Class StudySpec.ConvexAutomatedStoppingSpec.Builder (3.28.0)

public static final class StudySpec.ConvexAutomatedStoppingSpec.Builder extends GeneratedMessageV3.Builder<StudySpec.ConvexAutomatedStoppingSpec.Builder> implements StudySpec.ConvexAutomatedStoppingSpecOrBuilder

Configuration for ConvexAutomatedStoppingSpec. When there are enough completed trials (configured by min_measurement_count), for pending trials with enough measurements and steps, the policy first computes an overestimate of the objective value at max_num_steps according to the slope of the incomplete objective value curve. No prediction can be made if the curve is completely flat. If the overestimation is worse than the best objective value of the completed trials, this pending trial will be early-stopped, but a last measurement will be added to the pending trial with max_num_steps and predicted objective value from the autoregression model.

Protobuf type google.cloud.aiplatform.v1.StudySpec.ConvexAutomatedStoppingSpec

Static Methods

getDescriptor()

public static final Descriptors.Descriptor getDescriptor()
Returns
TypeDescription
Descriptor

Methods

addRepeatedField(Descriptors.FieldDescriptor field, Object value)

public StudySpec.ConvexAutomatedStoppingSpec.Builder addRepeatedField(Descriptors.FieldDescriptor field, Object value)
Parameters
NameDescription
fieldFieldDescriptor
valueObject
Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder
Overrides

build()

public StudySpec.ConvexAutomatedStoppingSpec build()
Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec

buildPartial()

public StudySpec.ConvexAutomatedStoppingSpec buildPartial()
Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec

clear()

public StudySpec.ConvexAutomatedStoppingSpec.Builder clear()
Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder
Overrides

clearField(Descriptors.FieldDescriptor field)

public StudySpec.ConvexAutomatedStoppingSpec.Builder clearField(Descriptors.FieldDescriptor field)
Parameter
NameDescription
fieldFieldDescriptor
Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder
Overrides

clearLearningRateParameterName()

public StudySpec.ConvexAutomatedStoppingSpec.Builder clearLearningRateParameterName()

The hyper-parameter name used in the tuning job that stands for learning rate. Leave it blank if learning rate is not in a parameter in tuning. The learning_rate is used to estimate the objective value of the ongoing trial.

string learning_rate_parameter_name = 4;

Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder

This builder for chaining.

clearMaxStepCount()

public StudySpec.ConvexAutomatedStoppingSpec.Builder clearMaxStepCount()

Steps used in predicting the final objective for early stopped trials. In general, it's set to be the same as the defined steps in training / tuning. If not defined, it will learn it from the completed trials. When use_steps is false, this field is set to the maximum elapsed seconds.

int64 max_step_count = 1;

Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder

This builder for chaining.

clearMinMeasurementCount()

public StudySpec.ConvexAutomatedStoppingSpec.Builder clearMinMeasurementCount()

The minimal number of measurements in a Trial. Early-stopping checks will not trigger if less than min_measurement_count+1 completed trials or pending trials with less than min_measurement_count measurements. If not defined, the default value is 5.

int64 min_measurement_count = 3;

Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder

This builder for chaining.

clearMinStepCount()

public StudySpec.ConvexAutomatedStoppingSpec.Builder clearMinStepCount()

Minimum number of steps for a trial to complete. Trials which do not have a measurement with step_count > min_step_count won't be considered for early stopping. It's ok to set it to 0, and a trial can be early stopped at any stage. By default, min_step_count is set to be one-tenth of the max_step_count. When use_elapsed_duration is true, this field is set to the minimum elapsed seconds.

int64 min_step_count = 2;

Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder

This builder for chaining.

clearOneof(Descriptors.OneofDescriptor oneof)

public StudySpec.ConvexAutomatedStoppingSpec.Builder clearOneof(Descriptors.OneofDescriptor oneof)
Parameter
NameDescription
oneofOneofDescriptor
Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder
Overrides

clearUpdateAllStoppedTrials()

public StudySpec.ConvexAutomatedStoppingSpec.Builder clearUpdateAllStoppedTrials()

ConvexAutomatedStoppingSpec by default only updates the trials that needs to be early stopped using a newly trained auto-regressive model. When this flag is set to True, all stopped trials from the beginning are potentially updated in terms of their final_measurement. Also, note that the training logic of autoregressive models is different in this case. Enabling this option has shown better results and this may be the default option in the future.

optional bool update_all_stopped_trials = 6;

Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder

This builder for chaining.

clearUseElapsedDuration()

public StudySpec.ConvexAutomatedStoppingSpec.Builder clearUseElapsedDuration()

This bool determines whether or not the rule is applied based on elapsed_secs or steps. If use_elapsed_duration==false, the early stopping decision is made according to the predicted objective values according to the target steps. If use_elapsed_duration==true, elapsed_secs is used instead of steps. Also, in this case, the parameters max_num_steps and min_num_steps are overloaded to contain max_elapsed_seconds and min_elapsed_seconds.

bool use_elapsed_duration = 5;

Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder

This builder for chaining.

clone()

public StudySpec.ConvexAutomatedStoppingSpec.Builder clone()
Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder
Overrides

getDefaultInstanceForType()

public StudySpec.ConvexAutomatedStoppingSpec getDefaultInstanceForType()
Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec

getDescriptorForType()

public Descriptors.Descriptor getDescriptorForType()
Returns
TypeDescription
Descriptor
Overrides

getLearningRateParameterName()

public String getLearningRateParameterName()

The hyper-parameter name used in the tuning job that stands for learning rate. Leave it blank if learning rate is not in a parameter in tuning. The learning_rate is used to estimate the objective value of the ongoing trial.

string learning_rate_parameter_name = 4;

Returns
TypeDescription
String

The learningRateParameterName.

getLearningRateParameterNameBytes()

public ByteString getLearningRateParameterNameBytes()

The hyper-parameter name used in the tuning job that stands for learning rate. Leave it blank if learning rate is not in a parameter in tuning. The learning_rate is used to estimate the objective value of the ongoing trial.

string learning_rate_parameter_name = 4;

Returns
TypeDescription
ByteString

The bytes for learningRateParameterName.

getMaxStepCount()

public long getMaxStepCount()

Steps used in predicting the final objective for early stopped trials. In general, it's set to be the same as the defined steps in training / tuning. If not defined, it will learn it from the completed trials. When use_steps is false, this field is set to the maximum elapsed seconds.

int64 max_step_count = 1;

Returns
TypeDescription
long

The maxStepCount.

getMinMeasurementCount()

public long getMinMeasurementCount()

The minimal number of measurements in a Trial. Early-stopping checks will not trigger if less than min_measurement_count+1 completed trials or pending trials with less than min_measurement_count measurements. If not defined, the default value is 5.

int64 min_measurement_count = 3;

Returns
TypeDescription
long

The minMeasurementCount.

getMinStepCount()

public long getMinStepCount()

Minimum number of steps for a trial to complete. Trials which do not have a measurement with step_count > min_step_count won't be considered for early stopping. It's ok to set it to 0, and a trial can be early stopped at any stage. By default, min_step_count is set to be one-tenth of the max_step_count. When use_elapsed_duration is true, this field is set to the minimum elapsed seconds.

int64 min_step_count = 2;

Returns
TypeDescription
long

The minStepCount.

getUpdateAllStoppedTrials()

public boolean getUpdateAllStoppedTrials()

ConvexAutomatedStoppingSpec by default only updates the trials that needs to be early stopped using a newly trained auto-regressive model. When this flag is set to True, all stopped trials from the beginning are potentially updated in terms of their final_measurement. Also, note that the training logic of autoregressive models is different in this case. Enabling this option has shown better results and this may be the default option in the future.

optional bool update_all_stopped_trials = 6;

Returns
TypeDescription
boolean

The updateAllStoppedTrials.

getUseElapsedDuration()

public boolean getUseElapsedDuration()

This bool determines whether or not the rule is applied based on elapsed_secs or steps. If use_elapsed_duration==false, the early stopping decision is made according to the predicted objective values according to the target steps. If use_elapsed_duration==true, elapsed_secs is used instead of steps. Also, in this case, the parameters max_num_steps and min_num_steps are overloaded to contain max_elapsed_seconds and min_elapsed_seconds.

bool use_elapsed_duration = 5;

Returns
TypeDescription
boolean

The useElapsedDuration.

hasUpdateAllStoppedTrials()

public boolean hasUpdateAllStoppedTrials()

ConvexAutomatedStoppingSpec by default only updates the trials that needs to be early stopped using a newly trained auto-regressive model. When this flag is set to True, all stopped trials from the beginning are potentially updated in terms of their final_measurement. Also, note that the training logic of autoregressive models is different in this case. Enabling this option has shown better results and this may be the default option in the future.

optional bool update_all_stopped_trials = 6;

Returns
TypeDescription
boolean

Whether the updateAllStoppedTrials field is set.

internalGetFieldAccessorTable()

protected GeneratedMessageV3.FieldAccessorTable internalGetFieldAccessorTable()
Returns
TypeDescription
FieldAccessorTable
Overrides

isInitialized()

public final boolean isInitialized()
Returns
TypeDescription
boolean
Overrides

mergeFrom(StudySpec.ConvexAutomatedStoppingSpec other)

public StudySpec.ConvexAutomatedStoppingSpec.Builder mergeFrom(StudySpec.ConvexAutomatedStoppingSpec other)
Parameter
NameDescription
otherStudySpec.ConvexAutomatedStoppingSpec
Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder

mergeFrom(CodedInputStream input, ExtensionRegistryLite extensionRegistry)

public StudySpec.ConvexAutomatedStoppingSpec.Builder mergeFrom(CodedInputStream input, ExtensionRegistryLite extensionRegistry)
Parameters
NameDescription
inputCodedInputStream
extensionRegistryExtensionRegistryLite
Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder
Overrides
Exceptions
TypeDescription
IOException

mergeFrom(Message other)

public StudySpec.ConvexAutomatedStoppingSpec.Builder mergeFrom(Message other)
Parameter
NameDescription
otherMessage
Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder
Overrides

mergeUnknownFields(UnknownFieldSet unknownFields)

public final StudySpec.ConvexAutomatedStoppingSpec.Builder mergeUnknownFields(UnknownFieldSet unknownFields)
Parameter
NameDescription
unknownFieldsUnknownFieldSet
Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder
Overrides

setField(Descriptors.FieldDescriptor field, Object value)

public StudySpec.ConvexAutomatedStoppingSpec.Builder setField(Descriptors.FieldDescriptor field, Object value)
Parameters
NameDescription
fieldFieldDescriptor
valueObject
Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder
Overrides

setLearningRateParameterName(String value)

public StudySpec.ConvexAutomatedStoppingSpec.Builder setLearningRateParameterName(String value)

The hyper-parameter name used in the tuning job that stands for learning rate. Leave it blank if learning rate is not in a parameter in tuning. The learning_rate is used to estimate the objective value of the ongoing trial.

string learning_rate_parameter_name = 4;

Parameter
NameDescription
valueString

The learningRateParameterName to set.

Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder

This builder for chaining.

setLearningRateParameterNameBytes(ByteString value)

public StudySpec.ConvexAutomatedStoppingSpec.Builder setLearningRateParameterNameBytes(ByteString value)

The hyper-parameter name used in the tuning job that stands for learning rate. Leave it blank if learning rate is not in a parameter in tuning. The learning_rate is used to estimate the objective value of the ongoing trial.

string learning_rate_parameter_name = 4;

Parameter
NameDescription
valueByteString

The bytes for learningRateParameterName to set.

Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder

This builder for chaining.

setMaxStepCount(long value)

public StudySpec.ConvexAutomatedStoppingSpec.Builder setMaxStepCount(long value)

Steps used in predicting the final objective for early stopped trials. In general, it's set to be the same as the defined steps in training / tuning. If not defined, it will learn it from the completed trials. When use_steps is false, this field is set to the maximum elapsed seconds.

int64 max_step_count = 1;

Parameter
NameDescription
valuelong

The maxStepCount to set.

Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder

This builder for chaining.

setMinMeasurementCount(long value)

public StudySpec.ConvexAutomatedStoppingSpec.Builder setMinMeasurementCount(long value)

The minimal number of measurements in a Trial. Early-stopping checks will not trigger if less than min_measurement_count+1 completed trials or pending trials with less than min_measurement_count measurements. If not defined, the default value is 5.

int64 min_measurement_count = 3;

Parameter
NameDescription
valuelong

The minMeasurementCount to set.

Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder

This builder for chaining.

setMinStepCount(long value)

public StudySpec.ConvexAutomatedStoppingSpec.Builder setMinStepCount(long value)

Minimum number of steps for a trial to complete. Trials which do not have a measurement with step_count > min_step_count won't be considered for early stopping. It's ok to set it to 0, and a trial can be early stopped at any stage. By default, min_step_count is set to be one-tenth of the max_step_count. When use_elapsed_duration is true, this field is set to the minimum elapsed seconds.

int64 min_step_count = 2;

Parameter
NameDescription
valuelong

The minStepCount to set.

Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder

This builder for chaining.

setRepeatedField(Descriptors.FieldDescriptor field, int index, Object value)

public StudySpec.ConvexAutomatedStoppingSpec.Builder setRepeatedField(Descriptors.FieldDescriptor field, int index, Object value)
Parameters
NameDescription
fieldFieldDescriptor
indexint
valueObject
Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder
Overrides

setUnknownFields(UnknownFieldSet unknownFields)

public final StudySpec.ConvexAutomatedStoppingSpec.Builder setUnknownFields(UnknownFieldSet unknownFields)
Parameter
NameDescription
unknownFieldsUnknownFieldSet
Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder
Overrides

setUpdateAllStoppedTrials(boolean value)

public StudySpec.ConvexAutomatedStoppingSpec.Builder setUpdateAllStoppedTrials(boolean value)

ConvexAutomatedStoppingSpec by default only updates the trials that needs to be early stopped using a newly trained auto-regressive model. When this flag is set to True, all stopped trials from the beginning are potentially updated in terms of their final_measurement. Also, note that the training logic of autoregressive models is different in this case. Enabling this option has shown better results and this may be the default option in the future.

optional bool update_all_stopped_trials = 6;

Parameter
NameDescription
valueboolean

The updateAllStoppedTrials to set.

Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder

This builder for chaining.

setUseElapsedDuration(boolean value)

public StudySpec.ConvexAutomatedStoppingSpec.Builder setUseElapsedDuration(boolean value)

This bool determines whether or not the rule is applied based on elapsed_secs or steps. If use_elapsed_duration==false, the early stopping decision is made according to the predicted objective values according to the target steps. If use_elapsed_duration==true, elapsed_secs is used instead of steps. Also, in this case, the parameters max_num_steps and min_num_steps are overloaded to contain max_elapsed_seconds and min_elapsed_seconds.

bool use_elapsed_duration = 5;

Parameter
NameDescription
valueboolean

The useElapsedDuration to set.

Returns
TypeDescription
StudySpec.ConvexAutomatedStoppingSpec.Builder

This builder for chaining.