v1beta1

An un-set value of this enum.

At most one label is allowed per example.

Multiple labels are allowed for one example.

Should not be used, an un-set enum has this value by default.

Model is deployed.

Model is not deployed.

Should not be used.

Document dimension is measured in inches.

Document dimension is measured in centimeters.

Document dimension is measured in points. 72 points = 1 inch.

Should not be used.

The text segment is a token. e.g. word.

The text segment is a paragraph.

The text segment is a form field.

The text segment is the name part of a form field. It will be treated as child of another FORM_FIELD TextSegment if its span is subspan of another TextSegment with type FORM_FIELD.

The text segment is the text content part of a form field. It will be treated as child of another FORM_FIELD TextSegment if its span is subspan of another TextSegment with type FORM_FIELD.

The text segment is a whole table, including headers, and all rows.

The text segment is a table's headers. It will be treated as child of another TABLE TextSegment if its span is subspan of another TextSegment with type TABLE.

The text segment is a row in table. It will be treated as child of another TABLE TextSegment if its span is subspan of another TextSegment with type TABLE.

The text segment is a cell in table. It will be treated as child of another TABLE_ROW TextSegment if its span is subspan of another TextSegment with type TABLE_ROW.

Not specified. Should not be used.

Encoded as number, or the strings "NaN", "Infinity", or "-Infinity".

Must be between 0AD and 9999AD. Encoded as string according to time_format, or, if that format is not set, then in RFC 3339 date-time format, where time-offset = "Z" (e.g. 1985-04-12T23:20:50.52Z).

Encoded as string.

Encoded as list, where the list elements are represented according to

list_element_type.

Encoded as struct, where field values are represented according to struct_type.

Values of this type are not further understood by AutoML, e.g. AutoML is unable to tell the order of values (as it could with FLOAT64), or is unable to say if one value contains another (as it could with STRING). Encoded as string (bytes should be base64-encoded, as described in RFC 4648, section 4).

Annotation details for translation.

This object should have the same structure as TranslationAnnotation

Annotation details for content or image classification.

This object should have the same structure as ClassificationAnnotation

Annotation details for image object detection.

This object should have the same structure as ImageObjectDetectionAnnotation

Annotation details for video classification. Returned for Video Classification predictions.

This object should have the same structure as VideoClassificationAnnotation

Annotation details for video object tracking.

This object should have the same structure as VideoObjectTrackingAnnotation

Annotation details for text extraction.

This object should have the same structure as TextExtractionAnnotation

Annotation details for text sentiment.

This object should have the same structure as TextSentimentAnnotation

Annotation details for Tables.

This object should have the same structure as TablesAnnotation

Output only . The resource ID of the annotation spec that this annotation pertains to. The annotation spec comes from either an ancestor dataset, or the dataset that was used to train the model in use.

Output only. The value of display_name when the model was trained. Because this field returns a value at model training time, for different models trained using the same dataset, the returned value could be different as model owner could update the display_name between any two model training.

Output only. Resource name of the annotation spec. Form:

'projects/{project_id}/locations/{location_id}/datasets/{dataset_id}/annotationSpecs/{annotation_spec_id}'

Required. The name of the annotation spec to show in the interface. The name can be up to 32 characters long and must match the regexp [a-zA-Z0-9_]+. (_), and ASCII digits 0-9.

Output only. The number of examples in the parent dataset labeled by the annotation spec.

Stats of all the values of all arrays, as if they were a single long series of data. The type depends on the element type of the array.

This object should have the same structure as DataStats

The Google Cloud Storage location for the input content.

This object should have the same structure as GcsSource

The BigQuery location for the input content.

This object should have the same structure as BigQuerySource

The Google Cloud Storage location of the directory where the output is to be written to.

This object should have the same structure as GcsDestination

The BigQuery location where the output is to be written to.

This object should have the same structure as BigQueryDestination

Name of the model requested to serve the batch prediction.

Required. The input configuration for batch prediction.

This object should have the same structure as BatchPredictInputConfig

Required. The Configuration specifying where output predictions should be written.

This object should have the same structure as BatchPredictOutputConfig

Additional domain-specific parameters for the predictions, any string must be up to 25000 characters long.

• For Text Classification:

  score_threshold - (float) A value from 0.0 to 1.0. When the model makes predictions for a text snippet, it will only produce results that have at least this confidence score. The default is 0.5.

• For Image Classification:

  score_threshold - (float) A value from 0.0 to 1.0. When the model makes predictions for an image, it will only produce results that have at least this confidence score. The default is 0.5.

• For Image Object Detection:

  score_threshold - (float) When Model detects objects on the image, it will only produce bounding boxes which have at least this confidence score. Value in 0 to 1 range, default is 0.5. max_bounding_box_count - (int64) No more than this number of bounding boxes will be produced per image. Default is 100, the requested value may be limited by server.

• For Video Classification : score_threshold - (float) A value from 0.0 to 1.0. When the model makes predictions for a video, it will only produce results that have at least this confidence score. The default is 0.5. segment_classification - (boolean) Set to true to request segment-level classification. AutoML Video Intelligence returns labels and their confidence scores for the entire segment of the video that user specified in the request configuration. The default is "true". shot_classification - (boolean) Set to true to request shot-level classification. AutoML Video Intelligence determines the boundaries for each camera shot in the entire segment of the video that user specified in the request configuration. AutoML Video Intelligence then returns labels and their confidence scores for each detected shot, along with the start and end time of the shot. WARNING: Model evaluation is not done for this classification type, the quality of it depends on training data, but there are no metrics provided to describe that quality. The default is "false". 1s_interval_classification - (boolean) Set to true to request classification for a video at one-second intervals. AutoML Video Intelligence returns labels and their confidence scores for each second of the entire segment of the video that user specified in the request configuration. WARNING: Model evaluation is not done for this classification type, the quality of it depends on training data, but there are no metrics provided to describe that quality. The default is "false".

• For Video Object Tracking: score_threshold - (float) When Model detects objects on video frames, it will only produce bounding boxes which have at least this confidence score. Value in 0 to 1 range, default is 0.5. max_bounding_box_count - (int64) No more than this number of bounding boxes will be returned per frame. Default is 100, the requested value may be limited by server. min_bounding_box_size - (float) Only bounding boxes with shortest edge at least that long as a relative value of video frame size will be returned. Value in 0 to 1 range. Default is 0.

Additional domain-specific prediction response metadata.

• For Image Object Detection: max_bounding_box_count - (int64) At most that many bounding boxes per image could have been returned.

• For Video Object Tracking: max_bounding_box_count - (int64) At most that many bounding boxes per frame could have been returned.

Required. BigQuery URI to a project, up to 2000 characters long. Accepted forms:

• BigQuery path e.g. bq://projectId

Required. BigQuery URI to a table, up to 2000 characters long. Accepted forms:

• BigQuery path e.g. bq://projectId.bqDatasetId.bqTableId

Output only. The intersection-over-union threshold value used to compute this metrics entry.

Output only. The mean average precision, most often close to au_prc.

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.

This object should have the same structure as ConfidenceMetricsEntry

Output only . The bounding polygon normalized vertices.

This object should have the same structure as NormalizedVertex

The statistics of the top 20 CATEGORY values, ordered by

count.

This object should have the same structure as SingleCategoryStats

Output only. A confidence estimate between 0.0 and 1.0. A higher value means greater confidence that the annotation is positive. If a user approves an annotation as negative or positive, the score value remains unchanged. If a user creates an annotation, the score is 0 for negative or 1 for positive.

Output only. The Area Under Precision-Recall Curve metric. Micro-averaged for the overall evaluation.

Output only. The Area Under Precision-Recall Curve metric based on priors. Micro-averaged for the overall evaluation. Deprecated.

Output only. The Area Under Receiver Operating Characteristic curve metric. Micro-averaged for the overall evaluation.

Output only. The Log Loss metric.

Output only. Metrics for each confidence_threshold in 0.00,0.05,0.10,...,0.95,0.96,0.97,0.98,0.99 and position_threshold = INT32_MAX_VALUE. ROC and precision-recall curves, and other aggregated metrics are derived from them. The confidence metrics entries may also be supplied for additional values of position_threshold, but from these no aggregated metrics are computed.

This object should have the same structure as ConfidenceMetricsEntry

Output only. Confusion matrix of the evaluation. Only set for MULTICLASS classification problems where number of labels is no more than 10. Only set for model level evaluation, not for evaluation per label.

This object should have the same structure as ConfusionMatrix

Output only. The annotation spec ids used for this evaluation.

Output only. The resource name of the column specs. Form:

projects/{project_id}/locations/{location_id}/datasets/{dataset_id}/tableSpecs/{table_spec_id}/columnSpecs/{column_spec_id}

The data type of elements stored in the column.

This object should have the same structure as DataType

Output only. The name of the column to show in the interface. The name can be up to 100 characters long and can consist only of ASCII Latin letters A-Z and a-z, ASCII digits 0-9, underscores(_), and forward slashes(/), and must start with a letter or a digit.

Output only. Stats of the series of values in the column. This field may be stale, see the ancestor's Dataset.tables_dataset_metadata.stats_update_time field for the timestamp at which these stats were last updated.

This object should have the same structure as DataStats

Deprecated.

This object should have the same structure as CorrelatedColumn

Used to perform consistent read-modify-write updates. If not set, a blind "overwrite" update happens.

Output only. Metrics are computed with an assumption that the model never returns predictions with score lower than this value.

Output only. 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.

Output only. Recall (True Positive Rate) for the given confidence threshold.

Output only. Precision for the given confidence threshold.

Output only. False Positive Rate for the given confidence threshold.

Output only. The harmonic mean of recall and precision.

Output only. 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.

Output only. The precision when only considering the label that has the highest prediction score and not below the confidence threshold for each example.

Output only. The False Positive Rate when only considering the label that has the highest prediction score and not below the confidence threshold for each example.

Output only. The harmonic mean of recall_at1 and precision_at1.

Output only. The number of model created labels that match a ground truth label.

Output only. The number of model created labels that do not match a ground truth label.

Output only. The number of ground truth labels that are not matched by a model created label.

Output only. The number of labels that were not created by the model, but if they would, they would not match a ground truth label.

Output only. The confidence threshold value used to compute the metrics. Only annotations with score of at least this threshold are considered to be ones the model would return.

Output only. Recall under the given confidence threshold.

Output only. Precision under the given confidence threshold.

Output only. The harmonic mean of recall and precision.

Output only. The confidence threshold value used to compute the metrics.

Output only. Recall under the given confidence threshold.

Output only. Precision under the given confidence threshold.

Output only. The harmonic mean of recall and precision.

Output only. IDs of the annotation specs used in the confusion matrix. For Tables CLASSIFICATION

prediction_type only list of annotation_spec_display_name-s is populated.

Output only. Display name of the annotation specs used in the confusion matrix, as they were at the moment of the evaluation. For Tables CLASSIFICATION

prediction_type-s, distinct values of the target column at the moment of the model evaluation are populated here.

Output only. Rows in the confusion matrix. The number of rows is equal to the size of annotation_spec_id. row[i].value[j] is the number of examples that have ground truth of the annotation_spec_id[i] and are predicted as annotation_spec_id[j] by the model being evaluated.

This object should have the same structure as Row

The column_spec_id of the correlated column, which belongs to the same table as the in-context column.

Correlation between this and the in-context column.

This object should have the same structure as CorrelationStats

The correlation value using the Cramer's V measure.

The resource name of the project to create the dataset for.

The dataset to create.

This object should have the same structure as Dataset

Resource name of the parent project where the model is being created.

The model to create.

This object should have the same structure as Model

Metadata for a dataset used for translation.

This object should have the same structure as TranslationDatasetMetadata

Metadata for a dataset used for image classification.

This object should have the same structure as ImageClassificationDatasetMetadata

Metadata for a dataset used for text classification.

This object should have the same structure as TextClassificationDatasetMetadata

Metadata for a dataset used for image object detection.

This object should have the same structure as ImageObjectDetectionDatasetMetadata

Metadata for a dataset used for video classification.

This object should have the same structure as VideoClassificationDatasetMetadata

Metadata for a dataset used for video object tracking.

This object should have the same structure as VideoObjectTrackingDatasetMetadata

Metadata for a dataset used for text extraction.

This object should have the same structure as TextExtractionDatasetMetadata

Metadata for a dataset used for text sentiment.

This object should have the same structure as TextSentimentDatasetMetadata

Metadata for a dataset used for Tables.

This object should have the same structure as TablesDatasetMetadata

Output only. The resource name of the dataset. Form: projects/{project_id}/locations/{location_id}/datasets/{dataset_id}

Required. The name of the dataset to show in the interface. The name can be up to 32 characters long and can consist only of ASCII Latin letters A-Z and a-z, underscores (_), and ASCII digits 0-9.

User-provided description of the dataset. The description can be up to 25000 characters long.

Output only. The number of examples in the dataset.

Output only. Timestamp when this dataset was created.

This object should have the same structure as Timestamp

Used to perform consistent read-modify-write updates. If not set, a blind "overwrite" update happens.

The statistics for FLOAT64 DataType.

This object should have the same structure as Float64Stats

The statistics for STRING DataType.

This object should have the same structure as StringStats

The statistics for TIMESTAMP DataType.

This object should have the same structure as TimestampStats

The statistics for ARRAY DataType.

This object should have the same structure as ArrayStats

The statistics for STRUCT DataType.

This object should have the same structure as StructStats

The statistics for CATEGORY DataType.

This object should have the same structure as CategoryStats

The number of distinct values.

The number of values that are null.

The number of values that are valid.

If type_code == ARRAY, then list_element_type is the type of the elements.

This object should have the same structure as DataType

If type_code == STRUCT, then struct_type provides type information for the struct's fields.

This object should have the same structure as StructType

If type_code == TIMESTAMP then time_format provides the format in which that time field is expressed. The time_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 of the Unix epoch); or be written in strftime syntax. If time_format is not set, then the default format as described on the type_code is used.

Required. The TypeCode for this type.

The number should be among the values of TypeCode

If true, this DataType can also be NULL. In .CSV files NULL value is expressed as an empty string.

The resource name of the dataset to delete.

Resource name of the model being deleted.

Model deployment metadata specific to Image Object Detection.

This object should have the same structure as ImageObjectDetectionModelDeploymentMetadata

Model deployment metadata specific to Image Classification.

This object should have the same structure as ImageClassificationModelDeploymentMetadata

Resource name of the model to deploy.

An input config specifying the content of the document.

This object should have the same structure as DocumentInputConfig

The plain text version of this document.

This object should have the same structure as TextSnippet

Describes the layout of the document. Sorted by page_number.

This object should have the same structure as Layout

The dimensions of the page in the document.

This object should have the same structure as DocumentDimensions

Number of pages in the document.

Unit of the dimension.

The number should be among the values of DocumentDimensionUnit

Width value of the document, works together with the unit.

Height value of the document, works together with the unit.

The Google Cloud Storage location of the document file. Only a single path should be given. Max supported size: 512MB. Supported extensions: .PDF.

This object should have the same structure as GcsSource

Start of the range, inclusive.

End of the range, exclusive.

Example image.

This object should have the same structure as Image

Example text.

This object should have the same structure as TextSnippet

Example document.

This object should have the same structure as Document

Example relational table row.

This object should have the same structure as Row

Required. The resource name of the dataset.

Required. The desired output location.

This object should have the same structure as OutputConfig

The BigQuery location where the output is to be written to.

This object should have the same structure as BigQueryDestination

Required. The resource name of the model whose evaluated examples are to be exported.

Required. The desired output location and configuration.

This object should have the same structure as ExportEvaluatedExamplesOutputConfig

Required. The resource name of the model to export.

Required. The desired output location and configuration.

This object should have the same structure as ModelExportOutputConfig

The mean of the series.

The standard deviation of the series.

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.

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".

This object should have the same structure as HistogramBucket

Required. Google Contained Registry URI of the new image, up to 2000 characters long. See

https: //cloud.google.com/container-registry/do // cs/pushing-and-pulling#pushing_an_image_to_a_registry Accepted forms:

• [HOSTNAME]/[PROJECT-ID]/[IMAGE]
• [HOSTNAME]/[PROJECT-ID]/[IMAGE]:[TAG]

The requesting user must have permission to push images the project.

Required. Google Cloud Storage URI to output directory, up to 2000 characters long. Accepted forms:

• Prefix path: gs://bucket/directory The requesting user must have write permission to the bucket. The directory is created if it doesn't exist.

Required. Google Cloud Storage URIs to input files, up to 2000 characters long. Accepted forms:

• Full object path, e.g. gs://bucket/directory/object.csv

The resource name of the annotation spec to retrieve.

The resource name of the column spec to retrieve.

Mask specifying which fields to read.

This object should have the same structure as FieldMask

The resource name of the dataset to retrieve.

Resource name for the model evaluation.

Resource name of the model.

The resource name of the table spec to retrieve.

Mask specifying which fields to read.

This object should have the same structure as FieldMask

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).

The minimum value of the bucket, inclusive.

The maximum value of the bucket, exclusive unless max = "Infinity", in which case it's inclusive.

The number of data values that are in the bucket, i.e. are between min and max values.

Image content represented as a stream of bytes. Note: As with all bytes fields, protobuffers use a pure binary representation, whereas JSON representations use base64.

An input config specifying the content of the image.

This object should have the same structure as InputConfig

Output only. HTTP URI to the thumbnail image.

Required. Type of the classification problem.

The number should be among the values of ClassificationType

Input only. The number of nodes to deploy the model on. A node is an abstraction of a machine resource, which can handle online prediction QPS as given in the model's node_qps. Must be between 1 and 100, inclusive on both ends.

Optional. The ID of the base model. If it is specified, the new model will be created based on the base model. Otherwise, the new model will be created from scratch. The base model must be in the same project and location as the new model to create, and have the same model_type.

Required. The train budget of creating this model, expressed in hours. The actual train_cost will be equal or less than this value.

Output only. The actual train cost of creating this model, expressed in hours. If this model is created from a base model, the train cost used to create the base model are not included.

Output only. The reason that this create model operation stopped, e.g. BUDGET_REACHED, MODEL_CONVERGED.

Optional. Type of the model. The available values are:

• cloud - Model to be used via prediction calls to AutoML API. This is the default value.
• mobile-low-latency-1 - A model that, in addition to providing prediction via AutoML API, can also be exported (see AutoMl.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 models.
• mobile-versatile-1 - A model that, in addition to providing prediction via AutoML API, can also be exported (see AutoMl.ExportModel) and used on a mobile or edge device with TensorFlow afterwards.
• mobile-high-accuracy-1 - A model that, in addition to providing prediction via AutoML API, can also be exported (see AutoMl.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 models.
• mobile-core-ml-low-latency-1 - A model that, in addition to providing prediction via AutoML API, can also be exported (see AutoMl.ExportModel) and used on a mobile device with Core ML afterwards. Expected to have low latency, but may have lower prediction quality than other models.
• mobile-core-ml-versatile-1 - A model that, in addition to providing prediction via AutoML API, can also be exported (see AutoMl.ExportModel) and used on a mobile device with Core ML afterwards.
• mobile-core-ml-high-accuracy-1 - A model that, in addition to providing prediction via AutoML API, can also be exported (see AutoMl.ExportModel) and used on a mobile device with Core ML afterwards. Expected to have a higher latency, but should also have a higher prediction quality than other models.

Output only. The rectangle representing the object location.

This object should have the same structure as BoundingPoly

Output only. The confidence that this annotation is positive for the parent example, value in [0, 1], higher means higher positivity confidence.

Output only. The total number of bounding boxes (i.e. summed over all images) the ground truth used to create this evaluation had.

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 and each label confidence threshold 0.05,0.10,...,0.95,0.96,0.97,0.98,0.99 pair.

This object should have the same structure as BoundingBoxMetricsEntry

Output only. The single metric for bounding boxes evaluation: the mean_average_precision averaged over all bounding_box_metrics_entries.

Input only. The number of nodes to deploy the model on. A node is an abstraction of a machine resource, which can handle online prediction QPS as given in the model's

qps_per_node. Must be between 1 and 100, inclusive on both ends.

Optional. Type of the model. The available values are:

• cloud-high-accuracy-1 - (default) A model to be used via prediction calls to AutoML API. Expected to have a higher latency, but should also have a higher prediction quality than other models.
• cloud-low-latency-1 - A model to be used via prediction calls to AutoML API. Expected to have low latency, but may have lower prediction quality than other models.

Output only. The number of nodes this model is deployed on. A node is an abstraction of a machine resource, which can handle online prediction QPS as given in the qps_per_node field.

Output only. An approximate number of online prediction QPS that can be supported by this model per each node on which it is deployed.

Output only. The reason that this create model operation stopped, e.g. BUDGET_REACHED, MODEL_CONVERGED.

The train budget of creating this model, expressed in milli node hours i.e. 1,000 value in this field means 1 node hour. The actual train_cost will be equal or less than this value. If further model training ceases to provide any improvements, it will stop without using full budget and the stop_reason will be MODEL_CONVERGED. Note, node_hour = actual_hour * number_of_nodes_invovled. For model type cloud-high-accuracy-1(default) and cloud-low-latency-1, the train budget must be between 20,000 and 2,000,000 milli node hours, inclusive. The default value is 216, 000 which represents one day in wall time. For model type mobile-low-latency-1, mobile-versatile-1, mobile-high-accuracy-1, mobile-core-ml-low-latency-1, mobile-core-ml-versatile-1, mobile-core-ml-high-accuracy-1, the train budget must be between 1,000 and 100,000 milli node hours, inclusive. The default value is 24, 000 which represents one day in wall time.

Output only. The actual train cost of creating this 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.

Required. Dataset name. Dataset must already exist. All imported annotations and examples will be added.

Required. The desired input location and its domain specific semantics, if any.

This object should have the same structure as InputConfig

The Google Cloud Storage location for the input content. In ImportData, the gcs_source points to a csv with structure described in the comment.

This object should have the same structure as GcsSource

The BigQuery location for the input content.

This object should have the same structure as BigQuerySource

Additional domain-specific parameters describing the semantic of the imported data, any string must be up to 25000 characters long.

• For Tables: schema_inference_version - (integer) Required. The version of the algorithm that should be used for the initial inference of the schema (columns' DataTypes) of the table the data is being imported into. Allowed values: "1".

Text Segment that represents a segment in document_text.

This object should have the same structure as TextSegment

Page number of the text_segment in the original document, starts from 1.

The position of the text_segment in the page. Contains exactly 4

normalized_vertices and they are connected by edges in the order provided, which will represent a rectangle parallel to the frame. The NormalizedVertex-s are relative to the page. Coordinates are based on top-left as point (0,0).

This object should have the same structure as BoundingPoly

The type of the text_segment in document.

The number should be among the values of TextSegmentType

The resource name of the table spec to list column specs from.

Mask specifying which fields to read.

This object should have the same structure as FieldMask

Filter expression, see go/filtering.

Requested page size. The server can return fewer results than requested. If unspecified, the server will pick a default size.

A token identifying a page of results for the server to return. Typically obtained from the ListColumnSpecsResponse.next_page_token field of the previous AutoMl.ListColumnSpecs call.

The column specs read.

This object should have the same structure as ColumnSpec

A token to retrieve next page of results. Pass to ListColumnSpecsRequest.page_token to obtain that page.

The resource name of the project from which to list datasets.

An expression for filtering the results of the request.

* `dataset_metadata` - for existence of the case (e.g.
          image_classification_dataset_metadata:*). Some examples of using the filter are:

* `translation_dataset_metadata:*` --> The dataset has
                                       translation_dataset_metadata.

Requested page size. Server may return fewer results than requested. If unspecified, server will pick a default size.

A token identifying a page of results for the server to return Typically obtained via ListDatasetsResponse.next_page_token of the previous AutoMl.ListDatasets call.

The datasets read.

This object should have the same structure as Dataset

A token to retrieve next page of results. Pass to ListDatasetsRequest.page_token to obtain that page.

Resource name of the model to list the model evaluations for. If modelId is set as "-", this will list model evaluations from across all models of the parent location.

An expression for filtering the results of the request.

* `annotation_spec_id` - for =, !=  or existence. See example below for
                       the last.

Some examples of using the filter are:

* `annotation_spec_id!=4` --> The model evaluation was done for
                          annotation spec with ID different than 4.
* `NOT annotation_spec_id:*` --> The model evaluation was done for
                             aggregate of all annotation specs.

Requested page size.

A token identifying a page of results for the server to return. Typically obtained via ListModelEvaluationsResponse.next_page_token of the previous AutoMl.ListModelEvaluations call.

List of model evaluations in the requested page.

This object should have the same structure as ModelEvaluation

A token to retrieve next page of results. Pass to the ListModelEvaluationsRequest.page_token field of a new AutoMl.ListModelEvaluations request to obtain that page.

Resource name of the project, from which to list the models.

An expression for filtering the results of the request.

* `model_metadata` - for existence of the case (e.g.
          video_classification_model_metadata:*).
* `dataset_id` - for = or !=. Some examples of using the filter are:

* `image_classification_model_metadata:*` --> The model has
                                     image_classification_model_metadata.
* `dataset_id=5` --> The model was created from a dataset with ID 5.

Requested page size.

A token identifying a page of results for the server to return Typically obtained via ListModelsResponse.next_page_token of the previous AutoMl.ListModels call.

List of models in the requested page.

This object should have the same structure as Model

A token to retrieve next page of results. Pass to ListModelsRequest.page_token to obtain that page.

The resource name of the dataset to list table specs from.

Mask specifying which fields to read.

This object should have the same structure as FieldMask

Filter expression, see go/filtering.

Requested page size. The server can return fewer results than requested. If unspecified, the server will pick a default size.

A token identifying a page of results for the server to return. Typically obtained from the ListTableSpecsResponse.next_page_token field of the previous AutoMl.ListTableSpecs call.

The table specs read.

This object should have the same structure as TableSpec

A token to retrieve next page of results. Pass to ListTableSpecsRequest.page_token to obtain that page.

Metadata for translation models.

This object should have the same structure as TranslationModelMetadata

Metadata for image classification models.

This object should have the same structure as ImageClassificationModelMetadata

Metadata for text classification models.

This object should have the same structure as TextClassificationModelMetadata

Metadata for image object detection models.

This object should have the same structure as ImageObjectDetectionModelMetadata

Metadata for video classification models.

This object should have the same structure as VideoClassificationModelMetadata

Metadata for video object tracking models.

This object should have the same structure as VideoObjectTrackingModelMetadata

Metadata for text extraction models.

This object should have the same structure as TextExtractionModelMetadata

Metadata for Tables models.

This object should have the same structure as TablesModelMetadata

Metadata for text sentiment models.

This object should have the same structure as TextSentimentModelMetadata

Output only. Resource name of the model. Format: projects/{project_id}/locations/{location_id}/models/{model_id}

Required. The name of the model to show in the interface. The name can be up to 32 characters long and can consist only of ASCII Latin letters A-Z and a-z, underscores (_), and ASCII digits 0-9. It must start with a letter.

Required. The resource ID of the dataset used to create the model. The dataset must come from the same ancestor project and location.

Output only. Timestamp when the model training finished and can be used for prediction.

This object should have the same structure as Timestamp

Output only. Timestamp when this model was last updated.

This object should have the same structure as Timestamp

Output only. Deployment state of the model. A model can only serve prediction requests after it gets deployed.

The number should be among the values of DeploymentState

Model evaluation metrics for image, text, video and tables classification. Tables problem is considered a classification when the target column is CATEGORY DataType.

This object should have the same structure as ClassificationEvaluationMetrics

Model evaluation metrics for Tables regression. Tables problem is considered a regression when the target column has FLOAT64 DataType.

This object should have the same structure as RegressionEvaluationMetrics

Model evaluation metrics for translation.

This object should have the same structure as TranslationEvaluationMetrics

Model evaluation metrics for image object detection.

This object should have the same structure as ImageObjectDetectionEvaluationMetrics

Model evaluation metrics for video object tracking.

This object should have the same structure as VideoObjectTrackingEvaluationMetrics

Evaluation metrics for text sentiment models.

This object should have the same structure as TextSentimentEvaluationMetrics

Evaluation metrics for text extraction models.

This object should have the same structure as TextExtractionEvaluationMetrics

Output only. Resource name of the model evaluation. Format:

projects/{project_id}/locations/{location_id}/models/{model_id}/modelEvaluations/{model_evaluation_id}

Output only. The ID of the annotation spec that the model evaluation applies to. The The ID is empty for the overall model evaluation. For Tables annotation specs in the dataset do not exist and this ID is always not set, but for CLASSIFICATION

prediction_type-s the display_name field is used.

Output only. The value of display_name at the moment when the model was trained. Because this field returns a value at model training time, for different models trained from the same dataset, the values may differ, since display names could had been changed between the two model's trainings. For Tables CLASSIFICATION

prediction_type-s distinct values of the target column at the moment of the model evaluation are populated here. The display_name is empty for the overall model evaluation.

Output only. Timestamp when this model evaluation was created.

This object should have the same structure as Timestamp

Output only. The number of examples used for model evaluation, i.e. for which ground truth from time of model creation is compared against the predicted annotations created by the model. For overall ModelEvaluation (i.e. with annotation_spec_id not set) this is the total number of all examples used for evaluation. Otherwise, this is the count of examples that according to the ground truth were annotated by the

annotation_spec_id.

The Google Cloud Storage location where the model is to be written to. This location may only be set for the following model formats: "tflite", "edgetpu_tflite", "core_ml", "docker".

Under the directory given as the destination a new one with name "model-export--", where timestamp is in YYYY-MM-DDThh:mm:ss.sssZ ISO-8601 format, will be created. Inside the model and any of its supporting files will be written.

This object should have the same structure as GcsDestination

The GCR location where model image is to be pushed to. This location may only be set for the following model formats: "docker".

The model image will be created under the given URI.

This object should have the same structure as GcrDestination

The format in which the model must be exported. The available, and default, formats depend on the problem and model type (if given problem and type combination doesn't have a format listed, it means its models are not exportable):

• For Image Classification mobile-low-latency-1, mobile-versatile-1, mobile-high-accuracy-1: "tflite" (default), "edgetpu_tflite", "tf_saved_model", "docker".

• For Image Classification mobile-core-ml-low-latency-1, mobile-core-ml-versatile-1, mobile-core-ml-high-accuracy-1: "core_ml" (default). Formats description:

• tflite - Used for Android mobile devices.

• edgetpu_tflite - Used for Edge TPU devices.

• tf_saved_model - A tensorflow model in SavedModel format.

• docker - 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. See more at [containers

quickstart](https: //cloud.google.com/vision/automl/docs/containers-gcs-quickstart)

• core_ml - Used for iOS mobile devices.

Additional model-type and format specific parameters describing the requirements for the to be exported model files, any string must be up to 25000 characters long.

• For docker format: cpu_architecture - (string) "x86_64" (default). gpu_architecture - (string) "none" (default), "nvidia".

Required. Horizontal coordinate.

Required. Vertical coordinate.

The Google Cloud Storage location where the output is to be written to. For Image Object Detection, Text Extraction, Video Classification and Tables, in the given directory a new directory will be created with name: export_data-- where timestamp is in YYYY-MM-DDThh:mm:ss.sssZ ISO-8601 format. All export output will be written into that directory.

This object should have the same structure as GcsDestination

The BigQuery location where the output is to be written to.

This object should have the same structure as BigQueryDestination

Name of the model requested to serve the prediction.

Required. Payload to perform a prediction on. The payload must match the problem type that the model was trained to solve.

This object should have the same structure as ExamplePayload

Additional domain-specific parameters, any string must be up to 25000 characters long.

• For Image Classification:

  score_threshold - (float) A value from 0.0 to 1.0. When the model makes predictions for an image, it will only produce results that have at least this confidence score. The default is 0.5.

• For Image Object Detection: score_threshold - (float) When Model detects objects on the image, it will only produce bounding boxes which have at least this confidence score. Value in 0 to 1 range, default is 0.5. max_bounding_box_count - (int64) No more than this number of bounding boxes will be returned in the response. Default is 100, the requested value may be limited by server.

• For Tables: feature_importance - (boolean) Whether

[feature_importance][[google.cloud.automl.v1beta1.TablesModelColumnInfo.feature_importance] should be populated in the returned

[TablesAnnotation(-s)][[google.cloud.automl.v1beta1.TablesAnnotation]. The default is false.

Prediction result. Translation and Text Sentiment will return precisely one payload.

This object should have the same structure as AnnotationPayload

The preprocessed example that AutoML actually makes prediction on. Empty if AutoML does not preprocess the input example.

• For Text Extraction: If the input is a .pdf file, the OCR'ed text will be provided in document_text.

This object should have the same structure as ExamplePayload

Additional domain-specific prediction response metadata.

• For Image Object Detection: max_bounding_box_count - (int64) At most that many bounding boxes per image could have been returned.

• For Text Sentiment: sentiment_score - (float, deprecated) A value between -1 and 1, -1 maps to least positive sentiment, while 1 maps to the most positive one and the higher the score, the more positive the sentiment in the document is. Yet these values are relative to the training data, so e.g. if all data was positive then -1 will be also positive (though the least). The sentiment_score shouldn't be confused with "score" or "magnitude" from the previous Natural Language Sentiment Analysis API.

Output only. Root Mean Squared Error (RMSE).

Output only. Mean Absolute Error (MAE).

Output only. Mean absolute percentage error. Only set if all ground truth values are are positive.

Output only. R squared.

Output only. Root mean squared log error.

Output only. 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 field or, if that one is not populated, length of the display_name field.

The resource IDs of the column specs describing the columns of the row. If set must contain, but possibly in a different order, all input feature

column_spec_ids of the Model this row is being passed to. Note: The below values field must match order of this field, if this field is set.

Required. The values of the row cells, given in the same order as the column_spec_ids, or, if not set, then in the same order as input feature

column_specs of the Model this row is being passed to.

This object should have the same structure as Value

The CATEGORY value.

The number of occurrences of this value in the series.

The statistics of the top 20 unigrams, ordered by count.

This object should have the same structure as UnigramStats

Map from a field name of the struct to data stats aggregated over series of all data in that field across all the structs.

Unordered map of struct field names to their data types. Fields cannot be added or removed via Update. Their names and data types are still mutable.

Output only. A confidence estimate between 0.0 and 1.0, inclusive. A higher value means greater confidence in the returned value. For

target_column_spec of FLOAT64 data type the score is not populated.

Output only. Only populated when

target_column_spec has FLOAT64 data type. An interval in which the exactly correct target value has 95% chance to be in.

This object should have the same structure as DoubleRange

The predicted value of the row's

target_column. The value depends on the column's DataType:

• CATEGORY - the predicted (with the above confidence score) CATEGORY value.

• FLOAT64 - the predicted (with above prediction_interval) FLOAT64 value.

This object should have the same structure as Value

Output only. Auxiliary information for each of the model's

input_feature_column_specs with respect to this particular prediction. If no other fields than

column_spec_name and

column_display_name would be populated, then this whole field is not.

This object should have the same structure as TablesModelColumnInfo

Output only. The table_spec_id of the primary table of this dataset.

column_spec_id of the primary table's column that should be used as the training & prediction target. This column must be non-nullable and have one of following data types (otherwise model creation will error):

• CATEGORY

• FLOAT64

If the type is CATEGORY , only up to 100 unique values may exist in that column across all rows.

NOTE: Updates of this field will instantly affect any other users concurrently working with the dataset.

column_spec_id of the primary table's column that should be used as the weight column, i.e. the higher the value the more important the row will be during model training. Required type: FLOAT64. Allowed values: 0 to 10000, inclusive on both ends; 0 means the row is ignored for training. If not set all rows are assumed to have equal weight of 1. NOTE: Updates of this field will instantly affect any other users concurrently working with the dataset.

column_spec_id of the primary table column which specifies a possible ML use of the row, i.e. the column will be used to split the rows into TRAIN, VALIDATE and TEST sets. Required type: STRING. This column, if set, must either have all of TRAIN, VALIDATE, TEST among its values, or only have TEST, UNASSIGNED values. In the latter case the rows with UNASSIGNED value will be assigned by AutoML. Note that if a given ml use distribution makes it impossible to create a "good" model, that call will error describing the issue. If both this column_spec_id and primary table's time_column_spec_id are not set, then all rows are treated as UNASSIGNED. NOTE: Updates of this field will instantly affect any other users concurrently working with the dataset.

Output only. Correlations between

TablesDatasetMetadata.target_column_spec_id, and other columns of the

TablesDatasetMetadataprimary_table. Only set if the target column is set. Mapping from other column spec id to its CorrelationStats with the target column. This field may be stale, see the stats_update_time field for for the timestamp at which these stats were last updated.

Output only. The most recent timestamp when target_column_correlations field and all descendant ColumnSpec.data_stats and ColumnSpec.top_correlated_columns fields were last (re-)generated. Any changes that happened to the dataset afterwards are not reflected in these fields values. The regeneration happens in the background on a best effort basis.

This object should have the same structure as Timestamp

Output only. The name of the ColumnSpec describing the column. Not populated when this proto is outputted to BigQuery.

Output only. The display name of the column (same as the display_name of its ColumnSpec).

Output only. When given as part of a Model (always populated): 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 (populated iff feature_importance param is set) or Batch Predict (populated iff feature_importance param is set): Measurement of how impactful for the prediction returned for the given row the value in this column was. A value between 0 and 1, higher means larger impact. These values are normalized - for all input feature columns of a single predicted row they add to 1.

Column spec of the dataset's primary table's column the model is predicting. Snapshotted when model creation started. Only 3 fields are used: name - May be set on CreateModel, if it's not then the ColumnSpec corresponding to the current target_column_spec_id of the dataset the model is trained from is used. If neither is set, CreateModel will error. display_name - Output only. data_type - Output only.

This object should have the same structure as ColumnSpec

Column specs of the dataset's primary table's columns, on which the model is trained and which are used as the input for predictions. The

target_column as well as, according to dataset's state upon model creation,

weight_column, and

ml_use_column must never be included here.

Only 3 fields are used:

• name - May be set on CreateModel, if set only the columns specified are used, otherwise all primary table's columns (except the ones listed above) are used for the training and prediction input.

• display_name - Output only.

• data_type - Output only.

This object should have the same structure as ColumnSpec

Objective function the model is optimizing towards. The training process creates a model that maximizes/minimizes the value of the objective function over the validation set.

The supported optimization objectives depend on the prediction type. If the field is not set, a default objective function is used.

CLASSIFICATION_BINARY: "MAXIMIZE_AU_ROC" (default) - Maximize the area under the receiver operating characteristic (ROC) curve. "MINIMIZE_LOG_LOSS" - Minimize log loss. "MAXIMIZE_AU_PRC" - Maximize the area under the precision-recall curve. "MAXIMIZE_PRECISION_AT_RECALL" - Maximize precision for a specified recall value. "MAXIMIZE_RECALL_AT_PRECISION" - Maximize recall for a specified precision 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-squared log error (RMSLE).

Output only. Auxiliary information for each of the input_feature_column_specs with respect to this particular model.

This object should have the same structure as TablesModelColumnInfo

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

The training cost of the model will not exceed this budget. The final cost will be attempted to be close to the budget, though may end up being (even) noticeably smaller - at the backend's discretion. This especially may happen when further model training ceases to provide any improvements.

If the budget is set to a value known to be insufficient to train a model for the given dataset, the training won't be attempted and will error.

The train budget must be between 1,000 and 72,000 milli node hours, inclusive.

Output only. 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.

Use the entire training budget. This disables the early stopping feature. By default, the early stopping feature is enabled, which means that AutoML Tables might stop training before the entire training budget has been used.

Output only. The resource name of the table spec. Form:

projects/{project_id}/locations/{location_id}/datasets/{dataset_id}/tableSpecs/{table_spec_id}

column_spec_id of the time column. Only used if the parent dataset's ml_use_column_spec_id is not set. Used to split rows into TRAIN, VALIDATE and TEST sets such that oldest rows go to TRAIN set, newest to TEST, and those in between to VALIDATE. Required type: TIMESTAMP. If both this column and ml_use_column are not set, then ML use of all rows will be assigned by AutoML. NOTE: Updates of this field will instantly affect any other users concurrently working with the dataset.

Output only. The number of rows (i.e. examples) in the table.

Output only. The number of valid rows (i.e. without values that don't match DataType-s of their columns).

Output only. The number of columns of the table. That is, the number of child ColumnSpec-s.

Output only. Input configs via which data currently residing in the table had been imported.

This object should have the same structure as InputConfig

Used to perform consistent read-modify-write updates. If not set, a blind "overwrite" update happens.

Required. Type of the classification problem.

The number should be among the values of ClassificationType

An entity annotation will set this, which is the part of the original text to which the annotation pertains.

This object should have the same structure as TextSegment

Output only. A confidence estimate between 0.0 and 1.0. A higher value means greater confidence in correctness of the annotation.

Output only. The Area under precision recall curve metric.

Output only. Metrics that have confidence thresholds. Precision-recall curve can be derived from it.

This object should have the same structure as ConfidenceMetricsEntry

Output only. The content of the TextSegment.

Required. Zero-based character index of the first character of the text segment (counting characters from the beginning of the text).

Required. Zero-based character index of the first character past the end of the text segment (counting character from the beginning of the text). The character at the end_offset is NOT included in the text segment.

Output only. The sentiment with the semantic, as given to the AutoMl.ImportData when populating the dataset from which the model used for the prediction had been trained. The sentiment values are between 0 and Dataset.text_sentiment_dataset_metadata.sentiment_max (inclusive), with higher value meaning more positive sentiment. They are completely relative, i.e. 0 means least positive sentiment and sentiment_max means the most positive from the sentiments present in the train data. Therefore e.g. if train data had only negative sentiment, then sentiment_max, would be still negative (although least negative). The sentiment shouldn't be confused with "score" or "magnitude" from the previous Natural Language Sentiment Analysis API.

Required. A sentiment is expressed as an integer ordinal, where higher value means a more positive sentiment. The range of sentiments that will be used is between 0 and sentiment_max (inclusive on both ends), and all the values in the range must be represented in the dataset before a model can be created. sentiment_max value must be between 1 and 10 (inclusive).

Output only. Precision.

Output only. Recall.

Output only. The harmonic mean of recall and precision.

Output only. Mean absolute error. Only set for the overall model evaluation, not for evaluation of a single annotation spec.

Output only. Mean squared error. Only set for the overall model evaluation, not for evaluation of a single annotation spec.

Output only. Linear weighted kappa. Only set for the overall model evaluation, not for evaluation of a single annotation spec.

Output only. Quadratic weighted kappa. Only set for the overall model evaluation, not for evaluation of a single annotation spec.

Output only. Confusion matrix of the evaluation. Only set for the overall model evaluation, not for evaluation of a single annotation spec.

This object should have the same structure as ConfusionMatrix

Output only. The annotation spec ids used for this evaluation. Deprecated .

Required. The content of the text snippet as a string. Up to 250000 characters long.

Optional. The format of content. Currently the only two allowed values are "text/html" and "text/plain". If left blank, the format is automatically determined from the type of the uploaded content.

Output only. HTTP URI where you can download the content.

Start of the time segment (inclusive), represented as the duration since the example start.

This object should have the same structure as Duration

End of the time segment (exclusive), represented as the duration since the example start.

This object should have the same structure as Duration

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).

Output only . The translated content.

This object should have the same structure as TextSnippet

Required. The BCP-47 language code of the source language.

Required. The BCP-47 language code of the target language.

Output only. BLEU score.

Output only. BLEU score for base model.

The resource name of the model to use as a baseline to train the custom model. If unset, we use the default base model provided by Google Translate. Format: projects/{project_id}/locations/{location_id}/models/{model_id}

Output only. Inferred from the dataset. The source languge (The BCP-47 language code) that is used for training.

Output only. The target languge (The BCP-47 language code) that is used for training.

Resource name of the model to undeploy.

The unigram.

The number of occurrences of this unigram in the series.

The column spec which replaces the resource on the server.

This object should have the same structure as ColumnSpec

The update mask applies to the resource.

This object should have the same structure as FieldMask

The dataset which replaces the resource on the server.

This object should have the same structure as Dataset

The update mask applies to the resource.

This object should have the same structure as FieldMask

The table spec which replaces the resource on the server.

This object should have the same structure as TableSpec

The update mask applies to the resource.

This object should have the same structure as FieldMask

Output only. Expresses the type of video classification. Possible values:

• segment - Classification done on a specified by user time segment of a video. AnnotationSpec is answered to be present in that time segment, if it is present in any part of it. The video ML model evaluations are done only for this type of classification.

• shot- Shot-level classification. AutoML Video Intelligence determines the boundaries for each camera shot in the entire segment of the video that user specified in the request configuration. AutoML Video Intelligence then returns labels and their confidence scores for each detected shot, along with the start and end time of the shot. WARNING: Model evaluation is not done for this classification type, the quality of it depends on training data, but there are no metrics provided to describe that quality.

• 1s_interval - AutoML Video Intelligence returns labels and their confidence scores for each second of the entire segment of the video that user specified in the request configuration. WARNING: Model evaluation is not done for this classification type, the quality of it depends on training data, but there are no metrics provided to describe that quality.

Output only . The classification details of this annotation.

This object should have the same structure as ClassificationAnnotation

Output only . The time segment of the video to which the annotation applies.

This object should have the same structure as TimeSegment

Optional. The instance of the object, expressed as a positive integer. Used to tell apart objects of the same type (i.e. AnnotationSpec) when multiple are present on a single example. NOTE: Instance ID prediction quality is not a part of model evaluation and is done as best effort. Especially in cases when an entity goes off-screen for a longer time (minutes), when it comes back it may be given a new instance ID.

Required. A time (frame) of a video to which this annotation pertains. Represented as the duration since the video's start.

This object should have the same structure as Duration

Required. The rectangle representing the object location on the frame (i.e. at the time_offset of the video).

This object should have the same structure as BoundingPoly

Output only. The confidence that this annotation is positive for the video at the time_offset, value in [0, 1], higher means higher positivity confidence. For annotations created by the user the score is 1. When user approves an annotation, the original float score is kept (and not changed to 1).

Output only. The number of video frames used to create this evaluation.

Output only. The total number of bounding boxes (i.e. summed over all frames) the ground truth used to create this evaluation had.

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 and each label confidence threshold 0.05,0.10,...,0.95,0.96,0.97,0.98,0.99 pair.

This object should have the same structure as BoundingBoxMetricsEntry

Output only. The single metric for bounding boxes evaluation: the mean_average_precision averaged over all bounding_box_metrics_entries.