XgboostMtModelEval¶

XGBoost multi-task model evaluation step


SageMaker step type	`Processing`
Node type	internal (consumes upstream, produces downstream)
Container entry point	`xgboost_mt_model_eval.py`
Interface file	`steps/interfaces/xgboost_mt_model_eval.step.yaml`

Compute¶


Compute kind	`framework`
SDK class	`PyTorch` (SageMaker DLC via `image_uris.retrieve`)

Functionality¶

XgboostMt multi-task model evaluation. Generates per-task and aggregate metrics with visualizations.

Inputs (dependencies)¶

Input	Type	Required	Compatible producers
`model_input`	`model_artifacts`	yes	XgboostMtTraining, XGBoostTraining, XgboostMtModel, XGBoostModel, DummyTraining
`processed_data`	`processing_output`	yes	TabularPreprocessing, CradleDataLoading, RiskTableMapping, CurrencyConversion, LabelRulesetExecution, BedrockBatchProcessing, BedrockProcessing, TemporalSplitPreprocessing

Outputs¶

Output	Type
`eval_output`	`processing_output`
`metrics_output`	`processing_output`

Consumers (downstream steps)¶

Steps that declare this step as a compatible input source:

ModelCalibration

Framework requirements¶

Package	Version
`pandas`	`>=1.2.0,<2.0.0`
`numpy`	`>=1.21.0`
`scikit-learn`	`>=0.23.2,<1.0.0`
`matplotlib`	`>=3.0.0`

← Back to the Step Catalog