Class IntegratedGradientsAttribution.Builder (0.4.0)

publicstaticfinalclass IntegratedGradientsAttribution.BuilderextendsGeneratedMessageV3.Builder<IntegratedGradientsAttribution.Builder>implementsIntegratedGradientsAttributionOrBuilder

An attribution method that computes the Aumann-Shapley value taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1703.01365

Protobuf type google.cloud.vertexai.v1.IntegratedGradientsAttribution

Inheritance

Object > AbstractMessageLite.Builder<MessageType,BuilderType> > AbstractMessage.Builder<BuilderType> > GeneratedMessageV3.Builder > IntegratedGradientsAttribution.Builder

Implements

IntegratedGradientsAttributionOrBuilder

Inherited Members

publicstaticfinalDescriptors.DescriptorgetDescriptor()

publicIntegratedGradientsAttribution.BuilderaddRepeatedField(Descriptors.FieldDescriptorfield,Objectvalue)

GeneratedMessageV3.Builder<BuilderType>.addRepeatedField(Descriptors.FieldDescriptor field, Object value)

publicIntegratedGradientsAttributionbuild()

publicIntegratedGradientsAttributionbuildPartial()

publicIntegratedGradientsAttribution.Builderclear()

GeneratedMessageV3.Builder<BuilderType>.clear()

publicIntegratedGradientsAttribution.BuilderclearBlurBaselineConfig()

Config for IG with blur baseline.

When enabled, a linear path from the maximally blurred image to the input image is created. Using a blurred baseline instead of zero (black image) is motivated by the BlurIG approach explained here: https://arxiv.org/abs/2004.03383

.google.cloud.vertexai.v1.BlurBaselineConfig blur_baseline_config = 3;

publicIntegratedGradientsAttribution.BuilderclearField(Descriptors.FieldDescriptorfield)

GeneratedMessageV3.Builder<BuilderType>.clearField(Descriptors.FieldDescriptor field)

publicIntegratedGradientsAttribution.BuilderclearOneof(Descriptors.OneofDescriptoroneof)

GeneratedMessageV3.Builder<BuilderType>.clearOneof(Descriptors.OneofDescriptor oneof)

publicIntegratedGradientsAttribution.BuilderclearSmoothGradConfig()

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

.google.cloud.vertexai.v1.SmoothGradConfig smooth_grad_config = 2;

publicIntegratedGradientsAttribution.BuilderclearStepCount()

Required. The number of steps for approximating the path integral. A good value to start is 50 and gradually increase until the sum to diff property is within the desired error range.

Valid range of its value is [1, 100], inclusively.

int32 step_count = 1 [(.google.api.field_behavior) = REQUIRED];

publicIntegratedGradientsAttribution.Builderclone()

GeneratedMessageV3.Builder<BuilderType>.clone()

publicBlurBaselineConfiggetBlurBaselineConfig()

Config for IG with blur baseline.

When enabled, a linear path from the maximally blurred image to the input image is created. Using a blurred baseline instead of zero (black image) is motivated by the BlurIG approach explained here: https://arxiv.org/abs/2004.03383

.google.cloud.vertexai.v1.BlurBaselineConfig blur_baseline_config = 3;

publicBlurBaselineConfig.BuildergetBlurBaselineConfigBuilder()

Config for IG with blur baseline.

When enabled, a linear path from the maximally blurred image to the input image is created. Using a blurred baseline instead of zero (black image) is motivated by the BlurIG approach explained here: https://arxiv.org/abs/2004.03383

.google.cloud.vertexai.v1.BlurBaselineConfig blur_baseline_config = 3;

publicBlurBaselineConfigOrBuildergetBlurBaselineConfigOrBuilder()

Config for IG with blur baseline.

When enabled, a linear path from the maximally blurred image to the input image is created. Using a blurred baseline instead of zero (black image) is motivated by the BlurIG approach explained here: https://arxiv.org/abs/2004.03383

.google.cloud.vertexai.v1.BlurBaselineConfig blur_baseline_config = 3;

publicIntegratedGradientsAttributiongetDefaultInstanceForType()

publicDescriptors.DescriptorgetDescriptorForType()

GeneratedMessageV3.Builder<BuilderType>.getDescriptorForType()

publicSmoothGradConfiggetSmoothGradConfig()

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

.google.cloud.vertexai.v1.SmoothGradConfig smooth_grad_config = 2;

publicSmoothGradConfig.BuildergetSmoothGradConfigBuilder()

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

.google.cloud.vertexai.v1.SmoothGradConfig smooth_grad_config = 2;

publicSmoothGradConfigOrBuildergetSmoothGradConfigOrBuilder()

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

.google.cloud.vertexai.v1.SmoothGradConfig smooth_grad_config = 2;

publicintgetStepCount()

Required. The number of steps for approximating the path integral. A good value to start is 50 and gradually increase until the sum to diff property is within the desired error range.

Valid range of its value is [1, 100], inclusively.

int32 step_count = 1 [(.google.api.field_behavior) = REQUIRED];

publicbooleanhasBlurBaselineConfig()

Config for IG with blur baseline.

When enabled, a linear path from the maximally blurred image to the input image is created. Using a blurred baseline instead of zero (black image) is motivated by the BlurIG approach explained here: https://arxiv.org/abs/2004.03383

.google.cloud.vertexai.v1.BlurBaselineConfig blur_baseline_config = 3;

publicbooleanhasSmoothGradConfig()

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

.google.cloud.vertexai.v1.SmoothGradConfig smooth_grad_config = 2;

protectedGeneratedMessageV3.FieldAccessorTableinternalGetFieldAccessorTable()

GeneratedMessageV3.Builder<BuilderType>.internalGetFieldAccessorTable()

publicfinalbooleanisInitialized()

GeneratedMessageV3.Builder<BuilderType>.isInitialized()

publicIntegratedGradientsAttribution.BuildermergeBlurBaselineConfig(BlurBaselineConfigvalue)

Config for IG with blur baseline.

When enabled, a linear path from the maximally blurred image to the input image is created. Using a blurred baseline instead of zero (black image) is motivated by the BlurIG approach explained here: https://arxiv.org/abs/2004.03383

.google.cloud.vertexai.v1.BlurBaselineConfig blur_baseline_config = 3;

publicIntegratedGradientsAttribution.BuildermergeFrom(IntegratedGradientsAttributionother)

publicIntegratedGradientsAttribution.BuildermergeFrom(CodedInputStreaminput,ExtensionRegistryLiteextensionRegistry)

AbstractMessage.Builder<BuilderType>.mergeFrom(CodedInputStream input, ExtensionRegistryLite extensionRegistry)

publicIntegratedGradientsAttribution.BuildermergeFrom(Messageother)

AbstractMessage.Builder<BuilderType>.mergeFrom(Message other)

publicIntegratedGradientsAttribution.BuildermergeSmoothGradConfig(SmoothGradConfigvalue)

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

.google.cloud.vertexai.v1.SmoothGradConfig smooth_grad_config = 2;

publicfinalIntegratedGradientsAttribution.BuildermergeUnknownFields(UnknownFieldSetunknownFields)

GeneratedMessageV3.Builder<BuilderType>.mergeUnknownFields(UnknownFieldSet unknownFields)

publicIntegratedGradientsAttribution.BuildersetBlurBaselineConfig(BlurBaselineConfigvalue)

Config for IG with blur baseline.

When enabled, a linear path from the maximally blurred image to the input image is created. Using a blurred baseline instead of zero (black image) is motivated by the BlurIG approach explained here: https://arxiv.org/abs/2004.03383

.google.cloud.vertexai.v1.BlurBaselineConfig blur_baseline_config = 3;

publicIntegratedGradientsAttribution.BuildersetBlurBaselineConfig(BlurBaselineConfig.BuilderbuilderForValue)

Config for IG with blur baseline.

When enabled, a linear path from the maximally blurred image to the input image is created. Using a blurred baseline instead of zero (black image) is motivated by the BlurIG approach explained here: https://arxiv.org/abs/2004.03383

.google.cloud.vertexai.v1.BlurBaselineConfig blur_baseline_config = 3;

publicIntegratedGradientsAttribution.BuildersetField(Descriptors.FieldDescriptorfield,Objectvalue)

GeneratedMessageV3.Builder<BuilderType>.setField(Descriptors.FieldDescriptor field, Object value)

publicIntegratedGradientsAttribution.BuildersetRepeatedField(Descriptors.FieldDescriptorfield,intindex,Objectvalue)

GeneratedMessageV3.Builder<BuilderType>.setRepeatedField(Descriptors.FieldDescriptor field, int index, Object value)

publicIntegratedGradientsAttribution.BuildersetSmoothGradConfig(SmoothGradConfigvalue)

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

.google.cloud.vertexai.v1.SmoothGradConfig smooth_grad_config = 2;

publicIntegratedGradientsAttribution.BuildersetSmoothGradConfig(SmoothGradConfig.BuilderbuilderForValue)

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

.google.cloud.vertexai.v1.SmoothGradConfig smooth_grad_config = 2;

publicIntegratedGradientsAttribution.BuildersetStepCount(intvalue)

Required. The number of steps for approximating the path integral. A good value to start is 50 and gradually increase until the sum to diff property is within the desired error range.

Valid range of its value is [1, 100], inclusively.

int32 step_count = 1 [(.google.api.field_behavior) = REQUIRED];

publicfinalIntegratedGradientsAttribution.BuildersetUnknownFields(UnknownFieldSetunknownFields)

GeneratedMessageV3.Builder<BuilderType>.setUnknownFields(UnknownFieldSet unknownFields)