def test_regression_output_tensors_roundtrip(self):
value = array_ops.placeholder(dtypes.float32, 1, name='value')
regression_output = export_lib.RegressionOutput(value)
self.assertSequenceEqual(
[value],
tpu_estimator._export_output_to_tensors(regression_output))
value_new = array_ops.placeholder(dtypes.float32, 1, name='value_new')
regression_output_new = (
tpu_estimator._clone_export_output_with_tensors(
regression_output, [value_new]))
self.assertEqual(value_new, regression_output_new.value)
def test_classification_output_tensors_roundtrip_scores_only(self):
scores = array_ops.placeholder(dtypes.float32, 1, name='scores')
classification_output = export_lib.ClassificationOutput(scores=scores)
classification_output_tensors = (
tpu_estimator._export_output_to_tensors(classification_output))
self.assertEqual(classification_output_tensors, [scores, None])
scores_new = array_ops.placeholder(dtypes.float32,
1,
name='scores_new')
classification_output_new = (
tpu_estimator._clone_export_output_with_tensors(
classification_output, [scores_new, None]))
self.assertEqual(classification_output_new.scores, scores_new)
def test_classification_output_tensors_roundtrip_classes_only(self):
classes = array_ops.placeholder(dtypes.string, 1, name='classes')
classification_output = export_lib.ClassificationOutput(
classes=classes)
classification_output_tensors = (
tpu_estimator._export_output_to_tensors(classification_output))
self.assertEqual(classification_output_tensors, [None, classes])
classes_new = array_ops.placeholder(dtypes.string,
1,
name='classes_new')
classification_output_new = (
tpu_estimator._clone_export_output_with_tensors(
classification_output, [None, classes_new]))
self.assertEqual(classification_output_new.classes, classes_new)
def test_classification_output_tensors_roundtrip_classify_both(self):
classes = array_ops.placeholder(dtypes.string, 1, name='classes')
scores = array_ops.placeholder(dtypes.float32, 1, name='scores')
classification_output = export_lib.ClassificationOutput(
scores, classes)
classification_output_tensors = (
tpu_estimator._export_output_to_tensors(classification_output))
self.assertSequenceEqual(classification_output_tensors,
[scores, classes])
classes_new = array_ops.placeholder(dtypes.string,
1,
name='classes_new')
scores_new = array_ops.placeholder(dtypes.float32,
1,
name='scores_new')
classification_output_new = (
tpu_estimator._clone_export_output_with_tensors(
classification_output, [scores_new, classes_new]))
self.assertEqual(classification_output_new.classes, classes_new)
self.assertEqual(classification_output_new.scores, scores_new)
def test_predict_output_tensors_roundtrip(self):
value1 = array_ops.placeholder(dtypes.float32, 1, name='value1')
value2 = array_ops.placeholder(dtypes.float32, 1, name='value2')
predict_output = export_lib.PredictOutput({
'value1': value1,
'value2': value2
})
export_output_tensors = tpu_estimator._export_output_to_tensors(
predict_output)
self.assertSameElements([value1, value2], export_output_tensors)
self.assertLen(export_output_tensors, 2)
tensors_new = [
array_ops.identity(t, name=t.name.split(':')[0] + '_new')
for t in export_output_tensors
]
predict_output_new = tpu_estimator._clone_export_output_with_tensors(
predict_output, tensors_new)
outputs = predict_output_new.outputs
self.assertLen(outputs, 2)
self.assertEqual(outputs['value1'].name, 'value1_new:0')
self.assertEqual(outputs['value2'].name, 'value2_new:0')