AWS sagemaker documentation change

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

File: sagemaker/latest/dg/debugger-configuration-for-profiling.md

Summary

Added SageMaker Python SDK v3 examples for profiler configuration and reorganized legacy SDK examples with framework-specific notes

Security assessment

The changes focus on profiler configuration examples and framework limitations. There is no evidence of security fixes or security feature documentation.

Diff

diff --git a/sagemaker/latest/dg/debugger-configuration-for-profiling.md b/sagemaker/latest/dg/debugger-configuration-for-profiling.md
index d84b05ac5..8f25453d7 100644
--- a//sagemaker/latest/dg/debugger-configuration-for-profiling.md
+++ b//sagemaker/latest/dg/debugger-configuration-for-profiling.md
@@ -35 +35 @@ The following code examples are not directly executable. Proceed to the next sec
-PyTorch
+SageMaker Python SDK v3
@@ -38,0 +39,61 @@ PyTorch
+    # An example of creating a training job with profiler configuration
+    import boto3
+    from sagemaker.core import image_uris
+    from sagemaker.core.resources import TrainingJob
+    from sagemaker.core.shapes import (
+        AlgorithmSpecification,
+        ResourceConfig,
+        OutputDataConfig,
+        StoppingCondition,
+        ProfilerConfig as CoreProfilerConfig,
+        ProfilerRuleConfiguration,
+    )
+    
+    session=boto3.session.Session()
+    region=session.region_name
+    
+    # Retrieve the training image for your framework
+    # Change framework to "tensorflow", "mxnet", "xgboost", etc. as needed
+    training_image = image_uris.retrieve(
+        framework="pytorch",  # or "tensorflow", "mxnet", "xgboost"
+        region=region,
+        version="1.12.0",
+        py_version="py37",
+        instance_type="ml.p3.2xlarge",
+        image_scope="training"
+    )
+    
+    profiler_config=CoreProfilerConfig(...)
+    profiler_rule_configurations=[
+        ProfilerRuleConfiguration(
+            rule_configuration_name="BuiltInRule",
+            rule_evaluator_image="rule-evaluator-image-uri",
+        )
+    ]
+    
+    TrainingJob.create(
+        training_job_name="debugger-profiling-demo",
+        algorithm_specification=AlgorithmSpecification(
+            training_image=training_image,
+            training_input_mode="File",
+        ),
+        role_arn="arn:aws:iam::123456789012:role/SageMakerRole",
+        resource_config=ResourceConfig(instance_type="ml.p3.2xlarge", 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),
+    
+        # SageMaker Debugger parameters
+        profiler_config=profiler_config,
+        profiler_rule_configurations=profiler_rule_configurations,
+    )
+
+###### Note
+
+For MXNet and XGBoost, when configuring the `profiler_config` parameter, you can only configure for system monitoring. Profiling framework metrics is not supported for MXNet or XGBoost.
+
+SageMaker Python SDK v2 (Legacy)
+    
+
+**PyTorch:**
+    
+    
@@ -69,2 +130 @@ PyTorch
-TensorFlow
-    
+**TensorFlow:**
@@ -103,2 +163 @@ TensorFlow
-MXNet
-    
+**MXNet:**
@@ -137,2 +196 @@ For MXNet, when configuring the `profiler_config` parameter, you can only config
-XGBoost
-    
+**XGBoost:**
@@ -170,2 +228 @@ For XGBoost, when configuring the `profiler_config` parameter, you can only conf
-Generic estimator
-    
+**Generic estimator:**