Эта страница переведена с помощью Cloud Translation API.

Демонстрация набора инструментов для моделей MLMD

Посмотреть на TensorFlow.org

Запускаем в Google Colab

Посмотреть на GitHub

Скачать блокнот

Задний план

В этой записной книжке показано, как создать карту модели с помощью набора инструментов Model Card Toolkit с конвейером MLMD и TFX в среде Jupyter/Colab. Вы можете узнать больше о модели карт на https://modelcards.withgoogle.com/about

Настраивать

Сначала нам нужно а) установить и импортировать необходимые пакеты и б) загрузить данные.

Обновите до Pip 20.2 и установите TFX

pip install -q --upgrade pip==20.2
pip install -q "tfx==0.26.0"
pip install -q model-card-toolkit

Вы перезапустили среду выполнения?

Если вы используете Google Colab, при первом запуске указанной выше ячейки необходимо перезапустить среду выполнения (Время выполнения > Перезапустить среду выполнения...). Это связано с тем, как Colab загружает пакеты.

Импорт пакетов

Мы импортируем необходимые пакеты, в том числе стандартные классы компонентов TFX и проверяем версии библиотек.

import os
import pprint
import tempfile
import urllib

import absl
import tensorflow as tf
import tensorflow_model_analysis as tfma
tf.get_logger().propagate = False
pp = pprint.PrettyPrinter()

import tfx
from tfx.components import CsvExampleGen
from tfx.components import Evaluator
from tfx.components import Pusher
from tfx.components import ResolverNode
from tfx.components import SchemaGen
from tfx.components import StatisticsGen
from tfx.components import Trainer
from tfx.components import Transform
from tfx.components.base import executor_spec
from tfx.components.trainer.executor import GenericExecutor
from tfx.dsl.experimental import latest_blessed_model_resolver
from tfx.orchestration import metadata
from tfx.orchestration import pipeline
from tfx.orchestration.experimental.interactive.interactive_context import InteractiveContext
from tfx.proto import pusher_pb2
from tfx.proto import trainer_pb2
from tfx.types import Channel
from tfx.types.standard_artifacts import Model
from tfx.types.standard_artifacts import ModelBlessing
from tfx.utils.dsl_utils import external_input

import ml_metadata as mlmd

WARNING:absl:RuntimeParameter is only supported on Cloud-based DAG runner currently.

print('TensorFlow version: {}'.format(tf.__version__))
print('TFX version: {}'.format(tfx.version.__version__))
print('MLMD version: {}'.format(mlmd.__version__))

TensorFlow version: 2.3.2
TFX version: 0.26.0
MLMD version: 0.26.0

Настройка путей конвейера

# This is the root directory for your TFX pip package installation.
_tfx_root = tfx.__path__

# Set up logging.
absl.logging.set_verbosity(absl.logging.INFO)

Скачать пример данных

Мы загружаем пример набора данных для использования в нашем конвейере TFX.

DATA_PATH = 'https://archive.ics.uci.edu/ml/machine-learning-databases/adult/' \
   'adult.data'
_data_root = tempfile.mkdtemp(prefix='tfx-data')
_data_filepath = os.path.join(_data_root, "data.csv")
urllib.request.urlretrieve(DATA_PATH, _data_filepath)

columns = [
  "Age", "Workclass", "fnlwgt", "Education", "Education-Num", "Marital-Status",
  "Occupation", "Relationship", "Race", "Sex", "Capital-Gain", "Capital-Loss",
  "Hours-per-week", "Country", "Over-50K"]

with open(_data_filepath, 'r') as f:
  content = f.read()
  content = content.replace(", <=50K", ', 0').replace(", >50K", ', 1')

with open(_data_filepath, 'w') as f:
  f.write(','.join(columns) + '\n' + content)

Взгляните на файл CSV.

head {_data_filepath}

Age,Workclass,fnlwgt,Education,Education-Num,Marital-Status,Occupation,Relationship,Race,Sex,Capital-Gain,Capital-Loss,Hours-per-week,Country,Over-50K
39, State-gov, 77516, Bachelors, 13, Never-married, Adm-clerical, Not-in-family, White, Male, 2174, 0, 40, United-States, 0
50, Self-emp-not-inc, 83311, Bachelors, 13, Married-civ-spouse, Exec-managerial, Husband, White, Male, 0, 0, 13, United-States, 0
38, Private, 215646, HS-grad, 9, Divorced, Handlers-cleaners, Not-in-family, White, Male, 0, 0, 40, United-States, 0
53, Private, 234721, 11th, 7, Married-civ-spouse, Handlers-cleaners, Husband, Black, Male, 0, 0, 40, United-States, 0
28, Private, 338409, Bachelors, 13, Married-civ-spouse, Prof-specialty, Wife, Black, Female, 0, 0, 40, Cuba, 0
37, Private, 284582, Masters, 14, Married-civ-spouse, Exec-managerial, Wife, White, Female, 0, 0, 40, United-States, 0
49, Private, 160187, 9th, 5, Married-spouse-absent, Other-service, Not-in-family, Black, Female, 0, 0, 16, Jamaica, 0
52, Self-emp-not-inc, 209642, HS-grad, 9, Married-civ-spouse, Exec-managerial, Husband, White, Male, 0, 0, 45, United-States, 1
31, Private, 45781, Masters, 14, Never-married, Prof-specialty, Not-in-family, White, Female, 14084, 0, 50, United-States, 1

Создайте интерактивный контекст

Наконец, мы создаем InteractiveContext, который позволит нам интерактивно запускать компоненты TFX в этой записной книжке.

# Here, we create an InteractiveContext using default parameters. This will
# use a temporary directory with an ephemeral ML Metadata database instance.
# To use your own pipeline root or database, the optional properties
# `pipeline_root` and `metadata_connection_config` may be passed to
# InteractiveContext. Calls to InteractiveContext are no-ops outside of the
# notebook.
context = InteractiveContext()

WARNING:absl:InteractiveContext pipeline_root argument not provided: using temporary directory /tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1 as root for pipeline outputs.
WARNING:absl:InteractiveContext metadata_connection_config not provided: using SQLite ML Metadata database at /tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/metadata.sqlite.

Интерактивный запуск компонентов TFX

В следующих ячейках мы создаем компоненты TFX один за другим, запускаем каждый из них и визуализируем их выходные артефакты. В этом ноутбуке, мы не будем предоставлять подробные объяснения каждого компонента TFX, но вы можете увидеть , что каждый делает в мастерской TFX Colab .

ExampleGen

Создание ExampleGen компоненты к данным расщепленных в учебные и оценочных наборы, конвертировать данные в tf.Example формат, а также копировать данные в _tfx_root каталог для других компонентов для доступа.

example_gen = CsvExampleGen(input=external_input(_data_root))
context.run(example_gen)

WARNING:absl:From <ipython-input-1-2e0190c2dd16>:1: external_input (from tfx.utils.dsl_utils) is deprecated and will be removed in a future version.
Instructions for updating:
external_input is deprecated, directly pass the uri to ExampleGen.
WARNING:absl:The "input" argument to the CsvExampleGen component has been deprecated by "input_base". Please update your usage as support for this argument will be removed soon.
INFO:absl:Running driver for CsvExampleGen
INFO:absl:MetadataStore with DB connection initialized
INFO:absl:select span and version = (0, None)
INFO:absl:latest span and version = (0, None)
INFO:absl:Running executor for CsvExampleGen
INFO:absl:Generating examples.
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.
INFO:absl:Processing input csv data /tmp/tfx-datajjx_v0dr/* to TFExample.
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:absl:Examples generated.
INFO:absl:Running publisher for CsvExampleGen
INFO:absl:MetadataStore with DB connection initialized

artifact = example_gen.outputs['examples'].get()[0]
print(artifact.split_names, artifact.uri)

["train", "eval"] /tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/CsvExampleGen/examples/1

Давайте взглянем на первые три обучающих примера:

# Get the URI of the output artifact representing the training examples, which is a directory
train_uri = os.path.join(example_gen.outputs['examples'].get()[0].uri, 'train')

# Get the list of files in this directory (all compressed TFRecord files)
tfrecord_filenames = [os.path.join(train_uri, name)
                      for name in os.listdir(train_uri)]

# Create a `TFRecordDataset` to read these files
dataset = tf.data.TFRecordDataset(tfrecord_filenames, compression_type="GZIP")

# Iterate over the first 3 records and decode them.
for tfrecord in dataset.take(3):
  serialized_example = tfrecord.numpy()
  example = tf.train.Example()
  example.ParseFromString(serialized_example)
  pp.pprint(example)

features {
  feature {
    key: "Age"
    value {
      int64_list {
        value: 39
      }
    }
  }
  feature {
    key: "Capital-Gain"
    value {
      int64_list {
        value: 2174
      }
    }
  }
  feature {
    key: "Capital-Loss"
    value {
      int64_list {
        value: 0
      }
    }
  }
  feature {
    key: "Country"
    value {
      bytes_list {
        value: " United-States"
      }
    }
  }
  feature {
    key: "Education"
    value {
      bytes_list {
        value: " Bachelors"
      }
    }
  }
  feature {
    key: "Education-Num"
    value {
      int64_list {
        value: 13
      }
    }
  }
  feature {
    key: "Hours-per-week"
    value {
      int64_list {
        value: 40
      }
    }
  }
  feature {
    key: "Marital-Status"
    value {
      bytes_list {
        value: " Never-married"
      }
    }
  }
  feature {
    key: "Occupation"
    value {
      bytes_list {
        value: " Adm-clerical"
      }
    }
  }
  feature {
    key: "Over-50K"
    value {
      int64_list {
        value: 0
      }
    }
  }
  feature {
    key: "Race"
    value {
      bytes_list {
        value: " White"
      }
    }
  }
  feature {
    key: "Relationship"
    value {
      bytes_list {
        value: " Not-in-family"
      }
    }
  }
  feature {
    key: "Sex"
    value {
      bytes_list {
        value: " Male"
      }
    }
  }
  feature {
    key: "Workclass"
    value {
      bytes_list {
        value: " State-gov"
      }
    }
  }
  feature {
    key: "fnlwgt"
    value {
      int64_list {
        value: 77516
      }
    }
  }
}

features {
  feature {
    key: "Age"
    value {
      int64_list {
        value: 50
      }
    }
  }
  feature {
    key: "Capital-Gain"
    value {
      int64_list {
        value: 0
      }
    }
  }
  feature {
    key: "Capital-Loss"
    value {
      int64_list {
        value: 0
      }
    }
  }
  feature {
    key: "Country"
    value {
      bytes_list {
        value: " United-States"
      }
    }
  }
  feature {
    key: "Education"
    value {
      bytes_list {
        value: " Bachelors"
      }
    }
  }
  feature {
    key: "Education-Num"
    value {
      int64_list {
        value: 13
      }
    }
  }
  feature {
    key: "Hours-per-week"
    value {
      int64_list {
        value: 13
      }
    }
  }
  feature {
    key: "Marital-Status"
    value {
      bytes_list {
        value: " Married-civ-spouse"
      }
    }
  }
  feature {
    key: "Occupation"
    value {
      bytes_list {
        value: " Exec-managerial"
      }
    }
  }
  feature {
    key: "Over-50K"
    value {
      int64_list {
        value: 0
      }
    }
  }
  feature {
    key: "Race"
    value {
      bytes_list {
        value: " White"
      }
    }
  }
  feature {
    key: "Relationship"
    value {
      bytes_list {
        value: " Husband"
      }
    }
  }
  feature {
    key: "Sex"
    value {
      bytes_list {
        value: " Male"
      }
    }
  }
  feature {
    key: "Workclass"
    value {
      bytes_list {
        value: " Self-emp-not-inc"
      }
    }
  }
  feature {
    key: "fnlwgt"
    value {
      int64_list {
        value: 83311
      }
    }
  }
}

features {
  feature {
    key: "Age"
    value {
      int64_list {
        value: 38
      }
    }
  }
  feature {
    key: "Capital-Gain"
    value {
      int64_list {
        value: 0
      }
    }
  }
  feature {
    key: "Capital-Loss"
    value {
      int64_list {
        value: 0
      }
    }
  }
  feature {
    key: "Country"
    value {
      bytes_list {
        value: " United-States"
      }
    }
  }
  feature {
    key: "Education"
    value {
      bytes_list {
        value: " HS-grad"
      }
    }
  }
  feature {
    key: "Education-Num"
    value {
      int64_list {
        value: 9
      }
    }
  }
  feature {
    key: "Hours-per-week"
    value {
      int64_list {
        value: 40
      }
    }
  }
  feature {
    key: "Marital-Status"
    value {
      bytes_list {
        value: " Divorced"
      }
    }
  }
  feature {
    key: "Occupation"
    value {
      bytes_list {
        value: " Handlers-cleaners"
      }
    }
  }
  feature {
    key: "Over-50K"
    value {
      int64_list {
        value: 0
      }
    }
  }
  feature {
    key: "Race"
    value {
      bytes_list {
        value: " White"
      }
    }
  }
  feature {
    key: "Relationship"
    value {
      bytes_list {
        value: " Not-in-family"
      }
    }
  }
  feature {
    key: "Sex"
    value {
      bytes_list {
        value: " Male"
      }
    }
  }
  feature {
    key: "Workclass"
    value {
      bytes_list {
        value: " Private"
      }
    }
  }
  feature {
    key: "fnlwgt"
    value {
      int64_list {
        value: 215646
      }
    }
  }
}

СтатистикаGen

StatisticsGen принимает в качестве входных данных набора данных , мы просто проглоченные с помощью ExampleGen и позволяет выполнить некоторый анализ набора данных с помощью TensorFlow Data Validation (TFDV).

statistics_gen = StatisticsGen(
    examples=example_gen.outputs['examples'])
context.run(statistics_gen)

INFO:absl:Excluding no splits because exclude_splits is not set.
INFO:absl:Running driver for StatisticsGen
INFO:absl:MetadataStore with DB connection initialized
INFO:absl:Running executor for StatisticsGen
INFO:absl:Generating statistics for split train.
INFO:absl:Statistics for split train written to /tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/StatisticsGen/statistics/2/train.
INFO:absl:Generating statistics for split eval.
INFO:absl:Statistics for split eval written to /tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/StatisticsGen/statistics/2/eval.
INFO:absl:Running publisher for StatisticsGen
INFO:absl:MetadataStore with DB connection initialized

После StatisticsGen завершения работы, мы можем визуализировать выведенную статистику. Попробуйте поиграть с разными сюжетами!

context.show(statistics_gen.outputs['statistics'])

WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.6/site-packages/tensorflow_data_validation/utils/stats_util.py:247: tf_record_iterator (from tensorflow.python.lib.io.tf_record) is deprecated and will be removed in a future version.
Instructions for updating:
Use eager execution and: 
`tf.data.TFRecordDataset(path)`

SchemaGen

SchemaGen будет принимать в качестве входных данных статистики , которые мы получили с StatisticsGen , глядя на учебный раскол по умолчанию.

schema_gen = SchemaGen(
    statistics=statistics_gen.outputs['statistics'],
    infer_feature_shape=False)
context.run(schema_gen)

INFO:absl:Excluding no splits because exclude_splits is not set.
INFO:absl:Running driver for SchemaGen
INFO:absl:MetadataStore with DB connection initialized
INFO:absl:Running executor for SchemaGen
INFO:absl:Processing schema from statistics for split train.
INFO:absl:Processing schema from statistics for split eval.
INFO:absl:Schema written to /tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/SchemaGen/schema/3/schema.pbtxt.
INFO:absl:Running publisher for SchemaGen
INFO:absl:MetadataStore with DB connection initialized

context.show(schema_gen.outputs['schema'])

/tmpfs/src/tf_docs_env/lib/python3.6/site-packages/tensorflow_data_validation/utils/display_util.py:151: FutureWarning: Passing a negative integer is deprecated in version 1.0 and will not be supported in future version. Instead, use None to not limit the column width.
  pd.set_option('max_colwidth', -1)

Чтобы узнать больше о схемах, см документации SchemaGen .

Трансформировать

Transform будет принимать в качестве входных данных из ExampleGen , схемы из SchemaGen , а также модуль, содержащий определенные пользователем преобразования кода.

Давайте рассмотрим пример определяемого пользователя преобразования кода ниже (для введения в TensorFlow Transform API - интерфейсы, можно найти в учебнике ).

_census_income_constants_module_file = 'census_income_constants.py'

%%writefile {_census_income_constants_module_file}

# Categorical features are assumed to each have a maximum value in the dataset.
MAX_CATEGORICAL_FEATURE_VALUES = [20]

CATEGORICAL_FEATURE_KEYS = ["Education-Num"]


DENSE_FLOAT_FEATURE_KEYS = ["Capital-Gain", "Hours-per-week", "Capital-Loss"]

# Number of buckets used by tf.transform for encoding each feature.
FEATURE_BUCKET_COUNT = 10

BUCKET_FEATURE_KEYS = ["Age"]

# Number of vocabulary terms used for encoding VOCAB_FEATURES by tf.transform
VOCAB_SIZE = 200

# Count of out-of-vocab buckets in which unrecognized VOCAB_FEATURES are hashed.
OOV_SIZE = 10

VOCAB_FEATURE_KEYS = ["Workclass", "Education", "Marital-Status", "Occupation", 
                      "Relationship", "Race", "Sex", "Country"]

# Keys
LABEL_KEY = "Over-50K"

def transformed_name(key):
  return key + '_xf'

Writing census_income_constants.py

_census_income_transform_module_file = 'census_income_transform.py'

%%writefile {_census_income_transform_module_file}

import tensorflow as tf
import tensorflow_transform as tft

import census_income_constants

_DENSE_FLOAT_FEATURE_KEYS = census_income_constants.DENSE_FLOAT_FEATURE_KEYS
_VOCAB_FEATURE_KEYS = census_income_constants.VOCAB_FEATURE_KEYS
_VOCAB_SIZE = census_income_constants.VOCAB_SIZE
_OOV_SIZE = census_income_constants.OOV_SIZE
_FEATURE_BUCKET_COUNT = census_income_constants.FEATURE_BUCKET_COUNT
_BUCKET_FEATURE_KEYS = census_income_constants.BUCKET_FEATURE_KEYS
_CATEGORICAL_FEATURE_KEYS = census_income_constants.CATEGORICAL_FEATURE_KEYS
_LABEL_KEY = census_income_constants.LABEL_KEY
_transformed_name = census_income_constants.transformed_name


def preprocessing_fn(inputs):
  """tf.transform's callback function for preprocessing inputs.
  Args:
    inputs: map from feature keys to raw not-yet-transformed features.
  Returns:
    Map from string feature key to transformed feature operations.
  """
  outputs = {}
  for key in _DENSE_FLOAT_FEATURE_KEYS:
    # Preserve this feature as a dense float, setting nan's to the mean.
    outputs[_transformed_name(key)] = tft.scale_to_z_score(
        _fill_in_missing(inputs[key]))

  for key in _VOCAB_FEATURE_KEYS:
    # Build a vocabulary for this feature.
    outputs[_transformed_name(key)] = tft.compute_and_apply_vocabulary(
        _fill_in_missing(inputs[key]),
        top_k=_VOCAB_SIZE,
        num_oov_buckets=_OOV_SIZE)

  for key in _BUCKET_FEATURE_KEYS:
    outputs[_transformed_name(key)] = tft.bucketize(
        _fill_in_missing(inputs[key]), _FEATURE_BUCKET_COUNT)

  for key in _CATEGORICAL_FEATURE_KEYS:
    outputs[_transformed_name(key)] = _fill_in_missing(inputs[key])

  label = _fill_in_missing(inputs[_LABEL_KEY])
  outputs[_transformed_name(_LABEL_KEY)] = label

  return outputs


def _fill_in_missing(x):
  """Replace missing values in a SparseTensor.
  Fills in missing values of `x` with '' or 0, and converts to a dense tensor.
  Args:
    x: A `SparseTensor` of rank 2.  Its dense shape should have size at most 1
      in the second dimension.
  Returns:
    A rank 1 tensor where missing values of `x` have been filled in.
  """
  default_value = '' if x.dtype == tf.string else 0
  return tf.squeeze(
      tf.sparse.to_dense(
          tf.SparseTensor(x.indices, x.values, [x.dense_shape[0], 1]),
          default_value),
      axis=1)

Writing census_income_transform.py

transform = Transform(
    examples=example_gen.outputs['examples'],
    schema=schema_gen.outputs['schema'],
    module_file=os.path.abspath(_census_income_transform_module_file))
context.run(transform)

INFO:absl:Running driver for Transform
INFO:absl:MetadataStore with DB connection initialized
INFO:absl:Running executor for Transform
INFO:absl:Analyze the 'train' split and transform all splits when splits_config is not set.
WARNING:absl:The default value of `force_tf_compat_v1` will change in a future release from `True` to `False`. Since this pipeline has TF 2 behaviors enabled, Transform will use native TF 2 at that point. You can test this behavior now by passing `force_tf_compat_v1=False` or disable it by explicitly setting `force_tf_compat_v1=True` in the Transform component.
WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.6/site-packages/tfx/components/transform/executor.py:541: Schema (from tensorflow_transform.tf_metadata.dataset_schema) is deprecated and will be removed in a future version.
Instructions for updating:
Schema is a deprecated, use schema_utils.schema_from_feature_spec to create a `Schema`
INFO:absl:Loading /tmpfs/src/temp/model_card_toolkit/documentation/examples/census_income_transform.py because it has not been loaded before.
INFO:absl:/tmpfs/src/temp/model_card_toolkit/documentation/examples/census_income_transform.py is already loaded.
INFO:absl:Feature Age has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Capital-Gain has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Capital-Loss has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Country has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Education has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Education-Num has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Hours-per-week has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Marital-Status has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Occupation has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Over-50K has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Race has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Relationship has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Sex has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Workclass has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature fnlwgt has no shape. Setting to VarLenSparseTensor.
WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.6/site-packages/tensorflow_transform/tf_utils.py:261: Tensor.experimental_ref (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use ref() instead.
INFO:absl:Feature Age has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Capital-Gain has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Capital-Loss has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Country has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Education has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Education-Num has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Hours-per-week has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Marital-Status has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Occupation has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Over-50K has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Race has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Relationship has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Sex has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Workclass has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature fnlwgt has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Age has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Capital-Gain has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Capital-Loss has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Country has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Education has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Education-Num has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Hours-per-week has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Marital-Status has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Occupation has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Over-50K has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Race has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Relationship has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Sex has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Workclass has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature fnlwgt has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Age has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Capital-Gain has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Capital-Loss has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Country has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Education has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Education-Num has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Hours-per-week has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Marital-Status has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Occupation has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Over-50K has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Race has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Relationship has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Sex has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Workclass has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature fnlwgt has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Age has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Capital-Gain has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Capital-Loss has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Country has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Education has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Education-Num has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Hours-per-week has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Marital-Status has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Occupation has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Over-50K has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Race has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Relationship has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Sex has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Workclass has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature fnlwgt has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Age has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Capital-Gain has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Capital-Loss has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Country has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Education has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Education-Num has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Hours-per-week has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Marital-Status has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Occupation has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Over-50K has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Race has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Relationship has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Sex has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Workclass has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature fnlwgt has no shape. Setting to VarLenSparseTensor.
WARNING:tensorflow:TFT beam APIs accept both the TFXIO format and the instance dict format now. There is no need to set use_tfxio any more and it will be removed soon.
WARNING:root:This output type hint will be ignored and not used for type-checking purposes. Typically, output type hints for a PTransform are single (or nested) types wrapped by a PCollection, PDone, or None. Got: Tuple[Dict[str, Union[NoneType, _Dataset]], Union[Dict[str, Dict[str, PCollection]], NoneType]] instead.
WARNING:root:This output type hint will be ignored and not used for type-checking purposes. Typically, output type hints for a PTransform are single (or nested) types wrapped by a PCollection, PDone, or None. Got: Tuple[Dict[str, Union[NoneType, _Dataset]], Union[Dict[str, Dict[str, PCollection]], NoneType]] instead.
WARNING:tensorflow:Tensorflow version (2.3.2) found. Note that Tensorflow Transform support for TF 2.0 is currently in beta, and features such as tf.function may not work as intended. 
WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.6/site-packages/tensorflow/python/saved_model/signature_def_utils_impl.py:201: 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:Assets added to graph.
INFO:tensorflow:No assets to write.
WARNING:tensorflow:Issue encountered when serializing tft_mapper_use.
Type is unsupported, or the types of the items don't match field type in CollectionDef. Note this is a warning and probably safe to ignore.
'Counter' object has no attribute 'name'
INFO:tensorflow:SavedModel written to: /tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/Transform/transform_graph/4/.temp_path/tftransform_tmp/259914d385c64e718981569626d0274c/saved_model.pb
INFO:tensorflow:Assets added to graph.
INFO:tensorflow:No assets to write.
WARNING:tensorflow:Issue encountered when serializing tft_mapper_use.
Type is unsupported, or the types of the items don't match field type in CollectionDef. Note this is a warning and probably safe to ignore.
'Counter' object has no attribute 'name'
INFO:tensorflow:SavedModel written to: /tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/Transform/transform_graph/4/.temp_path/tftransform_tmp/5a8d2f32189a42faa1a8db83c514c74e/saved_model.pb
INFO:absl:Feature Age has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Capital-Gain has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Capital-Loss has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Country has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Education has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Education-Num has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Hours-per-week has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Marital-Status has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Occupation has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Over-50K has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Race has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Relationship has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Sex has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Workclass has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature fnlwgt has no shape. Setting to VarLenSparseTensor.
WARNING:tensorflow:Tensorflow version (2.3.2) found. Note that Tensorflow Transform support for TF 2.0 is currently in beta, and features such as tf.function may not work as intended.
WARNING:apache_beam.typehints.typehints:Ignoring send_type hint: <class 'NoneType'>
WARNING:apache_beam.typehints.typehints:Ignoring return_type hint: <class 'NoneType'>
WARNING:apache_beam.typehints.typehints:Ignoring send_type hint: <class 'NoneType'>
WARNING:apache_beam.typehints.typehints:Ignoring return_type hint: <class 'NoneType'>
WARNING:apache_beam.typehints.typehints:Ignoring send_type hint: <class 'NoneType'>
WARNING:apache_beam.typehints.typehints:Ignoring return_type hint: <class 'NoneType'>
INFO:absl:Feature Age has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Capital-Gain has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Capital-Loss has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Country has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Education has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Education-Num has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Hours-per-week has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Marital-Status has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Occupation has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Over-50K has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Race has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Relationship has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Sex has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature Workclass has no shape. Setting to VarLenSparseTensor.
INFO:absl:Feature fnlwgt has no shape. Setting to VarLenSparseTensor.
WARNING:tensorflow:Tensorflow version (2.3.2) found. Note that Tensorflow Transform support for TF 2.0 is currently in beta, and features such as tf.function may not work as intended.
WARNING:apache_beam.typehints.typehints:Ignoring send_type hint: <class 'NoneType'>
WARNING:apache_beam.typehints.typehints:Ignoring return_type hint: <class 'NoneType'>
WARNING:apache_beam.typehints.typehints:Ignoring send_type hint: <class 'NoneType'>
WARNING:apache_beam.typehints.typehints:Ignoring return_type hint: <class 'NoneType'>
WARNING:apache_beam.typehints.typehints:Ignoring send_type hint: <class 'NoneType'>
WARNING:apache_beam.typehints.typehints:Ignoring return_type hint: <class 'NoneType'>
INFO:tensorflow:Saver not created because there are no variables in the graph to restore
INFO:tensorflow:Saver not created because there are no variables in the graph to restore
INFO:tensorflow:Assets added to graph.
INFO:tensorflow:Assets written to: /tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/Transform/transform_graph/4/.temp_path/tftransform_tmp/f4e3edcb37534b60889ae56fa82df09f/assets
INFO:tensorflow:SavedModel written to: /tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/Transform/transform_graph/4/.temp_path/tftransform_tmp/f4e3edcb37534b60889ae56fa82df09f/saved_model.pb
WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\013\n\tConst_2:0\022-vocab_compute_and_apply_vocabulary_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\013\n\tConst_4:0\022/vocab_compute_and_apply_vocabulary_1_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\013\n\tConst_6:0\022/vocab_compute_and_apply_vocabulary_2_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\013\n\tConst_8:0\022/vocab_compute_and_apply_vocabulary_3_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\014\n\nConst_10:0\022/vocab_compute_and_apply_vocabulary_4_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\014\n\nConst_12:0\022/vocab_compute_and_apply_vocabulary_5_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\014\n\nConst_14:0\022/vocab_compute_and_apply_vocabulary_6_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\014\n\nConst_16:0\022/vocab_compute_and_apply_vocabulary_7_vocabulary"

INFO:tensorflow:Saver not created because there are no variables in the graph to restore
WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\013\n\tConst_2:0\022-vocab_compute_and_apply_vocabulary_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\013\n\tConst_4:0\022/vocab_compute_and_apply_vocabulary_1_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\013\n\tConst_6:0\022/vocab_compute_and_apply_vocabulary_2_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\013\n\tConst_8:0\022/vocab_compute_and_apply_vocabulary_3_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\014\n\nConst_10:0\022/vocab_compute_and_apply_vocabulary_4_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\014\n\nConst_12:0\022/vocab_compute_and_apply_vocabulary_5_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\014\n\nConst_14:0\022/vocab_compute_and_apply_vocabulary_6_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\014\n\nConst_16:0\022/vocab_compute_and_apply_vocabulary_7_vocabulary"

INFO:tensorflow:Saver not created because there are no variables in the graph to restore
WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\013\n\tConst_2:0\022-vocab_compute_and_apply_vocabulary_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\013\n\tConst_4:0\022/vocab_compute_and_apply_vocabulary_1_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\013\n\tConst_6:0\022/vocab_compute_and_apply_vocabulary_2_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\013\n\tConst_8:0\022/vocab_compute_and_apply_vocabulary_3_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\014\n\nConst_10:0\022/vocab_compute_and_apply_vocabulary_4_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\014\n\nConst_12:0\022/vocab_compute_and_apply_vocabulary_5_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\014\n\nConst_14:0\022/vocab_compute_and_apply_vocabulary_6_vocabulary"

WARNING:tensorflow:Expected binary or unicode string, got type_url: "type.googleapis.com/tensorflow.AssetFileDef"
value: "\n\014\n\nConst_16:0\022/vocab_compute_and_apply_vocabulary_7_vocabulary"

INFO:tensorflow:Saver not created because there are no variables in the graph to restore
INFO:absl:Running publisher for Transform
INFO:absl:MetadataStore with DB connection initialized

transform.outputs

{'transform_graph': Channel(
    type_name: TransformGraph
    artifacts: [Artifact(artifact: id: 4
type_id: 11
uri: "/tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/Transform/transform_graph/4"
custom_properties {
  key: "name"
  value {
    string_value: "transform_graph"
  }
}
custom_properties {
  key: "producer_component"
  value {
    string_value: "Transform"
  }
}
custom_properties {
  key: "state"
  value {
    string_value: "published"
  }
}
state: LIVE
, artifact_type: id: 11
name: "TransformGraph"
)]
), 'transformed_examples': Channel(
    type_name: Examples
    artifacts: [Artifact(artifact: id: 5
type_id: 5
uri: "/tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/Transform/transformed_examples/4"
properties {
  key: "split_names"
  value {
    string_value: "[\"train\", \"eval\"]"
  }
}
custom_properties {
  key: "name"
  value {
    string_value: "transformed_examples"
  }
}
custom_properties {
  key: "producer_component"
  value {
    string_value: "Transform"
  }
}
custom_properties {
  key: "state"
  value {
    string_value: "published"
  }
}
state: LIVE
, artifact_type: id: 5
name: "Examples"
properties {
  key: "span"
  value: INT
}
properties {
  key: "split_names"
  value: STRING
}
properties {
  key: "version"
  value: INT
}
)]
), 'updated_analyzer_cache': Channel(
    type_name: TransformCache
    artifacts: [Artifact(artifact: id: 6
type_id: 12
uri: "/tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/Transform/updated_analyzer_cache/4"
custom_properties {
  key: "name"
  value {
    string_value: "updated_analyzer_cache"
  }
}
custom_properties {
  key: "producer_component"
  value {
    string_value: "Transform"
  }
}
custom_properties {
  key: "state"
  value {
    string_value: "published"
  }
}
state: LIVE
, artifact_type: id: 12
name: "TransformCache"
)]
)}

Тренер

Давайте рассмотрим пример пользовательского кода модели ниже (для введения в TensorFlow Keras APIs, можно найти в учебнике ):

_census_income_trainer_module_file = 'census_income_trainer.py'

%%writefile {_census_income_trainer_module_file}

from typing import List, Text

import os
import absl
import datetime
import tensorflow as tf
import tensorflow_transform as tft

from tfx.components.trainer.executor import TrainerFnArgs

import census_income_constants

_DENSE_FLOAT_FEATURE_KEYS = census_income_constants.DENSE_FLOAT_FEATURE_KEYS
_VOCAB_FEATURE_KEYS = census_income_constants.VOCAB_FEATURE_KEYS
_VOCAB_SIZE = census_income_constants.VOCAB_SIZE
_OOV_SIZE = census_income_constants.OOV_SIZE
_FEATURE_BUCKET_COUNT = census_income_constants.FEATURE_BUCKET_COUNT
_BUCKET_FEATURE_KEYS = census_income_constants.BUCKET_FEATURE_KEYS
_CATEGORICAL_FEATURE_KEYS = census_income_constants.CATEGORICAL_FEATURE_KEYS
_MAX_CATEGORICAL_FEATURE_VALUES = census_income_constants.MAX_CATEGORICAL_FEATURE_VALUES
_LABEL_KEY = census_income_constants.LABEL_KEY
_transformed_name = census_income_constants.transformed_name


def _transformed_names(keys):
  return [_transformed_name(key) for key in keys]


def _gzip_reader_fn(filenames):
  """Small utility returning a record reader that can read gzip'ed files."""
  return tf.data.TFRecordDataset(
      filenames,
      compression_type='GZIP')


def _get_serve_tf_examples_fn(model, tf_transform_output):
  """Returns a function that parses a serialized tf.Example and applies TFT."""

  model.tft_layer = tf_transform_output.transform_features_layer()

  @tf.function
  def serve_tf_examples_fn(serialized_tf_examples):
    """Returns the output to be used in the serving signature."""
    feature_spec = tf_transform_output.raw_feature_spec()
    feature_spec.pop(_LABEL_KEY)
    parsed_features = tf.io.parse_example(serialized_tf_examples, feature_spec)

    transformed_features = model.tft_layer(parsed_features)
    if _transformed_name(_LABEL_KEY) in transformed_features:
      transformed_features.pop(_transformed_name(_LABEL_KEY))

    return model(transformed_features)

  return serve_tf_examples_fn


def _input_fn(file_pattern: List[Text],
              tf_transform_output: tft.TFTransformOutput,
              batch_size: int = 200) -> tf.data.Dataset:
  """Generates features and label for tuning/training.

  Args:
    file_pattern: List of paths or patterns of input tfrecord files.
    tf_transform_output: A TFTransformOutput.
    batch_size: representing the number of consecutive elements of returned
      dataset to combine in a single batch

  Returns:
    A dataset that contains (features, indices) tuple where features is a
      dictionary of Tensors, and indices is a single Tensor of label indices.
  """
  transformed_feature_spec = (
      tf_transform_output.transformed_feature_spec().copy())

  dataset = tf.data.experimental.make_batched_features_dataset(
      file_pattern=file_pattern,
      batch_size=batch_size,
      features=transformed_feature_spec,
      reader=_gzip_reader_fn,
      label_key=_transformed_name(_LABEL_KEY))

  return dataset


def _build_keras_model(hidden_units: List[int] = None) -> tf.keras.Model:
  """Creates a DNN Keras model.

  Args:
    hidden_units: [int], the layer sizes of the DNN (input layer first).

  Returns:
    A keras Model.
  """
  real_valued_columns = [
      tf.feature_column.numeric_column(key, shape=())
      for key in _transformed_names(_DENSE_FLOAT_FEATURE_KEYS)
  ]
  categorical_columns = [
      tf.feature_column.categorical_column_with_identity(
          key, num_buckets=_VOCAB_SIZE + _OOV_SIZE, default_value=0)
      for key in _transformed_names(_VOCAB_FEATURE_KEYS)
  ]
  categorical_columns += [
      tf.feature_column.categorical_column_with_identity(
          key, num_buckets=_FEATURE_BUCKET_COUNT, default_value=0)
      for key in _transformed_names(_BUCKET_FEATURE_KEYS)
  ]
  categorical_columns += [
      tf.feature_column.categorical_column_with_identity(  # pylint: disable=g-complex-comprehension
          key,
          num_buckets=num_buckets,
          default_value=0) for key, num_buckets in zip(
              _transformed_names(_CATEGORICAL_FEATURE_KEYS),
              _MAX_CATEGORICAL_FEATURE_VALUES)
  ]
  indicator_column = [
      tf.feature_column.indicator_column(categorical_column)
      for categorical_column in categorical_columns
  ]

  model = _wide_and_deep_classifier(
      # TODO(b/139668410) replace with premade wide_and_deep keras model
      wide_columns=indicator_column,
      deep_columns=real_valued_columns,
      dnn_hidden_units=hidden_units or [100, 70, 50, 25])
  return model


def _wide_and_deep_classifier(wide_columns, deep_columns, dnn_hidden_units):
  """Build a simple keras wide and deep model.

  Args:
    wide_columns: Feature columns wrapped in indicator_column for wide (linear)
      part of the model.
    deep_columns: Feature columns for deep part of the model.
    dnn_hidden_units: [int], the layer sizes of the hidden DNN.

  Returns:
    A Wide and Deep Keras model
  """
  # Following values are hard coded for simplicity in this example,
  # However prefarably they should be passsed in as hparams.

  # Keras needs the feature definitions at compile time.
  # TODO(b/139081439): Automate generation of input layers from FeatureColumn.
  input_layers = {
      colname: tf.keras.layers.Input(name=colname, shape=(), dtype=tf.float32)
      for colname in _transformed_names(_DENSE_FLOAT_FEATURE_KEYS)
  }
  input_layers.update({
      colname: tf.keras.layers.Input(name=colname, shape=(), dtype='int32')
      for colname in _transformed_names(_VOCAB_FEATURE_KEYS)
  })
  input_layers.update({
      colname: tf.keras.layers.Input(name=colname, shape=(), dtype='int32')
      for colname in _transformed_names(_BUCKET_FEATURE_KEYS)
  })
  input_layers.update({
      colname: tf.keras.layers.Input(name=colname, shape=(), dtype='int32')
      for colname in _transformed_names(_CATEGORICAL_FEATURE_KEYS)
  })

  # TODO(b/161816639): SparseFeatures for feature columns + Keras.
  deep = tf.keras.layers.DenseFeatures(deep_columns)(input_layers)
  for numnodes in dnn_hidden_units:
    deep = tf.keras.layers.Dense(numnodes)(deep)
  wide = tf.keras.layers.DenseFeatures(wide_columns)(input_layers)

  output = tf.keras.layers.Dense(
      1, activation='sigmoid')(
          tf.keras.layers.concatenate([deep, wide]))

  model = tf.keras.Model(input_layers, output)
  model.compile(
      loss='binary_crossentropy',
      optimizer=tf.keras.optimizers.Adam(lr=0.001),
      metrics=[tf.keras.metrics.BinaryAccuracy()])
  model.summary(print_fn=absl.logging.info)
  return model


# TFX Trainer will call this function.
def run_fn(fn_args: TrainerFnArgs):
  """Train the model based on given args.

  Args:
    fn_args: Holds args used to train the model as name/value pairs.
  """
  # Number of nodes in the first layer of the DNN
  first_dnn_layer_size = 100
  num_dnn_layers = 4
  dnn_decay_factor = 0.7

  tf_transform_output = tft.TFTransformOutput(fn_args.transform_output)

  train_dataset = _input_fn(fn_args.train_files, tf_transform_output, 40)
  eval_dataset = _input_fn(fn_args.eval_files, tf_transform_output, 40)

  model = _build_keras_model(
      # Construct layers sizes with exponetial decay
      hidden_units=[
          max(2, int(first_dnn_layer_size * dnn_decay_factor**i))
          for i in range(num_dnn_layers)
      ])

  # This log path might change in the future.
  log_dir = os.path.join(os.path.dirname(fn_args.serving_model_dir), 'logs')
  tensorboard_callback = tf.keras.callbacks.TensorBoard(
      log_dir=log_dir, update_freq='batch')
  model.fit(
      train_dataset,
      steps_per_epoch=fn_args.train_steps,
      validation_data=eval_dataset,
      validation_steps=fn_args.eval_steps,
      callbacks=[tensorboard_callback])

  signatures = {
      'serving_default':
          _get_serve_tf_examples_fn(model,
                                    tf_transform_output).get_concrete_function(
                                        tf.TensorSpec(
                                            shape=[None],
                                            dtype=tf.string,
                                            name='examples')),
  }
  model.save(fn_args.serving_model_dir, save_format='tf', signatures=signatures)

Writing census_income_trainer.py

trainer = Trainer(
    module_file=os.path.abspath(_census_income_trainer_module_file),
    custom_executor_spec=executor_spec.ExecutorClassSpec(GenericExecutor),
    examples=transform.outputs['transformed_examples'],
    transform_graph=transform.outputs['transform_graph'],
    schema=schema_gen.outputs['schema'],
    train_args=trainer_pb2.TrainArgs(num_steps=100),
    eval_args=trainer_pb2.EvalArgs(num_steps=50))
context.run(trainer)

WARNING:absl:From <ipython-input-1-6ab3bbf2f5a0>:3: The name tfx.components.base.executor_spec.ExecutorClassSpec is deprecated. Please use tfx.dsl.components.base.executor_spec.ExecutorClassSpec instead.
INFO:absl:Running driver for Trainer
INFO:absl:MetadataStore with DB connection initialized
INFO:absl:Running executor for Trainer
INFO:absl:Train on the 'train' split when train_args.splits is not set.
INFO:absl:Evaluate on the 'eval' split when eval_args.splits is not set.
WARNING:absl:Examples artifact does not have payload_format custom property. Falling back to FORMAT_TF_EXAMPLE
WARNING:absl:Examples artifact does not have payload_format custom property. Falling back to FORMAT_TF_EXAMPLE
INFO:absl:Loading /tmpfs/src/temp/model_card_toolkit/documentation/examples/census_income_trainer.py because it has not been loaded before.
INFO:absl:Training model.
INFO:absl:Model: "functional_1"
INFO:absl:__________________________________________________________________________________________________
INFO:absl:Layer (type)                    Output Shape         Param #     Connected to                     
INFO:absl:==================================================================================================
INFO:absl:Age_xf (InputLayer)             [(None,)]            0                                            
INFO:absl:__________________________________________________________________________________________________
INFO:absl:Capital-Gain_xf (InputLayer)    [(None,)]            0                                            
INFO:absl:__________________________________________________________________________________________________
INFO:absl:Capital-Loss_xf (InputLayer)    [(None,)]            0                                            
INFO:absl:__________________________________________________________________________________________________
INFO:absl:Country_xf (InputLayer)         [(None,)]            0                                            
INFO:absl:__________________________________________________________________________________________________
INFO:absl:Education-Num_xf (InputLayer)   [(None,)]            0                                            
INFO:absl:__________________________________________________________________________________________________
INFO:absl:Education_xf (InputLayer)       [(None,)]            0                                            
INFO:absl:__________________________________________________________________________________________________
INFO:absl:Hours-per-week_xf (InputLayer)  [(None,)]            0                                            
INFO:absl:__________________________________________________________________________________________________
INFO:absl:Marital-Status_xf (InputLayer)  [(None,)]            0                                            
INFO:absl:__________________________________________________________________________________________________
INFO:absl:Occupation_xf (InputLayer)      [(None,)]            0                                            
INFO:absl:__________________________________________________________________________________________________
INFO:absl:Race_xf (InputLayer)            [(None,)]            0                                            
INFO:absl:__________________________________________________________________________________________________
INFO:absl:Relationship_xf (InputLayer)    [(None,)]            0                                            
INFO:absl:__________________________________________________________________________________________________
INFO:absl:Sex_xf (InputLayer)             [(None,)]            0                                            
INFO:absl:__________________________________________________________________________________________________
INFO:absl:Workclass_xf (InputLayer)       [(None,)]            0                                            
INFO:absl:__________________________________________________________________________________________________
INFO:absl:dense_features (DenseFeatures)  (None, 3)            0           Age_xf[0][0]                     
INFO:absl:                                                                 Capital-Gain_xf[0][0]            
INFO:absl:                                                                 Capital-Loss_xf[0][0]            
INFO:absl:                                                                 Country_xf[0][0]                 
INFO:absl:                                                                 Education-Num_xf[0][0]           
INFO:absl:                                                                 Education_xf[0][0]               
INFO:absl:                                                                 Hours-per-week_xf[0][0]          
INFO:absl:                                                                 Marital-Status_xf[0][0]          
INFO:absl:                                                                 Occupation_xf[0][0]              
INFO:absl:                                                                 Race_xf[0][0]                    
INFO:absl:                                                                 Relationship_xf[0][0]            
INFO:absl:                                                                 Sex_xf[0][0]                     
INFO:absl:                                                                 Workclass_xf[0][0]               
INFO:absl:__________________________________________________________________________________________________
INFO:absl:dense (Dense)                   (None, 100)          400         dense_features[0][0]             
INFO:absl:__________________________________________________________________________________________________
INFO:absl:dense_1 (Dense)                 (None, 70)           7070        dense[0][0]                      
INFO:absl:__________________________________________________________________________________________________
INFO:absl:dense_2 (Dense)                 (None, 48)           3408        dense_1[0][0]                    
INFO:absl:__________________________________________________________________________________________________
INFO:absl:dense_3 (Dense)                 (None, 34)           1666        dense_2[0][0]                    
INFO:absl:__________________________________________________________________________________________________
INFO:absl:dense_features_1 (DenseFeatures (None, 1710)         0           Age_xf[0][0]                     
INFO:absl:                                                                 Capital-Gain_xf[0][0]            
INFO:absl:                                                                 Capital-Loss_xf[0][0]            
INFO:absl:                                                                 Country_xf[0][0]                 
INFO:absl:                                                                 Education-Num_xf[0][0]           
INFO:absl:                                                                 Education_xf[0][0]               
INFO:absl:                                                                 Hours-per-week_xf[0][0]          
INFO:absl:                                                                 Marital-Status_xf[0][0]          
INFO:absl:                                                                 Occupation_xf[0][0]              
INFO:absl:                                                                 Race_xf[0][0]                    
INFO:absl:                                                                 Relationship_xf[0][0]            
INFO:absl:                                                                 Sex_xf[0][0]                     
INFO:absl:                                                                 Workclass_xf[0][0]               
INFO:absl:__________________________________________________________________________________________________
INFO:absl:concatenate (Concatenate)       (None, 1744)         0           dense_3[0][0]                    
INFO:absl:                                                                 dense_features_1[0][0]           
INFO:absl:__________________________________________________________________________________________________
INFO:absl:dense_4 (Dense)                 (None, 1)            1745        concatenate[0][0]                
INFO:absl:==================================================================================================
INFO:absl:Total params: 14,289
INFO:absl:Trainable params: 14,289
INFO:absl:Non-trainable params: 0
INFO:absl:__________________________________________________________________________________________________
1/100 [..............................] - ETA: 0s - loss: 0.7236 - binary_accuracy: 0.2250WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.6/site-packages/tensorflow/python/ops/summary_ops_v2.py:1277: stop (from tensorflow.python.eager.profiler) is deprecated and will be removed after 2020-07-01.
Instructions for updating:
use `tf.profiler.experimental.stop` instead.
WARNING:tensorflow:Callbacks method `on_train_batch_end` is slow compared to the batch time (batch time: 0.0029s vs `on_train_batch_end` time: 0.0135s). Check your callbacks.
100/100 [==============================] - 1s 9ms/step - loss: 0.5104 - binary_accuracy: 0.7710 - val_loss: 0.4469 - val_binary_accuracy: 0.8005
INFO:tensorflow:Saver not created because there are no variables in the graph to restore
WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.6/site-packages/tensorflow/python/training/tracking/tracking.py:111: Model.state_updates (from tensorflow.python.keras.engine.training) is deprecated and will be removed in a future version.
Instructions for updating:
This property should not be used in TensorFlow 2.0, as updates are applied automatically.
WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.6/site-packages/tensorflow/python/training/tracking/tracking.py:111: Layer.updates (from tensorflow.python.keras.engine.base_layer) is deprecated and will be removed in a future version.
Instructions for updating:
This property should not be used in TensorFlow 2.0, as updates are applied automatically.
INFO:tensorflow:Assets written to: /tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/Trainer/model/5/serving_model_dir/assets
INFO:absl:Training complete. Model written to /tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/Trainer/model/5/serving_model_dir. ModelRun written to /tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/Trainer/model_run/5
INFO:absl:Running publisher for Trainer
INFO:absl:MetadataStore with DB connection initialized

Оценщик

Evaluator компонент вычисляет модели метрик производительности по сравнению с набором оценки. Он использует TensorFlow модель анализа библиотеки.

Evaluator будет принимать в качестве входных данных из ExampleGen , обученной модели от Trainer и конфигурации нарезки. Конфигурация среза позволяет вам нарезать метрики по значениям функций. См. пример этой конфигурации ниже:

from google.protobuf.wrappers_pb2 import BoolValue

eval_config = tfma.EvalConfig(
    model_specs=[
        # This assumes a serving model with signature 'serving_default'. If
        # using estimator based EvalSavedModel, add signature_name: 'eval' and 
        # remove the label_key.
        tfma.ModelSpec(label_key="Over-50K")
    ],
    metrics_specs=[
        tfma.MetricsSpec(
            # The metrics added here are in addition to those saved with the
            # model (assuming either a keras model or EvalSavedModel is used).
            # Any metrics added into the saved model (for example using
            # model.compile(..., metrics=[...]), etc) will be computed
            # automatically.
            # To add validation thresholds for metrics saved with the model,
            # add them keyed by metric name to the thresholds map.
            metrics=[
                tfma.MetricConfig(class_name='ExampleCount'),
                tfma.MetricConfig(class_name='BinaryAccuracy'),
                tfma.MetricConfig(class_name='FairnessIndicators',
                                  config='{ "thresholds": [0.5] }'),
            ]
        )
    ],
    slicing_specs=[
        # An empty slice spec means the overall slice, i.e. the whole dataset.
        tfma.SlicingSpec(),
        # Data can be sliced along a feature column. In this case, data is
        # sliced by feature column Race and Sex.
        tfma.SlicingSpec(feature_keys=['Race']),
        tfma.SlicingSpec(feature_keys=['Sex']),
        tfma.SlicingSpec(feature_keys=['Race', 'Sex']),
    ],
    options = tfma.Options(compute_confidence_intervals=BoolValue(value=True))
)

# Use TFMA to compute a evaluation statistics over features of a model and
# validate them against a baseline.
evaluator = Evaluator(
    examples=example_gen.outputs['examples'],
    model=trainer.outputs['model'],
    eval_config=eval_config)
context.run(evaluator)

INFO:absl:Running driver for Evaluator
INFO:absl:MetadataStore with DB connection initialized
INFO:absl:Running executor for Evaluator
WARNING:absl:"maybe_add_baseline" and "maybe_remove_baseline" are deprecated,
        please use "has_baseline" instead.
INFO:absl:Request was made to ignore the baseline ModelSpec and any change thresholds. This is likely because a baseline model was not provided: updated_config=
model_specs {
  label_key: "Over-50K"
}
slicing_specs {
}
slicing_specs {
  feature_keys: "Race"
}
slicing_specs {
  feature_keys: "Sex"
}
slicing_specs {
  feature_keys: "Race"
  feature_keys: "Sex"
}
metrics_specs {
  metrics {
    class_name: "ExampleCount"
  }
  metrics {
    class_name: "BinaryAccuracy"
  }
  metrics {
    class_name: "FairnessIndicators"
    config: "{ \"thresholds\": [0.5] }"
  }
}
options {
  compute_confidence_intervals {
    value: true
  }
  confidence_intervals {
    method: JACKKNIFE
  }
}

INFO:absl:Using /tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/Trainer/model/5/serving_model_dir as  model.
INFO:absl:The 'example_splits' parameter is not set, using 'eval' split.
INFO:absl:Evaluating model.
INFO:absl:Request was made to ignore the baseline ModelSpec and any change thresholds. This is likely because a baseline model was not provided: updated_config=
model_specs {
  label_key: "Over-50K"
}
slicing_specs {
}
slicing_specs {
  feature_keys: "Race"
}
slicing_specs {
  feature_keys: "Sex"
}
slicing_specs {
  feature_keys: "Race"
  feature_keys: "Sex"
}
metrics_specs {
  metrics {
    class_name: "ExampleCount"
  }
  metrics {
    class_name: "BinaryAccuracy"
  }
  metrics {
    class_name: "FairnessIndicators"
    config: "{ \"thresholds\": [0.5] }"
  }
  model_names: ""
}
options {
  compute_confidence_intervals {
    value: true
  }
  confidence_intervals {
    method: JACKKNIFE
  }
}

INFO:absl:Request was made to ignore the baseline ModelSpec and any change thresholds. This is likely because a baseline model was not provided: updated_config=
model_specs {
  label_key: "Over-50K"
}
slicing_specs {
}
slicing_specs {
  feature_keys: "Race"
}
slicing_specs {
  feature_keys: "Sex"
}
slicing_specs {
  feature_keys: "Race"
  feature_keys: "Sex"
}
metrics_specs {
  metrics {
    class_name: "ExampleCount"
  }
  metrics {
    class_name: "BinaryAccuracy"
  }
  metrics {
    class_name: "FairnessIndicators"
    config: "{ \"thresholds\": [0.5] }"
  }
  model_names: ""
}
options {
  compute_confidence_intervals {
    value: true
  }
  confidence_intervals {
    method: JACKKNIFE
  }
}

INFO:absl:Request was made to ignore the baseline ModelSpec and any change thresholds. This is likely because a baseline model was not provided: updated_config=
model_specs {
  label_key: "Over-50K"
}
slicing_specs {
}
slicing_specs {
  feature_keys: "Race"
}
slicing_specs {
  feature_keys: "Sex"
}
slicing_specs {
  feature_keys: "Race"
  feature_keys: "Sex"
}
metrics_specs {
  metrics {
    class_name: "ExampleCount"
  }
  metrics {
    class_name: "BinaryAccuracy"
  }
  metrics {
    class_name: "FairnessIndicators"
    config: "{ \"thresholds\": [0.5] }"
  }
  model_names: ""
}
options {
  compute_confidence_intervals {
    value: true
  }
  confidence_intervals {
    method: JACKKNIFE
  }
}
WARNING:tensorflow:5 out of the last 5 calls to <function recreate_function.<locals>.restored_function_body at 0x7f77a02d2510> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.
INFO:absl:Evaluation complete. Results written to /tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/Evaluator/evaluation/6.
INFO:absl:No threshold configured, will not validate model.
INFO:absl:Running publisher for Evaluator
INFO:absl:MetadataStore with DB connection initialized

evaluator.outputs

{'evaluation': Channel(
    type_name: ModelEvaluation
    artifacts: [Artifact(artifact: id: 9
type_id: 17
uri: "/tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/Evaluator/evaluation/6"
custom_properties {
  key: "name"
  value {
    string_value: "evaluation"
  }
}
custom_properties {
  key: "producer_component"
  value {
    string_value: "Evaluator"
  }
}
custom_properties {
  key: "state"
  value {
    string_value: "published"
  }
}
state: LIVE
, artifact_type: id: 17
name: "ModelEvaluation"
)]
), 'blessing': Channel(
    type_name: ModelBlessing
    artifacts: [Artifact(artifact: id: 10
type_id: 18
uri: "/tmp/tfx-interactive-2021-02-05T22_01_17.129037-teowo3b1/Evaluator/blessing/6"
custom_properties {
  key: "name"
  value {
    string_value: "blessing"
  }
}
custom_properties {
  key: "producer_component"
  value {
    string_value: "Evaluator"
  }
}
custom_properties {
  key: "state"
  value {
    string_value: "published"
  }
}
state: LIVE
, artifact_type: id: 18
name: "ModelBlessing"
)]
)}

Используя evaluation вывода мы можем показать визуализацию по умолчанию глобальных показателей по всей совокупности оценок.

context.show(evaluator.outputs['evaluation'])

Заполнение свойств из ModelCard с помощью Model Card Toolkit

Теперь, когда мы настроили наш конвейер TFX, мы будем использовать набор инструментов Model Card Toolkit для извлечения ключевых артефактов из запуска и заполнения Model Card.

Подключитесь к хранилищу MLMD, используемому InteractiveContext.

from ml_metadata.metadata_store import metadata_store
from IPython import display

mlmd_store = metadata_store.MetadataStore(context.metadata_connection_config)
model_uri = trainer.outputs["model"].get()[0].uri

INFO:absl:MetadataStore with DB connection initialized

Используйте инструментарий карты модели

Инициализируйте набор инструментов Model Card.

from model_card_toolkit import ModelCardToolkit

mct = ModelCardToolkit(mlmd_store=mlmd_store, model_uri=model_uri)

Создайте рабочую область карты модели.

model_card = mct.scaffold_assets()

Аннотируйте больше информации в карточке модели.

Также важно документировать информацию о модели, которая может быть важна для последующих пользователей, например, ее ограничения, предполагаемые варианты использования, компромиссы и этические соображения. Для каждого из этих разделов мы можем напрямую добавить новые объекты JSON для представления этой информации.

model_card.model_details.name = 'Census Income Classifier'
model_card.model_details.overview = (
    'This is a wide and deep Keras model which aims to classify whether or not '
    'an individual has an income of over $50,000 based on various demographic '
    'features. The model is trained on the UCI Census Income Dataset. This is '
    'not a production model, and this dataset has traditionally only been used '
    'for research purposes. In this Model Card, you can review quantitative '
    'components of the model’s performance and data, as well as information '
    'about the model’s intended uses, limitations, and ethical considerations.'
)
model_card.model_details.owners = [
  {'name': 'Model Cards Team', 'contact': 'model-cards@google.com'}
]

model_card.considerations.use_cases = [
    'This dataset that this model was trained on was originally created to '
    'support the machine learning community in conducting empirical analysis '
    'of ML algorithms. The Adult Data Set can be used in fairness-related '
    'studies that compare inequalities across sex and race, based on '
    'people’s annual incomes.'
]
model_card.considerations.limitations = [
    'This is a class-imbalanced dataset across a variety of sensitive classes.'
    ' The ratio of male-to-female examples is about 2:1 and there are far more'
    ' examples with the “white” attribute than every other race combined. '
    'Furthermore, the ratio of $50,000 or less earners to $50,000 or more '
    'earners is just over 3:1. Due to the imbalance across income levels, we '
    'can see that our true negative rate seems quite high, while our true '
    'positive rate seems quite low. This is true to an even greater degree '
    'when we only look at the “female” sub-group, because there are even '
    'fewer female examples in the $50,000+ earner group, causing our model to '
    'overfit these examples. To avoid this, we can try various remediation '
    'strategies in future iterations (e.g. undersampling, hyperparameter '
    'tuning, etc), but we may not be able to fix all of the fairness issues.'
]
model_card.considerations.ethical_considerations = [{
    'name':
        'We risk expressing the viewpoint that the attributes in this dataset '
        'are the only ones that are predictive of someone’s income, even '
        'though we know this is not the case.',
    'mitigation_strategy':
        'As mentioned, some interventions may need to be performed to address '
        'the class imbalances in the dataset.'
}]

Фильтровать и добавлять графики.

Мы можем отфильтровать графики, созданные компонентами TFX, чтобы включить те, которые наиболее важны для карты модели, используя функцию, определенную ниже. В этом примере мы фильтруем для race и sex , два потенциально чувствительных атрибутов.

Каждая карточка модели будет иметь до трех разделов для графиков: статистика набора данных для обучения, статистика набора данных для оценки и количественный анализ производительности нашей модели.

# These are the graphs that will appear in the Quantiative Analysis portion of 
# the Model Card. Feel free to add or remove from this list. 
TARGET_EVAL_GRAPH_NAMES = [
  'fairness_indicators_metrics/false_positive_rate@0.5',
  'fairness_indicators_metrics/false_negative_rate@0.5',
  'binary_accuracy',
  'example_count | Race_X_Sex',
]

# These are the graphs that will appear in both the Train Set and Eval Set 
# portions of the Model Card. Feel free to add or remove from this list. 
TARGET_DATASET_GRAPH_NAMES = [
  'counts | Race',
  'counts | Sex',
]

def filter_graphs(graphics, target_graph_names):
  result = []
  for graph in graphics:
    for target_graph_name in target_graph_names:
      if graph.name.startswith(target_graph_name):
        result.append(graph)
  result.sort(key=lambda g: g.name)
  return result

# Populating the three different sections using the filter defined above. To 
# see all the graphs available in a section, we can iterate through each of the
# different collections. 
model_card.quantitative_analysis.graphics.collection = filter_graphs(
    model_card.quantitative_analysis.graphics.collection, TARGET_EVAL_GRAPH_NAMES)
model_card.model_parameters.data.eval.graphics.collection = filter_graphs(
    model_card.model_parameters.data.eval.graphics.collection, TARGET_DATASET_GRAPH_NAMES)
model_card.model_parameters.data.train.graphics.collection = filter_graphs(
    model_card.model_parameters.data.train.graphics.collection, TARGET_DATASET_GRAPH_NAMES)

Затем мы добавляем (необязательно) описания для каждого из каждого из разделов графика.

model_card.model_parameters.data.train.graphics.description = (
    'This section includes graphs displaying the class distribution for the '
    '“Race” and “Sex” attributes in our training dataset. We chose to '
    'show these graphs in particular because we felt it was important that '
    'users see the class imbalance.'
)
model_card.model_parameters.data.eval.graphics.description = (
    'Like the training set, we provide graphs showing the class distribution '
    'of the data we used to evaluate our model’s performance. '
)
model_card.quantitative_analysis.graphics.description = (
    'These graphs show how the model performs for data sliced by “Race”, '
    '“Sex” and the intersection of these attributes. The metrics we chose '
    'to display are “Accuracy”, “False Positive Rate”, and “False '
    'Negative Rate”, because we anticipated that the class imbalances might '
    'cause our model to underperform for certain groups.'
)

mct.update_model_card_json(model_card)

Создайте карточку модели.

Теперь мы можем отобразить карточку модели в формате HTML.

html = mct.export_format()

display.display(display.HTML(html))