Types for Google Cloud Aiplatform V1 Schema Trainingjob Definition v1 API

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageClassification(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

A TrainingJob that trains and uploads an AutoML ImageClassification Model.

inputs()

The input parameters of this TrainingJob.

metadata()

The metadata information.

inputs(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_image_classification.AutoMlImageClassificationInputs](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageClassificationInputs_ )

metadata(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_image_classification.AutoMlImageClassificationMetadata](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageClassificationMetadata_ )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageClassificationInputs(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

model_type()

base_model_id()

The ID of thebase model. If it is specified, the newmodel will be trained based on thebase model.Otherwise, the new model will be trained from scratch. Thebase model must be in the same Project and Location asthe new Model to train, and have the same modelType.

budget_milli_node_hours()

The training budget of creating this model, expressed inmilli node hours i.e. 1,000 value in this field means 1 nodehour. The actual metadata.costMilliNodeHours will be equalor less than this value. If further model training ceases toprovide any improvements, it will stop without using thefull budget and the metadata.successfulStopReason will bemodel-converged. Note, node_hour = actual_hour *number_of_nodes_involved. For modelTypecloud(default), the budget must be between 8,000 and800,000 milli node hours, inclusive. The default value is192,000 which represents one day in wall time, considering 8nodes are used. For model typesmobile-tf-low-latency-1,mobile-tf-versatile-1,mobile-tf-high-accuracy-1,the training budget must be between 1,000 and 100,000 millinode hours, inclusive. The default value is 24,000 whichrepresents one day in wall time on a single node that isused.

disable_early_stopping()

Use the entire training budget. This disablesthe early stopping feature. When false the earlystopping feature is enabled, which means thatAutoML Image Classification might stop trainingbefore the entire training budget has been used.

multi_label()

If false, a single-label (multi-class) Modelwill be trained (i.e. assuming that for eachimage just up to one annotation may beapplicable). If true, a multi-label Model willbe trained (i.e. assuming that for each imagemultiple annotations may be applicable).

class ModelType(value)

Bases:proto.enums.Enum

Values:

MODEL_TYPE_UNSPECIFIED (0):    Should not be set.CLOUD (1):    A Model best tailored to be used within    Google Cloud, and which cannot be exported.    Default.MOBILE_TF_LOW_LATENCY_1 (2):    A model that, in addition to being available    within Google Cloud, can also be exported (see    ModelService.ExportModel) as TensorFlow or Core    ML model and used on a mobile or edge device    afterwards. Expected to have low latency, but    may have lower prediction quality than other    mobile models.MOBILE_TF_VERSATILE_1 (3):    A model that, in addition to being available    within Google Cloud, can also be exported (see    ModelService.ExportModel) as TensorFlow or Core    ML model and used on a mobile or edge device    with afterwards.MOBILE_TF_HIGH_ACCURACY_1 (4):    A model that, in addition to being available    within Google Cloud, can also be exported (see    ModelService.ExportModel) as TensorFlow or Core    ML model and used on a mobile or edge device    afterwards. Expected to have a higher latency,    but should also have a higher prediction quality    than other mobile models.

CLOUD( = )

MOBILE_TF_HIGH_ACCURACY_1( = )

MOBILE_TF_LOW_LATENCY_1( = )

MOBILE_TF_VERSATILE_1( = )

MODEL_TYPE_UNSPECIFIED( = )

base_model_id(: [str](https://docs.python.org/3/library/stdtypes.html#str )

budget_milli_node_hours(: [int](https://docs.python.org/3/library/functions.html#int )

disable_early_stopping(: [bool](https://docs.python.org/3/library/functions.html#bool )

model_type(: [ModelType](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageClassificationInputs.ModelType_ )

multi_label(: [bool](https://docs.python.org/3/library/functions.html#bool )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageClassificationMetadata(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

cost_milli_node_hours()

The actual training cost of creating thismodel, expressed in milli node hours, i.e. 1,000value in this field means 1 node hour.Guaranteed to not exceedinputs.budgetMilliNodeHours.

successful_stop_reason()

For successful job completions, this is thereason why the job has finished.

class SuccessfulStopReason(value)

Bases:proto.enums.Enum

Values:

SUCCESSFUL_STOP_REASON_UNSPECIFIED (0):    Should not be set.BUDGET_REACHED (1):    The inputs.budgetMilliNodeHours had been    reached.MODEL_CONVERGED (2):    Further training of the Model ceased to    increase its quality, since it already has    converged.

BUDGET_REACHED( = )

MODEL_CONVERGED( = )

SUCCESSFUL_STOP_REASON_UNSPECIFIED( = )

cost_milli_node_hours(: [int](https://docs.python.org/3/library/functions.html#int )

successful_stop_reason(: [SuccessfulStopReason](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageClassificationMetadata.SuccessfulStopReason_ )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageObjectDetection(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

A TrainingJob that trains and uploads an AutoML Image ObjectDetection Model.

inputs()

The input parameters of this TrainingJob.

metadata()

The metadata information

inputs(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_image_object_detection.AutoMlImageObjectDetectionInputs](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageObjectDetectionInputs_ )

metadata(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_image_object_detection.AutoMlImageObjectDetectionMetadata](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageObjectDetectionMetadata_ )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageObjectDetectionInputs(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

model_type()

budget_milli_node_hours()

The training budget of creating this model, expressed inmilli node hours i.e. 1,000 value in this field means 1 nodehour. The actual metadata.costMilliNodeHours will be equalor less than this value. If further model training ceases toprovide any improvements, it will stop without using thefull budget and the metadata.successfulStopReason will bemodel-converged. Note, node_hour = actual_hour *number_of_nodes_involved. For modelTypecloud(default), the budget must be between 20,000 and900,000 milli node hours, inclusive. The default value is216,000 which represents one day in wall time, considering 9nodes are used. For model typesmobile-tf-low-latency-1,mobile-tf-versatile-1,mobile-tf-high-accuracy-1 thetraining budget must be between 1,000 and 100,000 milli nodehours, inclusive. The default value is 24,000 whichrepresents one day in wall time on a single node that isused.

disable_early_stopping()

Use the entire training budget. This disablesthe early stopping feature. When false the earlystopping feature is enabled, which means thatAutoML Image Object Detection might stoptraining before the entire training budget hasbeen used.

class ModelType(value)

Bases:proto.enums.Enum

Values:

MODEL_TYPE_UNSPECIFIED (0):    Should not be set.CLOUD_HIGH_ACCURACY_1 (1):    A model best tailored to be used within    Google Cloud, and which cannot be exported.    Expected to have a higher latency, but should    also have a higher prediction quality than other    cloud models.CLOUD_LOW_LATENCY_1 (2):    A model best tailored to be used within    Google Cloud, and which cannot be exported.    Expected to have a low latency, but may have    lower prediction quality than other cloud    models.MOBILE_TF_LOW_LATENCY_1 (3):    A model that, in addition to being available    within Google Cloud can also be exported (see    ModelService.ExportModel) and used on a mobile    or edge device with TensorFlow afterwards.    Expected to have low latency, but may have lower    prediction quality than other mobile models.MOBILE_TF_VERSATILE_1 (4):    A model that, in addition to being available    within Google Cloud can also be exported (see    ModelService.ExportModel) and used on a mobile    or edge device with TensorFlow afterwards.MOBILE_TF_HIGH_ACCURACY_1 (5):    A model that, in addition to being available    within Google Cloud, can also be exported (see    ModelService.ExportModel) and used on a mobile    or edge device with TensorFlow afterwards.    Expected to have a higher latency, but should    also have a higher prediction quality than other    mobile models.

CLOUD_HIGH_ACCURACY_1( = )

CLOUD_LOW_LATENCY_1( = )

MOBILE_TF_HIGH_ACCURACY_1( = )

MOBILE_TF_LOW_LATENCY_1( = )

MOBILE_TF_VERSATILE_1( = )

MODEL_TYPE_UNSPECIFIED( = )

budget_milli_node_hours(: [int](https://docs.python.org/3/library/functions.html#int )

disable_early_stopping(: [bool](https://docs.python.org/3/library/functions.html#bool )

model_type(: [ModelType](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageObjectDetectionInputs.ModelType_ )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageObjectDetectionMetadata(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

cost_milli_node_hours()

The actual training cost of creating thismodel, expressed in milli node hours, i.e. 1,000value in this field means 1 node hour.Guaranteed to not exceedinputs.budgetMilliNodeHours.

successful_stop_reason()

For successful job completions, this is thereason why the job has finished.

class SuccessfulStopReason(value)

Bases:proto.enums.Enum

Values:

SUCCESSFUL_STOP_REASON_UNSPECIFIED (0):    Should not be set.BUDGET_REACHED (1):    The inputs.budgetMilliNodeHours had been    reached.MODEL_CONVERGED (2):    Further training of the Model ceased to    increase its quality, since it already has    converged.

BUDGET_REACHED( = )

MODEL_CONVERGED( = )

SUCCESSFUL_STOP_REASON_UNSPECIFIED( = )

cost_milli_node_hours(: [int](https://docs.python.org/3/library/functions.html#int )

successful_stop_reason(: [SuccessfulStopReason](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageObjectDetectionMetadata.SuccessfulStopReason_ )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageSegmentation(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

A TrainingJob that trains and uploads an AutoML ImageSegmentation Model.

inputs()

The input parameters of this TrainingJob.

metadata()

The metadata information.

inputs(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_image_segmentation.AutoMlImageSegmentationInputs](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageSegmentationInputs_ )

metadata(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_image_segmentation.AutoMlImageSegmentationMetadata](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageSegmentationMetadata_ )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageSegmentationInputs(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

model_type()

budget_milli_node_hours()

The training budget of creating this model, expressed inmilli node hours i.e. 1,000 value in this field means 1 nodehour. The actual metadata.costMilliNodeHours will be equalor less than this value. If further model training ceases toprovide any improvements, it will stop without using thefull budget and the metadata.successfulStopReason will bemodel-converged. Note, node_hour = actual_hour *number_of_nodes_involved. Or actaul_wall_clock_hours =train_budget_milli_node_hours / (number_of_nodes_involved *1000) For modelTypecloud-high-accuracy-1(default),the budget must be between 20,000 and 2,000,000 milli nodehours, inclusive. The default value is 192,000 whichrepresents one day in wall time (1000 milli * 24 hours * 8nodes).

base_model_id()

The ID of thebase model. If it is specified, the newmodel will be trained based on thebase model.Otherwise, the new model will be trained from scratch. Thebase model must be in the same Project and Location asthe new Model to train, and have the same modelType.

class ModelType(value)

Bases:proto.enums.Enum

Values:

MODEL_TYPE_UNSPECIFIED (0):    Should not be set.CLOUD_HIGH_ACCURACY_1 (1):    A model to be used via prediction calls to    uCAIP API. Expected to have a higher latency,    but should also have a higher prediction quality    than other models.CLOUD_LOW_ACCURACY_1 (2):    A model to be used via prediction calls to    uCAIP API. Expected to have a lower latency but    relatively lower prediction quality.MOBILE_TF_LOW_LATENCY_1 (3):    A model that, in addition to being available    within Google Cloud, can also be exported (see    ModelService.ExportModel) as TensorFlow model    and used on a mobile or edge device afterwards.    Expected to have low latency, but may have lower    prediction quality than other mobile models.

CLOUD_HIGH_ACCURACY_1( = )

CLOUD_LOW_ACCURACY_1( = )

MOBILE_TF_LOW_LATENCY_1( = )

MODEL_TYPE_UNSPECIFIED( = )

base_model_id(: [str](https://docs.python.org/3/library/stdtypes.html#str )

budget_milli_node_hours(: [int](https://docs.python.org/3/library/functions.html#int )

model_type(: [ModelType](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageSegmentationInputs.ModelType_ )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageSegmentationMetadata(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

cost_milli_node_hours()

The actual training cost of creating thismodel, expressed in milli node hours, i.e. 1,000value in this field means 1 node hour.Guaranteed to not exceedinputs.budgetMilliNodeHours.

successful_stop_reason()

For successful job completions, this is thereason why the job has finished.

class SuccessfulStopReason(value)

Bases:proto.enums.Enum

Values:

SUCCESSFUL_STOP_REASON_UNSPECIFIED (0):    Should not be set.BUDGET_REACHED (1):    The inputs.budgetMilliNodeHours had been    reached.MODEL_CONVERGED (2):    Further training of the Model ceased to    increase its quality, since it already has    converged.

BUDGET_REACHED( = )

MODEL_CONVERGED( = )

SUCCESSFUL_STOP_REASON_UNSPECIFIED( = )

cost_milli_node_hours(: [int](https://docs.python.org/3/library/functions.html#int )

successful_stop_reason(: [SuccessfulStopReason](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlImageSegmentationMetadata.SuccessfulStopReason_ )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTables(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

A TrainingJob that trains and uploads an AutoML Tables Model.

inputs()

The input parameters of this TrainingJob.

metadata()

The metadata information.

inputs(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_tables.AutoMlTablesInputs](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTablesInputs_ )

metadata(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_tables.AutoMlTablesMetadata](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTablesMetadata_ )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTablesInputs(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

This message hasoneof fields (mutually exclusive fields).For each oneof, at most one member field can be set at the same time.Setting any member of the oneof automatically clears all othermembers.

optimization_objective_recall_value()

Required when optimization_objective is“maximize-precision-at-recall”. Must be between 0 and 1,inclusive.

This field is a member ofoneofadditional_optimization_objective_config.

optimization_objective_precision_value()

Required when optimization_objective is“maximize-recall-at-precision”. Must be between 0 and 1,inclusive.

This field is a member ofoneofadditional_optimization_objective_config.

prediction_type()

The type of prediction the Model is toproduce. “classification” - Predict one out ofmultiple target values ispicked for each row.

“regression” - Predict a value based on its

relation to other values. Thistype is available only to columns that containsemantically numeric values, i.e. integers orfloating point number, even ifstored as e.g. strings.

target_column()

The column name of the target column that themodel is to predict.

transformations()

Each transformation will apply transformfunction to given input column. And the resultwill be used for training. When creatingtransformation for BigQuery Struct column, thecolumn should be flattened using “.” as thedelimiter.

optimization_objective()

Objective function the model is optimizingtowards. The training process creates a modelthat maximizes/minimizes the value of theobjective function over the validation set.

The supported optimization objectives depend onthe prediction type. If the field is not set, adefault objective function is used.

classification (binary):

“maximize-au-roc” (default) - Maximize the

area under the receiveroperating characteristic (ROC) curve.“minimize-log-loss” - Minimize log loss.

“maximize-au-prc” - Maximize the area under

the precision-recall curve.“maximize-precision-at-recall” - Maximizeprecision for a specifiedrecall value. “maximize-recall-at-precision” -Maximize recall for a specifiedprecision value.

classification (multi-class):

“minimize-log-loss” (default) - Minimize log

loss.

regression:

“minimize-rmse” (default) - Minimize

root-mean-squared error (RMSE). “minimize-mae”

  • Minimize mean-absolute error (MAE).“minimize-rmsle” - Minimize root-mean-squaredlog error (RMSLE).

  • Type

    str

train_budget_milli_node_hours()

Required. The train budget of creating thismodel, expressed in milli node hours i.e. 1,000value in this field means 1 node hour.

The training cost of the model will not exceedthis budget. The final cost will be attempted tobe close to the budget, though may end up being(even) noticeably smaller - at the backend’sdiscretion. This especially may happen whenfurther model training ceases to provide anyimprovements.

If the budget is set to a value known to beinsufficient to train a model for the givendataset, the training won’t be attempted andwill error.

The train budget must be between 1,000 and72,000 milli node hours, inclusive.

disable_early_stopping()

Use the entire training budget. This disablesthe early stopping feature. By default, theearly stopping feature is enabled, which meansthat AutoML Tables might stop training beforethe entire training budget has been used.

weight_column_name()

Column name that should be used as the weightcolumn. Higher values in this column give moreimportance to the row during model training. Thecolumn must have numeric values between 0 and10000 inclusively; 0 means the row is ignoredfor training. If weight column field is not set,then all rows are assumed to have equal weightof 1.

export_evaluated_data_items_config()

Configuration for exporting test setpredictions to a BigQuery table. If thisconfiguration is absent, then the export is notperformed.

additional_experiments()

Additional experiment flags for the Tablestraining pipeline.

  • Type

    MutableSequence[str]

class Transformation(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

This message hasoneof fields (mutually exclusive fields).For each oneof, at most one member field can be set at the same time.Setting any member of the oneof automatically clears all othermembers.

auto()

This field is a member ofoneoftransformation_detail.

numeric()

This field is a member ofoneoftransformation_detail.

categorical()

This field is a member ofoneoftransformation_detail.

timestamp()

This field is a member ofoneoftransformation_detail.

text()

This field is a member ofoneoftransformation_detail.

repeated_numeric()

This field is a member ofoneoftransformation_detail.

repeated_categorical()

This field is a member ofoneoftransformation_detail.

repeated_text()

This field is a member ofoneoftransformation_detail.

class AutoTransformation(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Training pipeline will infer the proper transformation basedon the statistic of dataset.

column_name()

column_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

class CategoricalArrayTransformation(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Treats the column as categorical array and performs followingtransformation functions.

  • For each element in the array, convert the category name to adictionary lookup index and generate an embedding for each index.Combine the embedding of all elements into a single embeddingusing the mean.

  • Empty arrays treated as an embedding of zeroes.

column_name()

column_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

class CategoricalTransformation(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Training pipeline will perform following transformation functions.

  • The categorical string as is–no change to case, punctuation,spelling, tense, and so on.

  • Convert the category name to a dictionary lookup index andgenerate an embedding for each index.

  • Categories that appear less than 5 times in the training datasetare treated as the “unknown” category. The “unknown” categorygets its own special lookup index and resulting embedding.

column_name()

column_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

class NumericArrayTransformation(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Treats the column as numerical array and performs followingtransformation functions.

  • All transformations for Numerical types applied to the average ofthe all elements.

  • The average of empty arrays is treated as zero.

column_name()

invalid_values_allowed()

If invalid values is allowed, the trainingpipeline will create a boolean feature thatindicated whether the value is valid. Otherwise,the training pipeline will discard the input rowfrom trainining data.

column_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

invalid_values_allowed(: [bool](https://docs.python.org/3/library/functions.html#bool )

class NumericTransformation(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Training pipeline will perform following transformation functions.

  • The value converted to float32.

  • The z_score of the value.

  • log(value+1) when the value is greater than or equal to 0.Otherwise, this transformation is not applied and the value isconsidered a missing value.

  • z_score of log(value+1) when the value is greater than or equalto 0. Otherwise, this transformation is not applied and the valueis considered a missing value.

  • A boolean value that indicates whether the value is valid.

column_name()

invalid_values_allowed()

If invalid values is allowed, the trainingpipeline will create a boolean feature thatindicated whether the value is valid. Otherwise,the training pipeline will discard the input rowfrom trainining data.

column_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

invalid_values_allowed(: [bool](https://docs.python.org/3/library/functions.html#bool )

class TextArrayTransformation(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Treats the column as text array and performs followingtransformation functions.

  • Concatenate all text values in the array into a single text valueusing a space (” “) as a delimiter, and then treat the result asa single text value. Apply the transformations for Text columns.

  • Empty arrays treated as an empty text.

column_name()

column_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

class TextTransformation(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Training pipeline will perform following transformation functions.

  • The text as is–no change to case, punctuation, spelling, tense,and so on.

  • Tokenize text to words. Convert each words to a dictionary lookupindex and generate an embedding for each index. Combine theembedding of all elements into a single embedding using the mean.

  • Tokenization is based on unicode script boundaries.

  • Missing values get their own lookup index and resultingembedding.

  • Stop-words receive no special treatment and are not removed.

column_name()

column_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

class TimestampTransformation(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Training pipeline will perform following transformation functions.

  • Apply the transformation functions for Numerical columns.

  • Determine the year, month, day,and weekday. Treat each value fromthe

  • timestamp as a Categorical column.

  • Invalid numerical values (for example, values that fall outsideof a typical timestamp range, or are extreme values) receive nospecial treatment and are not removed.

column_name()

time_format()

The format in which that time field is expressed. Thetime_format must either be one of:

  • unix-seconds

  • unix-milliseconds

  • unix-microseconds

  • unix-nanoseconds (for respectively number of seconds,milliseconds, microseconds and nanoseconds since start ofthe Unix epoch); or be written instrftime syntax. Iftime_format is not set, then the default format is RFC3339date-time format, wheretime-offset ="Z" (e.g. 1985-04-12T23:20:50.52Z)

  • Type

    str

invalid_values_allowed()

If invalid values is allowed, the trainingpipeline will create a boolean feature thatindicated whether the value is valid. Otherwise,the training pipeline will discard the input rowfrom trainining data.

column_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

invalid_values_allowed(: [bool](https://docs.python.org/3/library/functions.html#bool )

time_format(: [str](https://docs.python.org/3/library/stdtypes.html#str )

auto(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_tables.AutoMlTablesInputs.Transformation.AutoTransformation](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTablesInputs.Transformation.AutoTransformation_ )

categorical(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_tables.AutoMlTablesInputs.Transformation.CategoricalTransformation](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTablesInputs.Transformation.CategoricalTransformation_ )

numeric(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_tables.AutoMlTablesInputs.Transformation.NumericTransformation](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTablesInputs.Transformation.NumericTransformation_ )

repeated_categorical(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_tables.AutoMlTablesInputs.Transformation.CategoricalArrayTransformation](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTablesInputs.Transformation.CategoricalArrayTransformation_ )

repeated_numeric(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_tables.AutoMlTablesInputs.Transformation.NumericArrayTransformation](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTablesInputs.Transformation.NumericArrayTransformation_ )

repeated_text(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_tables.AutoMlTablesInputs.Transformation.TextArrayTransformation](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTablesInputs.Transformation.TextArrayTransformation_ )

text(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_tables.AutoMlTablesInputs.Transformation.TextTransformation](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTablesInputs.Transformation.TextTransformation_ )

timestamp(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_tables.AutoMlTablesInputs.Transformation.TimestampTransformation](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTablesInputs.Transformation.TimestampTransformation_ )

additional_experiments(: MutableSequence[str )

disable_early_stopping(: [bool](https://docs.python.org/3/library/functions.html#bool )

export_evaluated_data_items_config(: gcastd_export_evaluated_data_items_config.ExportEvaluatedDataItemsConfi )

optimization_objective(: [str](https://docs.python.org/3/library/stdtypes.html#str )

optimization_objective_precision_value(: [float](https://docs.python.org/3/library/functions.html#float )

optimization_objective_recall_value(: [float](https://docs.python.org/3/library/functions.html#float )

prediction_type(: [str](https://docs.python.org/3/library/stdtypes.html#str )

target_column(: [str](https://docs.python.org/3/library/stdtypes.html#str )

train_budget_milli_node_hours(: [int](https://docs.python.org/3/library/functions.html#int )

transformations(: MutableSequence[[Transformation](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTablesInputs.Transformation)_ )

weight_column_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTablesMetadata(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Model metadata specific to AutoML Tables.

train_cost_milli_node_hours()

Output only. The actual training cost of themodel, expressed in milli node hours, i.e. 1,000value in this field means 1 node hour.Guaranteed to not exceed the train budget.

train_cost_milli_node_hours(: [int](https://docs.python.org/3/library/functions.html#int )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTextClassification(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

A TrainingJob that trains and uploads an AutoML TextClassification Model.

inputs()

The input parameters of this TrainingJob.

inputs(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_text_classification.AutoMlTextClassificationInputs](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTextClassificationInputs_ )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTextClassificationInputs(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

multi_label()

multi_label(: [bool](https://docs.python.org/3/library/functions.html#bool )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTextExtraction(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

A TrainingJob that trains and uploads an AutoML TextExtraction Model.

inputs()

The input parameters of this TrainingJob.

  • Type

    google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTextExtractionInputs

inputs(: google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_text_extraction.AutoMlTextExtractionInput )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTextExtractionInputs(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTextSentiment(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

A TrainingJob that trains and uploads an AutoML TextSentiment Model.

inputs()

The input parameters of this TrainingJob.

inputs(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_text_sentiment.AutoMlTextSentimentInputs](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTextSentimentInputs_ )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlTextSentimentInputs(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

sentiment_max()

A sentiment is expressed as an integerordinal, where higher value means a morepositive sentiment. The range of sentiments thatwill be used is between 0 and sentimentMax(inclusive on both ends), and all the values inthe range must be represented in the datasetbefore a model can be created.Only the Annotations with this sentimentMax willbe used for training. sentimentMax value must bebetween 1 and 10 (inclusive).

sentiment_max(: [int](https://docs.python.org/3/library/functions.html#int )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlVideoActionRecognition(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

A TrainingJob that trains and uploads an AutoML Video ActionRecognition Model.

inputs()

The input parameters of this TrainingJob.

inputs(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_video_action_recognition.AutoMlVideoActionRecognitionInputs](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlVideoActionRecognitionInputs_ )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlVideoActionRecognitionInputs(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

model_type()

class ModelType(value)

Bases:proto.enums.Enum

Values:

MODEL_TYPE_UNSPECIFIED (0):    Should not be set.CLOUD (1):    A model best tailored to be used within    Google Cloud, and which c annot be exported.    Default.MOBILE_VERSATILE_1 (2):    A model that, in addition to being available    within Google Cloud, can also be exported (see    ModelService.ExportModel) as a TensorFlow or    TensorFlow Lite model and used on a mobile or    edge device afterwards.MOBILE_JETSON_VERSATILE_1 (3):    A model that, in addition to being available    within Google Cloud, can also be exported (see    ModelService.ExportModel) to a Jetson device    afterwards.MOBILE_CORAL_VERSATILE_1 (4):    A model that, in addition to being available    within Google Cloud, can also be exported (see    ModelService.ExportModel) as a TensorFlow or    TensorFlow Lite model and used on a Coral device    afterwards.

CLOUD( = )

MOBILE_CORAL_VERSATILE_1( = )

MOBILE_JETSON_VERSATILE_1( = )

MOBILE_VERSATILE_1( = )

MODEL_TYPE_UNSPECIFIED( = )

model_type(: [ModelType](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlVideoActionRecognitionInputs.ModelType_ )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlVideoClassification(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

A TrainingJob that trains and uploads an AutoML VideoClassification Model.

inputs()

The input parameters of this TrainingJob.

inputs(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_video_classification.AutoMlVideoClassificationInputs](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlVideoClassificationInputs_ )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlVideoClassificationInputs(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

model_type()

class ModelType(value)

Bases:proto.enums.Enum

Values:

MODEL_TYPE_UNSPECIFIED (0):    Should not be set.CLOUD (1):    A model best tailored to be used within    Google Cloud, and which cannot be exported.    Default.MOBILE_VERSATILE_1 (2):    A model that, in addition to being available    within Google Cloud, can also be exported (see    ModelService.ExportModel) as a TensorFlow or    TensorFlow Lite model and used on a mobile or    edge device afterwards.MOBILE_JETSON_VERSATILE_1 (3):    A model that, in addition to being available    within Google Cloud, can also be exported (see    ModelService.ExportModel) to a Jetson device    afterwards.

CLOUD( = )

MOBILE_JETSON_VERSATILE_1( = )

MOBILE_VERSATILE_1( = )

MODEL_TYPE_UNSPECIFIED( = )

model_type(: [ModelType](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlVideoClassificationInputs.ModelType_ )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlVideoObjectTracking(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

A TrainingJob that trains and uploads an AutoML VideoObjectTracking Model.

inputs()

The input parameters of this TrainingJob.

inputs(: [google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.automl_video_object_tracking.AutoMlVideoObjectTrackingInputs](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlVideoObjectTrackingInputs_ )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlVideoObjectTrackingInputs(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

model_type()

class ModelType(value)

Bases:proto.enums.Enum

Values:

MODEL_TYPE_UNSPECIFIED (0):    Should not be set.CLOUD (1):    A model best tailored to be used within    Google Cloud, and which c annot be exported.    Default.MOBILE_VERSATILE_1 (2):    A model that, in addition to being available    within Google Cloud, can also be exported (see    ModelService.ExportModel) as a TensorFlow or    TensorFlow Lite model and used on a mobile or    edge device afterwards.MOBILE_CORAL_VERSATILE_1 (3):    A versatile model that is meant to be    exported (see ModelService.ExportModel) and used    on a Google Coral device.MOBILE_CORAL_LOW_LATENCY_1 (4):    A model that trades off quality for low    latency, to be exported (see    ModelService.ExportModel) and used on a Google    Coral device.MOBILE_JETSON_VERSATILE_1 (5):    A versatile model that is meant to be    exported (see ModelService.ExportModel) and used    on an NVIDIA Jetson device.MOBILE_JETSON_LOW_LATENCY_1 (6):    A model that trades off quality for low    latency, to be exported (see    ModelService.ExportModel) and used on an NVIDIA    Jetson device.

CLOUD( = )

MOBILE_CORAL_LOW_LATENCY_1( = )

MOBILE_CORAL_VERSATILE_1( = )

MOBILE_JETSON_LOW_LATENCY_1( = )

MOBILE_JETSON_VERSATILE_1( = )

MOBILE_VERSATILE_1( = )

MODEL_TYPE_UNSPECIFIED( = )

model_type(: [ModelType](../definition_v1/types.md#google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.AutoMlVideoObjectTrackingInputs.ModelType_ )

class google.cloud.aiplatform.v1.schema.trainingjob.definition_v1.types.ExportEvaluatedDataItemsConfig(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Configuration for exporting test set predictions to aBigQuery table.

destination_bigquery_uri()

URI of desired destination BigQuery table. Expected format:bq://<project_id>:<dataset_id>:

If not specified, then results are exported to the followingauto-created BigQuery table:<project_id>:export_evaluated_examples_<model_name>_<yyyy_MM_dd’T’HH_mm_ss_SSS’Z’>.evaluated_examples

override_existing_table()

If true and an export destination isspecified, then the contents of the destinationare overwritten. Otherwise, if the exportdestination already exists, then the exportoperation fails.

destination_bigquery_uri(: [str](https://docs.python.org/3/library/stdtypes.html#str )

override_existing_table(: [bool](https://docs.python.org/3/library/functions.html#bool )

Except as otherwise noted, the content of this page is licensed under theCreative Commons Attribution 4.0 License, and code samples are licensed under theApache 2.0 License. For details, see theGoogle Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.

Last updated 2025-10-30 UTC.