public static final class SmoothGradConfig.Builder extends GeneratedMessageV3.Builder<SmoothGradConfig.Builder> implements SmoothGradConfigOrBuilder
Config for SmoothGrad approximation of gradients.
When enabled, the gradients are approximated by averaging the gradients from
noisy samples in the vicinity of the inputs. Adding noise can help improve
the computed gradients. Refer to this paper for more details:
https://arxiv.org/pdf/1706.03825.pdf
Protobuf type google.cloud.aiplatform.v1.SmoothGradConfig
Inherited Members
com.google.protobuf.GeneratedMessageV3.Builder.getUnknownFieldSetBuilder()
com.google.protobuf.GeneratedMessageV3.Builder.mergeUnknownLengthDelimitedField(int,com.google.protobuf.ByteString)
com.google.protobuf.GeneratedMessageV3.Builder.mergeUnknownVarintField(int,int)
com.google.protobuf.GeneratedMessageV3.Builder.parseUnknownField(com.google.protobuf.CodedInputStream,com.google.protobuf.ExtensionRegistryLite,int)
com.google.protobuf.GeneratedMessageV3.Builder.setUnknownFieldSetBuilder(com.google.protobuf.UnknownFieldSet.Builder)
Static Methods
public static final Descriptors.Descriptor getDescriptor()
Methods
public SmoothGradConfig.Builder addRepeatedField(Descriptors.FieldDescriptor field, Object value)
Overrides
public SmoothGradConfig build()
public SmoothGradConfig buildPartial()
public SmoothGradConfig.Builder clear()
Overrides
public SmoothGradConfig.Builder clearFeatureNoiseSigma()
This is similar to
noise_sigma,
but provides additional flexibility. A separate noise sigma can be
provided for each feature, which is useful if their distributions are
different. No noise is added to features that are not set. If this field
is unset,
noise_sigma
will be used for all features.
.google.cloud.aiplatform.v1.FeatureNoiseSigma feature_noise_sigma = 2;
public SmoothGradConfig.Builder clearField(Descriptors.FieldDescriptor field)
Overrides
public SmoothGradConfig.Builder clearGradientNoiseSigma()
public SmoothGradConfig.Builder clearNoiseSigma()
This is a single float value and will be used to add noise to all the
features. Use this field when all features are normalized to have the
same distribution: scale to range [0, 1], [-1, 1] or z-scoring, where
features are normalized to have 0-mean and 1-variance. Learn more about
normalization.
For best results the recommended value is about 10% - 20% of the standard
deviation of the input feature. Refer to section 3.2 of the SmoothGrad
paper: https://arxiv.org/pdf/1706.03825.pdf. Defaults to 0.1.
If the distribution is different per feature, set
feature_noise_sigma
instead for each feature.
float noise_sigma = 1;
public SmoothGradConfig.Builder clearNoisySampleCount()
The number of gradient samples to use for
approximation. The higher this number, the more accurate the gradient
is, but the runtime complexity increases by this factor as well.
Valid range of its value is [1, 50]. Defaults to 3.
int32 noisy_sample_count = 3;
public SmoothGradConfig.Builder clearOneof(Descriptors.OneofDescriptor oneof)
Overrides
public SmoothGradConfig.Builder clone()
Overrides
public SmoothGradConfig getDefaultInstanceForType()
public Descriptors.Descriptor getDescriptorForType()
Overrides
public FeatureNoiseSigma getFeatureNoiseSigma()
This is similar to
noise_sigma,
but provides additional flexibility. A separate noise sigma can be
provided for each feature, which is useful if their distributions are
different. No noise is added to features that are not set. If this field
is unset,
noise_sigma
will be used for all features.
.google.cloud.aiplatform.v1.FeatureNoiseSigma feature_noise_sigma = 2;
public FeatureNoiseSigma.Builder getFeatureNoiseSigmaBuilder()
This is similar to
noise_sigma,
but provides additional flexibility. A separate noise sigma can be
provided for each feature, which is useful if their distributions are
different. No noise is added to features that are not set. If this field
is unset,
noise_sigma
will be used for all features.
.google.cloud.aiplatform.v1.FeatureNoiseSigma feature_noise_sigma = 2;
public FeatureNoiseSigmaOrBuilder getFeatureNoiseSigmaOrBuilder()
This is similar to
noise_sigma,
but provides additional flexibility. A separate noise sigma can be
provided for each feature, which is useful if their distributions are
different. No noise is added to features that are not set. If this field
is unset,
noise_sigma
will be used for all features.
.google.cloud.aiplatform.v1.FeatureNoiseSigma feature_noise_sigma = 2;
public SmoothGradConfig.GradientNoiseSigmaCase getGradientNoiseSigmaCase()
public float getNoiseSigma()
This is a single float value and will be used to add noise to all the
features. Use this field when all features are normalized to have the
same distribution: scale to range [0, 1], [-1, 1] or z-scoring, where
features are normalized to have 0-mean and 1-variance. Learn more about
normalization.
For best results the recommended value is about 10% - 20% of the standard
deviation of the input feature. Refer to section 3.2 of the SmoothGrad
paper: https://arxiv.org/pdf/1706.03825.pdf. Defaults to 0.1.
If the distribution is different per feature, set
feature_noise_sigma
instead for each feature.
float noise_sigma = 1;
Returns |
---|
Type | Description |
float | The noiseSigma.
|
public int getNoisySampleCount()
The number of gradient samples to use for
approximation. The higher this number, the more accurate the gradient
is, but the runtime complexity increases by this factor as well.
Valid range of its value is [1, 50]. Defaults to 3.
int32 noisy_sample_count = 3;
Returns |
---|
Type | Description |
int | The noisySampleCount.
|
public boolean hasFeatureNoiseSigma()
This is similar to
noise_sigma,
but provides additional flexibility. A separate noise sigma can be
provided for each feature, which is useful if their distributions are
different. No noise is added to features that are not set. If this field
is unset,
noise_sigma
will be used for all features.
.google.cloud.aiplatform.v1.FeatureNoiseSigma feature_noise_sigma = 2;
Returns |
---|
Type | Description |
boolean | Whether the featureNoiseSigma field is set.
|
public boolean hasNoiseSigma()
This is a single float value and will be used to add noise to all the
features. Use this field when all features are normalized to have the
same distribution: scale to range [0, 1], [-1, 1] or z-scoring, where
features are normalized to have 0-mean and 1-variance. Learn more about
normalization.
For best results the recommended value is about 10% - 20% of the standard
deviation of the input feature. Refer to section 3.2 of the SmoothGrad
paper: https://arxiv.org/pdf/1706.03825.pdf. Defaults to 0.1.
If the distribution is different per feature, set
feature_noise_sigma
instead for each feature.
float noise_sigma = 1;
Returns |
---|
Type | Description |
boolean | Whether the noiseSigma field is set.
|
protected GeneratedMessageV3.FieldAccessorTable internalGetFieldAccessorTable()
Overrides
public final boolean isInitialized()
Overrides
public SmoothGradConfig.Builder mergeFeatureNoiseSigma(FeatureNoiseSigma value)
This is similar to
noise_sigma,
but provides additional flexibility. A separate noise sigma can be
provided for each feature, which is useful if their distributions are
different. No noise is added to features that are not set. If this field
is unset,
noise_sigma
will be used for all features.
.google.cloud.aiplatform.v1.FeatureNoiseSigma feature_noise_sigma = 2;
public SmoothGradConfig.Builder mergeFrom(SmoothGradConfig other)
public SmoothGradConfig.Builder mergeFrom(CodedInputStream input, ExtensionRegistryLite extensionRegistry)
Overrides
public SmoothGradConfig.Builder mergeFrom(Message other)
Parameter |
---|
Name | Description |
other | Message
|
Overrides
public final SmoothGradConfig.Builder mergeUnknownFields(UnknownFieldSet unknownFields)
Overrides
public SmoothGradConfig.Builder setFeatureNoiseSigma(FeatureNoiseSigma value)
This is similar to
noise_sigma,
but provides additional flexibility. A separate noise sigma can be
provided for each feature, which is useful if their distributions are
different. No noise is added to features that are not set. If this field
is unset,
noise_sigma
will be used for all features.
.google.cloud.aiplatform.v1.FeatureNoiseSigma feature_noise_sigma = 2;
public SmoothGradConfig.Builder setFeatureNoiseSigma(FeatureNoiseSigma.Builder builderForValue)
This is similar to
noise_sigma,
but provides additional flexibility. A separate noise sigma can be
provided for each feature, which is useful if their distributions are
different. No noise is added to features that are not set. If this field
is unset,
noise_sigma
will be used for all features.
.google.cloud.aiplatform.v1.FeatureNoiseSigma feature_noise_sigma = 2;
public SmoothGradConfig.Builder setField(Descriptors.FieldDescriptor field, Object value)
Overrides
public SmoothGradConfig.Builder setNoiseSigma(float value)
This is a single float value and will be used to add noise to all the
features. Use this field when all features are normalized to have the
same distribution: scale to range [0, 1], [-1, 1] or z-scoring, where
features are normalized to have 0-mean and 1-variance. Learn more about
normalization.
For best results the recommended value is about 10% - 20% of the standard
deviation of the input feature. Refer to section 3.2 of the SmoothGrad
paper: https://arxiv.org/pdf/1706.03825.pdf. Defaults to 0.1.
If the distribution is different per feature, set
feature_noise_sigma
instead for each feature.
float noise_sigma = 1;
Parameter |
---|
Name | Description |
value | float
The noiseSigma to set.
|
public SmoothGradConfig.Builder setNoisySampleCount(int value)
The number of gradient samples to use for
approximation. The higher this number, the more accurate the gradient
is, but the runtime complexity increases by this factor as well.
Valid range of its value is [1, 50]. Defaults to 3.
int32 noisy_sample_count = 3;
Parameter |
---|
Name | Description |
value | int
The noisySampleCount to set.
|
public SmoothGradConfig.Builder setRepeatedField(Descriptors.FieldDescriptor field, int index, Object value)
Overrides
public final SmoothGradConfig.Builder setUnknownFields(UnknownFieldSet unknownFields)
Overrides