Add Chainer 4.1.0

CHANGELOG.rst

@@ -2,6 +2,12 @@
 CHANGELOG
 =========
 
+1.6.0
+=====
+
+* feature: Add Chainer 4.1.0 support
+* bug-fix: Use chainer_full_version fixture in Chainer integration tests
+
 1.5.4
 =====
 

README.rst

@@ -50,7 +50,7 @@ You can install from source by cloning this repository and issuing a pip install
 
     git clone https://github.com/aws/sagemaker-python-sdk.git
     python setup.py sdist
-    pip install dist/sagemaker-1.5.4.tar.gz
+    pip install dist/sagemaker-1.6.0.tar.gz
 
 Supported Python versions
 ~~~~~~~~~~~~~~~~~~~~~~~~~
@@ -236,7 +236,7 @@ Chainer SageMaker Estimators
 
 With Chainer Estimators, you can train and host Chainer models on Amazon SageMaker.
 
-Supported versions of Chainer: ``4.0.0``
+Supported versions of Chainer: ``4.0.0``, ``4.1.0``.
 
 You can visit the Chainer repository at https://github.com/chainer/chainer.
 

setup.py

@@ -23,7 +23,7 @@ def read(fname):
 
 
 setup(name="sagemaker",
-      version="1.5.4",
+      version="1.6.0",
       description="Open source library for training and deploying models on Amazon SageMaker.",
       packages=find_packages('src'),
       package_dir={'': 'src'},

src/sagemaker/chainer/defaults.py

@@ -12,4 +12,4 @@
 # language governing permissions and limitations under the License.
 from __future__ import absolute_import
 
-CHAINER_VERSION = '4.0.0'
+CHAINER_VERSION = '4.1.0'

tests/conftest.py

@@ -86,7 +86,7 @@ def pytorch_version(request):
     return request.param
 
 
-@pytest.fixture(scope='module', params=['4.0', '4.0.0'])
+@pytest.fixture(scope='module', params=['4.0', '4.0.0', '4.1', '4.1.0'])
 def chainer_version(request):
     return request.param
 
@@ -106,6 +106,6 @@ def pytorch_full_version(request):
     return request.param
 
 
-@pytest.fixture(scope='module', params=['4.0.0'])
+@pytest.fixture(scope='module', params=['4.0.0', '4.1.0'])
 def chainer_full_version(request):
     return request.param

tests/integ/test_chainer_train.py

@@ -18,6 +18,7 @@
 import pytest
 import numpy
 
+from sagemaker.chainer.defaults import CHAINER_VERSION
 from sagemaker.chainer.estimator import Chainer
 from sagemaker.chainer.model import ChainerModel
 from sagemaker.utils import sagemaker_timestamp
@@ -26,25 +27,26 @@
 
 
 @pytest.fixture(scope='module')
-def chainer_training_job(sagemaker_session):
-    return _run_mnist_training_job(sagemaker_session, "ml.c4.xlarge", 1)
+def chainer_training_job(sagemaker_session, chainer_full_version):
+    return _run_mnist_training_job(sagemaker_session, "ml.c4.xlarge", 1, chainer_full_version)
 
 
-def test_distributed_cpu_training(sagemaker_session):
-    _run_mnist_training_job(sagemaker_session, "ml.c4.xlarge", 2)
+def test_distributed_cpu_training(sagemaker_session, chainer_full_version):
+    _run_mnist_training_job(sagemaker_session, "ml.c4.xlarge", 2, chainer_full_version)
 
 
-def test_distributed_gpu_training(sagemaker_session):
-    _run_mnist_training_job(sagemaker_session, "ml.p2.xlarge", 2)
+def test_distributed_gpu_training(sagemaker_session, chainer_full_version):
+    _run_mnist_training_job(sagemaker_session, "ml.p2.xlarge", 2, chainer_full_version)
 
 
-def test_training_with_additional_hyperparameters(sagemaker_session):
+def test_training_with_additional_hyperparameters(sagemaker_session, chainer_full_version):
     with timeout(minutes=15):
         script_path = os.path.join(DATA_DIR, 'chainer_mnist', 'mnist.py')
         data_path = os.path.join(DATA_DIR, 'chainer_mnist')
 
         chainer = Chainer(entry_point=script_path, role='SageMakerRole',
                           train_instance_count=1, train_instance_type="ml.c4.xlarge",
+                          framework_version=chainer_full_version,
                           sagemaker_session=sagemaker_session, hyperparameters={'epochs': 1},
                           use_mpi=True,
                           num_processes=2,
@@ -75,8 +77,7 @@ def test_deploy_model(chainer_training_job, sagemaker_session):
         desc = sagemaker_session.sagemaker_client.describe_training_job(TrainingJobName=chainer_training_job)
         model_data = desc['ModelArtifacts']['S3ModelArtifacts']
         script_path = os.path.join(DATA_DIR, 'chainer_mnist', 'mnist.py')
-        model = ChainerModel(model_data, 'SageMakerRole', entry_point=script_path,
-                             sagemaker_session=sagemaker_session)
+        model = ChainerModel(model_data, 'SageMakerRole', entry_point=script_path, sagemaker_session=sagemaker_session)
         predictor = model.deploy(1, "ml.m4.xlarge", endpoint_name=endpoint_name)
         _predict_and_assert(predictor)
 
@@ -85,7 +86,8 @@ def test_async_fit(sagemaker_session):
     endpoint_name = 'test-chainer-attach-deploy-{}'.format(sagemaker_timestamp())
 
     with timeout(minutes=5):
-        training_job_name = _run_mnist_training_job(sagemaker_session, "ml.c4.xlarge", 1, wait=False)
+        training_job_name = _run_mnist_training_job(sagemaker_session, "ml.c4.xlarge", 1,
+                                                    chainer_full_version=CHAINER_VERSION, wait=False)
 
         print("Waiting to re-attach to the training job: %s" % training_job_name)
         time.sleep(20)
@@ -97,12 +99,13 @@ def test_async_fit(sagemaker_session):
         _predict_and_assert(predictor)
 
 
-def test_failed_training_job(sagemaker_session):
+def test_failed_training_job(sagemaker_session, chainer_full_version):
     with timeout(minutes=15):
         script_path = os.path.join(DATA_DIR, 'chainer_mnist', 'failure_script.py')
         data_path = os.path.join(DATA_DIR, 'chainer_mnist')
 
         chainer = Chainer(entry_point=script_path, role='SageMakerRole',
+                          framework_version=chainer_full_version,
                           train_instance_count=1, train_instance_type='ml.c4.xlarge',
                           sagemaker_session=sagemaker_session)
 
@@ -113,7 +116,8 @@ def test_failed_training_job(sagemaker_session):
             chainer.fit(train_input)
 
 
-def _run_mnist_training_job(sagemaker_session, instance_type, instance_count, wait=True):
+def _run_mnist_training_job(sagemaker_session, instance_type, instance_count,
+                            chainer_full_version, wait=True):
     with timeout(minutes=15):
 
         script_path = os.path.join(DATA_DIR, 'chainer_mnist', 'mnist.py') if instance_type == 1 else \
@@ -122,6 +126,7 @@ def _run_mnist_training_job(sagemaker_session, instance_type, instance_count, wa
         data_path = os.path.join(DATA_DIR, 'chainer_mnist')
 
         chainer = Chainer(entry_point=script_path, role='SageMakerRole',
+                          framework_version=chainer_full_version,
                           train_instance_count=instance_count, train_instance_type=instance_type,
                           sagemaker_session=sagemaker_session, hyperparameters={'epochs': 1})