ignore xgboost estimator

Author: NihalHarish

src/sagemaker/estimator.py

@@ -2223,15 +2223,16 @@ def _validate_and_set_debugger_configs(self):
 
         # Disable debugger if checkpointing is enabled by the customer
         if self.checkpoint_s3_uri and self.checkpoint_local_path and self.debugger_hook_config:
-            if self.instance_count > 1 \
-                    or (
-                        hasattr(self, "distribution") and self.distribution is not None  # pylint: disable=no-member
-                    ):
-                logger.info(
-                    "SMDebug Does Not Currently Support \
-                    Distributed Training Jobs With Checkpointing Enabled"
-                )
-                self.debugger_hook_config = False
+            if self._framework_name in {"mxnet", "pytorch", "tensorflow"}:
+                if self.instance_count > 1 or (
+                    hasattr(self, "distribution")
+                    and self.distribution is not None  # pylint: disable=no-member
+                ):
+                    logger.info(
+                        "SMDebug Does Not Currently Support \
+                        Distributed Training Jobs With Checkpointing Enabled"
+                    )
+                    self.debugger_hook_config = False
 
     def _stage_user_code_in_s3(self):
         """Upload the user training script to s3 and return the location.

tests/integ/test_debugger.py

@@ -25,6 +25,7 @@
 )
 from sagemaker.mxnet.estimator import MXNet
 from sagemaker.pytorch.estimator import PyTorch
+from sagemaker.xgboost.estimator import XGBoost
 from tests.integ import DATA_DIR, TRAINING_DEFAULT_TIMEOUT_MINUTES
 from tests.integ.retry import retries
 from tests.integ.timeout import timeout
@@ -400,7 +401,7 @@ def test_debug_hook_disabled_with_checkpointing(
             checkpoint_s3_uri=os.path.join(s3_output_path, "checkpoints"),
         )
         mx._prepare_for_training()
-        # Debug Hook should be enabled
+        # Debug Hook should be disabled
         assert mx.debugger_hook_config is False
 
         # Estimator with checkpointing enabled and Model Parallel Enabled
@@ -416,12 +417,30 @@ def test_debug_hook_disabled_with_checkpointing(
             sagemaker_session=sagemaker_session,
             # Training using SMDataParallel Distributed Training Framework
             distribution={"smdistributed": {"dataparallel": {"enabled": True}}},
-            debugger_hook_config=False,
+            checkpoint_local_path="/opt/ml/checkpoints",
+            checkpoint_s3_uri=os.path.join(s3_output_path, "checkpoints"),
         )
         pt._prepare_for_training()
-        # Debug Hook should be enabled
+        # Debug Hook should be disabled
         assert pt.debugger_hook_config is False
 
+        # Estimator with checkpointing enabled with Xgboost Estimator
+        xg = XGBoost(
+            base_job_name="test_xgboost",
+            entry_point=script_path,
+            role="SageMakerRole",
+            framework_version="1.2-1",
+            py_version="py3",
+            instance_count=2,
+            # For training with p3dn instance use - ml.p3dn.24xlarge, with p4dn instance use - ml.p4d.24xlarge
+            instance_type="ml.p3.16xlarge",
+            sagemaker_session=sagemaker_session,
+            # Training using SMDataParallel Distributed Training Framework
+        )
+        xg._prepare_for_training()
+        # Debug Hook should be enabled
+        assert xg.debugger_hook_config is not None
+
 
 def test_mxnet_with_rules_and_debugger_hook_config(
     sagemaker_session,