MLModelConfigOutput
ML model configuration for training pipelines
Fields
Author-owned XML source envelope used to reproduce and edit this model configuration; runtime results are stored separately.
Unique identifier
Tenant this config belongs to
Config name
Config description
Type of ML model
Possible values (3)
REGRESSIONRegression model (predicts numeric values)
CLASSIFICATIONClassification model (predicts categories)
FORECASTTime-series forecast model (predicts future values over a horizon)
Engine that trains/serves this model
Possible values (2)
XGBOOSTXGBoost gradient-boosted trees (tabular regression/classification)
AUGURSAugurs time-series engine (forecasting)
MicroSQL query for data retrieval. Each SQL page has a 30-second statement timeout; preview sampling does not limit the full training query.
Concept-specific configuration (tabular or time-series)
Status of the latest training run. A published version remains available for serving while retraining or after a failed run.
Possible values (4)
IDLECreated but never trained
TRAININGCurrently training
TRAINEDSuccessfully trained and ready for predictions
FAILEDLast training attempt failed
Blob storage key for the trained model
When the model was last trained (ISO8601)
Duration of last training run in milliseconds
Metrics from last training run
Error message if last training failed; SQL source pages have a 30-second statement timeout
Feature importance from trained model
Per-round evaluation history
Original CSV filename (if CSV data source)
Size of uploaded CSV in bytes
Whether a CSV file has been uploaded as training data
First few rows of CSV data for preview (header + up to 10 rows)
JSON Schema describing the model's input features (types, enum values). Generated after training.
The pipeline's emitted columns + types (cached from the last preview of the saved config)
ID of the version currently used for predictions
Version number of the published version (e.g. 1, 2, 3)
Whether new versions are automatically published after training
Whether this model is system-managed (auto-created and auto-retrained by the platform)
Creation timestamp (ISO8601)
Last update timestamp (ISO8601)
Referenced by (7)
Exact type SDL
"""ML model configuration for training pipelines"""
type MLModelConfigOutput @iac_resource(name: "ml_model_config") @iac_precondition(field: "authoringRevision", input: "expectedRevision") @iac_guide(ref: "ml-guide") @iac_ops(create: "upsertMLModelSource", read: "getMLModelConfig", update: "upsertMLModelSource", delete: "deleteMLModelConfig") {
"""
Author-owned XML source envelope used to reproduce and edit this model configuration; runtime results are stored separately.
"""
source: String! @value_validate(with: ["ml-model-source"]) @iac_artifact(language: XML, models: ["ml_model_document"], description: "Machine learning model document, including its data pipeline.") @iac_input(path: "sourceXML")
"""
Monotonic source/configuration revision used to reject stale authoring updates; independent of trained model versions.
"""
authoringRevision: Int! @iac_computed
"""Unique identifier"""
id: MLModelConfigID! @iac_identity(assignedBy: SERVER) @iac_computed
"""Tenant this config belongs to"""
tenantId: TenantID! @iac_computed
"""Config name"""
name: String! @iac_computed
"""Config description"""
description: String @iac_computed
"""Type of ML model"""
modelType: MLModelType! @iac_computed
"""Engine that trains/serves this model"""
backend: MLBackend! @iac_computed
"""
MicroSQL query for data retrieval. Each SQL page has a 30-second statement timeout; preview sampling does not limit the full training query.
"""
dataQuerySQL: String @iac_computed @iac_artifact(language: SQL)
"""Concept-specific configuration (tabular or time-series)"""
spec: MLModelSpec! @iac_computed @value_validate(with: ["ml-model-spec"])
"""
Status of the latest training run. A published version remains available for serving while retraining or after a failed run.
"""
status: MLModelTrainingStatus! @iac_computed
"""Blob storage key for the trained model"""
trainedModelID: String @iac_computed
"""When the model was last trained (ISO8601)"""
lastTrainedAt: String @iac_computed
"""Duration of last training run in milliseconds"""
trainingDurationMs: Int @iac_computed
"""Metrics from last training run"""
trainingMetrics: TrainingMetrics @iac_computed
"""
Error message if last training failed; SQL source pages have a 30-second statement timeout
"""
trainingError: String @iac_computed
"""Feature importance from trained model"""
featureImportance: [MLFeatureImportance!] @iac_computed
"""Per-round evaluation history"""
evaluationHistory: [MLEvaluationEntry!] @iac_computed
"""Original CSV filename (if CSV data source)"""
csvFilename: String @iac_computed
"""Size of uploaded CSV in bytes"""
csvSizeBytes: Int @iac_computed
"""Whether a CSV file has been uploaded as training data"""
hasTrainingCSV: Boolean! @iac_computed
"""First few rows of CSV data for preview (header + up to 10 rows)"""
csvPreviewRows: [[String!]!] @iac_computed
"""
JSON Schema describing the model's input features (types, enum values). Generated after training.
"""
inferenceSchema: String @iac_computed
"""
The pipeline's emitted columns + types (cached from the last preview of the saved config)
"""
emittedSchema: [MLEmittedColumn!] @iac_computed
"""ID of the version currently used for predictions"""
publishedVersionId: MLModelConfigVersionID @iac_writtenBy(operation: "publishMLModelVersion", argument: "versionId")
"""Version number of the published version (e.g. 1, 2, 3)"""
publishedVersionNumber: Int @iac_computed
"""Whether new versions are automatically published after training"""
autoPublish: Boolean! @iac_computed
"""
Whether this model is system-managed (auto-created and auto-retrained by the platform)
"""
isSystem: Boolean! @iac_computed
"""Creation timestamp (ISO8601)"""
createdAt: String! @iac_computed
"""Last update timestamp (ISO8601)"""
updatedAt: String! @iac_computed
}