google-apis-language_v2 0.3.0 → 0.5.0

data/lib/google/apis/language_v2/classes.rb

@@ -166,7 +166,7 @@ module Google
         # All available features. Setting each one to true will enable that specific
         # analysis for the input.
         # Corresponds to the JSON property `features`
-        # @return [Google::Apis::LanguageV2::Features]
+        # @return [Google::Apis::LanguageV2::AnnotateTextRequestFeatures]
         attr_accessor :features
       
         def initialize(**args)
@@ -181,6 +181,48 @@ module Google
         end
       end
       
+      # All available features. Setting each one to true will enable that specific
+      # analysis for the input.
+      class AnnotateTextRequestFeatures
+        include Google::Apis::Core::Hashable
+      
+        # Optional. Classify the full document into categories.
+        # Corresponds to the JSON property `classifyText`
+        # @return [Boolean]
+        attr_accessor :classify_text
+        alias_method :classify_text?, :classify_text
+      
+        # Optional. Extract document-level sentiment.
+        # Corresponds to the JSON property `extractDocumentSentiment`
+        # @return [Boolean]
+        attr_accessor :extract_document_sentiment
+        alias_method :extract_document_sentiment?, :extract_document_sentiment
+      
+        # Optional. Extract entities.
+        # Corresponds to the JSON property `extractEntities`
+        # @return [Boolean]
+        attr_accessor :extract_entities
+        alias_method :extract_entities?, :extract_entities
+      
+        # Optional. Moderate the document for harmful and sensitive categories.
+        # Corresponds to the JSON property `moderateText`
+        # @return [Boolean]
+        attr_accessor :moderate_text
+        alias_method :moderate_text?, :moderate_text
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @classify_text = args[:classify_text] if args.key?(:classify_text)
+          @extract_document_sentiment = args[:extract_document_sentiment] if args.key?(:extract_document_sentiment)
+          @extract_entities = args[:extract_entities] if args.key?(:extract_entities)
+          @moderate_text = args[:moderate_text] if args.key?(:moderate_text)
+        end
+      end
+      
       # The text annotations response message.
       class AnnotateTextResponse
         include Google::Apis::Core::Hashable
@@ -324,6 +366,165 @@ module Google
         end
       end
       
+      # Represents a color in the RGBA color space. This representation is designed
+      # for simplicity of conversion to and from color representations in various
+      # languages over compactness. For example, the fields of this representation can
+      # be trivially provided to the constructor of `java.awt.Color` in Java; it can
+      # also be trivially provided to UIColor's `+colorWithRed:green:blue:alpha`
+      # method in iOS; and, with just a little work, it can be easily formatted into a
+      # CSS `rgba()` string in JavaScript. This reference page doesn't have
+      # information about the absolute color space that should be used to interpret
+      # the RGB value—for example, sRGB, Adobe RGB, DCI-P3, and BT.2020. By default,
+      # applications should assume the sRGB color space. When color equality needs to
+      # be decided, implementations, unless documented otherwise, treat two colors as
+      # equal if all their red, green, blue, and alpha values each differ by at most `
+      # 1e-5`. Example (Java): import com.google.type.Color; // ... public static java.
+      # awt.Color fromProto(Color protocolor) ` float alpha = protocolor.hasAlpha() ?
+      # protocolor.getAlpha().getValue() : 1.0; return new java.awt.Color( protocolor.
+      # getRed(), protocolor.getGreen(), protocolor.getBlue(), alpha); ` public static
+      # Color toProto(java.awt.Color color) ` float red = (float) color.getRed();
+      # float green = (float) color.getGreen(); float blue = (float) color.getBlue();
+      # float denominator = 255.0; Color.Builder resultBuilder = Color .newBuilder() .
+      # setRed(red / denominator) .setGreen(green / denominator) .setBlue(blue /
+      # denominator); int alpha = color.getAlpha(); if (alpha != 255) ` result.
+      # setAlpha( FloatValue .newBuilder() .setValue(((float) alpha) / denominator) .
+      # build()); ` return resultBuilder.build(); ` // ... Example (iOS / Obj-C): // ..
+      # . static UIColor* fromProto(Color* protocolor) ` float red = [protocolor red];
+      # float green = [protocolor green]; float blue = [protocolor blue]; FloatValue*
+      # alpha_wrapper = [protocolor alpha]; float alpha = 1.0; if (alpha_wrapper !=
+      # nil) ` alpha = [alpha_wrapper value]; ` return [UIColor colorWithRed:red green:
+      # green blue:blue alpha:alpha]; ` static Color* toProto(UIColor* color) `
+      # CGFloat red, green, blue, alpha; if (![color getRed:&red green:&green blue:&
+      # blue alpha:&alpha]) ` return nil; ` Color* result = [[Color alloc] init]; [
+      # result setRed:red]; [result setGreen:green]; [result setBlue:blue]; if (alpha <
+      # = 0.9999) ` [result setAlpha:floatWrapperWithValue(alpha)]; ` [result
+      # autorelease]; return result; ` // ... Example (JavaScript): // ... var
+      # protoToCssColor = function(rgb_color) ` var redFrac = rgb_color.red || 0.0;
+      # var greenFrac = rgb_color.green || 0.0; var blueFrac = rgb_color.blue || 0.0;
+      # var red = Math.floor(redFrac * 255); var green = Math.floor(greenFrac * 255);
+      # var blue = Math.floor(blueFrac * 255); if (!('alpha' in rgb_color)) ` return
+      # rgbToCssColor(red, green, blue); ` var alphaFrac = rgb_color.alpha.value || 0.
+      # 0; var rgbParams = [red, green, blue].join(','); return ['rgba(', rgbParams, ',
+      # ', alphaFrac, ')'].join(''); `; var rgbToCssColor = function(red, green, blue)
+      # ` var rgbNumber = new Number((red << 16) | (green << 8) | blue); var hexString
+      # = rgbNumber.toString(16); var missingZeros = 6 - hexString.length; var
+      # resultBuilder = ['#']; for (var i = 0; i < missingZeros; i++) ` resultBuilder.
+      # push('0'); ` resultBuilder.push(hexString); return resultBuilder.join(''); `; /
+      # / ...
+      class Color
+        include Google::Apis::Core::Hashable
+      
+        # The fraction of this color that should be applied to the pixel. That is, the
+        # final pixel color is defined by the equation: `pixel color = alpha * (this
+        # color) + (1.0 - alpha) * (background color)` This means that a value of 1.0
+        # corresponds to a solid color, whereas a value of 0.0 corresponds to a
+        # completely transparent color. This uses a wrapper message rather than a simple
+        # float scalar so that it is possible to distinguish between a default value and
+        # the value being unset. If omitted, this color object is rendered as a solid
+        # color (as if the alpha value had been explicitly given a value of 1.0).
+        # Corresponds to the JSON property `alpha`
+        # @return [Float]
+        attr_accessor :alpha
+      
+        # The amount of blue in the color as a value in the interval [0, 1].
+        # Corresponds to the JSON property `blue`
+        # @return [Float]
+        attr_accessor :blue
+      
+        # The amount of green in the color as a value in the interval [0, 1].
+        # Corresponds to the JSON property `green`
+        # @return [Float]
+        attr_accessor :green
+      
+        # The amount of red in the color as a value in the interval [0, 1].
+        # Corresponds to the JSON property `red`
+        # @return [Float]
+        attr_accessor :red
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @alpha = args[:alpha] if args.key?(:alpha)
+          @blue = args[:blue] if args.key?(:blue)
+          @green = args[:green] if args.key?(:green)
+          @red = args[:red] if args.key?(:red)
+        end
+      end
+      
+      # Metric for billing reports.
+      class CpuMetric
+        include Google::Apis::Core::Hashable
+      
+        # Required. Number of CPU cores.
+        # Corresponds to the JSON property `coreNumber`
+        # @return [Fixnum]
+        attr_accessor :core_number
+      
+        # Required. Total seconds of core usage, e.g. 4.
+        # Corresponds to the JSON property `coreSec`
+        # @return [Fixnum]
+        attr_accessor :core_sec
+      
+        # Required. Type of cpu, e.g. N2.
+        # Corresponds to the JSON property `cpuType`
+        # @return [String]
+        attr_accessor :cpu_type
+      
+        # Required. Machine spec, e.g. N1_STANDARD_4.
+        # Corresponds to the JSON property `machineSpec`
+        # @return [String]
+        attr_accessor :machine_spec
+      
+        # Billing tracking labels. They do not contain any user data but only the labels
+        # set by Vertex Core Infra itself. Tracking labels' keys are defined with
+        # special format: goog-[\p`Ll`\p`N`]+ E.g. "key": "goog-k8s-cluster-name","value"
+        # : "us-east1-b4rk"
+        # Corresponds to the JSON property `trackingLabels`
+        # @return [Hash<String,String>]
+        attr_accessor :tracking_labels
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @core_number = args[:core_number] if args.key?(:core_number)
+          @core_sec = args[:core_sec] if args.key?(:core_sec)
+          @cpu_type = args[:cpu_type] if args.key?(:cpu_type)
+          @machine_spec = args[:machine_spec] if args.key?(:machine_spec)
+          @tracking_labels = args[:tracking_labels] if args.key?(:tracking_labels)
+        end
+      end
+      
+      # 
+      class DiskMetric
+        include Google::Apis::Core::Hashable
+      
+        # Required. Type of Disk, e.g. REGIONAL_SSD.
+        # Corresponds to the JSON property `diskType`
+        # @return [String]
+        attr_accessor :disk_type
+      
+        # Required. Seconds of physical disk usage, e.g. 3600.
+        # Corresponds to the JSON property `gibSec`
+        # @return [Fixnum]
+        attr_accessor :gib_sec
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @disk_type = args[:disk_type] if args.key?(:disk_type)
+          @gib_sec = args[:gib_sec] if args.key?(:gib_sec)
+        end
+      end
+      
       # Represents the input to API methods.
       class Document
         include Google::Apis::Core::Hashable
@@ -457,34 +658,32 @@ module Google
         end
       end
       
-      # All available features. Setting each one to true will enable that specific
-      # analysis for the input.
-      class Features
+      # 
+      class GpuMetric
         include Google::Apis::Core::Hashable
       
-        # Optional. Classify the full document into categories.
-        # Corresponds to the JSON property `classifyText`
-        # @return [Boolean]
-        attr_accessor :classify_text
-        alias_method :classify_text?, :classify_text
+        # Required. Seconds of GPU usage, e.g. 3600.
+        # Corresponds to the JSON property `gpuSec`
+        # @return [Fixnum]
+        attr_accessor :gpu_sec
       
-        # Optional. Extract document-level sentiment.
-        # Corresponds to the JSON property `extractDocumentSentiment`
-        # @return [Boolean]
-        attr_accessor :extract_document_sentiment
-        alias_method :extract_document_sentiment?, :extract_document_sentiment
+        # Required. Type of GPU, e.g. NVIDIA_TESLA_V100.
+        # Corresponds to the JSON property `gpuType`
+        # @return [String]
+        attr_accessor :gpu_type
       
-        # Optional. Extract entities.
-        # Corresponds to the JSON property `extractEntities`
-        # @return [Boolean]
-        attr_accessor :extract_entities
-        alias_method :extract_entities?, :extract_entities
+        # Required. Machine spec, e.g. N1_STANDARD_4.
+        # Corresponds to the JSON property `machineSpec`
+        # @return [String]
+        attr_accessor :machine_spec
       
-        # Optional. Moderate the document for harmful and sensitive categories.
-        # Corresponds to the JSON property `moderateText`
-        # @return [Boolean]
-        attr_accessor :moderate_text
-        alias_method :moderate_text?, :moderate_text
+        # Billing tracking labels. They do not contain any user data but only the labels
+        # set by Vertex Core Infra itself. Tracking labels' keys are defined with
+        # special format: goog-[\p`Ll`\p`N`]+ E.g. "key": "goog-k8s-cluster-name","value"
+        # : "us-east1-b4rk"
+        # Corresponds to the JSON property `trackingLabels`
+        # @return [Hash<String,String>]
+        attr_accessor :tracking_labels
       
         def initialize(**args)
            update!(**args)
@@ -492,10 +691,53 @@ module Google
       
         # Update properties of this object
         def update!(**args)
-          @classify_text = args[:classify_text] if args.key?(:classify_text)
-          @extract_document_sentiment = args[:extract_document_sentiment] if args.key?(:extract_document_sentiment)
-          @extract_entities = args[:extract_entities] if args.key?(:extract_entities)
-          @moderate_text = args[:moderate_text] if args.key?(:moderate_text)
+          @gpu_sec = args[:gpu_sec] if args.key?(:gpu_sec)
+          @gpu_type = args[:gpu_type] if args.key?(:gpu_type)
+          @machine_spec = args[:machine_spec] if args.key?(:machine_spec)
+          @tracking_labels = args[:tracking_labels] if args.key?(:tracking_labels)
+        end
+      end
+      
+      # Infra Usage of billing metrics. Next ID: 6
+      class InfraUsage
+        include Google::Apis::Core::Hashable
+      
+        # Aggregated core metrics since requested start_time.
+        # Corresponds to the JSON property `cpuMetrics`
+        # @return [Array<Google::Apis::LanguageV2::CpuMetric>]
+        attr_accessor :cpu_metrics
+      
+        # Aggregated persistent disk metrics since requested start_time.
+        # Corresponds to the JSON property `diskMetrics`
+        # @return [Array<Google::Apis::LanguageV2::DiskMetric>]
+        attr_accessor :disk_metrics
+      
+        # Aggregated gpu metrics since requested start_time.
+        # Corresponds to the JSON property `gpuMetrics`
+        # @return [Array<Google::Apis::LanguageV2::GpuMetric>]
+        attr_accessor :gpu_metrics
+      
+        # Aggregated ram metrics since requested start_time.
+        # Corresponds to the JSON property `ramMetrics`
+        # @return [Array<Google::Apis::LanguageV2::RamMetric>]
+        attr_accessor :ram_metrics
+      
+        # Aggregated tpu metrics since requested start_time.
+        # Corresponds to the JSON property `tpuMetrics`
+        # @return [Array<Google::Apis::LanguageV2::TpuMetric>]
+        attr_accessor :tpu_metrics
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @cpu_metrics = args[:cpu_metrics] if args.key?(:cpu_metrics)
+          @disk_metrics = args[:disk_metrics] if args.key?(:disk_metrics)
+          @gpu_metrics = args[:gpu_metrics] if args.key?(:gpu_metrics)
+          @ram_metrics = args[:ram_metrics] if args.key?(:ram_metrics)
+          @tpu_metrics = args[:tpu_metrics] if args.key?(:tpu_metrics)
         end
       end
       
@@ -553,6 +795,53 @@ module Google
         end
       end
       
+      # 
+      class RamMetric
+        include Google::Apis::Core::Hashable
+      
+        # Required. VM memory in Gigabyte second, e.g. 3600. Using int64 type to match
+        # billing metrics definition.
+        # Corresponds to the JSON property `gibSec`
+        # @return [Fixnum]
+        attr_accessor :gib_sec
+      
+        # Required. Machine spec, e.g. N1_STANDARD_4.
+        # Corresponds to the JSON property `machineSpec`
+        # @return [String]
+        attr_accessor :machine_spec
+      
+        # Required. VM memory in gb.
+        # Corresponds to the JSON property `memories`
+        # @return [Float]
+        attr_accessor :memories
+      
+        # Required. Type of ram.
+        # Corresponds to the JSON property `ramType`
+        # @return [String]
+        attr_accessor :ram_type
+      
+        # Billing tracking labels. They do not contain any user data but only the labels
+        # set by Vertex Core Infra itself. Tracking labels' keys are defined with
+        # special format: goog-[\p`Ll`\p`N`]+ E.g. "key": "goog-k8s-cluster-name","value"
+        # : "us-east1-b4rk"
+        # Corresponds to the JSON property `trackingLabels`
+        # @return [Hash<String,String>]
+        attr_accessor :tracking_labels
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @gib_sec = args[:gib_sec] if args.key?(:gib_sec)
+          @machine_spec = args[:machine_spec] if args.key?(:machine_spec)
+          @memories = args[:memories] if args.key?(:memories)
+          @ram_type = args[:ram_type] if args.key?(:ram_type)
+          @tracking_labels = args[:tracking_labels] if args.key?(:tracking_labels)
+        end
+      end
+      
       # Represents a sentence in the input document.
       class Sentence
         include Google::Apis::Core::Hashable
@@ -668,6 +957,4235 @@ module Google
           @content = args[:content] if args.key?(:content)
         end
       end
+      
+      # 
+      class TpuMetric
+        include Google::Apis::Core::Hashable
+      
+        # Required. Seconds of TPU usage, e.g. 3600.
+        # Corresponds to the JSON property `tpuSec`
+        # @return [Fixnum]
+        attr_accessor :tpu_sec
+      
+        # Required. Type of TPU, e.g. TPU_V2, TPU_V3_POD.
+        # Corresponds to the JSON property `tpuType`
+        # @return [String]
+        attr_accessor :tpu_type
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @tpu_sec = args[:tpu_sec] if args.key?(:tpu_sec)
+          @tpu_type = args[:tpu_type] if args.key?(:tpu_type)
+        end
+      end
+      
+      # The data statistics of a series of ARRAY values.
+      class XpsArrayStats
+        include Google::Apis::Core::Hashable
+      
+        # Common statistics for a column with a specified data type.
+        # Corresponds to the JSON property `commonStats`
+        # @return [Google::Apis::LanguageV2::XpsCommonStats]
+        attr_accessor :common_stats
+      
+        # The data statistics of a series of values that share the same DataType.
+        # Corresponds to the JSON property `memberStats`
+        # @return [Google::Apis::LanguageV2::XpsDataStats]
+        attr_accessor :member_stats
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @common_stats = args[:common_stats] if args.key?(:common_stats)
+          @member_stats = args[:member_stats] if args.key?(:member_stats)
+        end
+      end
+      
+      # 
+      class XpsBatchPredictResponse
+        include Google::Apis::Core::Hashable
+      
+        # Set of examples or input sources.
+        # Corresponds to the JSON property `exampleSet`
+        # @return [Google::Apis::LanguageV2::XpsExampleSet]
+        attr_accessor :example_set
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @example_set = args[:example_set] if args.key?(:example_set)
+        end
+      end
+      
+      # Bounding box matching model metrics for a single intersection-over-union
+      # threshold and multiple label match confidence thresholds.
+      class XpsBoundingBoxMetricsEntry
+        include Google::Apis::Core::Hashable
+      
+        # Metrics for each label-match confidence_threshold from 0.05,0.10,...,0.95,0.96,
+        # 0.97,0.98,0.99.
+        # Corresponds to the JSON property `confidenceMetricsEntries`
+        # @return [Array<Google::Apis::LanguageV2::XpsBoundingBoxMetricsEntryConfidenceMetricsEntry>]
+        attr_accessor :confidence_metrics_entries
+      
+        # The intersection-over-union threshold value used to compute this metrics entry.
+        # Corresponds to the JSON property `iouThreshold`
+        # @return [Float]
+        attr_accessor :iou_threshold
+      
+        # The mean average precision.
+        # Corresponds to the JSON property `meanAveragePrecision`
+        # @return [Float]
+        attr_accessor :mean_average_precision
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @confidence_metrics_entries = args[:confidence_metrics_entries] if args.key?(:confidence_metrics_entries)
+          @iou_threshold = args[:iou_threshold] if args.key?(:iou_threshold)
+          @mean_average_precision = args[:mean_average_precision] if args.key?(:mean_average_precision)
+        end
+      end
+      
+      # Metrics for a single confidence threshold.
+      class XpsBoundingBoxMetricsEntryConfidenceMetricsEntry
+        include Google::Apis::Core::Hashable
+      
+        # The confidence threshold value used to compute the metrics.
+        # Corresponds to the JSON property `confidenceThreshold`
+        # @return [Float]
+        attr_accessor :confidence_threshold
+      
+        # The harmonic mean of recall and precision.
+        # Corresponds to the JSON property `f1Score`
+        # @return [Float]
+        attr_accessor :f1_score
+      
+        # Precision for the given confidence threshold.
+        # Corresponds to the JSON property `precision`
+        # @return [Float]
+        attr_accessor :precision
+      
+        # Recall for the given confidence threshold.
+        # Corresponds to the JSON property `recall`
+        # @return [Float]
+        attr_accessor :recall
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @confidence_threshold = args[:confidence_threshold] if args.key?(:confidence_threshold)
+          @f1_score = args[:f1_score] if args.key?(:f1_score)
+          @precision = args[:precision] if args.key?(:precision)
+          @recall = args[:recall] if args.key?(:recall)
+        end
+      end
+      
+      # The data statistics of a series of CATEGORY values.
+      class XpsCategoryStats
+        include Google::Apis::Core::Hashable
+      
+        # Common statistics for a column with a specified data type.
+        # Corresponds to the JSON property `commonStats`
+        # @return [Google::Apis::LanguageV2::XpsCommonStats]
+        attr_accessor :common_stats
+      
+        # The statistics of the top 20 CATEGORY values, ordered by CategoryStats.
+        # SingleCategoryStats.count.
+        # Corresponds to the JSON property `topCategoryStats`
+        # @return [Array<Google::Apis::LanguageV2::XpsCategoryStatsSingleCategoryStats>]
+        attr_accessor :top_category_stats
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @common_stats = args[:common_stats] if args.key?(:common_stats)
+          @top_category_stats = args[:top_category_stats] if args.key?(:top_category_stats)
+        end
+      end
+      
+      # The statistics of a single CATEGORY value.
+      class XpsCategoryStatsSingleCategoryStats
+        include Google::Apis::Core::Hashable
+      
+        # The number of occurrences of this value in the series.
+        # Corresponds to the JSON property `count`
+        # @return [Fixnum]
+        attr_accessor :count
+      
+        # The CATEGORY value.
+        # Corresponds to the JSON property `value`
+        # @return [String]
+        attr_accessor :value
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @count = args[:count] if args.key?(:count)
+          @value = args[:value] if args.key?(:value)
+        end
+      end
+      
+      # Model evaluation metrics for classification problems. It can be used for image
+      # and video classification. Next tag: 9.
+      class XpsClassificationEvaluationMetrics
+        include Google::Apis::Core::Hashable
+      
+        # The Area under precision recall curve metric.
+        # Corresponds to the JSON property `auPrc`
+        # @return [Float]
+        attr_accessor :au_prc
+      
+        # The Area Under Receiver Operating Characteristic curve metric. Micro-averaged
+        # for the overall evaluation.
+        # Corresponds to the JSON property `auRoc`
+        # @return [Float]
+        attr_accessor :au_roc
+      
+        # The Area under precision recall curve metric based on priors.
+        # Corresponds to the JSON property `baseAuPrc`
+        # @return [Float]
+        attr_accessor :base_au_prc
+      
+        # Metrics that have confidence thresholds. Precision-recall curve can be derived
+        # from it.
+        # Corresponds to the JSON property `confidenceMetricsEntries`
+        # @return [Array<Google::Apis::LanguageV2::XpsConfidenceMetricsEntry>]
+        attr_accessor :confidence_metrics_entries
+      
+        # Confusion matrix of the model running the classification.
+        # Corresponds to the JSON property `confusionMatrix`
+        # @return [Google::Apis::LanguageV2::XpsConfusionMatrix]
+        attr_accessor :confusion_matrix
+      
+        # The number of examples used for model evaluation.
+        # Corresponds to the JSON property `evaluatedExamplesCount`
+        # @return [Fixnum]
+        attr_accessor :evaluated_examples_count
+      
+        # The Log Loss metric.
+        # Corresponds to the JSON property `logLoss`
+        # @return [Float]
+        attr_accessor :log_loss
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @au_prc = args[:au_prc] if args.key?(:au_prc)
+          @au_roc = args[:au_roc] if args.key?(:au_roc)
+          @base_au_prc = args[:base_au_prc] if args.key?(:base_au_prc)
+          @confidence_metrics_entries = args[:confidence_metrics_entries] if args.key?(:confidence_metrics_entries)
+          @confusion_matrix = args[:confusion_matrix] if args.key?(:confusion_matrix)
+          @evaluated_examples_count = args[:evaluated_examples_count] if args.key?(:evaluated_examples_count)
+          @log_loss = args[:log_loss] if args.key?(:log_loss)
+        end
+      end
+      
+      # Map from color to display name. Will only be used by Image Segmentation for
+      # uCAIP.
+      class XpsColorMap
+        include Google::Apis::Core::Hashable
+      
+        # Should be used during training.
+        # Corresponds to the JSON property `annotationSpecIdToken`
+        # @return [String]
+        attr_accessor :annotation_spec_id_token
+      
+        # Represents a color in the RGBA color space. This representation is designed
+        # for simplicity of conversion to and from color representations in various
+        # languages over compactness. For example, the fields of this representation can
+        # be trivially provided to the constructor of `java.awt.Color` in Java; it can
+        # also be trivially provided to UIColor's `+colorWithRed:green:blue:alpha`
+        # method in iOS; and, with just a little work, it can be easily formatted into a
+        # CSS `rgba()` string in JavaScript. This reference page doesn't have
+        # information about the absolute color space that should be used to interpret
+        # the RGB value—for example, sRGB, Adobe RGB, DCI-P3, and BT.2020. By default,
+        # applications should assume the sRGB color space. When color equality needs to
+        # be decided, implementations, unless documented otherwise, treat two colors as
+        # equal if all their red, green, blue, and alpha values each differ by at most `
+        # 1e-5`. Example (Java): import com.google.type.Color; // ... public static java.
+        # awt.Color fromProto(Color protocolor) ` float alpha = protocolor.hasAlpha() ?
+        # protocolor.getAlpha().getValue() : 1.0; return new java.awt.Color( protocolor.
+        # getRed(), protocolor.getGreen(), protocolor.getBlue(), alpha); ` public static
+        # Color toProto(java.awt.Color color) ` float red = (float) color.getRed();
+        # float green = (float) color.getGreen(); float blue = (float) color.getBlue();
+        # float denominator = 255.0; Color.Builder resultBuilder = Color .newBuilder() .
+        # setRed(red / denominator) .setGreen(green / denominator) .setBlue(blue /
+        # denominator); int alpha = color.getAlpha(); if (alpha != 255) ` result.
+        # setAlpha( FloatValue .newBuilder() .setValue(((float) alpha) / denominator) .
+        # build()); ` return resultBuilder.build(); ` // ... Example (iOS / Obj-C): // ..
+        # . static UIColor* fromProto(Color* protocolor) ` float red = [protocolor red];
+        # float green = [protocolor green]; float blue = [protocolor blue]; FloatValue*
+        # alpha_wrapper = [protocolor alpha]; float alpha = 1.0; if (alpha_wrapper !=
+        # nil) ` alpha = [alpha_wrapper value]; ` return [UIColor colorWithRed:red green:
+        # green blue:blue alpha:alpha]; ` static Color* toProto(UIColor* color) `
+        # CGFloat red, green, blue, alpha; if (![color getRed:&red green:&green blue:&
+        # blue alpha:&alpha]) ` return nil; ` Color* result = [[Color alloc] init]; [
+        # result setRed:red]; [result setGreen:green]; [result setBlue:blue]; if (alpha <
+        # = 0.9999) ` [result setAlpha:floatWrapperWithValue(alpha)]; ` [result
+        # autorelease]; return result; ` // ... Example (JavaScript): // ... var
+        # protoToCssColor = function(rgb_color) ` var redFrac = rgb_color.red || 0.0;
+        # var greenFrac = rgb_color.green || 0.0; var blueFrac = rgb_color.blue || 0.0;
+        # var red = Math.floor(redFrac * 255); var green = Math.floor(greenFrac * 255);
+        # var blue = Math.floor(blueFrac * 255); if (!('alpha' in rgb_color)) ` return
+        # rgbToCssColor(red, green, blue); ` var alphaFrac = rgb_color.alpha.value || 0.
+        # 0; var rgbParams = [red, green, blue].join(','); return ['rgba(', rgbParams, ',
+        # ', alphaFrac, ')'].join(''); `; var rgbToCssColor = function(red, green, blue)
+        # ` var rgbNumber = new Number((red << 16) | (green << 8) | blue); var hexString
+        # = rgbNumber.toString(16); var missingZeros = 6 - hexString.length; var
+        # resultBuilder = ['#']; for (var i = 0; i < missingZeros; i++) ` resultBuilder.
+        # push('0'); ` resultBuilder.push(hexString); return resultBuilder.join(''); `; /
+        # / ...
+        # Corresponds to the JSON property `color`
+        # @return [Google::Apis::LanguageV2::Color]
+        attr_accessor :color
+      
+        # Should be used during preprocessing.
+        # Corresponds to the JSON property `displayName`
+        # @return [String]
+        attr_accessor :display_name
+      
+        # RGB color and each channel is represented by an integer.
+        # Corresponds to the JSON property `intColor`
+        # @return [Google::Apis::LanguageV2::XpsColorMapIntColor]
+        attr_accessor :int_color
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @annotation_spec_id_token = args[:annotation_spec_id_token] if args.key?(:annotation_spec_id_token)
+          @color = args[:color] if args.key?(:color)
+          @display_name = args[:display_name] if args.key?(:display_name)
+          @int_color = args[:int_color] if args.key?(:int_color)
+        end
+      end
+      
+      # RGB color and each channel is represented by an integer.
+      class XpsColorMapIntColor
+        include Google::Apis::Core::Hashable
+      
+        # The value should be in range of [0, 255].
+        # Corresponds to the JSON property `blue`
+        # @return [Fixnum]
+        attr_accessor :blue
+      
+        # The value should be in range of [0, 255].
+        # Corresponds to the JSON property `green`
+        # @return [Fixnum]
+        attr_accessor :green
+      
+        # The value should be in range of [0, 255].
+        # Corresponds to the JSON property `red`
+        # @return [Fixnum]
+        attr_accessor :red
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @blue = args[:blue] if args.key?(:blue)
+          @green = args[:green] if args.key?(:green)
+          @red = args[:red] if args.key?(:red)
+        end
+      end
+      
+      # 
+      class XpsColumnSpec
+        include Google::Apis::Core::Hashable
+      
+        # The unique id of the column. When Preprocess, the Tables BE will popuate the
+        # order id of the column, which reflects the order of the column inside the
+        # table, i.e. 0 means the first column in the table, N-1 means the last column.
+        # AutoML BE will persist this order id in Spanner and set the order id here when
+        # calling RefreshTablesStats and Train. Note: it's different than the
+        # column_spec_id that is generated in AutoML BE.
+        # Corresponds to the JSON property `columnId`
+        # @return [Fixnum]
+        attr_accessor :column_id
+      
+        # The data statistics of a series of values that share the same DataType.
+        # Corresponds to the JSON property `dataStats`
+        # @return [Google::Apis::LanguageV2::XpsDataStats]
+        attr_accessor :data_stats
+      
+        # Indicated the type of data that can be stored in a structured data entity (e.g.
+        # a table).
+        # Corresponds to the JSON property `dataType`
+        # @return [Google::Apis::LanguageV2::XpsDataType]
+        attr_accessor :data_type
+      
+        # The display name of the column. It's outputed in Preprocess and a required
+        # input for RefreshTablesStats and Train.
+        # Corresponds to the JSON property `displayName`
+        # @return [String]
+        attr_accessor :display_name
+      
+        # =========================================================================== #
+        # The fields below are used exclusively for Forecasting.
+        # Corresponds to the JSON property `forecastingMetadata`
+        # @return [Google::Apis::LanguageV2::XpsColumnSpecForecastingMetadata]
+        attr_accessor :forecasting_metadata
+      
+        # It's outputed in RefreshTablesStats, and a required input in Train.
+        # Corresponds to the JSON property `topCorrelatedColumns`
+        # @return [Array<Google::Apis::LanguageV2::XpsColumnSpecCorrelatedColumn>]
+        attr_accessor :top_correlated_columns
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @column_id = args[:column_id] if args.key?(:column_id)
+          @data_stats = args[:data_stats] if args.key?(:data_stats)
+          @data_type = args[:data_type] if args.key?(:data_type)
+          @display_name = args[:display_name] if args.key?(:display_name)
+          @forecasting_metadata = args[:forecasting_metadata] if args.key?(:forecasting_metadata)
+          @top_correlated_columns = args[:top_correlated_columns] if args.key?(:top_correlated_columns)
+        end
+      end
+      
+      # Identifies a table's column, and its correlation with the column this
+      # ColumnSpec describes.
+      class XpsColumnSpecCorrelatedColumn
+        include Google::Apis::Core::Hashable
+      
+        # 
+        # Corresponds to the JSON property `columnId`
+        # @return [Fixnum]
+        attr_accessor :column_id
+      
+        # A correlation statistics between two series of DataType values. The series may
+        # have differing DataType-s, but within a single series the DataType must be the
+        # same.
+        # Corresponds to the JSON property `correlationStats`
+        # @return [Google::Apis::LanguageV2::XpsCorrelationStats]
+        attr_accessor :correlation_stats
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @column_id = args[:column_id] if args.key?(:column_id)
+          @correlation_stats = args[:correlation_stats] if args.key?(:correlation_stats)
+        end
+      end
+      
+      # =========================================================================== #
+      # The fields below are used exclusively for Forecasting.
+      class XpsColumnSpecForecastingMetadata
+        include Google::Apis::Core::Hashable
+      
+        # The type of the column for FORECASTING model training purposes.
+        # Corresponds to the JSON property `columnType`
+        # @return [String]
+        attr_accessor :column_type
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @column_type = args[:column_type] if args.key?(:column_type)
+        end
+      end
+      
+      # Common statistics for a column with a specified data type.
+      class XpsCommonStats
+        include Google::Apis::Core::Hashable
+      
+        # 
+        # Corresponds to the JSON property `distinctValueCount`
+        # @return [Fixnum]
+        attr_accessor :distinct_value_count
+      
+        # 
+        # Corresponds to the JSON property `nullValueCount`
+        # @return [Fixnum]
+        attr_accessor :null_value_count
+      
+        # 
+        # Corresponds to the JSON property `validValueCount`
+        # @return [Fixnum]
+        attr_accessor :valid_value_count
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @distinct_value_count = args[:distinct_value_count] if args.key?(:distinct_value_count)
+          @null_value_count = args[:null_value_count] if args.key?(:null_value_count)
+          @valid_value_count = args[:valid_value_count] if args.key?(:valid_value_count)
+        end
+      end
+      
+      # ConfidenceMetricsEntry includes generic precision, recall, f1 score etc. Next
+      # tag: 16.
+      class XpsConfidenceMetricsEntry
+        include Google::Apis::Core::Hashable
+      
+        # Metrics are computed with an assumption that the model never return
+        # predictions with score lower than this value.
+        # Corresponds to the JSON property `confidenceThreshold`
+        # @return [Float]
+        attr_accessor :confidence_threshold
+      
+        # The harmonic mean of recall and precision.
+        # Corresponds to the JSON property `f1Score`
+        # @return [Float]
+        attr_accessor :f1_score
+      
+        # The harmonic mean of recall_at1 and precision_at1.
+        # Corresponds to the JSON property `f1ScoreAt1`
+        # @return [Float]
+        attr_accessor :f1_score_at1
+      
+        # The number of ground truth labels that are not matched by a model created
+        # label.
+        # Corresponds to the JSON property `falseNegativeCount`
+        # @return [Fixnum]
+        attr_accessor :false_negative_count
+      
+        # The number of model created labels that do not match a ground truth label.
+        # Corresponds to the JSON property `falsePositiveCount`
+        # @return [Fixnum]
+        attr_accessor :false_positive_count
+      
+        # False Positive Rate for the given confidence threshold.
+        # Corresponds to the JSON property `falsePositiveRate`
+        # @return [Float]
+        attr_accessor :false_positive_rate
+      
+        # The False Positive Rate when only considering the label that has the highest
+        # prediction score and not below the confidence threshold for each example.
+        # Corresponds to the JSON property `falsePositiveRateAt1`
+        # @return [Float]
+        attr_accessor :false_positive_rate_at1
+      
+        # Metrics are computed with an assumption that the model always returns at most
+        # this many predictions (ordered by their score, descendingly), but they all
+        # still need to meet the confidence_threshold.
+        # Corresponds to the JSON property `positionThreshold`
+        # @return [Fixnum]
+        attr_accessor :position_threshold
+      
+        # Precision for the given confidence threshold.
+        # Corresponds to the JSON property `precision`
+        # @return [Float]
+        attr_accessor :precision
+      
+        # The precision when only considering the label that has the highest prediction
+        # score and not below the confidence threshold for each example.
+        # Corresponds to the JSON property `precisionAt1`
+        # @return [Float]
+        attr_accessor :precision_at1
+      
+        # Recall (true positive rate) for the given confidence threshold.
+        # Corresponds to the JSON property `recall`
+        # @return [Float]
+        attr_accessor :recall
+      
+        # The recall (true positive rate) when only considering the label that has the
+        # highest prediction score and not below the confidence threshold for each
+        # example.
+        # Corresponds to the JSON property `recallAt1`
+        # @return [Float]
+        attr_accessor :recall_at1
+      
+        # The number of labels that were not created by the model, but if they would,
+        # they would not match a ground truth label.
+        # Corresponds to the JSON property `trueNegativeCount`
+        # @return [Fixnum]
+        attr_accessor :true_negative_count
+      
+        # The number of model created labels that match a ground truth label.
+        # Corresponds to the JSON property `truePositiveCount`
+        # @return [Fixnum]
+        attr_accessor :true_positive_count
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @confidence_threshold = args[:confidence_threshold] if args.key?(:confidence_threshold)
+          @f1_score = args[:f1_score] if args.key?(:f1_score)
+          @f1_score_at1 = args[:f1_score_at1] if args.key?(:f1_score_at1)
+          @false_negative_count = args[:false_negative_count] if args.key?(:false_negative_count)
+          @false_positive_count = args[:false_positive_count] if args.key?(:false_positive_count)
+          @false_positive_rate = args[:false_positive_rate] if args.key?(:false_positive_rate)
+          @false_positive_rate_at1 = args[:false_positive_rate_at1] if args.key?(:false_positive_rate_at1)
+          @position_threshold = args[:position_threshold] if args.key?(:position_threshold)
+          @precision = args[:precision] if args.key?(:precision)
+          @precision_at1 = args[:precision_at1] if args.key?(:precision_at1)
+          @recall = args[:recall] if args.key?(:recall)
+          @recall_at1 = args[:recall_at1] if args.key?(:recall_at1)
+          @true_negative_count = args[:true_negative_count] if args.key?(:true_negative_count)
+          @true_positive_count = args[:true_positive_count] if args.key?(:true_positive_count)
+        end
+      end
+      
+      # Confusion matrix of the model running the classification.
+      class XpsConfusionMatrix
+        include Google::Apis::Core::Hashable
+      
+        # For the following three repeated fields, only one is intended to be set.
+        # annotation_spec_id_token is preferable to be set. ID tokens of the annotation
+        # specs used in the confusion matrix.
+        # Corresponds to the JSON property `annotationSpecIdToken`
+        # @return [Array<String>]
+        attr_accessor :annotation_spec_id_token
+      
+        # Category (mainly for segmentation). Set only for image segmentation models.
+        # Note: uCAIP Image Segmentation should use annotation_spec_id_token.
+        # Corresponds to the JSON property `category`
+        # @return [Array<Fixnum>]
+        attr_accessor :category
+      
+        # Rows in the confusion matrix. The number of rows is equal to the size of `
+        # annotation_spec_id_token`. `row[i].value[j]` is the number of examples that
+        # have ground truth of the `annotation_spec_id_token[i]` and are predicted as `
+        # annotation_spec_id_token[j]` by the model being evaluated.
+        # Corresponds to the JSON property `row`
+        # @return [Array<Google::Apis::LanguageV2::XpsConfusionMatrixRow>]
+        attr_accessor :row
+      
+        # Sentiment labels used in the confusion matrix. Set only for text sentiment
+        # models. For AutoML Text Revamp, use `annotation_spec_id_token` instead and
+        # leave this field empty.
+        # Corresponds to the JSON property `sentimentLabel`
+        # @return [Array<Fixnum>]
+        attr_accessor :sentiment_label
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @annotation_spec_id_token = args[:annotation_spec_id_token] if args.key?(:annotation_spec_id_token)
+          @category = args[:category] if args.key?(:category)
+          @row = args[:row] if args.key?(:row)
+          @sentiment_label = args[:sentiment_label] if args.key?(:sentiment_label)
+        end
+      end
+      
+      # A row in the confusion matrix.
+      class XpsConfusionMatrixRow
+        include Google::Apis::Core::Hashable
+      
+        # Same as above except intended to represent other counts (for e.g. for
+        # segmentation this is pixel count). NOTE(params): Only example_count or count
+        # is set (oneoff does not support repeated fields unless they are embedded
+        # inside another message).
+        # Corresponds to the JSON property `count`
+        # @return [Array<Fixnum>]
+        attr_accessor :count
+      
+        # Value of the specific cell in the confusion matrix. The number of values each
+        # row has (i.e. the length of the row) is equal to the length of the
+        # annotation_spec_id_token field.
+        # Corresponds to the JSON property `exampleCount`
+        # @return [Array<Fixnum>]
+        attr_accessor :example_count
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @count = args[:count] if args.key?(:count)
+          @example_count = args[:example_count] if args.key?(:example_count)
+        end
+      end
+      
+      # A model format used for iOS mobile devices.
+      class XpsCoreMlFormat
+        include Google::Apis::Core::Hashable
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+        end
+      end
+      
+      # A correlation statistics between two series of DataType values. The series may
+      # have differing DataType-s, but within a single series the DataType must be the
+      # same.
+      class XpsCorrelationStats
+        include Google::Apis::Core::Hashable
+      
+        # The correlation value using the Cramer's V measure.
+        # Corresponds to the JSON property `cramersV`
+        # @return [Float]
+        attr_accessor :cramers_v
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @cramers_v = args[:cramers_v] if args.key?(:cramers_v)
+        end
+      end
+      
+      # Different types of errors and the stats associatesd with each error.
+      class XpsDataErrors
+        include Google::Apis::Core::Hashable
+      
+        # Number of records having errors associated with the enum.
+        # Corresponds to the JSON property `count`
+        # @return [Fixnum]
+        attr_accessor :count
+      
+        # Type of the error.
+        # Corresponds to the JSON property `errorType`
+        # @return [String]
+        attr_accessor :error_type
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @count = args[:count] if args.key?(:count)
+          @error_type = args[:error_type] if args.key?(:error_type)
+        end
+      end
+      
+      # The data statistics of a series of values that share the same DataType.
+      class XpsDataStats
+        include Google::Apis::Core::Hashable
+      
+        # The data statistics of a series of ARRAY values.
+        # Corresponds to the JSON property `arrayStats`
+        # @return [Google::Apis::LanguageV2::XpsArrayStats]
+        attr_accessor :array_stats
+      
+        # The data statistics of a series of CATEGORY values.
+        # Corresponds to the JSON property `categoryStats`
+        # @return [Google::Apis::LanguageV2::XpsCategoryStats]
+        attr_accessor :category_stats
+      
+        # The number of distinct values.
+        # Corresponds to the JSON property `distinctValueCount`
+        # @return [Fixnum]
+        attr_accessor :distinct_value_count
+      
+        # The data statistics of a series of FLOAT64 values.
+        # Corresponds to the JSON property `float64Stats`
+        # @return [Google::Apis::LanguageV2::XpsFloat64Stats]
+        attr_accessor :float64_stats
+      
+        # The number of values that are null.
+        # Corresponds to the JSON property `nullValueCount`
+        # @return [Fixnum]
+        attr_accessor :null_value_count
+      
+        # The data statistics of a series of STRING values.
+        # Corresponds to the JSON property `stringStats`
+        # @return [Google::Apis::LanguageV2::XpsStringStats]
+        attr_accessor :string_stats
+      
+        # The data statistics of a series of STRUCT values.
+        # Corresponds to the JSON property `structStats`
+        # @return [Google::Apis::LanguageV2::XpsStructStats]
+        attr_accessor :struct_stats
+      
+        # The data statistics of a series of TIMESTAMP values.
+        # Corresponds to the JSON property `timestampStats`
+        # @return [Google::Apis::LanguageV2::XpsTimestampStats]
+        attr_accessor :timestamp_stats
+      
+        # The number of values that are valid.
+        # Corresponds to the JSON property `validValueCount`
+        # @return [Fixnum]
+        attr_accessor :valid_value_count
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @array_stats = args[:array_stats] if args.key?(:array_stats)
+          @category_stats = args[:category_stats] if args.key?(:category_stats)
+          @distinct_value_count = args[:distinct_value_count] if args.key?(:distinct_value_count)
+          @float64_stats = args[:float64_stats] if args.key?(:float64_stats)
+          @null_value_count = args[:null_value_count] if args.key?(:null_value_count)
+          @string_stats = args[:string_stats] if args.key?(:string_stats)
+          @struct_stats = args[:struct_stats] if args.key?(:struct_stats)
+          @timestamp_stats = args[:timestamp_stats] if args.key?(:timestamp_stats)
+          @valid_value_count = args[:valid_value_count] if args.key?(:valid_value_count)
+        end
+      end
+      
+      # Indicated the type of data that can be stored in a structured data entity (e.g.
+      # a table).
+      class XpsDataType
+        include Google::Apis::Core::Hashable
+      
+        # The highly compatible data types to this data type.
+        # Corresponds to the JSON property `compatibleDataTypes`
+        # @return [Array<Google::Apis::LanguageV2::XpsDataType>]
+        attr_accessor :compatible_data_types
+      
+        # Indicated the type of data that can be stored in a structured data entity (e.g.
+        # a table).
+        # Corresponds to the JSON property `listElementType`
+        # @return [Google::Apis::LanguageV2::XpsDataType]
+        attr_accessor :list_element_type
+      
+        # If true, this DataType can also be `null`.
+        # Corresponds to the JSON property `nullable`
+        # @return [Boolean]
+        attr_accessor :nullable
+        alias_method :nullable?, :nullable
+      
+        # `StructType` defines the DataType-s of a STRUCT type.
+        # Corresponds to the JSON property `structType`
+        # @return [Google::Apis::LanguageV2::XpsStructType]
+        attr_accessor :struct_type
+      
+        # If type_code == TIMESTAMP then `time_format` provides the format in which that
+        # time field is expressed. The time_format must be written in `strftime` syntax.
+        # If time_format is not set, then the default format as described on the field
+        # is used.
+        # Corresponds to the JSON property `timeFormat`
+        # @return [String]
+        attr_accessor :time_format
+      
+        # Required. The TypeCode for this type.
+        # Corresponds to the JSON property `typeCode`
+        # @return [String]
+        attr_accessor :type_code
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @compatible_data_types = args[:compatible_data_types] if args.key?(:compatible_data_types)
+          @list_element_type = args[:list_element_type] if args.key?(:list_element_type)
+          @nullable = args[:nullable] if args.key?(:nullable)
+          @struct_type = args[:struct_type] if args.key?(:struct_type)
+          @time_format = args[:time_format] if args.key?(:time_format)
+          @type_code = args[:type_code] if args.key?(:type_code)
+        end
+      end
+      
+      # A model format used for Docker containers. Use the params field to customize
+      # the container. The container is verified to work correctly on ubuntu 16.04
+      # operating system.
+      class XpsDockerFormat
+        include Google::Apis::Core::Hashable
+      
+        # Optional. Additional cpu information describing the requirements for the to be
+        # exported model files.
+        # Corresponds to the JSON property `cpuArchitecture`
+        # @return [String]
+        attr_accessor :cpu_architecture
+      
+        # Optional. Additional gpu information describing the requirements for the to be
+        # exported model files.
+        # Corresponds to the JSON property `gpuArchitecture`
+        # @return [String]
+        attr_accessor :gpu_architecture
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @cpu_architecture = args[:cpu_architecture] if args.key?(:cpu_architecture)
+          @gpu_architecture = args[:gpu_architecture] if args.key?(:gpu_architecture)
+        end
+      end
+      
+      # A model format used for [Edge TPU](https://cloud.google.com/edge-tpu/) devices.
+      class XpsEdgeTpuTfLiteFormat
+        include Google::Apis::Core::Hashable
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+        end
+      end
+      
+      # Contains xPS-specific model evaluation metrics either for a single annotation
+      # spec (label), or for the model overall. Next tag: 18.
+      class XpsEvaluationMetrics
+        include Google::Apis::Core::Hashable
+      
+        # The annotation_spec for which this evaluation metrics instance had been
+        # created. Empty iff this is an overall model evaluation (like Tables evaluation
+        # metrics), i.e. aggregated across all labels. The value comes from the input
+        # annotations in AnnotatedExample. For MVP product or for text sentiment models
+        # where annotation_spec_id_token is not available, set label instead.
+        # Corresponds to the JSON property `annotationSpecIdToken`
+        # @return [String]
+        attr_accessor :annotation_spec_id_token
+      
+        # The integer category label for which this evaluation metric instance had been
+        # created. Valid categories are 0 or higher. Overall model evaluation should set
+        # this to negative values (rather than implicit zero). Only used for Image
+        # Segmentation (prefer to set annotation_spec_id_token instead). Note: uCAIP
+        # Image Segmentation should use annotation_spec_id_token.
+        # Corresponds to the JSON property `category`
+        # @return [Fixnum]
+        attr_accessor :category
+      
+        # The number of examples used to create this evaluation metrics instance.
+        # Corresponds to the JSON property `evaluatedExampleCount`
+        # @return [Fixnum]
+        attr_accessor :evaluated_example_count
+      
+        # Model evaluation metrics for classification problems. It can be used for image
+        # and video classification. Next tag: 9.
+        # Corresponds to the JSON property `imageClassificationEvalMetrics`
+        # @return [Google::Apis::LanguageV2::XpsClassificationEvaluationMetrics]
+        attr_accessor :image_classification_eval_metrics
+      
+        # Model evaluation metrics for image object detection problems. Evaluates
+        # prediction quality of labeled bounding boxes.
+        # Corresponds to the JSON property `imageObjectDetectionEvalMetrics`
+        # @return [Google::Apis::LanguageV2::XpsImageObjectDetectionEvaluationMetrics]
+        attr_accessor :image_object_detection_eval_metrics
+      
+        # Model evaluation metrics for image segmentation problems. Next tag: 4.
+        # Corresponds to the JSON property `imageSegmentationEvalMetrics`
+        # @return [Google::Apis::LanguageV2::XpsImageSegmentationEvaluationMetrics]
+        attr_accessor :image_segmentation_eval_metrics
+      
+        # The label for which this evaluation metrics instance had been created. Empty
+        # iff this is an overall model evaluation (like Tables evaluation metrics), i.e.
+        # aggregated across all labels. The label maps to AnnotationSpec.display_name in
+        # Public API protos. Only used by MVP implementation and text sentiment FULL
+        # implementation.
+        # Corresponds to the JSON property `label`
+        # @return [String]
+        attr_accessor :label
+      
+        # Model evaluation metrics for regression problems. It can be used for Tables.
+        # Corresponds to the JSON property `regressionEvalMetrics`
+        # @return [Google::Apis::LanguageV2::XpsRegressionEvaluationMetrics]
+        attr_accessor :regression_eval_metrics
+      
+        # Model evaluation metrics for classification problems. It can be used for image
+        # and video classification. Next tag: 9.
+        # Corresponds to the JSON property `tablesClassificationEvalMetrics`
+        # @return [Google::Apis::LanguageV2::XpsClassificationEvaluationMetrics]
+        attr_accessor :tables_classification_eval_metrics
+      
+        # 
+        # Corresponds to the JSON property `tablesEvalMetrics`
+        # @return [Google::Apis::LanguageV2::XpsTablesEvaluationMetrics]
+        attr_accessor :tables_eval_metrics
+      
+        # Model evaluation metrics for classification problems. It can be used for image
+        # and video classification. Next tag: 9.
+        # Corresponds to the JSON property `textClassificationEvalMetrics`
+        # @return [Google::Apis::LanguageV2::XpsClassificationEvaluationMetrics]
+        attr_accessor :text_classification_eval_metrics
+      
+        # 
+        # Corresponds to the JSON property `textExtractionEvalMetrics`
+        # @return [Google::Apis::LanguageV2::XpsTextExtractionEvaluationMetrics]
+        attr_accessor :text_extraction_eval_metrics
+      
+        # Model evaluation metrics for text sentiment problems.
+        # Corresponds to the JSON property `textSentimentEvalMetrics`
+        # @return [Google::Apis::LanguageV2::XpsTextSentimentEvaluationMetrics]
+        attr_accessor :text_sentiment_eval_metrics
+      
+        # Evaluation metrics for the dataset.
+        # Corresponds to the JSON property `translationEvalMetrics`
+        # @return [Google::Apis::LanguageV2::XpsTranslationEvaluationMetrics]
+        attr_accessor :translation_eval_metrics
+      
+        # Model evaluation metrics for video action recognition.
+        # Corresponds to the JSON property `videoActionRecognitionEvalMetrics`
+        # @return [Google::Apis::LanguageV2::XpsVideoActionRecognitionEvaluationMetrics]
+        attr_accessor :video_action_recognition_eval_metrics
+      
+        # Model evaluation metrics for classification problems. It can be used for image
+        # and video classification. Next tag: 9.
+        # Corresponds to the JSON property `videoClassificationEvalMetrics`
+        # @return [Google::Apis::LanguageV2::XpsClassificationEvaluationMetrics]
+        attr_accessor :video_classification_eval_metrics
+      
+        # Model evaluation metrics for ObjectTracking problems. Next tag: 10.
+        # Corresponds to the JSON property `videoObjectTrackingEvalMetrics`
+        # @return [Google::Apis::LanguageV2::XpsVideoObjectTrackingEvaluationMetrics]
+        attr_accessor :video_object_tracking_eval_metrics
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @annotation_spec_id_token = args[:annotation_spec_id_token] if args.key?(:annotation_spec_id_token)
+          @category = args[:category] if args.key?(:category)
+          @evaluated_example_count = args[:evaluated_example_count] if args.key?(:evaluated_example_count)
+          @image_classification_eval_metrics = args[:image_classification_eval_metrics] if args.key?(:image_classification_eval_metrics)
+          @image_object_detection_eval_metrics = args[:image_object_detection_eval_metrics] if args.key?(:image_object_detection_eval_metrics)
+          @image_segmentation_eval_metrics = args[:image_segmentation_eval_metrics] if args.key?(:image_segmentation_eval_metrics)
+          @label = args[:label] if args.key?(:label)
+          @regression_eval_metrics = args[:regression_eval_metrics] if args.key?(:regression_eval_metrics)
+          @tables_classification_eval_metrics = args[:tables_classification_eval_metrics] if args.key?(:tables_classification_eval_metrics)
+          @tables_eval_metrics = args[:tables_eval_metrics] if args.key?(:tables_eval_metrics)
+          @text_classification_eval_metrics = args[:text_classification_eval_metrics] if args.key?(:text_classification_eval_metrics)
+          @text_extraction_eval_metrics = args[:text_extraction_eval_metrics] if args.key?(:text_extraction_eval_metrics)
+          @text_sentiment_eval_metrics = args[:text_sentiment_eval_metrics] if args.key?(:text_sentiment_eval_metrics)
+          @translation_eval_metrics = args[:translation_eval_metrics] if args.key?(:translation_eval_metrics)
+          @video_action_recognition_eval_metrics = args[:video_action_recognition_eval_metrics] if args.key?(:video_action_recognition_eval_metrics)
+          @video_classification_eval_metrics = args[:video_classification_eval_metrics] if args.key?(:video_classification_eval_metrics)
+          @video_object_tracking_eval_metrics = args[:video_object_tracking_eval_metrics] if args.key?(:video_object_tracking_eval_metrics)
+        end
+      end
+      
+      # Specifies location of model evaluation metrics.
+      class XpsEvaluationMetricsSet
+        include Google::Apis::Core::Hashable
+      
+        # Inline EvaluationMetrics - should be relatively small. For passing large
+        # quantities of exhaustive metrics, use file_spec.
+        # Corresponds to the JSON property `evaluationMetrics`
+        # @return [Array<Google::Apis::LanguageV2::XpsEvaluationMetrics>]
+        attr_accessor :evaluation_metrics
+      
+        # Spec of input and output files, on external file systems (CNS, GCS, etc).
+        # Corresponds to the JSON property `fileSpec`
+        # @return [Google::Apis::LanguageV2::XpsFileSpec]
+        attr_accessor :file_spec
+      
+        # Number of the evaluation metrics (usually one per label plus overall).
+        # Corresponds to the JSON property `numEvaluationMetrics`
+        # @return [Fixnum]
+        attr_accessor :num_evaluation_metrics
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @evaluation_metrics = args[:evaluation_metrics] if args.key?(:evaluation_metrics)
+          @file_spec = args[:file_spec] if args.key?(:file_spec)
+          @num_evaluation_metrics = args[:num_evaluation_metrics] if args.key?(:num_evaluation_metrics)
+        end
+      end
+      
+      # Set of examples or input sources.
+      class XpsExampleSet
+        include Google::Apis::Core::Hashable
+      
+        # Spec of input and output files, on external file systems (CNS, GCS, etc).
+        # Corresponds to the JSON property `fileSpec`
+        # @return [Google::Apis::LanguageV2::XpsFileSpec]
+        attr_accessor :file_spec
+      
+        # Fingerprint of the example set.
+        # Corresponds to the JSON property `fingerprint`
+        # @return [Fixnum]
+        attr_accessor :fingerprint
+      
+        # Number of examples.
+        # Corresponds to the JSON property `numExamples`
+        # @return [Fixnum]
+        attr_accessor :num_examples
+      
+        # Number of input sources.
+        # Corresponds to the JSON property `numInputSources`
+        # @return [Fixnum]
+        attr_accessor :num_input_sources
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @file_spec = args[:file_spec] if args.key?(:file_spec)
+          @fingerprint = args[:fingerprint] if args.key?(:fingerprint)
+          @num_examples = args[:num_examples] if args.key?(:num_examples)
+          @num_input_sources = args[:num_input_sources] if args.key?(:num_input_sources)
+        end
+      end
+      
+      # 
+      class XpsExportModelOutputConfig
+        include Google::Apis::Core::Hashable
+      
+        # A model format used for iOS mobile devices.
+        # Corresponds to the JSON property `coreMlFormat`
+        # @return [Google::Apis::LanguageV2::XpsCoreMlFormat]
+        attr_accessor :core_ml_format
+      
+        # A model format used for Docker containers. Use the params field to customize
+        # the container. The container is verified to work correctly on ubuntu 16.04
+        # operating system.
+        # Corresponds to the JSON property `dockerFormat`
+        # @return [Google::Apis::LanguageV2::XpsDockerFormat]
+        attr_accessor :docker_format
+      
+        # A model format used for [Edge TPU](https://cloud.google.com/edge-tpu/) devices.
+        # Corresponds to the JSON property `edgeTpuTfLiteFormat`
+        # @return [Google::Apis::LanguageV2::XpsEdgeTpuTfLiteFormat]
+        attr_accessor :edge_tpu_tf_lite_format
+      
+        # For any model and format: If true, will additionally export
+        # FirebaseExportedModelInfo in a firebase.txt file.
+        # Corresponds to the JSON property `exportFirebaseAuxiliaryInfo`
+        # @return [Boolean]
+        attr_accessor :export_firebase_auxiliary_info
+        alias_method :export_firebase_auxiliary_info?, :export_firebase_auxiliary_info
+      
+        # The Google Contained Registry (GCR) path the exported files to be pushed to.
+        # This location is set if the exported format is DOCKDER.
+        # Corresponds to the JSON property `outputGcrUri`
+        # @return [String]
+        attr_accessor :output_gcr_uri
+      
+        # The Google Cloud Storage (GCS) directory where XPS will output the exported
+        # models and related files. Format: gs://bucket/directory
+        # Corresponds to the JSON property `outputGcsUri`
+        # @return [String]
+        attr_accessor :output_gcs_uri
+      
+        # A [TensorFlow.js](https://www.tensorflow.org/js) model that can be used in the
+        # browser and in Node.js using JavaScript.
+        # Corresponds to the JSON property `tfJsFormat`
+        # @return [Google::Apis::LanguageV2::XpsTfJsFormat]
+        attr_accessor :tf_js_format
+      
+        # LINT.IfChange A model format used for mobile and IoT devices. See https://www.
+        # tensorflow.org/lite.
+        # Corresponds to the JSON property `tfLiteFormat`
+        # @return [Google::Apis::LanguageV2::XpsTfLiteFormat]
+        attr_accessor :tf_lite_format
+      
+        # A tensorflow model format in SavedModel format.
+        # Corresponds to the JSON property `tfSavedModelFormat`
+        # @return [Google::Apis::LanguageV2::XpsTfSavedModelFormat]
+        attr_accessor :tf_saved_model_format
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @core_ml_format = args[:core_ml_format] if args.key?(:core_ml_format)
+          @docker_format = args[:docker_format] if args.key?(:docker_format)
+          @edge_tpu_tf_lite_format = args[:edge_tpu_tf_lite_format] if args.key?(:edge_tpu_tf_lite_format)
+          @export_firebase_auxiliary_info = args[:export_firebase_auxiliary_info] if args.key?(:export_firebase_auxiliary_info)
+          @output_gcr_uri = args[:output_gcr_uri] if args.key?(:output_gcr_uri)
+          @output_gcs_uri = args[:output_gcs_uri] if args.key?(:output_gcs_uri)
+          @tf_js_format = args[:tf_js_format] if args.key?(:tf_js_format)
+          @tf_lite_format = args[:tf_lite_format] if args.key?(:tf_lite_format)
+          @tf_saved_model_format = args[:tf_saved_model_format] if args.key?(:tf_saved_model_format)
+        end
+      end
+      
+      # Spec of input and output files, on external file systems (CNS, GCS, etc).
+      class XpsFileSpec
+        include Google::Apis::Core::Hashable
+      
+        # Deprecated. Use file_spec.
+        # Corresponds to the JSON property `directoryPath`
+        # @return [String]
+        attr_accessor :directory_path
+      
+        # 
+        # Corresponds to the JSON property `fileFormat`
+        # @return [String]
+        attr_accessor :file_format
+      
+        # Single file path, or file pattern of format "/path/to/file@shard_count". E.g. /
+        # cns/cell-d/somewhere/file@2 is expanded to two files: /cns/cell-d/somewhere/
+        # file-00000-of-00002 and /cns/cell-d/somewhere/file-00001-of-00002.
+        # Corresponds to the JSON property `fileSpec`
+        # @return [String]
+        attr_accessor :file_spec
+      
+        # Deprecated. Use file_spec.
+        # Corresponds to the JSON property `singleFilePath`
+        # @return [String]
+        attr_accessor :single_file_path
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @directory_path = args[:directory_path] if args.key?(:directory_path)
+          @file_format = args[:file_format] if args.key?(:file_format)
+          @file_spec = args[:file_spec] if args.key?(:file_spec)
+          @single_file_path = args[:single_file_path] if args.key?(:single_file_path)
+        end
+      end
+      
+      # The data statistics of a series of FLOAT64 values.
+      class XpsFloat64Stats
+        include Google::Apis::Core::Hashable
+      
+        # Common statistics for a column with a specified data type.
+        # Corresponds to the JSON property `commonStats`
+        # @return [Google::Apis::LanguageV2::XpsCommonStats]
+        attr_accessor :common_stats
+      
+        # Histogram buckets of the data series. Sorted by the min value of the bucket,
+        # ascendingly, and the number of the buckets is dynamically generated. The
+        # buckets are non-overlapping and completely cover whole FLOAT64 range with min
+        # of first bucket being `"-Infinity"`, and max of the last one being `"Infinity"`
+        # .
+        # Corresponds to the JSON property `histogramBuckets`
+        # @return [Array<Google::Apis::LanguageV2::XpsFloat64StatsHistogramBucket>]
+        attr_accessor :histogram_buckets
+      
+        # The mean of the series.
+        # Corresponds to the JSON property `mean`
+        # @return [Float]
+        attr_accessor :mean
+      
+        # Ordered from 0 to k k-quantile values of the data series of n values. The
+        # value at index i is, approximately, the i*n/k-th smallest value in the series;
+        # for i = 0 and i = k these are, respectively, the min and max values.
+        # Corresponds to the JSON property `quantiles`
+        # @return [Array<Float>]
+        attr_accessor :quantiles
+      
+        # The standard deviation of the series.
+        # Corresponds to the JSON property `standardDeviation`
+        # @return [Float]
+        attr_accessor :standard_deviation
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @common_stats = args[:common_stats] if args.key?(:common_stats)
+          @histogram_buckets = args[:histogram_buckets] if args.key?(:histogram_buckets)
+          @mean = args[:mean] if args.key?(:mean)
+          @quantiles = args[:quantiles] if args.key?(:quantiles)
+          @standard_deviation = args[:standard_deviation] if args.key?(:standard_deviation)
+        end
+      end
+      
+      # A bucket of a histogram.
+      class XpsFloat64StatsHistogramBucket
+        include Google::Apis::Core::Hashable
+      
+        # The number of data values that are in the bucket, i.e. are between min and max
+        # values.
+        # Corresponds to the JSON property `count`
+        # @return [Fixnum]
+        attr_accessor :count
+      
+        # The maximum value of the bucket, exclusive unless max = `"Infinity"`, in which
+        # case it's inclusive.
+        # Corresponds to the JSON property `max`
+        # @return [Float]
+        attr_accessor :max
+      
+        # The minimum value of the bucket, inclusive.
+        # Corresponds to the JSON property `min`
+        # @return [Float]
+        attr_accessor :min
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @count = args[:count] if args.key?(:count)
+          @max = args[:max] if args.key?(:max)
+          @min = args[:min] if args.key?(:min)
+        end
+      end
+      
+      # 
+      class XpsImageClassificationTrainResponse
+        include Google::Apis::Core::Hashable
+      
+        # Total number of classes.
+        # Corresponds to the JSON property `classCount`
+        # @return [Fixnum]
+        attr_accessor :class_count
+      
+        # Information of downloadable models that are pre-generated as part of training
+        # flow and will be persisted in AutoMl backend. Upon receiving ExportModel
+        # request from user, AutoMl backend can serve the pre-generated models to user
+        # if exists (by copying the files from internal path to user provided location),
+        # otherwise, AutoMl backend will call xPS ExportModel API to generate the model
+        # on the fly with the requesting format.
+        # Corresponds to the JSON property `exportModelSpec`
+        # @return [Google::Apis::LanguageV2::XpsImageExportModelSpec]
+        attr_accessor :export_model_spec
+      
+        # Stores the locations and related metadata of the model artifacts. Populated
+        # for uCAIP requests only.
+        # Corresponds to the JSON property `modelArtifactSpec`
+        # @return [Google::Apis::LanguageV2::XpsImageModelArtifactSpec]
+        attr_accessor :model_artifact_spec
+      
+        # Serving specification for image models.
+        # Corresponds to the JSON property `modelServingSpec`
+        # @return [Google::Apis::LanguageV2::XpsImageModelServingSpec]
+        attr_accessor :model_serving_spec
+      
+        # Stop reason for training job, e.g. 'TRAIN_BUDGET_REACHED', 'MODEL_CONVERGED', '
+        # MODEL_EARLY_STOPPED'.
+        # Corresponds to the JSON property `stopReason`
+        # @return [String]
+        attr_accessor :stop_reason
+      
+        # The actual cost to create this model. - For edge type model, the cost is
+        # expressed in node hour. - For cloud type model,the cost is expressed in
+        # compute hour. - Populated for models created before GA. To be deprecated after
+        # GA.
+        # Corresponds to the JSON property `trainCostInNodeTime`
+        # @return [String]
+        attr_accessor :train_cost_in_node_time
+      
+        # The actual training cost, expressed in node seconds. Populated for models
+        # trained in node time.
+        # Corresponds to the JSON property `trainCostNodeSeconds`
+        # @return [Fixnum]
+        attr_accessor :train_cost_node_seconds
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @class_count = args[:class_count] if args.key?(:class_count)
+          @export_model_spec = args[:export_model_spec] if args.key?(:export_model_spec)
+          @model_artifact_spec = args[:model_artifact_spec] if args.key?(:model_artifact_spec)
+          @model_serving_spec = args[:model_serving_spec] if args.key?(:model_serving_spec)
+          @stop_reason = args[:stop_reason] if args.key?(:stop_reason)
+          @train_cost_in_node_time = args[:train_cost_in_node_time] if args.key?(:train_cost_in_node_time)
+          @train_cost_node_seconds = args[:train_cost_node_seconds] if args.key?(:train_cost_node_seconds)
+        end
+      end
+      
+      # Information of downloadable models that are pre-generated as part of training
+      # flow and will be persisted in AutoMl backend. Upon receiving ExportModel
+      # request from user, AutoMl backend can serve the pre-generated models to user
+      # if exists (by copying the files from internal path to user provided location),
+      # otherwise, AutoMl backend will call xPS ExportModel API to generate the model
+      # on the fly with the requesting format.
+      class XpsImageExportModelSpec
+        include Google::Apis::Core::Hashable
+      
+        # Contains the model format and internal location of the model files to be
+        # exported/downloaded. Use the GCS bucket name which is provided via
+        # TrainRequest.gcs_bucket_name to store the model files.
+        # Corresponds to the JSON property `exportModelOutputConfig`
+        # @return [Array<Google::Apis::LanguageV2::XpsExportModelOutputConfig>]
+        attr_accessor :export_model_output_config
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @export_model_output_config = args[:export_model_output_config] if args.key?(:export_model_output_config)
+        end
+      end
+      
+      # Stores the locations and related metadata of the model artifacts. Populated
+      # for uCAIP requests only.
+      class XpsImageModelArtifactSpec
+        include Google::Apis::Core::Hashable
+      
+        # A single model artifact item.
+        # Corresponds to the JSON property `checkpointArtifact`
+        # @return [Google::Apis::LanguageV2::XpsModelArtifactItem]
+        attr_accessor :checkpoint_artifact
+      
+        # The model binary files in different formats for model export.
+        # Corresponds to the JSON property `exportArtifact`
+        # @return [Array<Google::Apis::LanguageV2::XpsModelArtifactItem>]
+        attr_accessor :export_artifact
+      
+        # GCS uri of decoded labels file for model export 'dict.txt'.
+        # Corresponds to the JSON property `labelGcsUri`
+        # @return [String]
+        attr_accessor :label_gcs_uri
+      
+        # A single model artifact item.
+        # Corresponds to the JSON property `servingArtifact`
+        # @return [Google::Apis::LanguageV2::XpsModelArtifactItem]
+        attr_accessor :serving_artifact
+      
+        # GCS uri prefix of Tensorflow JavaScript binary files 'groupX-shardXofX.bin'
+        # Deprecated.
+        # Corresponds to the JSON property `tfJsBinaryGcsPrefix`
+        # @return [String]
+        attr_accessor :tf_js_binary_gcs_prefix
+      
+        # GCS uri of Tensorflow Lite metadata 'tflite_metadata.json'.
+        # Corresponds to the JSON property `tfLiteMetadataGcsUri`
+        # @return [String]
+        attr_accessor :tf_lite_metadata_gcs_uri
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @checkpoint_artifact = args[:checkpoint_artifact] if args.key?(:checkpoint_artifact)
+          @export_artifact = args[:export_artifact] if args.key?(:export_artifact)
+          @label_gcs_uri = args[:label_gcs_uri] if args.key?(:label_gcs_uri)
+          @serving_artifact = args[:serving_artifact] if args.key?(:serving_artifact)
+          @tf_js_binary_gcs_prefix = args[:tf_js_binary_gcs_prefix] if args.key?(:tf_js_binary_gcs_prefix)
+          @tf_lite_metadata_gcs_uri = args[:tf_lite_metadata_gcs_uri] if args.key?(:tf_lite_metadata_gcs_uri)
+        end
+      end
+      
+      # Serving specification for image models.
+      class XpsImageModelServingSpec
+        include Google::Apis::Core::Hashable
+      
+        # Populate under uCAIP request scope.
+        # Corresponds to the JSON property `modelThroughputEstimation`
+        # @return [Array<Google::Apis::LanguageV2::XpsImageModelServingSpecModelThroughputEstimation>]
+        attr_accessor :model_throughput_estimation
+      
+        # An estimated value of how much traffic a node can serve. Populated for AutoMl
+        # request only.
+        # Corresponds to the JSON property `nodeQps`
+        # @return [Float]
+        attr_accessor :node_qps
+      
+        # ## The fields below are only populated under uCAIP request scope. https://
+        # cloud.google.com/ml-engine/docs/runtime-version-list
+        # Corresponds to the JSON property `tfRuntimeVersion`
+        # @return [String]
+        attr_accessor :tf_runtime_version
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @model_throughput_estimation = args[:model_throughput_estimation] if args.key?(:model_throughput_estimation)
+          @node_qps = args[:node_qps] if args.key?(:node_qps)
+          @tf_runtime_version = args[:tf_runtime_version] if args.key?(:tf_runtime_version)
+        end
+      end
+      
+      # 
+      class XpsImageModelServingSpecModelThroughputEstimation
+        include Google::Apis::Core::Hashable
+      
+        # 
+        # Corresponds to the JSON property `computeEngineAcceleratorType`
+        # @return [String]
+        attr_accessor :compute_engine_accelerator_type
+      
+        # Estimated latency.
+        # Corresponds to the JSON property `latencyInMilliseconds`
+        # @return [Float]
+        attr_accessor :latency_in_milliseconds
+      
+        # The approximate qps a deployed node can serve.
+        # Corresponds to the JSON property `nodeQps`
+        # @return [Float]
+        attr_accessor :node_qps
+      
+        # 
+        # Corresponds to the JSON property `servomaticPartitionType`
+        # @return [String]
+        attr_accessor :servomatic_partition_type
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @compute_engine_accelerator_type = args[:compute_engine_accelerator_type] if args.key?(:compute_engine_accelerator_type)
+          @latency_in_milliseconds = args[:latency_in_milliseconds] if args.key?(:latency_in_milliseconds)
+          @node_qps = args[:node_qps] if args.key?(:node_qps)
+          @servomatic_partition_type = args[:servomatic_partition_type] if args.key?(:servomatic_partition_type)
+        end
+      end
+      
+      # Model evaluation metrics for image object detection problems. Evaluates
+      # prediction quality of labeled bounding boxes.
+      class XpsImageObjectDetectionEvaluationMetrics
+        include Google::Apis::Core::Hashable
+      
+        # The single metric for bounding boxes evaluation: the mean_average_precision
+        # averaged over all bounding_box_metrics_entries.
+        # Corresponds to the JSON property `boundingBoxMeanAveragePrecision`
+        # @return [Float]
+        attr_accessor :bounding_box_mean_average_precision
+      
+        # The bounding boxes match metrics for each Intersection-over-union threshold 0.
+        # 05,0.10,...,0.95,0.96,0.97,0.98,0.99 and each label confidence threshold 0.05,
+        # 0.10,...,0.95,0.96,0.97,0.98,0.99 pair.
+        # Corresponds to the JSON property `boundingBoxMetricsEntries`
+        # @return [Array<Google::Apis::LanguageV2::XpsBoundingBoxMetricsEntry>]
+        attr_accessor :bounding_box_metrics_entries
+      
+        # The total number of bounding boxes (i.e. summed over all images) the ground
+        # truth used to create this evaluation had.
+        # Corresponds to the JSON property `evaluatedBoundingBoxCount`
+        # @return [Fixnum]
+        attr_accessor :evaluated_bounding_box_count
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @bounding_box_mean_average_precision = args[:bounding_box_mean_average_precision] if args.key?(:bounding_box_mean_average_precision)
+          @bounding_box_metrics_entries = args[:bounding_box_metrics_entries] if args.key?(:bounding_box_metrics_entries)
+          @evaluated_bounding_box_count = args[:evaluated_bounding_box_count] if args.key?(:evaluated_bounding_box_count)
+        end
+      end
+      
+      # 
+      class XpsImageObjectDetectionModelSpec
+        include Google::Apis::Core::Hashable
+      
+        # Total number of classes.
+        # Corresponds to the JSON property `classCount`
+        # @return [Fixnum]
+        attr_accessor :class_count
+      
+        # Information of downloadable models that are pre-generated as part of training
+        # flow and will be persisted in AutoMl backend. Upon receiving ExportModel
+        # request from user, AutoMl backend can serve the pre-generated models to user
+        # if exists (by copying the files from internal path to user provided location),
+        # otherwise, AutoMl backend will call xPS ExportModel API to generate the model
+        # on the fly with the requesting format.
+        # Corresponds to the JSON property `exportModelSpec`
+        # @return [Google::Apis::LanguageV2::XpsImageExportModelSpec]
+        attr_accessor :export_model_spec
+      
+        # Max number of bounding box.
+        # Corresponds to the JSON property `maxBoundingBoxCount`
+        # @return [Fixnum]
+        attr_accessor :max_bounding_box_count
+      
+        # Stores the locations and related metadata of the model artifacts. Populated
+        # for uCAIP requests only.
+        # Corresponds to the JSON property `modelArtifactSpec`
+        # @return [Google::Apis::LanguageV2::XpsImageModelArtifactSpec]
+        attr_accessor :model_artifact_spec
+      
+        # Serving specification for image models.
+        # Corresponds to the JSON property `modelServingSpec`
+        # @return [Google::Apis::LanguageV2::XpsImageModelServingSpec]
+        attr_accessor :model_serving_spec
+      
+        # Stop reason for training job, e.g. 'TRAIN_BUDGET_REACHED', 'MODEL_CONVERGED'.
+        # Corresponds to the JSON property `stopReason`
+        # @return [String]
+        attr_accessor :stop_reason
+      
+        # The actual train cost of creating this model, expressed in node seconds, i.e.
+        # 3,600 value in this field means 1 node hour.
+        # Corresponds to the JSON property `trainCostNodeSeconds`
+        # @return [Fixnum]
+        attr_accessor :train_cost_node_seconds
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @class_count = args[:class_count] if args.key?(:class_count)
+          @export_model_spec = args[:export_model_spec] if args.key?(:export_model_spec)
+          @max_bounding_box_count = args[:max_bounding_box_count] if args.key?(:max_bounding_box_count)
+          @model_artifact_spec = args[:model_artifact_spec] if args.key?(:model_artifact_spec)
+          @model_serving_spec = args[:model_serving_spec] if args.key?(:model_serving_spec)
+          @stop_reason = args[:stop_reason] if args.key?(:stop_reason)
+          @train_cost_node_seconds = args[:train_cost_node_seconds] if args.key?(:train_cost_node_seconds)
+        end
+      end
+      
+      # Model evaluation metrics for image segmentation problems. Next tag: 4.
+      class XpsImageSegmentationEvaluationMetrics
+        include Google::Apis::Core::Hashable
+      
+        # Metrics that have confidence thresholds. Precision-recall curve can be derived
+        # from it.
+        # Corresponds to the JSON property `confidenceMetricsEntries`
+        # @return [Array<Google::Apis::LanguageV2::XpsImageSegmentationEvaluationMetricsConfidenceMetricsEntry>]
+        attr_accessor :confidence_metrics_entries
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @confidence_metrics_entries = args[:confidence_metrics_entries] if args.key?(:confidence_metrics_entries)
+        end
+      end
+      
+      # Metrics for a single confidence threshold.
+      class XpsImageSegmentationEvaluationMetricsConfidenceMetricsEntry
+        include Google::Apis::Core::Hashable
+      
+        # The confidence threshold value used to compute the metrics.
+        # Corresponds to the JSON property `confidenceThreshold`
+        # @return [Float]
+        attr_accessor :confidence_threshold
+      
+        # Confusion matrix of the model running the classification.
+        # Corresponds to the JSON property `confusionMatrix`
+        # @return [Google::Apis::LanguageV2::XpsConfusionMatrix]
+        attr_accessor :confusion_matrix
+      
+        # DSC or the F1 score: The harmonic mean of recall and precision.
+        # Corresponds to the JSON property `diceScoreCoefficient`
+        # @return [Float]
+        attr_accessor :dice_score_coefficient
+      
+        # IOU score.
+        # Corresponds to the JSON property `iouScore`
+        # @return [Float]
+        attr_accessor :iou_score
+      
+        # Precision for the given confidence threshold.
+        # Corresponds to the JSON property `precision`
+        # @return [Float]
+        attr_accessor :precision
+      
+        # Recall for the given confidence threshold.
+        # Corresponds to the JSON property `recall`
+        # @return [Float]
+        attr_accessor :recall
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @confidence_threshold = args[:confidence_threshold] if args.key?(:confidence_threshold)
+          @confusion_matrix = args[:confusion_matrix] if args.key?(:confusion_matrix)
+          @dice_score_coefficient = args[:dice_score_coefficient] if args.key?(:dice_score_coefficient)
+          @iou_score = args[:iou_score] if args.key?(:iou_score)
+          @precision = args[:precision] if args.key?(:precision)
+          @recall = args[:recall] if args.key?(:recall)
+        end
+      end
+      
+      # 
+      class XpsImageSegmentationTrainResponse
+        include Google::Apis::Core::Hashable
+      
+        # Color map of the model.
+        # Corresponds to the JSON property `colorMaps`
+        # @return [Array<Google::Apis::LanguageV2::XpsColorMap>]
+        attr_accessor :color_maps
+      
+        # Information of downloadable models that are pre-generated as part of training
+        # flow and will be persisted in AutoMl backend. Upon receiving ExportModel
+        # request from user, AutoMl backend can serve the pre-generated models to user
+        # if exists (by copying the files from internal path to user provided location),
+        # otherwise, AutoMl backend will call xPS ExportModel API to generate the model
+        # on the fly with the requesting format.
+        # Corresponds to the JSON property `exportModelSpec`
+        # @return [Google::Apis::LanguageV2::XpsImageExportModelSpec]
+        attr_accessor :export_model_spec
+      
+        # Stores the locations and related metadata of the model artifacts. Populated
+        # for uCAIP requests only.
+        # Corresponds to the JSON property `modelArtifactSpec`
+        # @return [Google::Apis::LanguageV2::XpsImageModelArtifactSpec]
+        attr_accessor :model_artifact_spec
+      
+        # Serving specification for image models.
+        # Corresponds to the JSON property `modelServingSpec`
+        # @return [Google::Apis::LanguageV2::XpsImageModelServingSpec]
+        attr_accessor :model_serving_spec
+      
+        # Stop reason for training job, e.g. 'TRAIN_BUDGET_REACHED', 'MODEL_CONVERGED'.
+        # Corresponds to the JSON property `stopReason`
+        # @return [String]
+        attr_accessor :stop_reason
+      
+        # The actual train cost of creating this model, expressed in node seconds, i.e.
+        # 3,600 value in this field means 1 node hour.
+        # Corresponds to the JSON property `trainCostNodeSeconds`
+        # @return [Fixnum]
+        attr_accessor :train_cost_node_seconds
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @color_maps = args[:color_maps] if args.key?(:color_maps)
+          @export_model_spec = args[:export_model_spec] if args.key?(:export_model_spec)
+          @model_artifact_spec = args[:model_artifact_spec] if args.key?(:model_artifact_spec)
+          @model_serving_spec = args[:model_serving_spec] if args.key?(:model_serving_spec)
+          @stop_reason = args[:stop_reason] if args.key?(:stop_reason)
+          @train_cost_node_seconds = args[:train_cost_node_seconds] if args.key?(:train_cost_node_seconds)
+        end
+      end
+      
+      # 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
+      class XpsIntegratedGradientsAttribution
+        include Google::Apis::Core::Hashable
+      
+        # 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.
+        # Corresponds to the JSON property `stepCount`
+        # @return [Fixnum]
+        attr_accessor :step_count
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @step_count = args[:step_count] if args.key?(:step_count)
+        end
+      end
+      
+      # 
+      class XpsMetricEntry
+        include Google::Apis::Core::Hashable
+      
+        # For billing metrics that are using legacy sku's, set the legacy billing metric
+        # id here. This will be sent to Chemist as the "cloudbilling.googleapis.com/
+        # argentum_metric_id" label. Otherwise leave empty.
+        # Corresponds to the JSON property `argentumMetricId`
+        # @return [String]
+        attr_accessor :argentum_metric_id
+      
+        # A double value.
+        # Corresponds to the JSON property `doubleValue`
+        # @return [Float]
+        attr_accessor :double_value
+      
+        # A signed 64-bit integer value.
+        # Corresponds to the JSON property `int64Value`
+        # @return [Fixnum]
+        attr_accessor :int64_value
+      
+        # The metric name defined in the service configuration.
+        # Corresponds to the JSON property `metricName`
+        # @return [String]
+        attr_accessor :metric_name
+      
+        # Billing system labels for this (metric, value) pair.
+        # Corresponds to the JSON property `systemLabels`
+        # @return [Array<Google::Apis::LanguageV2::XpsMetricEntryLabel>]
+        attr_accessor :system_labels
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @argentum_metric_id = args[:argentum_metric_id] if args.key?(:argentum_metric_id)
+          @double_value = args[:double_value] if args.key?(:double_value)
+          @int64_value = args[:int64_value] if args.key?(:int64_value)
+          @metric_name = args[:metric_name] if args.key?(:metric_name)
+          @system_labels = args[:system_labels] if args.key?(:system_labels)
+        end
+      end
+      
+      # 
+      class XpsMetricEntryLabel
+        include Google::Apis::Core::Hashable
+      
+        # The name of the label.
+        # Corresponds to the JSON property `labelName`
+        # @return [String]
+        attr_accessor :label_name
+      
+        # The value of the label.
+        # Corresponds to the JSON property `labelValue`
+        # @return [String]
+        attr_accessor :label_value
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @label_name = args[:label_name] if args.key?(:label_name)
+          @label_value = args[:label_value] if args.key?(:label_value)
+        end
+      end
+      
+      # A single model artifact item.
+      class XpsModelArtifactItem
+        include Google::Apis::Core::Hashable
+      
+        # The model artifact format.
+        # Corresponds to the JSON property `artifactFormat`
+        # @return [String]
+        attr_accessor :artifact_format
+      
+        # The Google Cloud Storage (GCS) uri that stores the model binary files.
+        # Corresponds to the JSON property `gcsUri`
+        # @return [String]
+        attr_accessor :gcs_uri
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @artifact_format = args[:artifact_format] if args.key?(:artifact_format)
+          @gcs_uri = args[:gcs_uri] if args.key?(:gcs_uri)
+        end
+      end
+      
+      # Next ID: 8
+      class XpsPreprocessResponse
+        include Google::Apis::Core::Hashable
+      
+        # Set of examples or input sources.
+        # Corresponds to the JSON property `outputExampleSet`
+        # @return [Google::Apis::LanguageV2::XpsExampleSet]
+        attr_accessor :output_example_set
+      
+        # 
+        # Corresponds to the JSON property `speechPreprocessResp`
+        # @return [Google::Apis::LanguageV2::XpsSpeechPreprocessResponse]
+        attr_accessor :speech_preprocess_resp
+      
+        # 
+        # Corresponds to the JSON property `tablesPreprocessResponse`
+        # @return [Google::Apis::LanguageV2::XpsTablesPreprocessResponse]
+        attr_accessor :tables_preprocess_response
+      
+        # Translation preprocess response.
+        # Corresponds to the JSON property `translationPreprocessResp`
+        # @return [Google::Apis::LanguageV2::XpsTranslationPreprocessResponse]
+        attr_accessor :translation_preprocess_resp
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @output_example_set = args[:output_example_set] if args.key?(:output_example_set)
+          @speech_preprocess_resp = args[:speech_preprocess_resp] if args.key?(:speech_preprocess_resp)
+          @tables_preprocess_response = args[:tables_preprocess_response] if args.key?(:tables_preprocess_response)
+          @translation_preprocess_resp = args[:translation_preprocess_resp] if args.key?(:translation_preprocess_resp)
+        end
+      end
+      
+      # Model evaluation metrics for regression problems. It can be used for Tables.
+      class XpsRegressionEvaluationMetrics
+        include Google::Apis::Core::Hashable
+      
+        # Mean Absolute Error (MAE).
+        # Corresponds to the JSON property `meanAbsoluteError`
+        # @return [Float]
+        attr_accessor :mean_absolute_error
+      
+        # Mean absolute percentage error. Only set if all ground truth values are
+        # positive.
+        # Corresponds to the JSON property `meanAbsolutePercentageError`
+        # @return [Float]
+        attr_accessor :mean_absolute_percentage_error
+      
+        # R squared.
+        # Corresponds to the JSON property `rSquared`
+        # @return [Float]
+        attr_accessor :r_squared
+      
+        # A list of actual versus predicted points for the model being evaluated.
+        # Corresponds to the JSON property `regressionMetricsEntries`
+        # @return [Array<Google::Apis::LanguageV2::XpsRegressionMetricsEntry>]
+        attr_accessor :regression_metrics_entries
+      
+        # Root Mean Squared Error (RMSE).
+        # Corresponds to the JSON property `rootMeanSquaredError`
+        # @return [Float]
+        attr_accessor :root_mean_squared_error
+      
+        # Root mean squared log error.
+        # Corresponds to the JSON property `rootMeanSquaredLogError`
+        # @return [Float]
+        attr_accessor :root_mean_squared_log_error
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @mean_absolute_error = args[:mean_absolute_error] if args.key?(:mean_absolute_error)
+          @mean_absolute_percentage_error = args[:mean_absolute_percentage_error] if args.key?(:mean_absolute_percentage_error)
+          @r_squared = args[:r_squared] if args.key?(:r_squared)
+          @regression_metrics_entries = args[:regression_metrics_entries] if args.key?(:regression_metrics_entries)
+          @root_mean_squared_error = args[:root_mean_squared_error] if args.key?(:root_mean_squared_error)
+          @root_mean_squared_log_error = args[:root_mean_squared_log_error] if args.key?(:root_mean_squared_log_error)
+        end
+      end
+      
+      # A pair of actual & observed values for the model being evaluated.
+      class XpsRegressionMetricsEntry
+        include Google::Apis::Core::Hashable
+      
+        # The observed value for a row in the dataset.
+        # Corresponds to the JSON property `predictedValue`
+        # @return [Float]
+        attr_accessor :predicted_value
+      
+        # The actual target value for a row in the dataset.
+        # Corresponds to the JSON property `trueValue`
+        # @return [Float]
+        attr_accessor :true_value
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @predicted_value = args[:predicted_value] if args.key?(:predicted_value)
+          @true_value = args[:true_value] if args.key?(:true_value)
+        end
+      end
+      
+      # 
+      class XpsReportingMetrics
+        include Google::Apis::Core::Hashable
+      
+        # The effective time training used. If set, this is used for quota management
+        # and billing. Deprecated. AutoML BE doesn't use this. Don't set.
+        # Corresponds to the JSON property `effectiveTrainingDuration`
+        # @return [String]
+        attr_accessor :effective_training_duration
+      
+        # One entry per metric name. The values must be aggregated per metric name.
+        # Corresponds to the JSON property `metricEntries`
+        # @return [Array<Google::Apis::LanguageV2::XpsMetricEntry>]
+        attr_accessor :metric_entries
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @effective_training_duration = args[:effective_training_duration] if args.key?(:effective_training_duration)
+          @metric_entries = args[:metric_entries] if args.key?(:metric_entries)
+        end
+      end
+      
+      # 
+      class XpsResponseExplanationMetadata
+        include Google::Apis::Core::Hashable
+      
+        # Metadata of the input.
+        # Corresponds to the JSON property `inputs`
+        # @return [Hash<String,Google::Apis::LanguageV2::XpsResponseExplanationMetadataInputMetadata>]
+        attr_accessor :inputs
+      
+        # Metadata of the output.
+        # Corresponds to the JSON property `outputs`
+        # @return [Hash<String,Google::Apis::LanguageV2::XpsResponseExplanationMetadataOutputMetadata>]
+        attr_accessor :outputs
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @inputs = args[:inputs] if args.key?(:inputs)
+          @outputs = args[:outputs] if args.key?(:outputs)
+        end
+      end
+      
+      # Metadata of the input of a feature.
+      class XpsResponseExplanationMetadataInputMetadata
+        include Google::Apis::Core::Hashable
+      
+        # Name of the input tensor for this model. Only needed in train response.
+        # Corresponds to the JSON property `inputTensorName`
+        # @return [String]
+        attr_accessor :input_tensor_name
+      
+        # Modality of the feature. Valid values are: numeric, image. Defaults to numeric.
+        # Corresponds to the JSON property `modality`
+        # @return [String]
+        attr_accessor :modality
+      
+        # Visualization configurations for image explanation.
+        # Corresponds to the JSON property `visualizationConfig`
+        # @return [Google::Apis::LanguageV2::XpsVisualization]
+        attr_accessor :visualization_config
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @input_tensor_name = args[:input_tensor_name] if args.key?(:input_tensor_name)
+          @modality = args[:modality] if args.key?(:modality)
+          @visualization_config = args[:visualization_config] if args.key?(:visualization_config)
+        end
+      end
+      
+      # Metadata of the prediction output to be explained.
+      class XpsResponseExplanationMetadataOutputMetadata
+        include Google::Apis::Core::Hashable
+      
+        # Name of the output tensor. Only needed in train response.
+        # Corresponds to the JSON property `outputTensorName`
+        # @return [String]
+        attr_accessor :output_tensor_name
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @output_tensor_name = args[:output_tensor_name] if args.key?(:output_tensor_name)
+        end
+      end
+      
+      # 
+      class XpsResponseExplanationParameters
+        include Google::Apis::Core::Hashable
+      
+        # 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
+        # Corresponds to the JSON property `integratedGradientsAttribution`
+        # @return [Google::Apis::LanguageV2::XpsIntegratedGradientsAttribution]
+        attr_accessor :integrated_gradients_attribution
+      
+        # An explanation method that redistributes Integrated Gradients attributions to
+        # segmented regions, taking advantage of the model's fully differentiable
+        # structure. Refer to this paper for more details: https://arxiv.org/abs/1906.
+        # 02825 Only supports image Models (modality is IMAGE).
+        # Corresponds to the JSON property `xraiAttribution`
+        # @return [Google::Apis::LanguageV2::XpsXraiAttribution]
+        attr_accessor :xrai_attribution
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @integrated_gradients_attribution = args[:integrated_gradients_attribution] if args.key?(:integrated_gradients_attribution)
+          @xrai_attribution = args[:xrai_attribution] if args.key?(:xrai_attribution)
+        end
+      end
+      
+      # Specification of Model explanation. Feature-based XAI in AutoML Vision ICN is
+      # deprecated, see b/288407203 for context.
+      class XpsResponseExplanationSpec
+        include Google::Apis::Core::Hashable
+      
+        # Explanation type. For AutoML Image Classification models, possible values are:
+        # * `image-integrated-gradients` * `image-xrai`
+        # Corresponds to the JSON property `explanationType`
+        # @return [String]
+        attr_accessor :explanation_type
+      
+        # Metadata describing the Model's input and output for explanation.
+        # Corresponds to the JSON property `metadata`
+        # @return [Google::Apis::LanguageV2::XpsResponseExplanationMetadata]
+        attr_accessor :metadata
+      
+        # Parameters that configure explaining of the Model's predictions.
+        # Corresponds to the JSON property `parameters`
+        # @return [Google::Apis::LanguageV2::XpsResponseExplanationParameters]
+        attr_accessor :parameters
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @explanation_type = args[:explanation_type] if args.key?(:explanation_type)
+          @metadata = args[:metadata] if args.key?(:metadata)
+          @parameters = args[:parameters] if args.key?(:parameters)
+        end
+      end
+      
+      # 
+      class XpsRow
+        include Google::Apis::Core::Hashable
+      
+        # The ids of the columns. Note: The below `values` field must match order of
+        # this field, if this field is set.
+        # Corresponds to the JSON property `columnIds`
+        # @return [Array<Fixnum>]
+        attr_accessor :column_ids
+      
+        # The values of the row cells, given in the same order as the column_ids. If
+        # column_ids is not set, then in the same order as the input_feature_column_ids
+        # in TablesModelMetadata.
+        # Corresponds to the JSON property `values`
+        # @return [Array<Object>]
+        attr_accessor :values
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @column_ids = args[:column_ids] if args.key?(:column_ids)
+          @values = args[:values] if args.key?(:values)
+        end
+      end
+      
+      # 
+      class XpsSpeechEvaluationMetrics
+        include Google::Apis::Core::Hashable
+      
+        # Evaluation metrics for all submodels contained in this model.
+        # Corresponds to the JSON property `subModelEvaluationMetrics`
+        # @return [Array<Google::Apis::LanguageV2::XpsSpeechEvaluationMetricsSubModelEvaluationMetric>]
+        attr_accessor :sub_model_evaluation_metrics
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @sub_model_evaluation_metrics = args[:sub_model_evaluation_metrics] if args.key?(:sub_model_evaluation_metrics)
+        end
+      end
+      
+      # 
+      class XpsSpeechEvaluationMetricsSubModelEvaluationMetric
+        include Google::Apis::Core::Hashable
+      
+        # Type of the biasing model.
+        # Corresponds to the JSON property `biasingModelType`
+        # @return [String]
+        attr_accessor :biasing_model_type
+      
+        # If true then it means we have an enhanced version of the biasing models.
+        # Corresponds to the JSON property `isEnhancedModel`
+        # @return [Boolean]
+        attr_accessor :is_enhanced_model
+        alias_method :is_enhanced_model?, :is_enhanced_model
+      
+        # 
+        # Corresponds to the JSON property `numDeletions`
+        # @return [Fixnum]
+        attr_accessor :num_deletions
+      
+        # 
+        # Corresponds to the JSON property `numInsertions`
+        # @return [Fixnum]
+        attr_accessor :num_insertions
+      
+        # 
+        # Corresponds to the JSON property `numSubstitutions`
+        # @return [Fixnum]
+        attr_accessor :num_substitutions
+      
+        # Number of utterances used in the wer computation.
+        # Corresponds to the JSON property `numUtterances`
+        # @return [Fixnum]
+        attr_accessor :num_utterances
+      
+        # Number of words over which the word error rate was computed.
+        # Corresponds to the JSON property `numWords`
+        # @return [Fixnum]
+        attr_accessor :num_words
+      
+        # Below fields are used for debugging purposes
+        # Corresponds to the JSON property `sentenceAccuracy`
+        # @return [Float]
+        attr_accessor :sentence_accuracy
+      
+        # Word error rate (standard error metric used for speech recognition).
+        # Corresponds to the JSON property `wer`
+        # @return [Float]
+        attr_accessor :wer
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @biasing_model_type = args[:biasing_model_type] if args.key?(:biasing_model_type)
+          @is_enhanced_model = args[:is_enhanced_model] if args.key?(:is_enhanced_model)
+          @num_deletions = args[:num_deletions] if args.key?(:num_deletions)
+          @num_insertions = args[:num_insertions] if args.key?(:num_insertions)
+          @num_substitutions = args[:num_substitutions] if args.key?(:num_substitutions)
+          @num_utterances = args[:num_utterances] if args.key?(:num_utterances)
+          @num_words = args[:num_words] if args.key?(:num_words)
+          @sentence_accuracy = args[:sentence_accuracy] if args.key?(:sentence_accuracy)
+          @wer = args[:wer] if args.key?(:wer)
+        end
+      end
+      
+      # 
+      class XpsSpeechModelSpec
+        include Google::Apis::Core::Hashable
+      
+        # Required for speech xps backend. Speech xps has to use dataset_id and model_id
+        # as the primary key in db so that speech API can query the db directly.
+        # Corresponds to the JSON property `datasetId`
+        # @return [Fixnum]
+        attr_accessor :dataset_id
+      
+        # 
+        # Corresponds to the JSON property `language`
+        # @return [String]
+        attr_accessor :language
+      
+        # Model specs for all submodels contained in this model.
+        # Corresponds to the JSON property `subModelSpecs`
+        # @return [Array<Google::Apis::LanguageV2::XpsSpeechModelSpecSubModelSpec>]
+        attr_accessor :sub_model_specs
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @dataset_id = args[:dataset_id] if args.key?(:dataset_id)
+          @language = args[:language] if args.key?(:language)
+          @sub_model_specs = args[:sub_model_specs] if args.key?(:sub_model_specs)
+        end
+      end
+      
+      # 
+      class XpsSpeechModelSpecSubModelSpec
+        include Google::Apis::Core::Hashable
+      
+        # Type of the biasing model.
+        # Corresponds to the JSON property `biasingModelType`
+        # @return [String]
+        attr_accessor :biasing_model_type
+      
+        # In S3, Recognition ClientContextId.client_id
+        # Corresponds to the JSON property `clientId`
+        # @return [String]
+        attr_accessor :client_id
+      
+        # In S3, Recognition ClientContextId.context_id
+        # Corresponds to the JSON property `contextId`
+        # @return [String]
+        attr_accessor :context_id
+      
+        # If true then it means we have an enhanced version of the biasing models.
+        # Corresponds to the JSON property `isEnhancedModel`
+        # @return [Boolean]
+        attr_accessor :is_enhanced_model
+        alias_method :is_enhanced_model?, :is_enhanced_model
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @biasing_model_type = args[:biasing_model_type] if args.key?(:biasing_model_type)
+          @client_id = args[:client_id] if args.key?(:client_id)
+          @context_id = args[:context_id] if args.key?(:context_id)
+          @is_enhanced_model = args[:is_enhanced_model] if args.key?(:is_enhanced_model)
+        end
+      end
+      
+      # 
+      class XpsSpeechPreprocessResponse
+        include Google::Apis::Core::Hashable
+      
+        # Location od shards of sstables (test data) of DataUtterance protos.
+        # Corresponds to the JSON property `cnsTestDataPath`
+        # @return [String]
+        attr_accessor :cns_test_data_path
+      
+        # Location of shards of sstables (training data) of DataUtterance protos.
+        # Corresponds to the JSON property `cnsTrainDataPath`
+        # @return [String]
+        attr_accessor :cns_train_data_path
+      
+        # The metrics for prebuilt speech models. They are included here because there
+        # is no prebuilt speech models stored in the AutoML.
+        # Corresponds to the JSON property `prebuiltModelEvaluationMetrics`
+        # @return [Google::Apis::LanguageV2::XpsSpeechEvaluationMetrics]
+        attr_accessor :prebuilt_model_evaluation_metrics
+      
+        # Stats associated with the data.
+        # Corresponds to the JSON property `speechPreprocessStats`
+        # @return [Google::Apis::LanguageV2::XpsSpeechPreprocessStats]
+        attr_accessor :speech_preprocess_stats
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @cns_test_data_path = args[:cns_test_data_path] if args.key?(:cns_test_data_path)
+          @cns_train_data_path = args[:cns_train_data_path] if args.key?(:cns_train_data_path)
+          @prebuilt_model_evaluation_metrics = args[:prebuilt_model_evaluation_metrics] if args.key?(:prebuilt_model_evaluation_metrics)
+          @speech_preprocess_stats = args[:speech_preprocess_stats] if args.key?(:speech_preprocess_stats)
+        end
+      end
+      
+      # 
+      class XpsSpeechPreprocessStats
+        include Google::Apis::Core::Hashable
+      
+        # Different types of data errors and the counts associated with them.
+        # Corresponds to the JSON property `dataErrors`
+        # @return [Array<Google::Apis::LanguageV2::XpsDataErrors>]
+        attr_accessor :data_errors
+      
+        # The number of rows marked HUMAN_LABELLED
+        # Corresponds to the JSON property `numHumanLabeledExamples`
+        # @return [Fixnum]
+        attr_accessor :num_human_labeled_examples
+      
+        # The number of samples found in the previously recorded logs data.
+        # Corresponds to the JSON property `numLogsExamples`
+        # @return [Fixnum]
+        attr_accessor :num_logs_examples
+      
+        # The number of rows marked as MACHINE_TRANSCRIBED
+        # Corresponds to the JSON property `numMachineTranscribedExamples`
+        # @return [Fixnum]
+        attr_accessor :num_machine_transcribed_examples
+      
+        # The number of examples labelled as TEST by Speech xps server.
+        # Corresponds to the JSON property `testExamplesCount`
+        # @return [Fixnum]
+        attr_accessor :test_examples_count
+      
+        # The number of sentences in the test data set.
+        # Corresponds to the JSON property `testSentencesCount`
+        # @return [Fixnum]
+        attr_accessor :test_sentences_count
+      
+        # The number of words in the test data set.
+        # Corresponds to the JSON property `testWordsCount`
+        # @return [Fixnum]
+        attr_accessor :test_words_count
+      
+        # The number of examples labeled as TRAIN by Speech xps server.
+        # Corresponds to the JSON property `trainExamplesCount`
+        # @return [Fixnum]
+        attr_accessor :train_examples_count
+      
+        # The number of sentences in the training data set.
+        # Corresponds to the JSON property `trainSentencesCount`
+        # @return [Fixnum]
+        attr_accessor :train_sentences_count
+      
+        # The number of words in the training data set.
+        # Corresponds to the JSON property `trainWordsCount`
+        # @return [Fixnum]
+        attr_accessor :train_words_count
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @data_errors = args[:data_errors] if args.key?(:data_errors)
+          @num_human_labeled_examples = args[:num_human_labeled_examples] if args.key?(:num_human_labeled_examples)
+          @num_logs_examples = args[:num_logs_examples] if args.key?(:num_logs_examples)
+          @num_machine_transcribed_examples = args[:num_machine_transcribed_examples] if args.key?(:num_machine_transcribed_examples)
+          @test_examples_count = args[:test_examples_count] if args.key?(:test_examples_count)
+          @test_sentences_count = args[:test_sentences_count] if args.key?(:test_sentences_count)
+          @test_words_count = args[:test_words_count] if args.key?(:test_words_count)
+          @train_examples_count = args[:train_examples_count] if args.key?(:train_examples_count)
+          @train_sentences_count = args[:train_sentences_count] if args.key?(:train_sentences_count)
+          @train_words_count = args[:train_words_count] if args.key?(:train_words_count)
+        end
+      end
+      
+      # The data statistics of a series of STRING values.
+      class XpsStringStats
+        include Google::Apis::Core::Hashable
+      
+        # Common statistics for a column with a specified data type.
+        # Corresponds to the JSON property `commonStats`
+        # @return [Google::Apis::LanguageV2::XpsCommonStats]
+        attr_accessor :common_stats
+      
+        # The statistics of the top 20 unigrams, ordered by StringStats.UnigramStats.
+        # count.
+        # Corresponds to the JSON property `topUnigramStats`
+        # @return [Array<Google::Apis::LanguageV2::XpsStringStatsUnigramStats>]
+        attr_accessor :top_unigram_stats
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @common_stats = args[:common_stats] if args.key?(:common_stats)
+          @top_unigram_stats = args[:top_unigram_stats] if args.key?(:top_unigram_stats)
+        end
+      end
+      
+      # The statistics of a unigram.
+      class XpsStringStatsUnigramStats
+        include Google::Apis::Core::Hashable
+      
+        # The number of occurrences of this unigram in the series.
+        # Corresponds to the JSON property `count`
+        # @return [Fixnum]
+        attr_accessor :count
+      
+        # The unigram.
+        # Corresponds to the JSON property `value`
+        # @return [String]
+        attr_accessor :value
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @count = args[:count] if args.key?(:count)
+          @value = args[:value] if args.key?(:value)
+        end
+      end
+      
+      # The data statistics of a series of STRUCT values.
+      class XpsStructStats
+        include Google::Apis::Core::Hashable
+      
+        # Common statistics for a column with a specified data type.
+        # Corresponds to the JSON property `commonStats`
+        # @return [Google::Apis::LanguageV2::XpsCommonStats]
+        attr_accessor :common_stats
+      
+        # Map from a field name of the struct to data stats aggregated over series of
+        # all data in that field across all the structs.
+        # Corresponds to the JSON property `fieldStats`
+        # @return [Hash<String,Google::Apis::LanguageV2::XpsDataStats>]
+        attr_accessor :field_stats
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @common_stats = args[:common_stats] if args.key?(:common_stats)
+          @field_stats = args[:field_stats] if args.key?(:field_stats)
+        end
+      end
+      
+      # `StructType` defines the DataType-s of a STRUCT type.
+      class XpsStructType
+        include Google::Apis::Core::Hashable
+      
+        # Unordered map of struct field names to their data types.
+        # Corresponds to the JSON property `fields`
+        # @return [Hash<String,Google::Apis::LanguageV2::XpsDataType>]
+        attr_accessor :fields
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @fields = args[:fields] if args.key?(:fields)
+        end
+      end
+      
+      # 
+      class XpsTableSpec
+        include Google::Apis::Core::Hashable
+      
+        # Mapping from column id to column spec.
+        # Corresponds to the JSON property `columnSpecs`
+        # @return [Hash<String,Google::Apis::LanguageV2::XpsColumnSpec>]
+        attr_accessor :column_specs
+      
+        # The total size of imported data of the table.
+        # Corresponds to the JSON property `importedDataSizeInBytes`
+        # @return [Fixnum]
+        attr_accessor :imported_data_size_in_bytes
+      
+        # The number of rows in the table.
+        # Corresponds to the JSON property `rowCount`
+        # @return [Fixnum]
+        attr_accessor :row_count
+      
+        # The id of the time column.
+        # Corresponds to the JSON property `timeColumnId`
+        # @return [Fixnum]
+        attr_accessor :time_column_id
+      
+        # The number of valid rows.
+        # Corresponds to the JSON property `validRowCount`
+        # @return [Fixnum]
+        attr_accessor :valid_row_count
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @column_specs = args[:column_specs] if args.key?(:column_specs)
+          @imported_data_size_in_bytes = args[:imported_data_size_in_bytes] if args.key?(:imported_data_size_in_bytes)
+          @row_count = args[:row_count] if args.key?(:row_count)
+          @time_column_id = args[:time_column_id] if args.key?(:time_column_id)
+          @valid_row_count = args[:valid_row_count] if args.key?(:valid_row_count)
+        end
+      end
+      
+      # Metrics for Tables classification problems.
+      class XpsTablesClassificationMetrics
+        include Google::Apis::Core::Hashable
+      
+        # Metrics building a curve.
+        # Corresponds to the JSON property `curveMetrics`
+        # @return [Array<Google::Apis::LanguageV2::XpsTablesClassificationMetricsCurveMetrics>]
+        attr_accessor :curve_metrics
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @curve_metrics = args[:curve_metrics] if args.key?(:curve_metrics)
+        end
+      end
+      
+      # Metrics curve data point for a single value.
+      class XpsTablesClassificationMetricsCurveMetrics
+        include Google::Apis::Core::Hashable
+      
+        # The area under the precision-recall curve.
+        # Corresponds to the JSON property `aucPr`
+        # @return [Float]
+        attr_accessor :auc_pr
+      
+        # The area under receiver operating characteristic curve.
+        # Corresponds to the JSON property `aucRoc`
+        # @return [Float]
+        attr_accessor :auc_roc
+      
+        # Metrics that have confidence thresholds. Precision-recall curve and ROC curve
+        # can be derived from them.
+        # Corresponds to the JSON property `confidenceMetricsEntries`
+        # @return [Array<Google::Apis::LanguageV2::XpsTablesConfidenceMetricsEntry>]
+        attr_accessor :confidence_metrics_entries
+      
+        # The Log loss metric.
+        # Corresponds to the JSON property `logLoss`
+        # @return [Float]
+        attr_accessor :log_loss
+      
+        # The position threshold value used to compute the metrics.
+        # Corresponds to the JSON property `positionThreshold`
+        # @return [Fixnum]
+        attr_accessor :position_threshold
+      
+        # The CATEGORY row value (for ARRAY unnested) the curve metrics are for.
+        # Corresponds to the JSON property `value`
+        # @return [String]
+        attr_accessor :value
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @auc_pr = args[:auc_pr] if args.key?(:auc_pr)
+          @auc_roc = args[:auc_roc] if args.key?(:auc_roc)
+          @confidence_metrics_entries = args[:confidence_metrics_entries] if args.key?(:confidence_metrics_entries)
+          @log_loss = args[:log_loss] if args.key?(:log_loss)
+          @position_threshold = args[:position_threshold] if args.key?(:position_threshold)
+          @value = args[:value] if args.key?(:value)
+        end
+      end
+      
+      # Metrics for a single confidence threshold.
+      class XpsTablesConfidenceMetricsEntry
+        include Google::Apis::Core::Hashable
+      
+        # The confidence threshold value used to compute the metrics.
+        # Corresponds to the JSON property `confidenceThreshold`
+        # @return [Float]
+        attr_accessor :confidence_threshold
+      
+        # The harmonic mean of recall and precision. (2 * precision * recall) / (
+        # precision + recall)
+        # Corresponds to the JSON property `f1Score`
+        # @return [Float]
+        attr_accessor :f1_score
+      
+        # False negative count.
+        # Corresponds to the JSON property `falseNegativeCount`
+        # @return [Fixnum]
+        attr_accessor :false_negative_count
+      
+        # False positive count.
+        # Corresponds to the JSON property `falsePositiveCount`
+        # @return [Fixnum]
+        attr_accessor :false_positive_count
+      
+        # FPR = #false positives / (#false positives + #true negatives)
+        # Corresponds to the JSON property `falsePositiveRate`
+        # @return [Float]
+        attr_accessor :false_positive_rate
+      
+        # Precision = #true positives / (#true positives + #false positives).
+        # Corresponds to the JSON property `precision`
+        # @return [Float]
+        attr_accessor :precision
+      
+        # Recall = #true positives / (#true positives + #false negatives).
+        # Corresponds to the JSON property `recall`
+        # @return [Float]
+        attr_accessor :recall
+      
+        # True negative count.
+        # Corresponds to the JSON property `trueNegativeCount`
+        # @return [Fixnum]
+        attr_accessor :true_negative_count
+      
+        # True positive count.
+        # Corresponds to the JSON property `truePositiveCount`
+        # @return [Fixnum]
+        attr_accessor :true_positive_count
+      
+        # TPR = #true positives / (#true positives + #false negatvies)
+        # Corresponds to the JSON property `truePositiveRate`
+        # @return [Float]
+        attr_accessor :true_positive_rate
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @confidence_threshold = args[:confidence_threshold] if args.key?(:confidence_threshold)
+          @f1_score = args[:f1_score] if args.key?(:f1_score)
+          @false_negative_count = args[:false_negative_count] if args.key?(:false_negative_count)
+          @false_positive_count = args[:false_positive_count] if args.key?(:false_positive_count)
+          @false_positive_rate = args[:false_positive_rate] if args.key?(:false_positive_rate)
+          @precision = args[:precision] if args.key?(:precision)
+          @recall = args[:recall] if args.key?(:recall)
+          @true_negative_count = args[:true_negative_count] if args.key?(:true_negative_count)
+          @true_positive_count = args[:true_positive_count] if args.key?(:true_positive_count)
+          @true_positive_rate = args[:true_positive_rate] if args.key?(:true_positive_rate)
+        end
+      end
+      
+      # Metadata for a dataset used for AutoML Tables. Next ID: 6
+      class XpsTablesDatasetMetadata
+        include Google::Apis::Core::Hashable
+      
+        # Id the column to split the table.
+        # Corresponds to the JSON property `mlUseColumnId`
+        # @return [Fixnum]
+        attr_accessor :ml_use_column_id
+      
+        # Primary table.
+        # Corresponds to the JSON property `primaryTableSpec`
+        # @return [Google::Apis::LanguageV2::XpsTableSpec]
+        attr_accessor :primary_table_spec
+      
+        # (the column id : its CorrelationStats with target column).
+        # Corresponds to the JSON property `targetColumnCorrelations`
+        # @return [Hash<String,Google::Apis::LanguageV2::XpsCorrelationStats>]
+        attr_accessor :target_column_correlations
+      
+        # Id of the primary table column that should be used as the training label.
+        # Corresponds to the JSON property `targetColumnId`
+        # @return [Fixnum]
+        attr_accessor :target_column_id
+      
+        # Id of the primary table column that should be used as the weight column.
+        # Corresponds to the JSON property `weightColumnId`
+        # @return [Fixnum]
+        attr_accessor :weight_column_id
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @ml_use_column_id = args[:ml_use_column_id] if args.key?(:ml_use_column_id)
+          @primary_table_spec = args[:primary_table_spec] if args.key?(:primary_table_spec)
+          @target_column_correlations = args[:target_column_correlations] if args.key?(:target_column_correlations)
+          @target_column_id = args[:target_column_id] if args.key?(:target_column_id)
+          @weight_column_id = args[:weight_column_id] if args.key?(:weight_column_id)
+        end
+      end
+      
+      # 
+      class XpsTablesEvaluationMetrics
+        include Google::Apis::Core::Hashable
+      
+        # Metrics for Tables classification problems.
+        # Corresponds to the JSON property `classificationMetrics`
+        # @return [Google::Apis::LanguageV2::XpsTablesClassificationMetrics]
+        attr_accessor :classification_metrics
+      
+        # Metrics for Tables regression problems.
+        # Corresponds to the JSON property `regressionMetrics`
+        # @return [Google::Apis::LanguageV2::XpsTablesRegressionMetrics]
+        attr_accessor :regression_metrics
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @classification_metrics = args[:classification_metrics] if args.key?(:classification_metrics)
+          @regression_metrics = args[:regression_metrics] if args.key?(:regression_metrics)
+        end
+      end
+      
+      # An information specific to given column and Tables Model, in context of the
+      # Model and the predictions created by it.
+      class XpsTablesModelColumnInfo
+        include Google::Apis::Core::Hashable
+      
+        # The ID of the column.
+        # Corresponds to the JSON property `columnId`
+        # @return [Fixnum]
+        attr_accessor :column_id
+      
+        # When given as part of a Model: Measurement of how much model predictions
+        # correctness on the TEST data depend on values in this column. A value between
+        # 0 and 1, higher means higher influence. These values are normalized - for all
+        # input feature columns of a given model they add to 1. When given back by
+        # Predict or Batch Predict: Measurement of how impactful for the prediction
+        # returned for the given row the value in this column was. Specifically, the
+        # feature importance specifies the marginal contribution that the feature made
+        # to the prediction score compared to the baseline score. These values are
+        # computed using the Sampled Shapley method.
+        # Corresponds to the JSON property `featureImportance`
+        # @return [Float]
+        attr_accessor :feature_importance
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @column_id = args[:column_id] if args.key?(:column_id)
+          @feature_importance = args[:feature_importance] if args.key?(:feature_importance)
+        end
+      end
+      
+      # A description of Tables model structure.
+      class XpsTablesModelStructure
+        include Google::Apis::Core::Hashable
+      
+        # A list of models.
+        # Corresponds to the JSON property `modelParameters`
+        # @return [Array<Google::Apis::LanguageV2::XpsTablesModelStructureModelParameters>]
+        attr_accessor :model_parameters
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @model_parameters = args[:model_parameters] if args.key?(:model_parameters)
+        end
+      end
+      
+      # Model hyper-parameters for a model.
+      class XpsTablesModelStructureModelParameters
+        include Google::Apis::Core::Hashable
+      
+        # 
+        # Corresponds to the JSON property `hyperparameters`
+        # @return [Array<Google::Apis::LanguageV2::XpsTablesModelStructureModelParametersParameter>]
+        attr_accessor :hyperparameters
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @hyperparameters = args[:hyperparameters] if args.key?(:hyperparameters)
+        end
+      end
+      
+      # 
+      class XpsTablesModelStructureModelParametersParameter
+        include Google::Apis::Core::Hashable
+      
+        # Float type parameter value.
+        # Corresponds to the JSON property `floatValue`
+        # @return [Float]
+        attr_accessor :float_value
+      
+        # Integer type parameter value.
+        # Corresponds to the JSON property `intValue`
+        # @return [Fixnum]
+        attr_accessor :int_value
+      
+        # Parameter name.
+        # Corresponds to the JSON property `name`
+        # @return [String]
+        attr_accessor :name
+      
+        # String type parameter value.
+        # Corresponds to the JSON property `stringValue`
+        # @return [String]
+        attr_accessor :string_value
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @float_value = args[:float_value] if args.key?(:float_value)
+          @int_value = args[:int_value] if args.key?(:int_value)
+          @name = args[:name] if args.key?(:name)
+          @string_value = args[:string_value] if args.key?(:string_value)
+        end
+      end
+      
+      # 
+      class XpsTablesPreprocessResponse
+        include Google::Apis::Core::Hashable
+      
+        # Metadata for a dataset used for AutoML Tables. Next ID: 6
+        # Corresponds to the JSON property `tablesDatasetMetadata`
+        # @return [Google::Apis::LanguageV2::XpsTablesDatasetMetadata]
+        attr_accessor :tables_dataset_metadata
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @tables_dataset_metadata = args[:tables_dataset_metadata] if args.key?(:tables_dataset_metadata)
+        end
+      end
+      
+      # Metrics for Tables regression problems.
+      class XpsTablesRegressionMetrics
+        include Google::Apis::Core::Hashable
+      
+        # Mean absolute error.
+        # Corresponds to the JSON property `meanAbsoluteError`
+        # @return [Float]
+        attr_accessor :mean_absolute_error
+      
+        # Mean absolute percentage error, only set if all of the target column's values
+        # are positive.
+        # Corresponds to the JSON property `meanAbsolutePercentageError`
+        # @return [Float]
+        attr_accessor :mean_absolute_percentage_error
+      
+        # R squared.
+        # Corresponds to the JSON property `rSquared`
+        # @return [Float]
+        attr_accessor :r_squared
+      
+        # A list of actual versus predicted points for the model being evaluated.
+        # Corresponds to the JSON property `regressionMetricsEntries`
+        # @return [Array<Google::Apis::LanguageV2::XpsRegressionMetricsEntry>]
+        attr_accessor :regression_metrics_entries
+      
+        # Root mean squared error.
+        # Corresponds to the JSON property `rootMeanSquaredError`
+        # @return [Float]
+        attr_accessor :root_mean_squared_error
+      
+        # Root mean squared log error.
+        # Corresponds to the JSON property `rootMeanSquaredLogError`
+        # @return [Float]
+        attr_accessor :root_mean_squared_log_error
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @mean_absolute_error = args[:mean_absolute_error] if args.key?(:mean_absolute_error)
+          @mean_absolute_percentage_error = args[:mean_absolute_percentage_error] if args.key?(:mean_absolute_percentage_error)
+          @r_squared = args[:r_squared] if args.key?(:r_squared)
+          @regression_metrics_entries = args[:regression_metrics_entries] if args.key?(:regression_metrics_entries)
+          @root_mean_squared_error = args[:root_mean_squared_error] if args.key?(:root_mean_squared_error)
+          @root_mean_squared_log_error = args[:root_mean_squared_log_error] if args.key?(:root_mean_squared_log_error)
+        end
+      end
+      
+      # 
+      class XpsTablesTrainResponse
+        include Google::Apis::Core::Hashable
+      
+        # A description of Tables model structure.
+        # Corresponds to the JSON property `modelStructure`
+        # @return [Google::Apis::LanguageV2::XpsTablesModelStructure]
+        attr_accessor :model_structure
+      
+        # Sample rows from the dataset this model was trained.
+        # Corresponds to the JSON property `predictionSampleRows`
+        # @return [Array<Google::Apis::LanguageV2::XpsRow>]
+        attr_accessor :prediction_sample_rows
+      
+        # Output only. Auxiliary information for each of the input_feature_column_specs,
+        # with respect to this particular model.
+        # Corresponds to the JSON property `tablesModelColumnInfo`
+        # @return [Array<Google::Apis::LanguageV2::XpsTablesModelColumnInfo>]
+        attr_accessor :tables_model_column_info
+      
+        # The actual training cost of the model, expressed in milli node hours, i.e. 1,
+        # 000 value in this field means 1 node hour. Guaranteed to not exceed the train
+        # budget.
+        # Corresponds to the JSON property `trainCostMilliNodeHours`
+        # @return [Fixnum]
+        attr_accessor :train_cost_milli_node_hours
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @model_structure = args[:model_structure] if args.key?(:model_structure)
+          @prediction_sample_rows = args[:prediction_sample_rows] if args.key?(:prediction_sample_rows)
+          @tables_model_column_info = args[:tables_model_column_info] if args.key?(:tables_model_column_info)
+          @train_cost_milli_node_hours = args[:train_cost_milli_node_hours] if args.key?(:train_cost_milli_node_hours)
+        end
+      end
+      
+      # 
+      class XpsTablesTrainingOperationMetadata
+        include Google::Apis::Core::Hashable
+      
+        # Current stage of creating model.
+        # Corresponds to the JSON property `createModelStage`
+        # @return [String]
+        attr_accessor :create_model_stage
+      
+        # The optimization objective for model.
+        # Corresponds to the JSON property `optimizationObjective`
+        # @return [String]
+        attr_accessor :optimization_objective
+      
+        # This field is for training. When the operation is terminated successfully,
+        # AutoML Backend post this field to operation metadata in spanner. If the
+        # metadata has no trials returned, the training operation is supposed to be a
+        # failure.
+        # Corresponds to the JSON property `topTrials`
+        # @return [Array<Google::Apis::LanguageV2::XpsTuningTrial>]
+        attr_accessor :top_trials
+      
+        # Creating model budget.
+        # Corresponds to the JSON property `trainBudgetMilliNodeHours`
+        # @return [Fixnum]
+        attr_accessor :train_budget_milli_node_hours
+      
+        # This field records the training objective value with respect to time, giving
+        # insight into how the model architecture search is performing as training time
+        # elapses.
+        # Corresponds to the JSON property `trainingObjectivePoints`
+        # @return [Array<Google::Apis::LanguageV2::XpsTrainingObjectivePoint>]
+        attr_accessor :training_objective_points
+      
+        # Timestamp when training process starts.
+        # Corresponds to the JSON property `trainingStartTime`
+        # @return [String]
+        attr_accessor :training_start_time
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @create_model_stage = args[:create_model_stage] if args.key?(:create_model_stage)
+          @optimization_objective = args[:optimization_objective] if args.key?(:optimization_objective)
+          @top_trials = args[:top_trials] if args.key?(:top_trials)
+          @train_budget_milli_node_hours = args[:train_budget_milli_node_hours] if args.key?(:train_budget_milli_node_hours)
+          @training_objective_points = args[:training_objective_points] if args.key?(:training_objective_points)
+          @training_start_time = args[:training_start_time] if args.key?(:training_start_time)
+        end
+      end
+      
+      # Component model. Next ID: 10
+      class XpsTextComponentModel
+        include Google::Apis::Core::Hashable
+      
+        # The Cloud Storage resource path to hold batch prediction model.
+        # Corresponds to the JSON property `batchPredictionModelGcsUri`
+        # @return [String]
+        attr_accessor :batch_prediction_model_gcs_uri
+      
+        # The Cloud Storage resource path to hold online prediction model.
+        # Corresponds to the JSON property `onlinePredictionModelGcsUri`
+        # @return [String]
+        attr_accessor :online_prediction_model_gcs_uri
+      
+        # The partition where the model is deployed. Populated by uCAIP BE as part of
+        # online PredictRequest.
+        # Corresponds to the JSON property `partition`
+        # @return [String]
+        attr_accessor :partition
+      
+        # A single model artifact item.
+        # Corresponds to the JSON property `servingArtifact`
+        # @return [Google::Apis::LanguageV2::XpsModelArtifactItem]
+        attr_accessor :serving_artifact
+      
+        # The name of servo model. Populated by uCAIP BE as part of online
+        # PredictRequest.
+        # Corresponds to the JSON property `servoModelName`
+        # @return [String]
+        attr_accessor :servo_model_name
+      
+        # The name of the trained NL submodel.
+        # Corresponds to the JSON property `submodelName`
+        # @return [String]
+        attr_accessor :submodel_name
+      
+        # The type of trained NL submodel
+        # Corresponds to the JSON property `submodelType`
+        # @return [String]
+        attr_accessor :submodel_type
+      
+        # ## The fields below are only populated under uCAIP request scope. https://
+        # cloud.google.com/ml-engine/docs/runtime-version-list
+        # Corresponds to the JSON property `tfRuntimeVersion`
+        # @return [String]
+        attr_accessor :tf_runtime_version
+      
+        # The servomatic model version number. Populated by uCAIP BE as part of online
+        # PredictRequest.
+        # Corresponds to the JSON property `versionNumber`
+        # @return [Fixnum]
+        attr_accessor :version_number
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @batch_prediction_model_gcs_uri = args[:batch_prediction_model_gcs_uri] if args.key?(:batch_prediction_model_gcs_uri)
+          @online_prediction_model_gcs_uri = args[:online_prediction_model_gcs_uri] if args.key?(:online_prediction_model_gcs_uri)
+          @partition = args[:partition] if args.key?(:partition)
+          @serving_artifact = args[:serving_artifact] if args.key?(:serving_artifact)
+          @servo_model_name = args[:servo_model_name] if args.key?(:servo_model_name)
+          @submodel_name = args[:submodel_name] if args.key?(:submodel_name)
+          @submodel_type = args[:submodel_type] if args.key?(:submodel_type)
+          @tf_runtime_version = args[:tf_runtime_version] if args.key?(:tf_runtime_version)
+          @version_number = args[:version_number] if args.key?(:version_number)
+        end
+      end
+      
+      # 
+      class XpsTextExtractionEvaluationMetrics
+        include Google::Apis::Core::Hashable
+      
+        # ConfidenceMetricsEntry includes generic precision, recall, f1 score etc. Next
+        # tag: 16.
+        # Corresponds to the JSON property `bestF1ConfidenceMetrics`
+        # @return [Google::Apis::LanguageV2::XpsConfidenceMetricsEntry]
+        attr_accessor :best_f1_confidence_metrics
+      
+        # If the enclosing EvaluationMetrics.label is empty, confidence_metrics_entries
+        # is an evaluation of the entire model across all labels. If the enclosing
+        # EvaluationMetrics.label is set, confidence_metrics_entries applies to that
+        # label.
+        # Corresponds to the JSON property `confidenceMetricsEntries`
+        # @return [Array<Google::Apis::LanguageV2::XpsConfidenceMetricsEntry>]
+        attr_accessor :confidence_metrics_entries
+      
+        # Confusion matrix of the model running the classification.
+        # Corresponds to the JSON property `confusionMatrix`
+        # @return [Google::Apis::LanguageV2::XpsConfusionMatrix]
+        attr_accessor :confusion_matrix
+      
+        # Only recall, precision, and f1_score will be set.
+        # Corresponds to the JSON property `perLabelConfidenceMetrics`
+        # @return [Hash<String,Google::Apis::LanguageV2::XpsConfidenceMetricsEntry>]
+        attr_accessor :per_label_confidence_metrics
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @best_f1_confidence_metrics = args[:best_f1_confidence_metrics] if args.key?(:best_f1_confidence_metrics)
+          @confidence_metrics_entries = args[:confidence_metrics_entries] if args.key?(:confidence_metrics_entries)
+          @confusion_matrix = args[:confusion_matrix] if args.key?(:confusion_matrix)
+          @per_label_confidence_metrics = args[:per_label_confidence_metrics] if args.key?(:per_label_confidence_metrics)
+        end
+      end
+      
+      # Model evaluation metrics for text sentiment problems.
+      class XpsTextSentimentEvaluationMetrics
+        include Google::Apis::Core::Hashable
+      
+        # Confusion matrix of the model running the classification.
+        # Corresponds to the JSON property `confusionMatrix`
+        # @return [Google::Apis::LanguageV2::XpsConfusionMatrix]
+        attr_accessor :confusion_matrix
+      
+        # Output only. The harmonic mean of recall and precision.
+        # Corresponds to the JSON property `f1Score`
+        # @return [Float]
+        attr_accessor :f1_score
+      
+        # Output only. Linear weighted kappa. Only set for the overall model evaluation,
+        # not for evaluation of a single annotation spec.
+        # Corresponds to the JSON property `linearKappa`
+        # @return [Float]
+        attr_accessor :linear_kappa
+      
+        # Output only. Mean absolute error. Only set for the overall model evaluation,
+        # not for evaluation of a single annotation spec.
+        # Corresponds to the JSON property `meanAbsoluteError`
+        # @return [Float]
+        attr_accessor :mean_absolute_error
+      
+        # Output only. Mean squared error. Only set for the overall model evaluation,
+        # not for evaluation of a single annotation spec.
+        # Corresponds to the JSON property `meanSquaredError`
+        # @return [Float]
+        attr_accessor :mean_squared_error
+      
+        # Output only. Precision.
+        # Corresponds to the JSON property `precision`
+        # @return [Float]
+        attr_accessor :precision
+      
+        # Output only. Quadratic weighted kappa. Only set for the overall model
+        # evaluation, not for evaluation of a single annotation spec.
+        # Corresponds to the JSON property `quadraticKappa`
+        # @return [Float]
+        attr_accessor :quadratic_kappa
+      
+        # Output only. Recall.
+        # Corresponds to the JSON property `recall`
+        # @return [Float]
+        attr_accessor :recall
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @confusion_matrix = args[:confusion_matrix] if args.key?(:confusion_matrix)
+          @f1_score = args[:f1_score] if args.key?(:f1_score)
+          @linear_kappa = args[:linear_kappa] if args.key?(:linear_kappa)
+          @mean_absolute_error = args[:mean_absolute_error] if args.key?(:mean_absolute_error)
+          @mean_squared_error = args[:mean_squared_error] if args.key?(:mean_squared_error)
+          @precision = args[:precision] if args.key?(:precision)
+          @quadratic_kappa = args[:quadratic_kappa] if args.key?(:quadratic_kappa)
+          @recall = args[:recall] if args.key?(:recall)
+        end
+      end
+      
+      # TextToSpeech train response
+      class XpsTextToSpeechTrainResponse
+        include Google::Apis::Core::Hashable
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+        end
+      end
+      
+      # 
+      class XpsTextTrainResponse
+        include Google::Apis::Core::Hashable
+      
+        # Component submodels.
+        # Corresponds to the JSON property `componentModel`
+        # @return [Array<Google::Apis::LanguageV2::XpsTextComponentModel>]
+        attr_accessor :component_model
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @component_model = args[:component_model] if args.key?(:component_model)
+        end
+      end
+      
+      # A [TensorFlow.js](https://www.tensorflow.org/js) model that can be used in the
+      # browser and in Node.js using JavaScript.
+      class XpsTfJsFormat
+        include Google::Apis::Core::Hashable
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+        end
+      end
+      
+      # LINT.IfChange A model format used for mobile and IoT devices. See https://www.
+      # tensorflow.org/lite.
+      class XpsTfLiteFormat
+        include Google::Apis::Core::Hashable
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+        end
+      end
+      
+      # A tensorflow model format in SavedModel format.
+      class XpsTfSavedModelFormat
+        include Google::Apis::Core::Hashable
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+        end
+      end
+      
+      # The data statistics of a series of TIMESTAMP values.
+      class XpsTimestampStats
+        include Google::Apis::Core::Hashable
+      
+        # Common statistics for a column with a specified data type.
+        # Corresponds to the JSON property `commonStats`
+        # @return [Google::Apis::LanguageV2::XpsCommonStats]
+        attr_accessor :common_stats
+      
+        # The string key is the pre-defined granularity. Currently supported:
+        # hour_of_day, day_of_week, month_of_year. Granularities finer that the
+        # granularity of timestamp data are not populated (e.g. if timestamps are at day
+        # granularity, then hour_of_day is not populated).
+        # Corresponds to the JSON property `granularStats`
+        # @return [Hash<String,Google::Apis::LanguageV2::XpsTimestampStatsGranularStats>]
+        attr_accessor :granular_stats
+      
+        # 
+        # Corresponds to the JSON property `medianTimestampNanos`
+        # @return [Fixnum]
+        attr_accessor :median_timestamp_nanos
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @common_stats = args[:common_stats] if args.key?(:common_stats)
+          @granular_stats = args[:granular_stats] if args.key?(:granular_stats)
+          @median_timestamp_nanos = args[:median_timestamp_nanos] if args.key?(:median_timestamp_nanos)
+        end
+      end
+      
+      # Stats split by a defined in context granularity.
+      class XpsTimestampStatsGranularStats
+        include Google::Apis::Core::Hashable
+      
+        # A map from granularity key to example count for that key. E.g. for hour_of_day
+        # `13` means 1pm, or for month_of_year `5` means May).
+        # Corresponds to the JSON property `buckets`
+        # @return [Hash<String,Fixnum>]
+        attr_accessor :buckets
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @buckets = args[:buckets] if args.key?(:buckets)
+        end
+      end
+      
+      # Track matching model metrics for a single track match threshold and multiple
+      # label match confidence thresholds. Next tag: 6.
+      class XpsTrackMetricsEntry
+        include Google::Apis::Core::Hashable
+      
+        # Output only. Metrics for each label-match confidence_threshold from 0.05,0.10,.
+        # ..,0.95,0.96,0.97,0.98,0.99. Precision-recall curve is derived from them.
+        # Corresponds to the JSON property `confidenceMetricsEntries`
+        # @return [Array<Google::Apis::LanguageV2::XpsTrackMetricsEntryConfidenceMetricsEntry>]
+        attr_accessor :confidence_metrics_entries
+      
+        # Output only. The intersection-over-union threshold value between bounding
+        # boxes across frames used to compute this metric entry.
+        # Corresponds to the JSON property `iouThreshold`
+        # @return [Float]
+        attr_accessor :iou_threshold
+      
+        # Output only. The mean bounding box iou over all confidence thresholds.
+        # Corresponds to the JSON property `meanBoundingBoxIou`
+        # @return [Float]
+        attr_accessor :mean_bounding_box_iou
+      
+        # Output only. The mean mismatch rate over all confidence thresholds.
+        # Corresponds to the JSON property `meanMismatchRate`
+        # @return [Float]
+        attr_accessor :mean_mismatch_rate
+      
+        # Output only. The mean average precision over all confidence thresholds.
+        # Corresponds to the JSON property `meanTrackingAveragePrecision`
+        # @return [Float]
+        attr_accessor :mean_tracking_average_precision
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @confidence_metrics_entries = args[:confidence_metrics_entries] if args.key?(:confidence_metrics_entries)
+          @iou_threshold = args[:iou_threshold] if args.key?(:iou_threshold)
+          @mean_bounding_box_iou = args[:mean_bounding_box_iou] if args.key?(:mean_bounding_box_iou)
+          @mean_mismatch_rate = args[:mean_mismatch_rate] if args.key?(:mean_mismatch_rate)
+          @mean_tracking_average_precision = args[:mean_tracking_average_precision] if args.key?(:mean_tracking_average_precision)
+        end
+      end
+      
+      # Metrics for a single confidence threshold. Next tag: 6.
+      class XpsTrackMetricsEntryConfidenceMetricsEntry
+        include Google::Apis::Core::Hashable
+      
+        # Output only. Bounding box intersection-over-union precision. Measures how well
+        # the bounding boxes overlap between each other (e.g. complete overlap or just
+        # barely above iou_threshold).
+        # Corresponds to the JSON property `boundingBoxIou`
+        # @return [Float]
+        attr_accessor :bounding_box_iou
+      
+        # Output only. The confidence threshold value used to compute the metrics.
+        # Corresponds to the JSON property `confidenceThreshold`
+        # @return [Float]
+        attr_accessor :confidence_threshold
+      
+        # Output only. Mismatch rate, which measures the tracking consistency, i.e.
+        # correctness of instance ID continuity.
+        # Corresponds to the JSON property `mismatchRate`
+        # @return [Float]
+        attr_accessor :mismatch_rate
+      
+        # Output only. Tracking precision.
+        # Corresponds to the JSON property `trackingPrecision`
+        # @return [Float]
+        attr_accessor :tracking_precision
+      
+        # Output only. Tracking recall.
+        # Corresponds to the JSON property `trackingRecall`
+        # @return [Float]
+        attr_accessor :tracking_recall
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @bounding_box_iou = args[:bounding_box_iou] if args.key?(:bounding_box_iou)
+          @confidence_threshold = args[:confidence_threshold] if args.key?(:confidence_threshold)
+          @mismatch_rate = args[:mismatch_rate] if args.key?(:mismatch_rate)
+          @tracking_precision = args[:tracking_precision] if args.key?(:tracking_precision)
+          @tracking_recall = args[:tracking_recall] if args.key?(:tracking_recall)
+        end
+      end
+      
+      # Next ID: 18
+      class XpsTrainResponse
+        include Google::Apis::Core::Hashable
+      
+        # Estimated model size in bytes once deployed.
+        # Corresponds to the JSON property `deployedModelSizeBytes`
+        # @return [Fixnum]
+        attr_accessor :deployed_model_size_bytes
+      
+        # Optional vision model error analysis configuration. The field is set when
+        # model error analysis is enabled in the training request. The results of error
+        # analysis will be binded together with evaluation results (in the format of
+        # AnnotatedExample).
+        # Corresponds to the JSON property `errorAnalysisConfigs`
+        # @return [Array<Google::Apis::LanguageV2::XpsVisionErrorAnalysisConfig>]
+        attr_accessor :error_analysis_configs
+      
+        # Set of examples or input sources.
+        # Corresponds to the JSON property `evaluatedExampleSet`
+        # @return [Google::Apis::LanguageV2::XpsExampleSet]
+        attr_accessor :evaluated_example_set
+      
+        # Specifies location of model evaluation metrics.
+        # Corresponds to the JSON property `evaluationMetricsSet`
+        # @return [Google::Apis::LanguageV2::XpsEvaluationMetricsSet]
+        attr_accessor :evaluation_metrics_set
+      
+        # VisionExplanationConfig for XAI on test set. Optional for when XAI is enable
+        # in training request.
+        # Corresponds to the JSON property `explanationConfigs`
+        # @return [Array<Google::Apis::LanguageV2::XpsResponseExplanationSpec>]
+        attr_accessor :explanation_configs
+      
+        # 
+        # Corresponds to the JSON property `imageClassificationTrainResp`
+        # @return [Google::Apis::LanguageV2::XpsImageClassificationTrainResponse]
+        attr_accessor :image_classification_train_resp
+      
+        # 
+        # Corresponds to the JSON property `imageObjectDetectionTrainResp`
+        # @return [Google::Apis::LanguageV2::XpsImageObjectDetectionModelSpec]
+        attr_accessor :image_object_detection_train_resp
+      
+        # 
+        # Corresponds to the JSON property `imageSegmentationTrainResp`
+        # @return [Google::Apis::LanguageV2::XpsImageSegmentationTrainResponse]
+        attr_accessor :image_segmentation_train_resp
+      
+        # Token that represents the trained model. This is considered immutable and is
+        # persisted in AutoML. xPS can put their own proto in the byte string, to e.g.
+        # point to the model checkpoints. The token is passed to other xPS APIs to refer
+        # to the model.
+        # Corresponds to the JSON property `modelToken`
+        # NOTE: Values are automatically base64 encoded/decoded in the client library.
+        # @return [String]
+        attr_accessor :model_token
+      
+        # 
+        # Corresponds to the JSON property `speechTrainResp`
+        # @return [Google::Apis::LanguageV2::XpsSpeechModelSpec]
+        attr_accessor :speech_train_resp
+      
+        # 
+        # Corresponds to the JSON property `tablesTrainResp`
+        # @return [Google::Apis::LanguageV2::XpsTablesTrainResponse]
+        attr_accessor :tables_train_resp
+      
+        # TextToSpeech train response
+        # Corresponds to the JSON property `textToSpeechTrainResp`
+        # @return [Google::Apis::LanguageV2::XpsTextToSpeechTrainResponse]
+        attr_accessor :text_to_speech_train_resp
+      
+        # Will only be needed for uCAIP from Beta.
+        # Corresponds to the JSON property `textTrainResp`
+        # @return [Google::Apis::LanguageV2::XpsTextTrainResponse]
+        attr_accessor :text_train_resp
+      
+        # Train response for translation.
+        # Corresponds to the JSON property `translationTrainResp`
+        # @return [Google::Apis::LanguageV2::XpsTranslationTrainResponse]
+        attr_accessor :translation_train_resp
+      
+        # 
+        # Corresponds to the JSON property `videoActionRecognitionTrainResp`
+        # @return [Google::Apis::LanguageV2::XpsVideoActionRecognitionTrainResponse]
+        attr_accessor :video_action_recognition_train_resp
+      
+        # 
+        # Corresponds to the JSON property `videoClassificationTrainResp`
+        # @return [Google::Apis::LanguageV2::XpsVideoClassificationTrainResponse]
+        attr_accessor :video_classification_train_resp
+      
+        # 
+        # Corresponds to the JSON property `videoObjectTrackingTrainResp`
+        # @return [Google::Apis::LanguageV2::XpsVideoObjectTrackingTrainResponse]
+        attr_accessor :video_object_tracking_train_resp
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @deployed_model_size_bytes = args[:deployed_model_size_bytes] if args.key?(:deployed_model_size_bytes)
+          @error_analysis_configs = args[:error_analysis_configs] if args.key?(:error_analysis_configs)
+          @evaluated_example_set = args[:evaluated_example_set] if args.key?(:evaluated_example_set)
+          @evaluation_metrics_set = args[:evaluation_metrics_set] if args.key?(:evaluation_metrics_set)
+          @explanation_configs = args[:explanation_configs] if args.key?(:explanation_configs)
+          @image_classification_train_resp = args[:image_classification_train_resp] if args.key?(:image_classification_train_resp)
+          @image_object_detection_train_resp = args[:image_object_detection_train_resp] if args.key?(:image_object_detection_train_resp)
+          @image_segmentation_train_resp = args[:image_segmentation_train_resp] if args.key?(:image_segmentation_train_resp)
+          @model_token = args[:model_token] if args.key?(:model_token)
+          @speech_train_resp = args[:speech_train_resp] if args.key?(:speech_train_resp)
+          @tables_train_resp = args[:tables_train_resp] if args.key?(:tables_train_resp)
+          @text_to_speech_train_resp = args[:text_to_speech_train_resp] if args.key?(:text_to_speech_train_resp)
+          @text_train_resp = args[:text_train_resp] if args.key?(:text_train_resp)
+          @translation_train_resp = args[:translation_train_resp] if args.key?(:translation_train_resp)
+          @video_action_recognition_train_resp = args[:video_action_recognition_train_resp] if args.key?(:video_action_recognition_train_resp)
+          @video_classification_train_resp = args[:video_classification_train_resp] if args.key?(:video_classification_train_resp)
+          @video_object_tracking_train_resp = args[:video_object_tracking_train_resp] if args.key?(:video_object_tracking_train_resp)
+        end
+      end
+      
+      # 
+      class XpsTrainingObjectivePoint
+        include Google::Apis::Core::Hashable
+      
+        # The time at which this point was recorded.
+        # Corresponds to the JSON property `createTime`
+        # @return [String]
+        attr_accessor :create_time
+      
+        # The objective value when this point was recorded.
+        # Corresponds to the JSON property `value`
+        # @return [Float]
+        attr_accessor :value
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @create_time = args[:create_time] if args.key?(:create_time)
+          @value = args[:value] if args.key?(:value)
+        end
+      end
+      
+      # Evaluation metrics for the dataset.
+      class XpsTranslationEvaluationMetrics
+        include Google::Apis::Core::Hashable
+      
+        # BLEU score for base model.
+        # Corresponds to the JSON property `baseBleuScore`
+        # @return [Float]
+        attr_accessor :base_bleu_score
+      
+        # BLEU score.
+        # Corresponds to the JSON property `bleuScore`
+        # @return [Float]
+        attr_accessor :bleu_score
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @base_bleu_score = args[:base_bleu_score] if args.key?(:base_bleu_score)
+          @bleu_score = args[:bleu_score] if args.key?(:bleu_score)
+        end
+      end
+      
+      # Translation preprocess response.
+      class XpsTranslationPreprocessResponse
+        include Google::Apis::Core::Hashable
+      
+        # Total example count parsed.
+        # Corresponds to the JSON property `parsedExampleCount`
+        # @return [Fixnum]
+        attr_accessor :parsed_example_count
+      
+        # Total valid example count.
+        # Corresponds to the JSON property `validExampleCount`
+        # @return [Fixnum]
+        attr_accessor :valid_example_count
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @parsed_example_count = args[:parsed_example_count] if args.key?(:parsed_example_count)
+          @valid_example_count = args[:valid_example_count] if args.key?(:valid_example_count)
+        end
+      end
+      
+      # Train response for translation.
+      class XpsTranslationTrainResponse
+        include Google::Apis::Core::Hashable
+      
+        # Type of the model.
+        # Corresponds to the JSON property `modelType`
+        # @return [String]
+        attr_accessor :model_type
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @model_type = args[:model_type] if args.key?(:model_type)
+        end
+      end
+      
+      # Metrics for a tuning job generated, will get forwarded to Stackdriver as model
+      # tuning logs. Setting this as a standalone message out of CreateModelMetadata
+      # to avoid confusion as we expose this message only to users.
+      class XpsTuningTrial
+        include Google::Apis::Core::Hashable
+      
+        # A description of Tables model structure.
+        # Corresponds to the JSON property `modelStructure`
+        # @return [Google::Apis::LanguageV2::XpsTablesModelStructure]
+        attr_accessor :model_structure
+      
+        # The optimization objective evaluation of the eval split data.
+        # Corresponds to the JSON property `trainingObjectivePoint`
+        # @return [Google::Apis::LanguageV2::XpsTrainingObjectivePoint]
+        attr_accessor :training_objective_point
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @model_structure = args[:model_structure] if args.key?(:model_structure)
+          @training_objective_point = args[:training_objective_point] if args.key?(:training_objective_point)
+        end
+      end
+      
+      # The Evaluation metrics entry given a specific precision_window_length.
+      class XpsVideoActionMetricsEntry
+        include Google::Apis::Core::Hashable
+      
+        # Metrics for each label-match confidence_threshold from 0.05,0.10,...,0.95,0.96,
+        # 0.97,0.98,0.99.
+        # Corresponds to the JSON property `confidenceMetricsEntries`
+        # @return [Array<Google::Apis::LanguageV2::XpsVideoActionMetricsEntryConfidenceMetricsEntry>]
+        attr_accessor :confidence_metrics_entries
+      
+        # The mean average precision.
+        # Corresponds to the JSON property `meanAveragePrecision`
+        # @return [Float]
+        attr_accessor :mean_average_precision
+      
+        # This VideoActionMetricsEntry is calculated based on this prediction window
+        # length. If the predicted action's timestamp is inside the time window whose
+        # center is the ground truth action's timestamp with this specific length, the
+        # prediction result is treated as a true positive.
+        # Corresponds to the JSON property `precisionWindowLength`
+        # @return [String]
+        attr_accessor :precision_window_length
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @confidence_metrics_entries = args[:confidence_metrics_entries] if args.key?(:confidence_metrics_entries)
+          @mean_average_precision = args[:mean_average_precision] if args.key?(:mean_average_precision)
+          @precision_window_length = args[:precision_window_length] if args.key?(:precision_window_length)
+        end
+      end
+      
+      # Metrics for a single confidence threshold.
+      class XpsVideoActionMetricsEntryConfidenceMetricsEntry
+        include Google::Apis::Core::Hashable
+      
+        # Output only. The confidence threshold value used to compute the metrics.
+        # Corresponds to the JSON property `confidenceThreshold`
+        # @return [Float]
+        attr_accessor :confidence_threshold
+      
+        # Output only. The harmonic mean of recall and precision.
+        # Corresponds to the JSON property `f1Score`
+        # @return [Float]
+        attr_accessor :f1_score
+      
+        # Output only. Precision for the given confidence threshold.
+        # Corresponds to the JSON property `precision`
+        # @return [Float]
+        attr_accessor :precision
+      
+        # Output only. Recall for the given confidence threshold.
+        # Corresponds to the JSON property `recall`
+        # @return [Float]
+        attr_accessor :recall
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @confidence_threshold = args[:confidence_threshold] if args.key?(:confidence_threshold)
+          @f1_score = args[:f1_score] if args.key?(:f1_score)
+          @precision = args[:precision] if args.key?(:precision)
+          @recall = args[:recall] if args.key?(:recall)
+        end
+      end
+      
+      # Model evaluation metrics for video action recognition.
+      class XpsVideoActionRecognitionEvaluationMetrics
+        include Google::Apis::Core::Hashable
+      
+        # Output only. The number of ground truth actions used to create this evaluation.
+        # Corresponds to the JSON property `evaluatedActionCount`
+        # @return [Fixnum]
+        attr_accessor :evaluated_action_count
+      
+        # Output only. The metric entries for precision window lengths: 1s,2s,3s,4s, 5s.
+        # Corresponds to the JSON property `videoActionMetricsEntries`
+        # @return [Array<Google::Apis::LanguageV2::XpsVideoActionMetricsEntry>]
+        attr_accessor :video_action_metrics_entries
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @evaluated_action_count = args[:evaluated_action_count] if args.key?(:evaluated_action_count)
+          @video_action_metrics_entries = args[:video_action_metrics_entries] if args.key?(:video_action_metrics_entries)
+        end
+      end
+      
+      # 
+      class XpsVideoActionRecognitionTrainResponse
+        include Google::Apis::Core::Hashable
+      
+        # ## The fields below are only populated under uCAIP request scope.
+        # Corresponds to the JSON property `modelArtifactSpec`
+        # @return [Google::Apis::LanguageV2::XpsVideoModelArtifactSpec]
+        attr_accessor :model_artifact_spec
+      
+        # The actual train cost of creating this model, expressed in node seconds, i.e.
+        # 3,600 value in this field means 1 node hour.
+        # Corresponds to the JSON property `trainCostNodeSeconds`
+        # @return [Fixnum]
+        attr_accessor :train_cost_node_seconds
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @model_artifact_spec = args[:model_artifact_spec] if args.key?(:model_artifact_spec)
+          @train_cost_node_seconds = args[:train_cost_node_seconds] if args.key?(:train_cost_node_seconds)
+        end
+      end
+      
+      # 
+      class XpsVideoBatchPredictOperationMetadata
+        include Google::Apis::Core::Hashable
+      
+        # All the partial batch prediction results that are completed at the moment.
+        # Output examples are sorted by completion time. The order will not be changed.
+        # Each output example should be the path of a single RecordIO file of
+        # AnnotatedExamples.
+        # Corresponds to the JSON property `outputExamples`
+        # @return [Array<String>]
+        attr_accessor :output_examples
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @output_examples = args[:output_examples] if args.key?(:output_examples)
+        end
+      end
+      
+      # 
+      class XpsVideoClassificationTrainResponse
+        include Google::Apis::Core::Hashable
+      
+        # ## The fields below are only populated under uCAIP request scope.
+        # Corresponds to the JSON property `modelArtifactSpec`
+        # @return [Google::Apis::LanguageV2::XpsVideoModelArtifactSpec]
+        attr_accessor :model_artifact_spec
+      
+        # The actual train cost of creating this model, expressed in node seconds, i.e.
+        # 3,600 value in this field means 1 node hour.
+        # Corresponds to the JSON property `trainCostNodeSeconds`
+        # @return [Fixnum]
+        attr_accessor :train_cost_node_seconds
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @model_artifact_spec = args[:model_artifact_spec] if args.key?(:model_artifact_spec)
+          @train_cost_node_seconds = args[:train_cost_node_seconds] if args.key?(:train_cost_node_seconds)
+        end
+      end
+      
+      # Information of downloadable models that are pre-generated as part of training
+      # flow and will be persisted in AutoMl backend. Upon receiving ExportModel
+      # request from user, AutoMl backend can serve the pre-generated models to user
+      # if exists (by copying the files from internal path to user provided location),
+      # otherwise, AutoMl backend will call xPS ExportModel API to generate the model
+      # on the fly with the requesting format.
+      class XpsVideoExportModelSpec
+        include Google::Apis::Core::Hashable
+      
+        # Contains the model format and internal location of the model files to be
+        # exported/downloaded. Use the GCS bucket name which is provided via
+        # TrainRequest.gcs_bucket_name to store the model files.
+        # Corresponds to the JSON property `exportModelOutputConfig`
+        # @return [Array<Google::Apis::LanguageV2::XpsExportModelOutputConfig>]
+        attr_accessor :export_model_output_config
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @export_model_output_config = args[:export_model_output_config] if args.key?(:export_model_output_config)
+        end
+      end
+      
+      # 
+      class XpsVideoModelArtifactSpec
+        include Google::Apis::Core::Hashable
+      
+        # The model binary files in different formats for model export.
+        # Corresponds to the JSON property `exportArtifact`
+        # @return [Array<Google::Apis::LanguageV2::XpsModelArtifactItem>]
+        attr_accessor :export_artifact
+      
+        # A single model artifact item.
+        # Corresponds to the JSON property `servingArtifact`
+        # @return [Google::Apis::LanguageV2::XpsModelArtifactItem]
+        attr_accessor :serving_artifact
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @export_artifact = args[:export_artifact] if args.key?(:export_artifact)
+          @serving_artifact = args[:serving_artifact] if args.key?(:serving_artifact)
+        end
+      end
+      
+      # Model evaluation metrics for ObjectTracking problems. Next tag: 10.
+      class XpsVideoObjectTrackingEvaluationMetrics
+        include Google::Apis::Core::Hashable
+      
+        # Output only. The single metric for bounding boxes evaluation: the
+        # mean_average_precision averaged over all bounding_box_metrics_entries.
+        # Corresponds to the JSON property `boundingBoxMeanAveragePrecision`
+        # @return [Float]
+        attr_accessor :bounding_box_mean_average_precision
+      
+        # Output only. The bounding boxes match metrics for each Intersection-over-union
+        # threshold 0.05,0.10,...,0.95,0.96,0.97,0.98,0.99.
+        # Corresponds to the JSON property `boundingBoxMetricsEntries`
+        # @return [Array<Google::Apis::LanguageV2::XpsBoundingBoxMetricsEntry>]
+        attr_accessor :bounding_box_metrics_entries
+      
+        # The number of bounding boxes used for model evaluation.
+        # Corresponds to the JSON property `evaluatedBoundingboxCount`
+        # @return [Fixnum]
+        attr_accessor :evaluated_boundingbox_count
+      
+        # The number of video frames used for model evaluation.
+        # Corresponds to the JSON property `evaluatedFrameCount`
+        # @return [Fixnum]
+        attr_accessor :evaluated_frame_count
+      
+        # The number of tracks used for model evaluation.
+        # Corresponds to the JSON property `evaluatedTrackCount`
+        # @return [Fixnum]
+        attr_accessor :evaluated_track_count
+      
+        # Output only. The single metric for tracks accuracy evaluation: the
+        # mean_average_precision averaged over all track_metrics_entries.
+        # Corresponds to the JSON property `trackMeanAveragePrecision`
+        # @return [Float]
+        attr_accessor :track_mean_average_precision
+      
+        # Output only. The single metric for tracks bounding box iou evaluation: the
+        # mean_bounding_box_iou averaged over all track_metrics_entries.
+        # Corresponds to the JSON property `trackMeanBoundingBoxIou`
+        # @return [Float]
+        attr_accessor :track_mean_bounding_box_iou
+      
+        # Output only. The single metric for tracking consistency evaluation: the
+        # mean_mismatch_rate averaged over all track_metrics_entries.
+        # Corresponds to the JSON property `trackMeanMismatchRate`
+        # @return [Float]
+        attr_accessor :track_mean_mismatch_rate
+      
+        # Output only. The tracks match metrics for each Intersection-over-union
+        # threshold 0.05,0.10,...,0.95,0.96,0.97,0.98,0.99.
+        # Corresponds to the JSON property `trackMetricsEntries`
+        # @return [Array<Google::Apis::LanguageV2::XpsTrackMetricsEntry>]
+        attr_accessor :track_metrics_entries
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @bounding_box_mean_average_precision = args[:bounding_box_mean_average_precision] if args.key?(:bounding_box_mean_average_precision)
+          @bounding_box_metrics_entries = args[:bounding_box_metrics_entries] if args.key?(:bounding_box_metrics_entries)
+          @evaluated_boundingbox_count = args[:evaluated_boundingbox_count] if args.key?(:evaluated_boundingbox_count)
+          @evaluated_frame_count = args[:evaluated_frame_count] if args.key?(:evaluated_frame_count)
+          @evaluated_track_count = args[:evaluated_track_count] if args.key?(:evaluated_track_count)
+          @track_mean_average_precision = args[:track_mean_average_precision] if args.key?(:track_mean_average_precision)
+          @track_mean_bounding_box_iou = args[:track_mean_bounding_box_iou] if args.key?(:track_mean_bounding_box_iou)
+          @track_mean_mismatch_rate = args[:track_mean_mismatch_rate] if args.key?(:track_mean_mismatch_rate)
+          @track_metrics_entries = args[:track_metrics_entries] if args.key?(:track_metrics_entries)
+        end
+      end
+      
+      # 
+      class XpsVideoObjectTrackingTrainResponse
+        include Google::Apis::Core::Hashable
+      
+        # Information of downloadable models that are pre-generated as part of training
+        # flow and will be persisted in AutoMl backend. Upon receiving ExportModel
+        # request from user, AutoMl backend can serve the pre-generated models to user
+        # if exists (by copying the files from internal path to user provided location),
+        # otherwise, AutoMl backend will call xPS ExportModel API to generate the model
+        # on the fly with the requesting format.
+        # Corresponds to the JSON property `exportModelSpec`
+        # @return [Google::Apis::LanguageV2::XpsVideoExportModelSpec]
+        attr_accessor :export_model_spec
+      
+        # ## The fields below are only populated under uCAIP request scope.
+        # Corresponds to the JSON property `modelArtifactSpec`
+        # @return [Google::Apis::LanguageV2::XpsVideoModelArtifactSpec]
+        attr_accessor :model_artifact_spec
+      
+        # The actual train cost of creating this model, expressed in node seconds, i.e.
+        # 3,600 value in this field means 1 node hour.
+        # Corresponds to the JSON property `trainCostNodeSeconds`
+        # @return [Fixnum]
+        attr_accessor :train_cost_node_seconds
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @export_model_spec = args[:export_model_spec] if args.key?(:export_model_spec)
+          @model_artifact_spec = args[:model_artifact_spec] if args.key?(:model_artifact_spec)
+          @train_cost_node_seconds = args[:train_cost_node_seconds] if args.key?(:train_cost_node_seconds)
+        end
+      end
+      
+      # 
+      class XpsVideoTrainingOperationMetadata
+        include Google::Apis::Core::Hashable
+      
+        # This is an estimation of the node hours necessary for training a model,
+        # expressed in milli node hours (i.e. 1,000 value in this field means 1 node
+        # hour). A node hour represents the time a virtual machine spends running your
+        # training job. The cost of one node running for one hour is a node hour.
+        # Corresponds to the JSON property `trainCostMilliNodeHour`
+        # @return [Fixnum]
+        attr_accessor :train_cost_milli_node_hour
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @train_cost_milli_node_hour = args[:train_cost_milli_node_hour] if args.key?(:train_cost_milli_node_hour)
+        end
+      end
+      
+      # The vision model error analysis configuration. Next tag: 3
+      class XpsVisionErrorAnalysisConfig
+        include Google::Apis::Core::Hashable
+      
+        # The number of query examples in error analysis.
+        # Corresponds to the JSON property `exampleCount`
+        # @return [Fixnum]
+        attr_accessor :example_count
+      
+        # The query type used in retrieval. The enum values are frozen in the
+        # foreseeable future.
+        # Corresponds to the JSON property `queryType`
+        # @return [String]
+        attr_accessor :query_type
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @example_count = args[:example_count] if args.key?(:example_count)
+          @query_type = args[:query_type] if args.key?(:query_type)
+        end
+      end
+      
+      # 
+      class XpsVisionTrainingOperationMetadata
+        include Google::Apis::Core::Hashable
+      
+        # Infra Usage of billing metrics. Next ID: 6
+        # Corresponds to the JSON property `explanationUsage`
+        # @return [Google::Apis::LanguageV2::InfraUsage]
+        attr_accessor :explanation_usage
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @explanation_usage = args[:explanation_usage] if args.key?(:explanation_usage)
+        end
+      end
+      
+      # Visualization configurations for image explanation.
+      class XpsVisualization
+        include Google::Apis::Core::Hashable
+      
+        # Excludes attributions below the specified percentile, from the highlighted
+        # areas. Defaults to 62.
+        # Corresponds to the JSON property `clipPercentLowerbound`
+        # @return [Float]
+        attr_accessor :clip_percent_lowerbound
+      
+        # Excludes attributions above the specified percentile from the highlighted
+        # areas. Using the clip_percent_upperbound and clip_percent_lowerbound together
+        # can be useful for filtering out noise and making it easier to see areas of
+        # strong attribution. Defaults to 99.9.
+        # Corresponds to the JSON property `clipPercentUpperbound`
+        # @return [Float]
+        attr_accessor :clip_percent_upperbound
+      
+        # The color scheme used for the highlighted areas. Defaults to PINK_GREEN for
+        # Integrated Gradients attribution, which shows positive attributions in green
+        # and negative in pink. Defaults to VIRIDIS for XRAI attribution, which
+        # highlights the most influential regions in yellow and the least influential in
+        # blue.
+        # Corresponds to the JSON property `colorMap`
+        # @return [String]
+        attr_accessor :color_map
+      
+        # How the original image is displayed in the visualization. Adjusting the
+        # overlay can help increase visual clarity if the original image makes it
+        # difficult to view the visualization. Defaults to NONE.
+        # Corresponds to the JSON property `overlayType`
+        # @return [String]
+        attr_accessor :overlay_type
+      
+        # Whether to only highlight pixels with positive contributions, negative or both.
+        # Defaults to POSITIVE.
+        # Corresponds to the JSON property `polarity`
+        # @return [String]
+        attr_accessor :polarity
+      
+        # Type of the image visualization. Only applicable to Integrated Gradients
+        # attribution. OUTLINES shows regions of attribution, while PIXELS shows per-
+        # pixel attribution. Defaults to OUTLINES.
+        # Corresponds to the JSON property `type`
+        # @return [String]
+        attr_accessor :type
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @clip_percent_lowerbound = args[:clip_percent_lowerbound] if args.key?(:clip_percent_lowerbound)
+          @clip_percent_upperbound = args[:clip_percent_upperbound] if args.key?(:clip_percent_upperbound)
+          @color_map = args[:color_map] if args.key?(:color_map)
+          @overlay_type = args[:overlay_type] if args.key?(:overlay_type)
+          @polarity = args[:polarity] if args.key?(:polarity)
+          @type = args[:type] if args.key?(:type)
+        end
+      end
+      
+      # 
+      class XpsXpsOperationMetadata
+        include Google::Apis::Core::Hashable
+      
+        # Optional. XPS server can opt to provide example count of the long running
+        # operation (e.g. training, data importing, batch prediction).
+        # Corresponds to the JSON property `exampleCount`
+        # @return [Fixnum]
+        attr_accessor :example_count
+      
+        # Metrics for the operation. By the time the operation is terminated (whether
+        # succeeded or failed) as returned from XPS, AutoML BE assumes the metrics are
+        # finalized. AutoML BE transparently posts the metrics to Chemist if it's not
+        # empty, regardless of the response content or error type. If user is supposed
+        # to be charged in case of cancellation/error, this field should be set. In the
+        # case where the type of LRO doesn't require any billing, this field should be
+        # left unset.
+        # Corresponds to the JSON property `reportingMetrics`
+        # @return [Google::Apis::LanguageV2::XpsReportingMetrics]
+        attr_accessor :reporting_metrics
+      
+        # 
+        # Corresponds to the JSON property `tablesTrainingOperationMetadata`
+        # @return [Google::Apis::LanguageV2::XpsTablesTrainingOperationMetadata]
+        attr_accessor :tables_training_operation_metadata
+      
+        # 
+        # Corresponds to the JSON property `videoBatchPredictOperationMetadata`
+        # @return [Google::Apis::LanguageV2::XpsVideoBatchPredictOperationMetadata]
+        attr_accessor :video_batch_predict_operation_metadata
+      
+        # 
+        # Corresponds to the JSON property `videoTrainingOperationMetadata`
+        # @return [Google::Apis::LanguageV2::XpsVideoTrainingOperationMetadata]
+        attr_accessor :video_training_operation_metadata
+      
+        # 
+        # Corresponds to the JSON property `visionTrainingOperationMetadata`
+        # @return [Google::Apis::LanguageV2::XpsVisionTrainingOperationMetadata]
+        attr_accessor :vision_training_operation_metadata
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @example_count = args[:example_count] if args.key?(:example_count)
+          @reporting_metrics = args[:reporting_metrics] if args.key?(:reporting_metrics)
+          @tables_training_operation_metadata = args[:tables_training_operation_metadata] if args.key?(:tables_training_operation_metadata)
+          @video_batch_predict_operation_metadata = args[:video_batch_predict_operation_metadata] if args.key?(:video_batch_predict_operation_metadata)
+          @video_training_operation_metadata = args[:video_training_operation_metadata] if args.key?(:video_training_operation_metadata)
+          @vision_training_operation_metadata = args[:vision_training_operation_metadata] if args.key?(:vision_training_operation_metadata)
+        end
+      end
+      
+      # An explanation method that redistributes Integrated Gradients attributions to
+      # segmented regions, taking advantage of the model's fully differentiable
+      # structure. Refer to this paper for more details: https://arxiv.org/abs/1906.
+      # 02825 Only supports image Models (modality is IMAGE).
+      class XpsXraiAttribution
+        include Google::Apis::Core::Hashable
+      
+        # 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 met within the
+        # desired error range. Valid range of its value is [1, 100], inclusively.
+        # Corresponds to the JSON property `stepCount`
+        # @return [Fixnum]
+        attr_accessor :step_count
+      
+        def initialize(**args)
+           update!(**args)
+        end
+      
+        # Update properties of this object
+        def update!(**args)
+          @step_count = args[:step_count] if args.key?(:step_count)
+        end
+      end
     end
   end
 end