experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Update keras feature column test to explicitly use keras symbols.

Browse Source 
PiperOrigin-RevId: 334662640
Change-Id: Iaf0b99006622b2d9d0549cb9be7e59f915c0102c
This commit is contained in:
Scott Zhu 2020-09-30 13:07:29 -07:00 committed by TensorFlower Gardener
parent f98fc1387d
commit 9a8b1c6d50
1 changed files with 9 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
tensorflow/python/keras/engine/feature_columns_integration_test.py
View File

@ -26,6 +26,7 @@ from tensorflow.python.feature_column import feature_column_lib as fc
from tensorflow.python.keras import keras_parameterized from tensorflow.python.keras import keras_parameterized
from tensorflow.python.keras import metrics as metrics_module from tensorflow.python.keras import metrics as metrics_module
from tensorflow.python.keras import testing_utils from tensorflow.python.keras import testing_utils
from tensorflow.python.keras.feature_column import dense_features as df
from tensorflow.python.keras.utils import np_utils from tensorflow.python.keras.utils import np_utils
from tensorflow.python.platform import test from tensorflow.python.platform import test
@ -34,7 +35,7 @@ class TestDNNModel(keras.models.Model):
def __init__(self, feature_columns, units, name=None, **kwargs): def __init__(self, feature_columns, units, name=None, **kwargs):
super(TestDNNModel, self).__init__(name=name, **kwargs) super(TestDNNModel, self).__init__(name=name, **kwargs)
self._input_layer = fc.DenseFeatures(feature_columns, name='input_layer') self._input_layer = df.DenseFeatures(feature_columns, name='input_layer')
self._dense_layer = keras.layers.Dense(units, name='dense_layer') self._dense_layer = keras.layers.Dense(units, name='dense_layer')
def call(self, features): def call(self, features):
@ -52,7 +53,7 @@ class FeatureColumnsIntegrationTest(keras_parameterized.TestCase):
def test_sequential_model(self): def test_sequential_model(self):
columns = [fc.numeric_column('a')] columns = [fc.numeric_column('a')]
model = keras.models.Sequential([ model = keras.models.Sequential([
fc.DenseFeatures(columns), df.DenseFeatures(columns),
keras.layers.Dense(64, activation='relu'), keras.layers.Dense(64, activation='relu'),
keras.layers.Dense(20, activation='softmax') keras.layers.Dense(20, activation='softmax')
]) ])
@ -74,7 +75,7 @@ class FeatureColumnsIntegrationTest(keras_parameterized.TestCase):
def test_sequential_model_with_ds_input(self): def test_sequential_model_with_ds_input(self):
columns = [fc.numeric_column('a')] columns = [fc.numeric_column('a')]
model = keras.models.Sequential([ model = keras.models.Sequential([
fc.DenseFeatures(columns), df.DenseFeatures(columns),
keras.layers.Dense(64, activation='relu'), keras.layers.Dense(64, activation='relu'),
keras.layers.Dense(20, activation='softmax') keras.layers.Dense(20, activation='softmax')
]) ])
@ -112,7 +113,7 @@ class FeatureColumnsIntegrationTest(keras_parameterized.TestCase):
crossed_feature = fc.indicator_column(crossed_feature) crossed_feature = fc.indicator_column(crossed_feature)
feature_columns.append(crossed_feature) feature_columns.append(crossed_feature)
feature_layer = fc.DenseFeatures(feature_columns) feature_layer = df.DenseFeatures(feature_columns)
model = keras.models.Sequential([ model = keras.models.Sequential([
feature_layer, feature_layer,
@ -185,7 +186,7 @@ class FeatureColumnsIntegrationTest(keras_parameterized.TestCase):
col_a = fc.numeric_column('a') col_a = fc.numeric_column('a')
col_b = fc.numeric_column('b') col_b = fc.numeric_column('b')
feature_layer = fc.DenseFeatures([col_a, col_b], name='fc') feature_layer = df.DenseFeatures([col_a, col_b], name='fc')
dense = keras.layers.Dense(4) dense = keras.layers.Dense(4)
# This seems problematic.... We probably need something for DenseFeatures # This seems problematic.... We probably need something for DenseFeatures
@ -213,8 +214,8 @@ class FeatureColumnsIntegrationTest(keras_parameterized.TestCase):
col_b = fc.numeric_column('b') col_b = fc.numeric_column('b')
col_c = fc.numeric_column('c') col_c = fc.numeric_column('c')
fc1 = fc.DenseFeatures([col_a, col_b], name='fc1') fc1 = df.DenseFeatures([col_a, col_b], name='fc1')
fc2 = fc.DenseFeatures([col_b, col_c], name='fc2') fc2 = df.DenseFeatures([col_b, col_c], name='fc2')
dense = keras.layers.Dense(4) dense = keras.layers.Dense(4)
# This seems problematic.... We probably need something for DenseFeatures # This seems problematic.... We probably need something for DenseFeatures
@ -289,7 +290,7 @@ class FeatureColumnsIntegrationTest(keras_parameterized.TestCase):
categorical_cols = [fc.categorical_column_with_hash_bucket('cabin', 10)] categorical_cols = [fc.categorical_column_with_hash_bucket('cabin', 10)]
feature_cols = ([fc.numeric_column('age')] feature_cols = ([fc.numeric_column('age')]
+ [fc.indicator_column(cc) for cc in categorical_cols]) + [fc.indicator_column(cc) for cc in categorical_cols])
layers = [fc.DenseFeatures(feature_cols), layers = [df.DenseFeatures(feature_cols),
keras.layers.Dense(128), keras.layers.Dense(128),
keras.layers.Dense(1)] keras.layers.Dense(1)]