Merge branch 'master' into ole

Author: xiaoyi-cheng

CHANGELOG.md

@@ -4,7 +4,7 @@
 
 ### Features
 
- * Support for SageMaker Intelligent Defaults
+ * Support for SageMaker SDK Defaults
 
 ### Bug Fixes and Other Changes
 

src/sagemaker/debugger/profiler_config.py

@@ -13,10 +13,14 @@
 """Configuration for collecting system and framework metrics in SageMaker training jobs."""
 from __future__ import absolute_import
 
+import logging
 from typing import Optional, Union
 
 from sagemaker.debugger.framework_profile import FrameworkProfile
 from sagemaker.workflow.entities import PipelineVariable
+from sagemaker.deprecations import deprecation_warn_base
+
+logger = logging.getLogger(__name__)
 
 
 class ProfilerConfig(object):
@@ -81,6 +85,11 @@ class and SageMaker Framework estimators.
         self.framework_profile_params = framework_profile_params
         self.disable_profiler = disable_profiler
 
+        if self.framework_profile_params is not None:
+            deprecation_warn_base(
+                "Framework profiling will be deprecated from tensorflow 2.12 and pytorch 2.0"
+            )
+
     def _to_request_dict(self):
         """Generate a request dictionary using the parameters provided when initializing the object.
 

src/sagemaker/deprecations.py

@@ -64,6 +64,15 @@ def deprecation_warn(name, date, msg=None):
     _warn(f"{name} will be deprecated on {date}.{msg}")
 
 
+def deprecation_warn_base(msg):
+    """Raise a warning for soon to be deprecated feature in sagemaker>=2
+
+    Args:
+        msg (str): the warning message.
+    """
+    _warn(msg)
+
+
 def deprecation_warning(date, msg=None):
     """Decorator for raising deprecation warning for a feature in sagemaker>=2
 

src/sagemaker/fw_utils.py

@@ -22,6 +22,7 @@
 import tempfile
 from collections import namedtuple
 from typing import Optional, Union, Dict
+from packaging import version
 
 import sagemaker.image_uris
 from sagemaker.session_settings import SessionSettings
@@ -30,6 +31,7 @@
 
 from sagemaker.deprecations import renamed_warning, renamed_kwargs
 from sagemaker.workflow.entities import PipelineVariable
+from sagemaker.deprecations import deprecation_warn_base
 
 logger = logging.getLogger(__name__)
 
@@ -638,6 +640,35 @@ def warn_if_parameter_server_with_multi_gpu(training_instance_type, distribution
         logger.warning(PARAMETER_SERVER_MULTI_GPU_WARNING)
 
 
+def profiler_config_deprecation_warning(
+    profiler_config, image_uri, framework_name, framework_version
+):
+    """Put out a deprecation message for if framework profiling is specified TF >= 2.12 and PT >= 2.0"""
+    if profiler_config is None or profiler_config.framework_profile_params is None:
+        return
+
+    if framework_name not in ("pytorch", "tensorflow"):
+        return
+
+    if framework_version is None:
+        framework_name, _, image_tag, _ = framework_name_from_image(image_uri)
+
+        if image_tag is not None:
+            framework_version = framework_version_from_tag(image_tag)
+
+    if framework_version is not None:
+        framework_profile_thresh = (
+            version.parse("2.0") if framework_name == "pytorch" else version.parse("2.12")
+        )
+        framework_profile = version.parse(framework_version)
+        if framework_profile >= framework_profile_thresh:
+            deprecation_warn_base(
+                f"Framework profiling is deprecated from\
+                 {framework_name} version {framework_version}.\
+                 No framework metrics will be collected"
+            )
+
+
 def validate_smdistributed(
     instance_type, framework_name, framework_version, py_version, distribution, image_uri=None
 ):

src/sagemaker/pytorch/estimator.py

@@ -25,6 +25,7 @@
     python_deprecation_warning,
     validate_version_or_image_args,
     validate_distribution,
+    profiler_config_deprecation_warning,
 )
 from sagemaker.pytorch import defaults
 from sagemaker.pytorch.model import PyTorchModel
@@ -298,6 +299,11 @@ def __init__(
             )
         self.compiler_config = compiler_config
 
+        if "profiler_config" in kwargs:
+            profiler_config_deprecation_warning(
+                kwargs["profiler_config"], image_uri, self._framework_name, framework_version
+            )
+
     def _pytorch_distribution_configuration(self, distribution):
         """Returns a dict of distribution config for PyTorch training
 

src/sagemaker/tensorflow/estimator.py

@@ -235,6 +235,11 @@ def __init__(
                 compiler_config.validate(self)
         self.compiler_config = compiler_config
 
+        if "profiler_config" in kwargs:
+            fw.profiler_config_deprecation_warning(
+                kwargs["profiler_config"], image_uri, self._framework_name, framework_version
+            )
+
     def _validate_args(self, py_version):
         """Placeholder docstring"""
 

src/sagemaker/tuner.py

@@ -559,6 +559,7 @@ def to_input_req(self):
         """
         completion_criteria_config = {}
         if self.max_number_of_training_jobs_not_improving is not None:
+            completion_criteria_config[BEST_OBJECTIVE_NOT_IMPROVING] = {}
             completion_criteria_config[BEST_OBJECTIVE_NOT_IMPROVING][
                 MAX_NUMBER_OF_TRAINING_JOBS_NOT_IMPROVING
             ] = self.max_number_of_training_jobs_not_improving
@@ -569,6 +570,7 @@ def to_input_req(self):
             ] = self.target_objective_metric_value
 
         if self.complete_on_convergence is not None:
+            completion_criteria_config[CONVERGENCE_DETECTED] = {}
             completion_criteria_config[CONVERGENCE_DETECTED][COMPLETE_ON_CONVERGENCE_DETECTED] = (
                 "Enabled" if self.complete_on_convergence else "Disabled"
             )

tests/unit/test_profiler_config.py

@@ -12,11 +12,17 @@
 # language governing permissions and limitations under the License.
 from __future__ import absolute_import
 
+import os
 import pytest
 import re
 import time
+import warnings
+from packaging import version
 
-
+from sagemaker import image_uris
+import sagemaker.fw_utils as fw
+from sagemaker.pytorch import PyTorch
+from sagemaker.tensorflow import TensorFlow
 from sagemaker.debugger.profiler_config import ProfilerConfig, FrameworkProfile
 
 from sagemaker.debugger.metrics_config import (
@@ -643,3 +649,162 @@ def test_validation():
 
     with pytest.raises(AssertionError, match=ErrorMessages.INVALID_CPROFILE_TIMER.value):
         PythonProfilingConfig(cprofile_timer="bad_cprofile_timer")
+
+
+DATA_DIR = os.path.join(os.path.dirname(__file__), "..", "data")
+SCRIPT_PATH = os.path.join(DATA_DIR, "dummy_script.py")
+INSTANCE_COUNT = 1
+INSTANCE_TYPE = "ml.p3.2xlarge"
+ROLE = "Dummy"
+REGION = "us-west-2"
+
+
+def _check_framework_profile_deprecation_warning(framework_version, framework_name, warn_list):
+    """Check the collected warnings for a framework fromfile DeprecationWarning"""
+
+    thresh = version.parse("2.12") if framework_name == "tensorflow" else version.parse("2.0")
+    actual = version.parse(framework_version)
+
+    if actual >= thresh:
+        # should find a Framework profiling deprecation warning
+        for w in warn_list:
+            if issubclass(w.category, DeprecationWarning):
+                if "Framework profiling" in str(w.message):
+                    return
+        assert 0  # Should have found a deprecation and exited above
+
+
+def test_create_pytorch_estimator_with_framework_profile(
+    sagemaker_session,
+    pytorch_inference_version,
+    pytorch_inference_py_version,
+    default_framework_profile,
+):
+    profiler_config = ProfilerConfig(framework_profile_params=default_framework_profile)
+
+    with warnings.catch_warnings(record=True) as warn_list:
+        warnings.simplefilter("always")
+        framework_version = pytorch_inference_version
+        pytorch = PyTorch(
+            entry_point=SCRIPT_PATH,
+            framework_version=framework_version,
+            py_version=pytorch_inference_py_version,
+            role=ROLE,
+            sagemaker_session=sagemaker_session,
+            instance_count=INSTANCE_COUNT,
+            instance_type=INSTANCE_TYPE,
+            base_job_name="job",
+            profiler_config=profiler_config,
+        )
+
+        _check_framework_profile_deprecation_warning(
+            framework_version, pytorch._framework_name, warn_list
+        )
+
+
+def test_create_pytorch_estimator_w_image_with_framework_profile(
+    sagemaker_session,
+    pytorch_inference_version,
+    pytorch_inference_py_version,
+    gpu_pytorch_instance_type,
+    default_framework_profile,
+):
+    image_uri = image_uris.retrieve(
+        "pytorch",
+        REGION,
+        version=pytorch_inference_version,
+        py_version=pytorch_inference_py_version,
+        instance_type=gpu_pytorch_instance_type,
+        image_scope="inference",
+    )
+
+    profiler_config = ProfilerConfig(framework_profile_params=default_framework_profile)
+
+    with warnings.catch_warnings(record=True) as warn_list:
+        warnings.simplefilter("always")
+        pytorch = PyTorch(
+            entry_point=SCRIPT_PATH,
+            role=ROLE,
+            sagemaker_session=sagemaker_session,
+            instance_count=INSTANCE_COUNT,
+            instance_type=gpu_pytorch_instance_type,
+            image_uri=image_uri,
+            profiler_config=profiler_config,
+        )
+
+        framework_version = None
+        _, _, image_tag, _ = fw.framework_name_from_image(image_uri)
+
+        if image_tag is not None:
+            framework_version = fw.framework_version_from_tag(image_tag)
+
+        if framework_version is not None:
+            _check_framework_profile_deprecation_warning(
+                framework_version, pytorch._framework_name, warn_list
+            )
+
+
+def test_create_tf_estimator_with_framework_profile_212(
+    sagemaker_session,
+    default_framework_profile,
+):
+    profiler_config = ProfilerConfig(framework_profile_params=default_framework_profile)
+
+    with warnings.catch_warnings(record=True) as warn_list:
+        warnings.simplefilter("always")
+        framework_version = "2.12"
+        tf = TensorFlow(
+            entry_point=SCRIPT_PATH,
+            role=ROLE,
+            framework_version=framework_version,
+            py_version="py39",
+            sagemaker_session=sagemaker_session,
+            instance_count=INSTANCE_COUNT,
+            instance_type=INSTANCE_TYPE,
+            profiler_config=profiler_config,
+        )
+
+        _check_framework_profile_deprecation_warning(
+            framework_version, tf._framework_name, warn_list
+        )
+
+
+def test_create_tf_estimator_w_image_with_framework_profile(
+    sagemaker_session,
+    tensorflow_inference_version,
+    tensorflow_inference_py_version,
+    default_framework_profile,
+):
+    image_uri = image_uris.retrieve(
+        "tensorflow",
+        REGION,
+        version=tensorflow_inference_version,
+        py_version=tensorflow_inference_py_version,
+        instance_type=INSTANCE_TYPE,
+        image_scope="inference",
+    )
+
+    profiler_config = ProfilerConfig(framework_profile_params=default_framework_profile)
+
+    with warnings.catch_warnings(record=True) as warn_list:
+        warnings.simplefilter("always")
+        tf = TensorFlow(
+            entry_point=SCRIPT_PATH,
+            role=ROLE,
+            sagemaker_session=sagemaker_session,
+            instance_count=INSTANCE_COUNT,
+            instance_type=INSTANCE_TYPE,
+            image_uri=image_uri,
+            profiler_config=profiler_config,
+        )
+
+        framework_version = None
+        _, _, image_tag, _ = fw.framework_name_from_image(image_uri)
+
+        if image_tag is not None:
+            framework_version = fw.framework_version_from_tag(image_tag)
+
+        if framework_version is not None:
+            _check_framework_profile_deprecation_warning(
+                framework_version, tf._framework_name, warn_list
+            )