def test_string_gauge(self):
gauge = monitoring.StringGauge('test/gauge', 'test gauge')
gauge.get_cell().set('left')
self.assertEqual(gauge.get_cell().value(), 'left')
gauge.get_cell().set('right')
self.assertEqual(gauge.get_cell().value(), 'right')
gauge1 = monitoring.StringGauge('test/gauge1', 'test gauge1', 'label1')
gauge1.get_cell('foo').set('start')
self.assertEqual(gauge1.get_cell('foo').value(), 'start')
class TFLiteMetrics(metrics_interface.TFLiteMetricsInterface):
"""TFLite metrics helper for prod (borg) environment.
Attributes:
model_hash: A string containing the hash of the model binary.
model_path: A string containing the path of the model for debugging
purposes.
"""
_counter_debugger_creation = monitoring.Counter(
'/tensorflow/lite/quantization_debugger/created',
'Counter for the number of debugger created.')
_counter_interpreter_creation = monitoring.Counter(
'/tensorflow/lite/interpreter/created',
'Counter for number of interpreter created in Python.', 'language')
# The following are conversion metrics. Attempt and success are kept separated
# instead of using a single metric with a label because the converter may
# raise exceptions if conversion failed. That may lead to cases when we are
# unable to capture the conversion attempt. Increasing attempt count at the
# beginning of conversion process and the success count at the end is more
# suitable in these cases.
_counter_conversion_attempt = monitoring.Counter(
'/tensorflow/lite/convert/attempt',
'Counter for number of conversion attempts.')
_counter_conversion_success = monitoring.Counter(
'/tensorflow/lite/convert/success',
'Counter for number of successful conversions.')
_gauge_conversion_params = monitoring.StringGauge(
'/tensorflow/lite/convert/params',
'Gauge for keeping conversion parameters.', 'name')
def __init__(self,
model_hash: Optional[Text] = None,
model_path: Optional[Text] = None) -> None:
del self # Temporarily removing self until parameter logic is implemented.
if model_hash and not model_path or not model_hash and model_path:
raise ValueError('Both model metadata(model_hash, model_path) should be '
'given at the same time.')
if model_hash:
# TODO(b/180400857): Create stub once the service is implemented.
pass
def increase_counter_debugger_creation(self):
self._counter_debugger_creation.get_cell().increase_by(1)
def increase_counter_interpreter_creation(self):
self._counter_interpreter_creation.get_cell('python').increase_by(1)
def increase_counter_converter_attempt(self):
self._counter_conversion_attempt.get_cell().increase_by(1)
def increase_counter_converter_success(self):
self._counter_conversion_success.get_cell().increase_by(1)
def set_converter_param(self, name, value):
self._gauge_conversion_params.get_cell(name).set(value)
from tensorflow.python.eager import monitoring
from tensorflow.python.framework import dtypes
from tensorflow.python.framework import errors
from tensorflow.python.framework import ops
from tensorflow.python.framework import sparse_tensor
from tensorflow.python.keras import backend as K
from tensorflow.python.keras.engine import training_generator
from tensorflow.python.keras.engine.base_layer import Layer
from tensorflow.python.keras.utils import tf_utils
from tensorflow.python.ops import sparse_ops
from tensorflow.python.ops.ragged import ragged_tensor
from tensorflow.python.util.tf_export import keras_export
_kpl_gauge = monitoring.StringGauge(
'/tensorflow/api/keras/layers/preprocessing',
'keras preprocessing layers usage', 'TFVersion')
@keras_export('keras.layers.experimental.preprocessing.PreprocessingLayer')
@six.add_metaclass(abc.ABCMeta)
class PreprocessingLayer(Layer):
"""Base class for PreprocessingLayers."""
_must_restore_from_config = True
def adapt(self, data, reset_state=True):
# TODO(momernick): Add examples.
"""Fits the state of the preprocessing layer to the data being passed.
Arguments:
data: The data to train on. It can be passed either as a tf.data
# The following are conversion metrics. Attempt and success are kept separated
# instead of using a single metric with a label because the converter may
# raise exceptions if conversion failed. That may lead to cases when we are
# unable to capture the conversion attempt. Increasing attempt count at the
# beginning of conversion process and the success count at the end is more
# suitable in these cases.
_counter_conversion_attempt = monitoring.Counter(
'/tensorflow/lite/convert/attempt',
'Counter for number of conversion attempts.')
_counter_conversion_success = monitoring.Counter(
'/tensorflow/lite/convert/success',
'Counter for number of successful conversions.')
_gauge_conversion_params = monitoring.StringGauge(
'/tensorflow/lite/convert/params',
'Gauge for keeping conversion parameters.', 'name')
_gauge_conversion_errors = monitoring.StringGauge(
'/tensorflow/lite/convert/errors',
'Gauge for collecting conversion errors. The value represents the error '
'message.', 'component', 'subcomponent', 'op_name', 'error_code')
_gauge_conversion_latency = monitoring.IntGauge(
'/tensorflow/lite/convert/latency', 'Conversion latency in ms.')
class TFLiteMetrics(metrics_interface.TFLiteMetricsInterface):
"""TFLite metrics helper for prod (borg) environment.
Attributes:
# The following are conversion metrics. Attempt and success are kept separated
# instead of using a single metric with a label because the converter may
# raise exceptions if conversion failed. That may lead to cases when we are
# unable to capture the conversion attempt. Increasing attempt count at the
# beginning of conversion process and the success count at the end is more
# suitable in these cases.
_counter_conversion_attempt = monitoring.Counter(
'/tensorflow/lite/convert/attempt',
'Counter for number of conversion attempts.')
_counter_conversion_success = monitoring.Counter(
'/tensorflow/lite/convert/success',
'Counter for number of successful conversions.')
_gauge_conversion_params = monitoring.StringGauge(
'/tensorflow/lite/convert/params',
'Gauge for keeping conversion parameters.', 'name')
class TFLiteMetrics(metrics_interface.TFLiteMetricsInterface):
"""TFLite metrics helper for prod (borg) environment.
Attributes:
model_hash: A string containing the hash of the model binary.
model_path: A string containing the path of the model for debugging
purposes.
"""
def __init__(self,
model_hash: Optional[Text] = None,
model_path: Optional[Text] = None) -> None:
del self # Temporarily removing self until parameter logic is implemented.