AWS sagemaker documentation change

Service: sagemaker · 2026-06-28 · Documentation low

File: sagemaker/latest/dg/define-train-metrics.md

Summary

Added documentation for SageMaker Python SDK v3 showing how to define training metrics using the new core API, including a code example for TrainingJob.create() with metric definitions.

Security assessment

The change adds API usage examples for the new SDK version without mentioning security vulnerabilities, configurations, or features. It focuses solely on metric definition syntax.

Diff

diff --git a/sagemaker/latest/dg/define-train-metrics.md b/sagemaker/latest/dg/define-train-metrics.md
index 64cd70f7b..676551112 100644
--- a//sagemaker/latest/dg/define-train-metrics.md
+++ b//sagemaker/latest/dg/define-train-metrics.md
@@ -46,0 +47,40 @@ In the regex for the `train:error` metric defined in the preceding example, the
+SageMaker Python SDK v3
+    
+
+Define the metrics that you want to send to CloudWatch by specifying a list of metric definitions in the `AlgorithmSpecification` when you create a `TrainingJob` using `sagemaker-core`. For example, if you want to monitor both the `train:error` and `validation:error` metrics in CloudWatch, your `TrainingJob.create()` call would look like the following example:
+    
+    
+    from sagemaker.core.resources import TrainingJob
+    from sagemaker.core.shapes import (
+        AlgorithmSpecification,
+        MetricDefinition,
+        ResourceConfig,
+        OutputDataConfig,
+        StoppingCondition,
+    )
+    
+    TrainingJob.create(
+        training_job_name="my-training-job",
+        role_arn="arn:aws:iam::123456789012:role/SageMakerRole",
+        algorithm_specification=AlgorithmSpecification(
+            training_image="your-own-image-uri",
+            training_input_mode="File",
+            metric_definitions=[
+                MetricDefinition(name="train:error", regex="Train_error=(.*?);"),
+                MetricDefinition(name="validation:error", regex="Valid_error=(.*?);"),
+            ],
+        ),
+        resource_config=ResourceConfig(
+            instance_type="ml.c4.xlarge",
+            instance_count=1,
+            volume_size_in_gb=30,
+        ),
+        output_data_config=OutputDataConfig(s3_output_path="s3://bucket/output"),
+        stopping_condition=StoppingCondition(max_runtime_in_seconds=3600),
+    )
+
+For more information about training by using [Amazon SageMaker Python SDK](https://sagemaker.readthedocs.io/en/stable) ModelTrainers, see[ Sagemaker Python SDK](https://github.com/aws/sagemaker-python-sdk#sagemaker-python-sdk-overview) on GitHub. 
+
+SageMaker Python SDK v2 (Legacy)
+    
+