Skip to content

Commit d48e989

Browse files
qidewenwhenDewen Qidanabensyzhu0
authored and
claytonparnell
committed
feature: Add SageMaker Experiment (aws#3536)
* feature: Add experiment plus Run class (aws#691) * feature: Add Experiment helper classes (aws#646) * feature: Add Experiment helper classes feature: Add helper class _RunEnvironment * change: Change sleep retry to backoff retry for get TC * minor fixes in backoff retry Co-authored-by: Dewen Qi <[email protected]> * feature: Add helper classes and methods for Run class (aws#660) * feature: Add helper classes and methods for Run class * Add Parent class to address comment * fix docstyle check * Add arg docstrings in _helper Co-authored-by: Dewen Qi <[email protected]> * feature: Add Experiment Run class (aws#651) Co-authored-by: Dewen Qi <[email protected]> * change: Add integ tests for Run (aws#673) Co-authored-by: Dewen Qi <[email protected]> * Update run log metric to use MetricsManager (aws#678) * Update run.log_metric to use _MetricsManager * fix several metrics issues * Add doc strings to metrics.py Co-authored-by: Dana Benson <[email protected]> Co-authored-by: Dana Benson <[email protected]> Co-authored-by: Dewen Qi <[email protected]> Co-authored-by: Dewen Qi <[email protected]> Co-authored-by: Dana Benson <[email protected]> Co-authored-by: Dana Benson <[email protected]> * change: Simplify exp plus integ test configuration (aws#694) Co-authored-by: Dewen Qi <[email protected]> * feature: add RunName to expeirment_config (aws#696) * change: Update Run init and add Run load and _RunContext (aws#707) * change: Update Run init and add Run load Add exp name and run group name to load and address comments * Address nit comments Co-authored-by: Dewen Qi <[email protected]> * fix: Fix run name uniqueness issue (aws#730) Co-authored-by: Dewen Qi <[email protected]> * change: Update integ tests for Exp Plus M1 changes (aws#741) Co-authored-by: Dewen Qi <[email protected]> * add metrics client to session object (aws#745) Co-authored-by: Dewen Qi <[email protected]> Co-authored-by: Dana Benson <[email protected]> Co-authored-by: Dana Benson <[email protected]> Co-authored-by: qidewenwhen <[email protected]> * change: Add integ test for using Run in Transform Job (aws#749) Co-authored-by: Dewen Qi <[email protected]> * Add async metrics sink (aws#739) Co-authored-by: Dewen Qi <[email protected]> Co-authored-by: Dana Benson <[email protected]> Co-authored-by: Dana Benson <[email protected]> Co-authored-by: qidewenwhen <[email protected]> * use metrics client provided by session (aws#754) * fix flaky metrics test (aws#753) * change: Change Run.init and Run.load to constructor and module method respectively (aws#752) Co-authored-by: Dewen Qi <[email protected]> * feature: Add latest metric service model (aws#757) Co-authored-by: Dewen Qi <[email protected]> Co-authored-by: qidewenwhen <[email protected]> * fix: lowercase run name (aws#767) * Change: Minimize use of lower case tc name (aws#769) * change: Clean up test resources to remove model files (aws#756) * change: Clean up test resources to remove model files * fix: Change experiment enums to upper case * change: Upgrade boto3 and update test to validate mixed case name * fix: Update as per latest botocore release and backend change Co-authored-by: Dewen Qi <[email protected]> * lowercase trial component name (aws#776) * change: Expose sagemaker experiment doc strings * fix: Fix exp name mixed case in issue Co-authored-by: Dewen Qi <[email protected]> Co-authored-by: Dana Benson <[email protected]> Co-authored-by: Dana Benson <[email protected]> Co-authored-by: Yifei Zhu <[email protected]>
1 parent 1115423 commit d48e989

File tree

82 files changed

+7894
-263
lines changed

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

82 files changed

+7894
-263
lines changed

‎.gitignore‎

Lines changed: 3 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -30,5 +30,6 @@ env/
3030
.vscode/
3131
**/tmp
3232
.python-version
33-
**/_repack_model.py
34-
**/_repack_script_launcher.sh
33+
**/_repack_script_launcher.sh
34+
tests/data/**/_repack_model.py
35+
tests/data/experiment/sagemaker-dev-1.0.tar.gz

‎doc/experiments/index.rst‎

Lines changed: 10 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,10 @@
1+
############################
2+
Amazon SageMaker Experiments
3+
############################
4+
5+
The SageMaker Python SDK supports to track and organize your machine learning workflow across SageMaker with jobs, such as Processing, Training and Transform, or locally.
6+
7+
.. toctree::
8+
:maxdepth: 2
9+
10+
sagemaker.experiments

‎doc/experiments/sagemaker.experiments.rst‎

Lines changed: 20 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,20 @@
1+
Experiments
2+
============
3+
4+
Run
5+
-------------
6+
7+
.. autoclass:: sagemaker.experiments.Run
8+
:members:
9+
10+
.. automethod:: sagemaker.experiments.load_run
11+
12+
.. automethod:: sagemaker.experiments.list_runs
13+
14+
.. autoclass:: sagemaker.experiments.SortByType
15+
:members:
16+
:undoc-members:
17+
18+
.. autoclass:: sagemaker.experiments.SortOrderType
19+
:members:
20+
:undoc-members:

‎doc/index.rst‎

Lines changed: 10 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -60,6 +60,16 @@ Orchestrate your SageMaker training and inference workflows with Airflow and Kub
6060
workflows/index
6161

6262

63+
****************************
64+
Amazon SageMaker Experiments
65+
****************************
66+
You can use Amazon SageMaker Experiments to track machine learning experiments.
67+
68+
.. toctree::
69+
:maxdepth: 2
70+
71+
experiments/index
72+
6373
*************************
6474
Amazon SageMaker Debugger
6575
*************************

‎requirements/extras/test_requirements.txt‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -20,3 +20,4 @@ requests==2.27.1
2020
sagemaker-experiments==0.1.35
2121
Jinja2==3.0.3
2222
pandas>=1.3.5,<1.5
23+
scikit-learn==1.0.2

‎setup.py‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -48,7 +48,7 @@ def read_requirements(filename):
4848
# Declare minimal set for installation
4949
required_packages = [
5050
"attrs>=20.3.0,<23",
51-
"boto3>=1.26.20,<2.0",
51+
"boto3>=1.26.28,<2.0",
5252
"google-pasta",
5353
"numpy>=1.9.0,<2.0",
5454
"protobuf>=3.1,<4.0",

‎src/sagemaker/amazon/amazon_estimator.py‎

Lines changed: 4 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -27,7 +27,7 @@
2727
from sagemaker.deprecations import renamed_warning
2828
from sagemaker.estimator import EstimatorBase, _TrainingJob
2929
from sagemaker.inputs import FileSystemInput, TrainingInput
30-
from sagemaker.utils import sagemaker_timestamp
30+
from sagemaker.utils import sagemaker_timestamp, check_and_get_run_experiment_config
3131
from sagemaker.workflow.entities import PipelineVariable
3232
from sagemaker.workflow.pipeline_context import runnable_by_pipeline
3333
from sagemaker.workflow import is_pipeline_variable
@@ -242,8 +242,8 @@ def fit(
242242
generates a default job name, based on the training image name
243243
and current timestamp.
244244
experiment_config (dict[str, str]): Experiment management configuration.
245-
Optionally, the dict can contain three keys:
246-
'ExperimentName', 'TrialName', and 'TrialComponentDisplayName'.
245+
Optionally, the dict can contain four keys:
246+
'ExperimentName', 'TrialName', 'TrialComponentDisplayName' and 'RunName'.
247247
The behavior of setting these keys is as follows:
248248
* If `ExperimentName` is supplied but `TrialName` is not a Trial will be
249249
automatically created and the job's Trial Component associated with the Trial.
@@ -255,6 +255,7 @@ def fit(
255255
"""
256256
self._prepare_for_training(records, job_name=job_name, mini_batch_size=mini_batch_size)
257257

258+
experiment_config = check_and_get_run_experiment_config(experiment_config)
258259
self.latest_training_job = _TrainingJob.start_new(
259260
self, records, experiment_config=experiment_config
260261
)

‎src/sagemaker/apiutils/_base_types.py‎

Lines changed: 4 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -173,8 +173,10 @@ def _search(
173173
search_items = search_method_response.get("Results", [])
174174
next_token = search_method_response.get(boto_next_token_name)
175175
for item in search_items:
176-
if cls.__name__ in item:
177-
yield search_item_factory(item[cls.__name__])
176+
# _TrialComponent class in experiments module is not public currently
177+
class_name = cls.__name__.lstrip("_")
178+
if class_name in item:
179+
yield search_item_factory(item[class_name])
178180
if not next_token:
179181
break
180182
except StopIteration:

‎src/sagemaker/apiutils/_boto_functions.py‎

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -68,7 +68,9 @@ def from_boto(boto_dict, boto_name_to_member_name, member_name_to_type):
6868
api_type, is_collection = member_name_to_type[member_name]
6969
if is_collection:
7070
if isinstance(boto_value, dict):
71-
member_value = api_type.from_boto(boto_value)
71+
member_value = {
72+
key: api_type.from_boto(value) for key, value in boto_value.items()
73+
}
7274
else:
7375
member_value = [api_type.from_boto(item) for item in boto_value]
7476
else:

‎src/sagemaker/dataset_definition/inputs.py‎

Lines changed: 4 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -124,8 +124,10 @@ class DatasetDefinition(ApiObject):
124124
"""DatasetDefinition input."""
125125

126126
_custom_boto_types = {
127-
"redshift_dataset_definition": (RedshiftDatasetDefinition, True),
128-
"athena_dataset_definition": (AthenaDatasetDefinition, True),
127+
# RedshiftDatasetDefinition and AthenaDatasetDefinition are not collection
128+
# Instead they are singleton objects. Thus, set the is_collection flag to False.
129+
"redshift_dataset_definition": (RedshiftDatasetDefinition, False),
130+
"athena_dataset_definition": (AthenaDatasetDefinition, False),
129131
}
130132

131133
def __init__(

0 commit comments

Comments
 (0)