensorflow:global_step/sec: 264.989
INFO:tensorflow:loss = 44.000027, step = 3300 (0.377 sec)
INFO:tensorflow:loss = 44.000027, step = 3300 (0.377 sec)
INFO:tensorflow:global_step/sec: 263.182
INFO:tensorflow:global_step/sec: 263.182
INFO:tensorflow:loss = 41.87715, step = 3400 (0.380 sec)
INFO:tensorflow:loss = 41.87715, step = 3400 (0.380 sec)
INFO:tensorflow:global_step/sec: 257.451
INFO:tensorflow:global_step/sec: 257.451
INFO:tensorflow:loss = 48.69912, step = 3500 (0.389 sec)
INFO:tensorflow:loss = 48.69912, step = 3500 (0.389 sec)
INFO:tensorflow:global_step/sec: 256.176
INFO:tensorflow:global_step/sec: 256.176
INFO:tensorflow:loss = 37.857075, step = 3600 (0.390 sec)
INFO:tensorflow:loss = 37.857075, step = 3600 (0.390 sec)
INFO:tensorflow:global_step/sec: 254.605
INFO:tensorflow:global_step/sec: 254.605
INFO:tensorflow:loss = 39.954582, step = 3700 (0.393 sec)
INFO:tensorflow:loss = 39.954582, step = 3700 (0.393 sec)
INFO:tensorflow:global_step/sec: 255.509
INFO:tensorflow:global_step/sec: 255.509
INFO:tensorflow:loss = 44.554424, step = 3800 (0.391 sec)
INFO:tensorflow:loss = 44.554424, step = 3800 (0.391 sec)
INFO:tensorflow:global_step/sec: 256.213
INFO:tensorflow:global_step/sec: 256.213
INFO:tensorflow:loss = 40.80271, step = 3900 (0.390 sec)
INFO:tensorflow:loss = 40.80271, step = 3900 (0.390 sec)
INFO:tensorflow:global_step/sec: 254.035
INFO:tensorflow:global_step/sec: 254.035
INFO:tensorflow:loss = 46.22782, step = 4000 (0.394 sec)
INFO:tensorflow:loss = 46.22782, step = 4000 (0.394 sec)
INFO:tensorflow:global_step/sec: 256.84
INFO:tensorflow:global_step/sec: 256.84
INFO:tensorflow:loss = 36.978947, step = 4100 (0.389 sec)
INFO:tensorflow:loss = 36.978947, step = 4100 (0.389 sec)
INFO:tensorflow:global_step/sec: 262.038
INFO:tensorflow:global_step/sec: 262.038
INFO:tensorflow:loss = 50.231255, step = 4200 (0.381 sec)
INFO:tensorflow:loss = 50.231255, step = 4200 (0.381 sec)
INFO:tensorflow:global_step/sec: 259.654
INFO:tensorflow:global_step/sec: 259.654
INFO:tensorflow:loss = 37.547096, step = 4300 (0.385 sec)
INFO:tensorflow:loss = 37.547096, step = 4300 (0.385 sec)
INFO:tensorflow:global_step/sec: 263.51
INFO:tensorflow:global_step/sec: 263.51
INFO:tensorflow:loss = 37.258484, step = 4400 (0.379 sec)
INFO:tensorflow:loss = 37.258484, step = 4400 (0.379 sec)
INFO:tensorflow:global_step/sec: 256.973
INFO:tensorflow:global_step/sec: 256.973
INFO:tensorflow:loss = 36.201225, step = 4500 (0.389 sec)
INFO:tensorflow:loss = 36.201225, step = 4500 (0.389 sec)
INFO:tensorflow:global_step/sec: 257.055
INFO:tensorflow:global_step/sec: 257.055
INFO:tensorflow:loss = 35.487583, step = 4600 (0.389 sec)
INFO:tensorflow:loss = 35.487583, step = 4600 (0.389 sec)
INFO:tensorflow:global_step/sec: 257.264
INFO:tensorflow:global_step/sec: 257.264
INFO:tensorflow:loss = 38.158768, step = 4700 (0.389 sec)
INFO:tensorflow:loss = 38.158768, step = 4700 (0.389 sec)
INFO:tensorflow:global_step/sec: 260.964
INFO:tensorflow:global_step/sec: 260.964
INFO:tensorflow:loss = 34.199955, step = 4800 (0.383 sec)
INFO:tensorflow:loss = 34.199955, step = 4800 (0.383 sec)
INFO:tensorflow:global_step/sec: 261.005
INFO:tensorflow:global_step/sec: 261.005
INFO:tensorflow:loss = 42.03213, step = 4900 (0.383 sec)
INFO:tensorflow:loss = 42.03213, step = 4900 (0.383 sec)
INFO:tensorflow:global_step/sec: 262.81
INFO:tensorflow:global_step/sec: 262.81
INFO:tensorflow:loss = 34.91445, step = 5000 (0.380 sec)
INFO:tensorflow:loss = 34.91445, step = 5000 (0.380 sec)
INFO:tensorflow:Calling checkpoint listeners before saving checkpoint 5088...
INFO:tensorflow:Calling checkpoint listeners before saving checkpoint 5088...
INFO:tensorflow:Saving checkpoints for 5088 into /tmpfs/tmp/tmpos6nccqo/model.ckpt.
INFO:tensorflow:Saving checkpoints for 5088 into /tmpfs/tmp/tmpos6nccqo/model.ckpt.
INFO:tensorflow:Calling checkpoint listeners after saving checkpoint 5088...
INFO:tensorflow:Calling checkpoint listeners after saving checkpoint 5088...
INFO:tensorflow:Loss for final step: 52.16411.
INFO:tensorflow:Loss for final step: 52.16411.
INFO:tensorflow:struct2tensor is not available.
INFO:tensorflow:struct2tensor is not available.
INFO:tensorflow:tensorflow_decision_forests is not available.
INFO:tensorflow:tensorflow_decision_forests is not available.
INFO:tensorflow:tensorflow_text is not available.
INFO:tensorflow:tensorflow_text is not available.
WARNING:tensorflow:Loading a TF2 SavedModel but eager mode seems disabled.
WARNING:tensorflow:Loading a TF2 SavedModel but eager mode seems disabled.
INFO:tensorflow:Calling model_fn.
INFO:tensorflow:Calling model_fn.
INFO:tensorflow:Done calling model_fn.
INFO:tensorflow:Done calling model_fn.
WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.9/site-packages/tensorflow/python/saved_model/signature_def_utils_impl.py:146: build_tensor_info (from tensorflow.python.saved_model.utils_impl) is deprecated and will be removed in a future version.
Instructions for updating:
This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.utils.build_tensor_info or tf.compat.v1.saved_model.build_tensor_info.
WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.9/site-packages/tensorflow/python/saved_model/signature_def_utils_impl.py:146: build_tensor_info (from tensorflow.python.saved_model.utils_impl) is deprecated and will be removed in a future version.
Instructions for updating:
This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.utils.build_tensor_info or tf.compat.v1.saved_model.build_tensor_info.
INFO:tensorflow:Signatures INCLUDED in export for Classify: ['serving_default', 'classification']
INFO:tensorflow:Signatures INCLUDED in export for Classify: ['serving_default', 'classification']
INFO:tensorflow:Signatures INCLUDED in export for Regress: ['regression']
INFO:tensorflow:Signatures INCLUDED in export for Regress: ['regression']
INFO:tensorflow:Signatures INCLUDED in export for Predict: ['predict']
INFO:tensorflow:Signatures INCLUDED in export for Predict: ['predict']
INFO:tensorflow:Signatures INCLUDED in export for Train: None
INFO:tensorflow:Signatures INCLUDED in export for Train: None
INFO:tensorflow:Signatures INCLUDED in export for Eval: None
INFO:tensorflow:Signatures INCLUDED in export for Eval: None
INFO:tensorflow:Restoring parameters from /tmpfs/tmp/tmpos6nccqo/model.ckpt-5088
INFO:tensorflow:Restoring parameters from /tmpfs/tmp/tmpos6nccqo/model.ckpt-5088
INFO:tensorflow:Assets added to graph.
INFO:tensorflow:Assets added to graph.
INFO:tensorflow:Assets written to: /tmpfs/tmp/tmp444vbvpg/estimator/exported_model_dir/temp-1654074911/assets
INFO:tensorflow:Assets written to: /tmpfs/tmp/tmp444vbvpg/estimator/exported_model_dir/temp-1654074911/assets
INFO:tensorflow:SavedModel written to: /tmpfs/tmp/tmp444vbvpg/estimator/exported_model_dir/temp-1654074911/saved_model.pb
INFO:tensorflow:SavedModel written to: /tmpfs/tmp/tmp444vbvpg/estimator/exported_model_dir/temp-1654074911/saved_model.pb
INFO:tensorflow:Calling model_fn.
INFO:tensorflow:Calling model_fn.
INFO:tensorflow:Done calling model_fn.
INFO:tensorflow:Done calling model_fn.
INFO:tensorflow:Starting evaluation at 2022-06-01T09:15:13
INFO:tensorflow:Starting evaluation at 2022-06-01T09:15:13
INFO:tensorflow:Graph was finalized.
INFO:tensorflow:Graph was finalized.
INFO:tensorflow:Restoring parameters from /tmpfs/tmp/tmpos6nccqo/model.ckpt-5088
INFO:tensorflow:Restoring parameters from /tmpfs/tmp/tmpos6nccqo/model.ckpt-5088
INFO:tensorflow:Running local_init_op.
INFO:tensorflow:Running local_init_op.
INFO:tensorflow:Done running local_init_op.
INFO:tensorflow:Done running local_init_op.
INFO:tensorflow:Evaluation [1628/16280]
INFO:tensorflow:Evaluation [1628/16280]
INFO:tensorflow:Evaluation [3256/16280]
INFO:tensorflow:Evaluation [3256/16280]
INFO:tensorflow:Evaluation [4884/16280]
INFO:tensorflow:Evaluation [4884/16280]
INFO:tensorflow:Evaluation [6512/16280]
INFO:tensorflow:Evaluation [6512/16280]
INFO:tensorflow:Evaluation [8140/16280]
INFO:tensorflow:Evaluation [8140/16280]
INFO:tensorflow:Evaluation [9768/16280]
INFO:tensorflow:Evaluation [9768/16280]
INFO:tensorflow:Evaluation [11396/16280]
INFO:tensorflow:Evaluation [11396/16280]
INFO:tensorflow:Evaluation [13024/16280]
INFO:tensorflow:Evaluation [13024/16280]
INFO:tensorflow:Evaluation [14652/16280]
INFO:tensorflow:Evaluation [14652/16280]
INFO:tensorflow:Evaluation [16280/16280]
INFO:tensorflow:Evaluation [16280/16280]
INFO:tensorflow:Inference Time : 59.30933s
INFO:tensorflow:Inference Time : 59.30933s
INFO:tensorflow:Finished evaluation at 2022-06-01-09:16:12
INFO:tensorflow:Finished evaluation at 2022-06-01-09:16:12
INFO:tensorflow:Saving dict for global step 5088: accuracy = 0.8509214, accuracy_baseline = 0.7637592, auc = 0.9023326, auc_precision_recall = 0.96733487, average_loss = 0.3238075, global_step = 5088, label/mean = 0.7637592, loss = 0.3238075, precision = 0.8871779, prediction/mean = 0.7502781, recall = 0.92206854
INFO:tensorflow:Saving dict for global step 5088: accuracy = 0.8509214, accuracy_baseline = 0.7637592, auc = 0.9023326, auc_precision_recall = 0.96733487, average_loss = 0.3238075, global_step = 5088, label/mean = 0.7637592, loss = 0.3238075, precision = 0.8871779, prediction/mean = 0.7502781, recall = 0.92206854
INFO:tensorflow:Saving 'checkpoint_path' summary for global step 5088: /tmpfs/tmp/tmpos6nccqo/model.ckpt-5088
INFO:tensorflow:Saving 'checkpoint_path' summary for global step 5088: /tmpfs/tmp/tmpos6nccqo/model.ckpt-5088
pprint.pprint(results)
{'accuracy': 0.8509214,
 'accuracy_baseline': 0.7637592,
 'auc': 0.9023326,
 'auc_precision_recall': 0.96733487,
 'average_loss': 0.3238075,
 'global_step': 5088,
 'label/mean': 0.7637592,
 'loss': 0.3238075,
 'precision': 0.8871779,
 'prediction/mean': 0.7502781,
 'recall': 0.92206854}