deploy-arm-templates api version update (#1458)

* updating api versions * addressing comments * black formatting Co-authored-by: Raghu Ramaswamy <raghura@microsoft.com>
2022-07-05 15:36:27 -07:00 · 2022-07-05 15:36:27 -07:00 · d75447d655
--- a/.github/kubernetes-compute/convert.py
+++ b/.github/kubernetes-compute/convert.py
@ -98,12 +98,7 @@ if __name__ == "__main__":
        required=False,
        help="Output AMLARC-compatible file, if not provides, " "replace file inplace",
    )
-    parser.add_argument(
-        "-c",
-        "--compute-target",
-        required=False,
-        help="Compute target",
-    )
+    parser.add_argument("-c", "--compute-target", required=False, help="Compute target")
    parser.add_argument("-it", "--instance-type", required=False, help="Instance type")
    parser.add_argument(
        "-cr",
--- a/.github/kubernetes-compute/papermill_count_failed_case.py
+++ b/.github/kubernetes-compute/papermill_count_failed_case.py
@ -41,12 +41,7 @@ def check_test_case(input_file):
 if __name__ == "__main__":
    # Parse command line arguments
    parser = argparse.ArgumentParser(description="Check all papermill output files.")
-    parser.add_argument(
-        "-i",
-        "--input",
-        required=True,
-        help="job ipynb file list",
-    )
+    parser.add_argument("-i", "--input", required=True, help="job ipynb file list")

    args = parser.parse_args()

--- a/.github/kubernetes-compute/sdk_testcase_collector.py
+++ b/.github/kubernetes-compute/sdk_testcase_collector.py
@ -34,10 +34,7 @@ if __name__ == "__main__":
    )
    parser.add_argument("-r", "--regex", required=True, help="test case name selector")
    parser.add_argument(
-        "-o",
-        "--output",
-        required=False,
-        help="the file selected test case send to",
+        "-o", "--output", required=False, help="the file selected test case send to"
    )

    args = parser.parse_args()
--- a/arm-templates/code-version.json
+++ b/arm-templates/code-version.json
@ -42,7 +42,7 @@
  "resources": [
    {
      "type": "Microsoft.MachineLearningServices/workspaces/codes/versions",
-      "apiVersion": "2021-10-01",
+      "apiVersion": "2022-05-01",
      "name": "[concat(parameters('workspaceName'), '/', parameters('codeAssetName'), '/', parameters('codeAssetVersion'))]",
      "properties": {
        "description": "[parameters('codeAssetDescription')]",
--- a/arm-templates/environment-version.json
+++ b/arm-templates/environment-version.json
@ -38,7 +38,7 @@
  "resources": [
    {
      "type": "Microsoft.MachineLearningServices/workspaces/environments/versions",
-      "apiVersion": "2021-10-01",
+      "apiVersion": "2022-05-01",
      "name": "[concat(parameters('workspaceName'), '/', parameters('environmentAssetName'), '/', parameters('environmentAssetVersion'))]",
      "properties": {
        "isAnonymous": "[parameters('isAnonymous')]",
--- a/arm-templates/model-version.json
+++ b/arm-templates/model-version.json
@ -39,7 +39,7 @@
    "resources": [
      {
        "type": "Microsoft.MachineLearningServices/workspaces/models/versions",
-        "apiVersion": "2021-10-01",
+        "apiVersion": "2022-05-01",
        "name": "[concat(parameters('workspaceName'), '/', parameters('modelAssetName'), '/', parameters('modelAssetVersion'))]",
        "properties": {
          "description": "[parameters('modelDescription')]",
--- a/arm-templates/online-endpoint-deployment.json
+++ b/arm-templates/online-endpoint-deployment.json
@ -82,7 +82,7 @@
  "resources": [
    {
      "type": "Microsoft.MachineLearningServices/workspaces/onlineEndpoints/deployments",
-      "apiVersion": "2021-10-01",
+      "apiVersion": "2022-05-01",
      "name": "[concat(parameters('workspaceName'), '/', parameters('onlineEndpointName'),'/', parameters('onlineDeploymentName'))]",
      "location": "[parameters('location')]",
      "tags": "[parameters('onlineDeploymentTags')]",
--- a/arm-templates/online-endpoint.json
+++ b/arm-templates/online-endpoint.json
@ -55,7 +55,7 @@
  "resources": [
    {
      "type": "Microsoft.MachineLearningServices/workspaces/onlineEndpoints",
-      "apiVersion": "2021-10-01",
+      "apiVersion": "2022-05-01",
      "name": "[concat(parameters('workspaceName'), '/', parameters('onlineEndpointName'))]",
      "location": "[parameters('location')]",
      "tags": "[parameters('onlineEndpointTags')]",
--- a/cli/jobs/automl-standalone-jobs/cli-automl-image-classification-multiclass-task-fridge-items/prepare_data.py
+++ b/cli/jobs/automl-standalone-jobs/cli-automl-image-classification-multiclass-task-fridge-items/prepare_data.py
@ -29,10 +29,7 @@ def create_jsonl_files(uri_folder_data_path):
    )

    # Baseline of json line dictionary
-    json_line_sample = {
-        "image_url": uri_folder_data_path,
-        "label": "",
-    }
+    json_line_sample = {"image_url": uri_folder_data_path, "label": ""}

    index = 0
    # Scan each sub directary and generate a jsonl line per image, distributed on train and valid JSONL files
--- a/cli/jobs/automl-standalone-jobs/cli-automl-image-classification-multilablel-task-fridge-items/prepare_data.py
+++ b/cli/jobs/automl-standalone-jobs/cli-automl-image-classification-multilablel-task-fridge-items/prepare_data.py
@ -29,10 +29,7 @@ def create_jsonl_files(uri_folder_data_path):
    )

    # Baseline of json line dictionary
-    json_line_sample = {
-        "image_url": uri_folder_data_path,
-        "label": [],
-    }
+    json_line_sample = {"image_url": uri_folder_data_path, "label": []}

    # Path to the labels file.
    labelFile = os.path.join(src_images, "labels.csv")
--- a/cli/jobs/pipelines-with-components/image_classification_with_densenet/image_cnn_train/image_classification/dataloaders.py
+++ b/cli/jobs/pipelines-with-components/image_classification_with_densenet/image_cnn_train/image_classification/dataloaders.py
@ -235,8 +235,9 @@ def get_dali_train_loader(dali_cpu=False):
            pipe, size=int(pipe.epoch_size("Reader") / world_size)
        )

-        return DALIWrapper(train_loader, num_classes, one_hot), int(
-            pipe.epoch_size("Reader") / (world_size * batch_size)
+        return (
+            DALIWrapper(train_loader, num_classes, one_hot),
+            int(pipe.epoch_size("Reader") / (world_size * batch_size)),
        )

    return gdtl
@ -275,8 +276,9 @@ def get_dali_val_loader():
            pipe, size=int(pipe.epoch_size("Reader") / world_size)
        )

-        return DALIWrapper(val_loader, num_classes, one_hot), int(
-            pipe.epoch_size("Reader") / (world_size * batch_size)
+        return (
+            DALIWrapper(val_loader, num_classes, one_hot),
+            int(pipe.epoch_size("Reader") / (world_size * batch_size)),
        )

    return gdvl
@ -385,10 +387,7 @@ def get_pytorch_train_loader(
    train_dataset = datasets.ImageFolder(
        traindir,
        transforms.Compose(
-            [
-                transforms.RandomResizedCrop(224),
-                transforms.RandomHorizontalFlip(),
-            ]
+            [transforms.RandomResizedCrop(224), transforms.RandomHorizontalFlip()]
        ),
    )

@ -409,8 +408,9 @@ def get_pytorch_train_loader(
        drop_last=True,
    )

-    return PrefetchedWrapper(train_loader, num_classes, fp16, one_hot), len(
-        train_loader
+    return (
+        PrefetchedWrapper(train_loader, num_classes, fp16, one_hot),
+        len(train_loader),
    )


@ -425,13 +425,7 @@ def get_pytorch_val_loader(
 ):
    valdir = os.path.join(data_path, "val")
    val_dataset = datasets.ImageFolder(
-        valdir,
-        transforms.Compose(
-            [
-                transforms.Resize(256),
-                transforms.CenterCrop(224),
-            ]
-        ),
+        valdir, transforms.Compose([transforms.Resize(256), transforms.CenterCrop(224)])
    )

    if torch.distributed.is_initialized():
--- a/cli/jobs/pipelines-with-components/image_classification_with_densenet/image_cnn_train/image_classification/training.py
+++ b/cli/jobs/pipelines-with-components/image_classification_with_densenet/image_cnn_train/image_classification/training.py
@ -257,7 +257,7 @@ def lr_exponential_policy(
            lr = base_lr * (epoch + 1) / warmup_length
        else:
            e = epoch - warmup_length
-            lr = base_lr * (epoch_decay**e)
+            lr = base_lr * (epoch_decay ** e)
        return lr

    return lr_policy(_lr_fn, logger=logger)
--- a/cli/jobs/pipelines-with-components/nyc_taxi_data_regression/prep_src/prep.py
+++ b/cli/jobs/pipelines-with-components/nyc_taxi_data_regression/prep_src/prep.py
@ -17,10 +17,7 @@ args = parser.parse_args()

 print("hello training world...")

-lines = [
-    f"Raw data path: {args.raw_data}",
-    f"Data output path: {args.prep_data}",
-]
+lines = [f"Raw data path: {args.raw_data}", f"Data output path: {args.prep_data}"]

 for line in lines:
    print(line)
--- a/cli/jobs/pipelines-with-components/rai_pipeline_adult_analyse/components/rai_analyse/gather_rai_insights.py
+++ b/cli/jobs/pipelines-with-components/rai_pipeline_adult_analyse/components/rai_analyse/gather_rai_insights.py
@ -72,12 +72,7 @@ def main(args):
        create_rai_tool_directories(incoming_dir)
        _logger.info("Saved empty RAI Insights input to temporary directory")

-        insight_paths = [
-            args.insight_1,
-            args.insight_2,
-            args.insight_3,
-            args.insight_4,
-        ]
+        insight_paths = [args.insight_1, args.insight_2, args.insight_3, args.insight_4]

        included_tools: Dict[str, bool] = {
            RAIToolType.CAUSAL: False,
--- a/cli/jobs/pipelines-with-components/rai_pipeline_adult_analyse/components/rai_analyse/rai_component_utilities.py
+++ b/cli/jobs/pipelines-with-components/rai_pipeline_adult_analyse/components/rai_analyse/rai_component_utilities.py
@ -160,10 +160,7 @@ def copy_insight_to_raiinsights(

    src_dir = insight_dir / tool_dir_name / tool_dir_items[0].parts[-1]
    dst_dir = rai_insights_dir / tool_dir_name / tool_dir_items[0].parts[-1]
-    shutil.copytree(
-        src=src_dir,
-        dst=dst_dir,
-    )
+    shutil.copytree(src=src_dir, dst=dst_dir)

    _logger.info("Copy complete")
    return tool_type
@ -186,9 +183,7 @@ def save_to_output_port(rai_i: RAIInsights, output_port_path: str, tool_type: st

        _logger.info("Starting copy")
        shutil.copytree(
-            pathlib.Path(tmpdirname) / tool_dir_name,
-            target_path,
-            dirs_exist_ok=True,
+            pathlib.Path(tmpdirname) / tool_dir_name, target_path, dirs_exist_ok=True
        )
    _logger.info("Copied to output")

--- a/cli/jobs/pipelines/nyc-taxi/src/prep/prep.py
+++ b/cli/jobs/pipelines/nyc-taxi/src/prep/prep.py
@ -17,10 +17,7 @@ args = parser.parse_args()

 print("hello training world...")

-lines = [
-    f"Raw data path: {args.raw_data}",
-    f"Data output path: {args.prep_data}",
-]
+lines = [f"Raw data path: {args.raw_data}", f"Data output path: {args.prep_data}"]

 for line in lines:
    print(line)
--- a/cli/jobs/single-step/dask/nyctaxi/src/prep-nyctaxi.py
+++ b/cli/jobs/single-step/dask/nyctaxi/src/prep-nyctaxi.py
@ -31,8 +31,7 @@ mlflow.log_text(str(c), "dask_cluster1")
 # read in the data from the provided file dataset (which is mounted at the same
 # location on all nodes of the job)
 df = dd.read_csv(
-    f"{dataset}/*.csv",
-    parse_dates=["tpep_pickup_datetime", "tpep_dropoff_datetime"],
+    f"{dataset}/*.csv", parse_dates=["tpep_pickup_datetime", "tpep_dropoff_datetime"]
 )

 # as an alternative, the below would be using abfs
--- a/cli/readme.py
+++ b/cli/readme.py
@ -7,17 +7,8 @@ import argparse
 # define constants
 EXCLUDED_JOBS = ["java"]
 EXCLUDED_ENDPOINTS = ["batch", "online", "amlarc"]
-EXCLUDED_RESOURCES = [
-    "workspace",
-    "datastore",
-    "vm-attach",
-    "instance",
-    "connections",
-]
-EXCLUDED_ASSETS = [
-    "conda-yamls",
-    "mlflow-models",
-]
+EXCLUDED_RESOURCES = ["workspace", "datastore", "vm-attach", "instance", "connections"]
+EXCLUDED_ASSETS = ["conda-yamls", "mlflow-models"]
 EXCLUDED_SCRIPTS = ["setup", "cleanup", "run-job"]
 BRANCH = "main"  # default - do not change
 # BRANCH = "sdk-preview"  # this should be deleted when this branch is merged to main
--- a/deploy-arm-templates-az-cli.sh
+++ b/deploy-arm-templates-az-cli.sh
@ -6,9 +6,9 @@ TOKEN=$(az account get-access-token --query accessToken -o tsv)

 # <create_variables>
 SUBSCRIPTION_ID=$(az account show --query id | tr -d '\r"')
+LOCATION=$(az ml workspace show --query location | tr -d '\r"')
 RESOURCE_GROUP=$(az group show --query name | tr -d '\r"')
 WORKSPACE=$(az configure -l | jq -r '.[] | select(.name=="workspace") | .value')
-LOCATION=$(az ml workspace show| jq -r '.location')
 #</create_variables>

 # <set_endpoint_name>
@ -16,7 +16,7 @@ export ENDPOINT_NAME=endpoint-`echo $RANDOM`
 # </set_endpoint_name>

 #<api_version>
-API_VERSION="2021-10-01"
+API_VERSION="2022-05-01"
 #</api_version>

 echo -e "Using:\nSUBSCRIPTION_ID=$SUBSCRIPTION_ID\nLOCATION=$LOCATION\nRESOURCE_GROUP=$RESOURCE_GROUP\nWORKSPACE=$WORKSPACE"
@ -58,7 +58,7 @@ export AZURE_STORAGE_ACCOUNT=$(echo $response | jq -r '.value[0].properties.acco
 # </get_storage_details>

 # <upload_code>
-az storage blob upload-batch -d $AZUREML_DEFAULT_CONTAINER/score -s endpoints/online/model-1/onlinescoring --account-name $AZURE_STORAGE_ACCOUNT
+az storage blob upload-batch -d $AZUREML_DEFAULT_CONTAINER/score -s cli/endpoints/online/model-1/onlinescoring --account-name $AZURE_STORAGE_ACCOUNT
 # </upload_code>

 # <create_code>
@ -71,7 +71,7 @@ codeUri="https://$AZURE_STORAGE_ACCOUNT.blob.core.windows.net/$AZUREML_DEFAULT_C
 # </create_code>

 # <upload_model>
-az storage blob upload-batch -d $AZUREML_DEFAULT_CONTAINER/model -s endpoints/online/model-1/model --account-name $AZURE_STORAGE_ACCOUNT
+az storage blob upload-batch -d $AZUREML_DEFAULT_CONTAINER/model -s cli/endpoints/online/model-1/model --account-name $AZURE_STORAGE_ACCOUNT
 # </upload_model>

 # <create_model>
@ -84,7 +84,7 @@ modelUri="azureml://subscriptions/$SUBSCRIPTION_ID/resourceGroups/$RESOURCE_GROU
 # </create_model>

 # <read_condafile>
-CONDA_FILE=$(cat endpoints/online/model-1/environment/conda.yml)
+CONDA_FILE=$(cat cli/endpoints/online/model-1/environment/conda.yml)
 # <read_condafile>

 # <create_environment>
@ -131,7 +131,7 @@ az deployment group create -g $RESOURCE_GROUP \
 codeId="$resourceScope/workspaces/$WORKSPACE/codes/score-sklearn/versions/1" \
 scoringScript=score.py \
 environmentId="$resourceScope/workspaces/$WORKSPACE/environments/sklearn-env/versions/$ENV_VERSION" \
- model="$resourceScope/workspaces/$WORKSPACE/models/sklearn/versions/1" \
+ model="$resourceScope/workspaces/$WORKSPACE/models/score-sklearn/versions/1" \
 endpointComputeType=Managed \
 skuName=Standard_F2s_v2 \
 skuCapacity=1
@ -158,7 +158,7 @@ accessToken=$(echo $response | jq -r '.accessToken')
 curl --location --request POST $scoringUri \
 --header "Authorization: Bearer $accessToken" \
 --header "Content-Type: application/json" \
--data-raw @endpoints/online/model-1/sample-request.json
+--data-raw @cli/endpoints/online/model-1/sample-request.json
 # </score_endpoint>

 # <get_deployment_logs>
--- a/deploy-arm-templates-rest.sh
+++ b/deploy-arm-templates-rest.sh
@ -11,7 +11,7 @@ schema='$schema'
 echo -e "Using:\nSUBSCRIPTION_ID=$SUBSCRIPTION_ID\nLOCATION=$LOCATION\nRESOURCE_GROUP=$RESOURCE_GROUP\nWORKSPACE=$WORKSPACE"

 # <read_condafile>
-CONDA_FILE=$(< model/environment/conda.yml)
+CONDA_FILE=$(< cli/endpoints/online/model-1/environment/conda.yml)
 # <read_condafile>

 #<get_access_token>
@ -23,7 +23,7 @@ export ENDPOINT_NAME=endpt-`echo $RANDOM`
 # </set_endpoint_name>

 #<api_version>
-API_VERSION="2021-10-01"
+API_VERSION="2022-05-01"
 #</api_version>

 # define how to wait
@ -62,7 +62,7 @@ export AZURE_STORAGE_ACCOUNT=$(echo $response | jq -r '.value[0].properties.acco
 # </get_storage_details>

 # <upload_code>
-az storage blob upload-batch -d $AZUREML_DEFAULT_CONTAINER/score -s endpoints/online/model-1/onlinescoring --account-name $AZURE_STORAGE_ACCOUNT
+az storage blob upload-batch -d $AZUREML_DEFAULT_CONTAINER/score -s cli/endpoints/online/model-1/onlinescoring --account-name $AZURE_STORAGE_ACCOUNT
 # </upload_code>

 # <create_code>
@ -117,7 +117,7 @@ curl --location --request PUT "https://management.azure.com/subscriptions/$SUBSC
        [
            {          
                \"type\": \"Microsoft.MachineLearningServices/workspaces/codes/versions\",
-                \"apiVersion\": \"2021-10-01\",
+                \"apiVersion\": \"$API_VERSION\",
                \"name\": \"[concat(parameters(\'workspaceName\'), \'/\', parameters(\'codeAssetName\'), \'/\', parameters(\'codeAssetVersion\'))]\",
                \"properties\": {
                    \"description\": \"[parameters(\'codeAssetDescription\')]\",
@ -148,7 +148,7 @@ curl --location --request PUT "https://management.azure.com/subscriptions/$SUBSC
 # <\create_code>

 # <upload_model>
-az storage blob upload-batch -d $AZUREML_DEFAULT_CONTAINER/model -s endpoints/online/model-1/model --account-name $AZURE_STORAGE_ACCOUNT
+az storage blob upload-batch -d $AZUREML_DEFAULT_CONTAINER/model -s cli/endpoints/online/model-1/model --account-name $AZURE_STORAGE_ACCOUNT
 # <\upload_model>

 # <create_model>
@ -234,7 +234,7 @@ curl --location --request PUT "https://management.azure.com/subscriptions/$SUBSC
 # <\create_model>

 # <read_condafile>
-CONDA_FILE=$(cat endpoints/online/model-1/environment/conda.yml)
+CONDA_FILE=$(cat cli/endpoints/online/model-1/environment/conda.yml)
 # <read_condafile>

 # <create_environment>
@ -296,7 +296,7 @@ curl --location --request PUT "https://management.azure.com/subscriptions/$SUBSC
        [
            {          
                \"type\": \"Microsoft.MachineLearningServices/workspaces/environments/versions\",
-                \"apiVersion\": \"2021-10-01\",
+                \"apiVersion\": \"$API_VERSION\",
                \"name\": \"[concat(parameters(\'workspaceName\'), \'/\', parameters(\'environmentAssetName\'), \'/\', parameters(\'environmentAssetVersion\'))]\",
                \"properties\": {
                    \"description\": \"[parameters(\'environmentAssetDescription\')]\",
@ -392,7 +392,7 @@ curl --location --request PUT "https://management.azure.com/subscriptions/$SUBSC
            [
              {
                \"type\": \"Microsoft.MachineLearningServices/workspaces/onlineEndpoints\",
-                \"apiVersion\": \"2021-10-01\",
+                \"apiVersion\": \"$API_VERSION\",
                \"name\": \"[concat(parameters(\'workspaceName\'), \'/\', parameters(\'onlineEndpointName\'))]\",
                \"location\": \"[parameters(\'location\')]\",
                \"tags\": \"[parameters(\'onlineEndpointTags\')]\",
@ -527,7 +527,7 @@ curl --location --request PUT "https://management.azure.com/subscriptions/$SUBSC
        \"resources\": [
          {
            \"type\": \"Microsoft.MachineLearningServices/workspaces/onlineEndpoints/deployments\",
-            \"apiVersion\": \"2021-10-01\",
+            \"apiVersion\": \"$API_VERSION\",
            \"name\": \"[concat(parameters(\'workspaceName\'), \'/\', parameters(\'onlineEndpointName\'), \'/\', parameters(\'onlineDeploymentName\'))]\",
            \"location\": \"[parameters(\'location\')]\",
            \"tags\": \"[parameters(\'onlineDeploymentTags\')]\",
@ -572,7 +572,7 @@ curl --location --request PUT "https://management.azure.com/subscriptions/$SUBSC
            \"value\": \"$resourceScope/workspaces/$WORKSPACE/environments/sklearn-env/versions/$ENV_VERSION\"
        },
        \"model\": {
-            \"value\": \"$resourceScope/workspaces/$WORKSPACE/models/sklearn/versions/1\"
+            \"value\": \"$resourceScope/workspaces/$WORKSPACE/models/score-sklearn/versions/1\"
        },
        \"endpointComputeType\": {
            \"value\": \"Managed\"
@ -610,7 +610,7 @@ accessToken=$(echo $response | jq -r '.accessToken')
 curl --location --request POST $scoringUri \
 --header "Authorization: Bearer $accessToken" \
 --header "Content-Type: application/json" \
--data @endpoints/online/model-1/sample-request.json
+--data @cli/endpoints/online/model-1/sample-request.json
 # </score_endpoint>

 # <get_deployment_logs>
--- a/notebooks/using-mlflow/no-code-deployment/score.py
+++ b/notebooks/using-mlflow/no-code-deployment/score.py
@ -27,15 +27,11 @@ input_sample = pd.DataFrame(
            "slope": 3,
            "ca": 0,
            "thal": 2,
-        },
+        }
    ]
 )

-output_sample = np.ndarray(
-    [
-        1,
-    ]
-)
+output_sample = np.ndarray([1])

 MODEL = None

--- a/notebooks/using-mlflow/samples/mnist-pytorch/src/trainer.py
+++ b/notebooks/using-mlflow/samples/mnist-pytorch/src/trainer.py
@ -100,16 +100,8 @@ def driver():
        test(args, model, device, test_loader)

    if args.save_model:
-        input_schema = Schema(
-            [
-                TensorSpec(np.dtype(np.uint8), (-1, 28, 28)),
-            ]
-        )
-        output_schema = Schema(
-            [
-                TensorSpec(np.dtype(np.float32), (-1, 10)),
-            ]
-        )
+        input_schema = Schema([TensorSpec(np.dtype(np.uint8), (-1, 28, 28))])
+        output_schema = Schema([TensorSpec(np.dtype(np.float32), (-1, 10))])
        signature = ModelSignature(inputs=input_schema, outputs=output_schema)
        mlflow.pytorch.log_model(model, "model", signature=signature)

--- a/python-sdk/experimental/deploy-triton/src/test_service.py
+++ b/python-sdk/experimental/deploy-triton/src/test_service.py
@ -40,10 +40,7 @@ if __name__ == "__main__":
    }

    file_name = os.path.join(
-        os.path.abspath(os.path.dirname(__file__)),
-        "..",
-        "data",
-        args.data_file,
+        os.path.abspath(os.path.dirname(__file__)), "..", "data", args.data_file
    )
    test_sample = open(file_name, "rb").read()
    resp = requests.post(aks_service.scoring_uri, test_sample, headers=headers)
--- a/python-sdk/experimental/deploy-triton/src/tritonhttpclient.py
+++ b/python-sdk/experimental/deploy-triton/src/tritonhttpclient.py
@ -756,11 +756,7 @@ class InferenceServerClient:
        """
        request_uri = "v2/systemsharedmemory/region/{}/register".format(quote(name))

-        register_request = {
-            "key": key,
-            "offset": offset,
-            "byte_size": byte_size,
-        }
+        register_request = {"key": key, "offset": offset, "byte_size": byte_size}
        request_body = json.dumps(register_request)

        response = self._post(
@ -864,13 +860,7 @@ class InferenceServerClient:
        return json.loads(content)

    def register_cuda_shared_memory(
-        self,
-        name,
-        raw_handle,
-        device_id,
-        byte_size,
-        headers=None,
-        query_params=None,
+        self, name, raw_handle, device_id, byte_size, headers=None, query_params=None
    ):
        """Request the server to register a system shared memory with the
        following specification.
--- a/python-sdk/experimental/deploy-triton/src/utils.py
+++ b/python-sdk/experimental/deploy-triton/src/utils.py
@ -25,12 +25,9 @@ def get_model_info():
    for model in repo_index:
        model_name = model["name"]
        model_version = model["version"]
-        (
-            input_meta,
-            input_config,
-            output_meta,
-            output_config,
-        ) = parse_model_http(model_name=model_name, model_version=model_version)
+        (input_meta, input_config, output_meta, output_config) = parse_model_http(
+            model_name=model_name, model_version=model_version
+        )
        print(
            f"Found model: {model_name}, version: {model_version}, \
              input meta: {input_meta}, input config: {input_config}, \
@ -69,11 +66,7 @@ def parse_model_http(model_name, model_version=""):


 def triton_infer(
-    input_mapping,
-    model_name,
-    binary_data=False,
-    binary_output=False,
-    class_count=0,
+    input_mapping, model_name, binary_data=False, binary_output=False, class_count=0
 ):
    """Helper function for setting Triton inputs and executing a request

--- a/python-sdk/setup-workspace.py
+++ b/python-sdk/setup-workspace.py
@ -65,14 +65,8 @@ v100computes = {
 }

 akscomputes = {
-    "aks-cpu-deploy": {
-        "vm_size": "STANDARD_DS3_V2",
-        "agent_count": 3,
-    },
-    "aks-gpu-deploy": {
-        "vm_size": "STANDARD_NC6S_V3",
-        "agent_count": 3,
-    },
+    "aks-cpu-deploy": {"vm_size": "STANDARD_DS3_V2", "agent_count": 3},
+    "aks-gpu-deploy": {"vm_size": "STANDARD_NC6S_V3", "agent_count": 3},
 }

 # create or get Workspace
--- a/python-sdk/tutorials/automl-with-azureml/forecasting-backtest-single-model/pipeline_helper.py
+++ b/python-sdk/tutorials/automl-with-azureml/forecasting-backtest-single-model/pipeline_helper.py
@ -149,12 +149,7 @@ def get_backtest_pipeline(
        inputs=[forecasts.as_mount()],
        outputs=[data_results],
        source_directory=PROJECT_FOLDER,
-        arguments=[
-            "--forecasts",
-            forecasts,
-            "--output-dir",
-            data_results,
-        ],
+        arguments=["--forecasts", forecasts, "--output-dir", data_results],
        runconfig=run_config,
        compute_target=compute_target,
        allow_reuse=False,
--- a/python-sdk/tutorials/automl-with-azureml/forecasting-pipelines/scripts/infer.py
+++ b/python-sdk/tutorials/automl-with-azureml/forecasting-pipelines/scripts/infer.py
@ -23,11 +23,7 @@ except ImportError:


 def infer_forecasting_dataset_tcn(
-    X_test,
-    y_test,
-    model,
-    output_path,
-    output_dataset_name="results",
+    X_test, y_test, model, output_path, output_dataset_name="results"
 ):

    y_pred, df_all = model.forecast(X_test, y_test)
@ -71,10 +67,7 @@ def get_model(model_path, model_file_name):
 def get_args():
    parser = argparse.ArgumentParser()
    parser.add_argument(
-        "--model_name",
-        type=str,
-        dest="model_name",
-        help="Model to be loaded",
+        "--model_name", type=str, dest="model_name", help="Model to be loaded"
    )

    parser.add_argument(
@ -108,12 +101,7 @@ def get_args():
    return args


-def get_data(
-    run,
-    fitted_model,
-    target_column_name,
-    test_dataset_name,
-):
+def get_data(run, fitted_model, target_column_name, test_dataset_name):

    # get input dataset by name
    test_dataset = Dataset.get_by_name(run.experiment.workspace, test_dataset_name)
@ -159,10 +147,7 @@ if __name__ == "__main__":
    fitted_model = get_model(model_path, model_file_name)

    X_test_df, y_test = get_data(
-        run,
-        fitted_model,
-        target_column_name,
-        test_dataset_name,
+        run, fitted_model, target_column_name, test_dataset_name
    )

    infer_forecasting_dataset_tcn(
--- a/python-sdk/workflows/deploy/scikit-learn/diabetes/job.py
+++ b/python-sdk/workflows/deploy/scikit-learn/diabetes/job.py
@ -68,10 +68,7 @@ key1, Key2 = webservice.get_keys()
 # access_token, refresh_after = webservice.get_token()

 # If (key) auth is enabled, don't forget to add key to the HTTP header.
-headers = {
-    "Content-Type": "application/json",
-    "Authorization": "Bearer " + key1,
-}
+headers = {"Content-Type": "application/json", "Authorization": "Bearer " + key1}

 # # If token auth is enabled, don't forget to add token to the HTTP header.
 # headers = {'Content-Type':'application/json', 'Authorization': 'Bearer ' + access_token}
--- a/python-sdk/workflows/train/deepspeed/transformers/job.py
+++ b/python-sdk/workflows/train/deepspeed/transformers/job.py
@ -10,11 +10,7 @@ from azureml.core import Workspace, ScriptRunConfig, Environment, Experiment
 from azureml.core.runconfig import PyTorchConfiguration


-TARGET_GPU_COUNT = {
-    "gpu-V100-1": 1,
-    "gpu-V100-2": 2,
-    "gpu-V100-4": 4,
-}
+TARGET_GPU_COUNT = {"gpu-V100-1": 1, "gpu-V100-2": 2, "gpu-V100-4": 4}


@dataclass
@ -92,7 +88,7 @@ if __name__ == "__main__":
    target_names = [
        # "gpu-V100-1",  # single GPU
        # "gpu-V100-2",  # two GPUs
-        "gpu-V100-4",  # four GPUs
+        "gpu-V100-4"  # four GPUs
    ]

    # https://huggingface.co/transformers/pretrained_models.html
--- a/python-sdk/workflows/train/deepspeed/transformers/src/glue_datasets.py
+++ b/python-sdk/workflows/train/deepspeed/transformers/src/glue_datasets.py
@ -13,10 +13,7 @@ from typing import Any, Union, Dict, Callable
 from datasets import load_dataset, load_metric
 from datasets import DatasetDict, Dataset, Metric  # used for typing
 from torch.utils.data.dataset import Dataset
-from transformers import (
-    PreTrainedTokenizerBase,
-    AutoTokenizer,
-)
+from transformers import PreTrainedTokenizerBase, AutoTokenizer


 logger = logging.getLogger(__name__)
@ -113,8 +110,7 @@ def load_raw_glue_dataset(task: str) -> Union[DatasetDict, Dataset]:


 def load_encoded_glue_dataset(
-    task: str,
-    tokenizer: PreTrainedTokenizerBase,
+    task: str, tokenizer: PreTrainedTokenizerBase
 ) -> Union[DatasetDict, Dataset]:
    """Load GLUE data, apply tokenizer and split into train/validation."""
    tokenizer_func = construct_tokenizer_function(tokenizer=tokenizer, task=task)
--- a/python-sdk/workflows/train/pytorch/mnist/src/train.py
+++ b/python-sdk/workflows/train/pytorch/mnist/src/train.py
@ -154,9 +154,7 @@ def main():

    kwargs = {"batch_size": args.batch_size}
    if use_cuda:
-        kwargs.update(
-            {"num_workers": 1, "pin_memory": True, "shuffle": True},
-        )
+        kwargs.update({"num_workers": 1, "pin_memory": True, "shuffle": True})

    transform = transforms.Compose(
        [transforms.ToTensor(), transforms.Normalize((0.1307,), (0.3081,))]
--- a/python-sdk/workflows/train/transformers/glue/1-aml-finetune-job.py
+++ b/python-sdk/workflows/train/transformers/glue/1-aml-finetune-job.py
@ -61,8 +61,7 @@ def transformers_environment(use_gpu=True):

        env_name = "transformers-gpu"
        env = Environment.from_pip_requirements(
-            name=env_name,
-            file_path=pip_requirements_path,
+            name=env_name, file_path=pip_requirements_path
        )
        env.docker.base_image = (
            "mcr.microsoft.com/azureml/intelmpi2018.3-cuda10.0-cudnn7-ubuntu16.04"
@ -72,8 +71,7 @@ def transformers_environment(use_gpu=True):

        env_name = "transformers-cpu"
        env = Environment.from_pip_requirements(
-            name=env_name,
-            file_path=pip_requirements_path,
+            name=env_name, file_path=pip_requirements_path
        )

    return env
--- a/python-sdk/workflows/train/transformers/glue/2-aml-comparison-of-sku-job.py
+++ b/python-sdk/workflows/train/transformers/glue/2-aml-comparison-of-sku-job.py
@ -53,8 +53,7 @@ def transformers_environment(use_gpu=True):

        env_name = "transformers-gpu"
        env = Environment.from_pip_requirements(
-            name=env_name,
-            file_path=pip_requirements_path,
+            name=env_name, file_path=pip_requirements_path
        )
        env.docker.base_image = (
            "mcr.microsoft.com/azureml/intelmpi2018.3-cuda10.0-cudnn7-ubuntu16.04"
@ -64,8 +63,7 @@ def transformers_environment(use_gpu=True):

        env_name = "transformers-cpu"
        env = Environment.from_pip_requirements(
-            name=env_name,
-            file_path=pip_requirements_path,
+            name=env_name, file_path=pip_requirements_path
        )

    return env
--- a/python-sdk/workflows/train/transformers/glue/3-aml-hyperdrive-job.py
+++ b/python-sdk/workflows/train/transformers/glue/3-aml-hyperdrive-job.py
@ -50,8 +50,7 @@ def transformers_environment(use_gpu=True):

        env_name = "transformers-gpu"
        env = Environment.from_pip_requirements(
-            name=env_name,
-            file_path=pip_requirements_path,
+            name=env_name, file_path=pip_requirements_path
        )
        env.docker.base_image = (
            "mcr.microsoft.com/azureml/intelmpi2018.3-cuda10.0-cudnn7-ubuntu16.04"
@ -61,8 +60,7 @@ def transformers_environment(use_gpu=True):

        env_name = "transformers-cpu"
        env = Environment.from_pip_requirements(
-            name=env_name,
-            file_path=pip_requirements_path,
+            name=env_name, file_path=pip_requirements_path
        )

    return env
--- a/python-sdk/workflows/train/transformers/glue/src/glue_datasets.py
+++ b/python-sdk/workflows/train/transformers/glue/src/glue_datasets.py
@ -13,10 +13,7 @@ from typing import Any, Union, Dict, Callable
 from datasets import load_dataset, load_metric
 from datasets import DatasetDict, Dataset, Metric  # used for typing
 from torch.utils.data.dataset import Dataset
-from transformers import (
-    PreTrainedTokenizerBase,
-    AutoTokenizer,
-)
+from transformers import PreTrainedTokenizerBase, AutoTokenizer


 logger = logging.getLogger(__name__)
@ -113,8 +110,7 @@ def load_raw_glue_dataset(task: str) -> Union[DatasetDict, Dataset]:


 def load_encoded_glue_dataset(
-    task: str,
-    tokenizer: PreTrainedTokenizerBase,
+    task: str, tokenizer: PreTrainedTokenizerBase
 ) -> Union[DatasetDict, Dataset]:
    """Load GLUE data, apply tokenizer and split into train/validation."""
    tokenizer_func = construct_tokenizer_function(tokenizer=tokenizer, task=task)
--- a/python-sdk/workflows/train/xgboost/iris/job.py
+++ b/python-sdk/workflows/train/xgboost/iris/job.py
@ -15,10 +15,7 @@ prefix = Path(__file__).parent
 script_dir = str(prefix.joinpath("src"))
 script_name = "train.py"

-arguments = [
-    "--compute",
-    "CPU",  # set to GPU for accelerated training
-]
+arguments = ["--compute", "CPU"]  # set to GPU for accelerated training

 # environment file
 environment_file = str(prefix.joinpath("requirements.txt"))
--- a/scripts/validation/check_experiment_result.py
+++ b/scripts/validation/check_experiment_result.py
@ -174,15 +174,11 @@ def checkVisionTrainRun(child_runs, expected_min_score, expected_max_score):
        )
        print("Primary metric value of {}: {}".format(hd_run.id, best_metric))

-        lower_err_msg = (
-            "Primary metric value was lower than the expected min value of {}".format(
-                expected_min_score
-            )
+        lower_err_msg = "Primary metric value was lower than the expected min value of {}".format(
+            expected_min_score
        )
-        higher_err_msg = (
-            "Primary metric value was higher than the expected max value of {}".format(
-                expected_max_score
-            )
+        higher_err_msg = "Primary metric value was higher than the expected max value of {}".format(
+            expected_max_score
        )
        assert best_metric >= expected_min_score, lower_err_msg
        assert best_metric <= expected_max_score, higher_err_msg
--- a/sdk/jobs/pipelines/1b_pipeline_with_python_function_components/src/components.py
+++ b/sdk/jobs/pipelines/1b_pipeline_with_python_function_components/src/components.py
@ -79,13 +79,9 @@ def score_data(
    (Path(score_output) / "score").write_text("scored with {}".format(model))


-@command_component(
-    display_name="Evaluate",
-    environment=conda_env,
-)
+@command_component(display_name="Evaluate", environment=conda_env)
 def eval_model(
-    scoring_result: Input(type="uri_folder"),
-    eval_output: Output(type="uri_folder"),
+    scoring_result: Input(type="uri_folder"), eval_output: Output(type="uri_folder")
 ):
    """A dummy evaluate component."""

--- a/sdk/jobs/pipelines/1e_pipeline_with_registered_components/eval_src/component.py
+++ b/sdk/jobs/pipelines/1e_pipeline_with_registered_components/eval_src/component.py
@ -1,13 +1,9 @@
 from mldesigner import command_component, Input, Output


-@command_component(
-    display_name="Eval Model",
-    version="0.0.9",
-)
+@command_component(display_name="Eval Model", version="0.0.9")
 def eval_model(
-    scoring_result: Input(type="uri_folder"),
-    eval_output: Output(type="uri_folder"),
+    scoring_result: Input(type="uri_folder"), eval_output: Output(type="uri_folder")
 ):
    """A dummy eval component defined by dsl component."""

--- a/sdk/jobs/pipelines/2c_nyc_taxi_data_regression/prep_src/prep.py
+++ b/sdk/jobs/pipelines/2c_nyc_taxi_data_regression/prep_src/prep.py
@ -17,10 +17,7 @@ args = parser.parse_args()

 print("hello training world...")

-lines = [
-    f"Raw data path: {args.raw_data}",
-    f"Data output path: {args.prep_data}",
-]
+lines = [f"Raw data path: {args.raw_data}", f"Data output path: {args.prep_data}"]

 for line in lines:
    print(line)
--- a/sdk/jobs/pipelines/2d_image_classification_with_densenet/imagecnn_train/image_classification/dataloaders.py
+++ b/sdk/jobs/pipelines/2d_image_classification_with_densenet/imagecnn_train/image_classification/dataloaders.py
@ -235,8 +235,9 @@ def get_dali_train_loader(dali_cpu=False):
            pipe, size=int(pipe.epoch_size("Reader") / world_size)
        )

-        return DALIWrapper(train_loader, num_classes, one_hot), int(
-            pipe.epoch_size("Reader") / (world_size * batch_size)
+        return (
+            DALIWrapper(train_loader, num_classes, one_hot),
+            int(pipe.epoch_size("Reader") / (world_size * batch_size)),
        )

    return gdtl
@ -275,8 +276,9 @@ def get_dali_val_loader():
            pipe, size=int(pipe.epoch_size("Reader") / world_size)
        )

-        return DALIWrapper(val_loader, num_classes, one_hot), int(
-            pipe.epoch_size("Reader") / (world_size * batch_size)
+        return (
+            DALIWrapper(val_loader, num_classes, one_hot),
+            int(pipe.epoch_size("Reader") / (world_size * batch_size)),
        )

    return gdvl
@ -385,10 +387,7 @@ def get_pytorch_train_loader(
    train_dataset = datasets.ImageFolder(
        traindir,
        transforms.Compose(
-            [
-                transforms.RandomResizedCrop(224),
-                transforms.RandomHorizontalFlip(),
-            ]
+            [transforms.RandomResizedCrop(224), transforms.RandomHorizontalFlip()]
        ),
    )

@ -409,8 +408,9 @@ def get_pytorch_train_loader(
        drop_last=True,
    )

-    return PrefetchedWrapper(train_loader, num_classes, fp16, one_hot), len(
-        train_loader
+    return (
+        PrefetchedWrapper(train_loader, num_classes, fp16, one_hot),
+        len(train_loader),
    )


@ -425,13 +425,7 @@ def get_pytorch_val_loader(
 ):
    valdir = os.path.join(data_path, "val")
    val_dataset = datasets.ImageFolder(
-        valdir,
-        transforms.Compose(
-            [
-                transforms.Resize(256),
-                transforms.CenterCrop(224),
-            ]
-        ),
+        valdir, transforms.Compose([transforms.Resize(256), transforms.CenterCrop(224)])
    )

    if torch.distributed.is_initialized():
--- a/sdk/jobs/pipelines/2d_image_classification_with_densenet/imagecnn_train/image_classification/training.py
+++ b/sdk/jobs/pipelines/2d_image_classification_with_densenet/imagecnn_train/image_classification/training.py
@ -257,7 +257,7 @@ def lr_exponential_policy(
            lr = base_lr * (epoch + 1) / warmup_length
        else:
            e = epoch - warmup_length
-            lr = base_lr * (epoch_decay**e)
+            lr = base_lr * (epoch_decay ** e)
        return lr

    return lr_policy(_lr_fn, logger=logger)
--- a/tutorials/e2e-distributed-pytorch-image/src/pytorch_dl_train/image_io.py
+++ b/tutorials/e2e-distributed-pytorch-image/src/pytorch_dl_train/image_io.py
@ -44,9 +44,7 @@ def find_image_subfolder(current_root):


 def build_image_datasets(
-    train_images_dir: str,
-    valid_images_dir: str,
-    input_size: int = 224,
+    train_images_dir: str, valid_images_dir: str, input_size: int = 224
 ):
    """
    Args:
--- a/tutorials/e2e-distributed-pytorch-image/src/pytorch_dl_train/model.py
+++ b/tutorials/e2e-distributed-pytorch-image/src/pytorch_dl_train/model.py
@ -10,9 +10,7 @@ import torch.nn as nn
 import torch.optim as optim
 import torchvision.models as models

-MODEL_ARCH_LIST = [
-    "resnet18",
-]
+MODEL_ARCH_LIST = ["resnet18"]


 def load_model(model_arch: str, output_dimension: int = 1, pretrained: bool = True):
--- a/tutorials/e2e-distributed-pytorch-image/src/pytorch_dl_train/train.py
+++ b/tutorials/e2e-distributed-pytorch-image/src/pytorch_dl_train/train.py
@ -613,8 +613,7 @@ def run(args):
    # PROFILER: here we use a helper class to enable profiling
    # see profiling.py for the implementation details
    training_profiler = PyTorchProfilerHandler(
-        enabled=bool(args.enable_profiling),
-        rank=training_handler.world_rank,
+        enabled=bool(args.enable_profiling), rank=training_handler.world_rank
    )
    # PROFILER: set profiler in trainer to call profiler.step() during training
    training_handler.profiler = training_profiler.start_profiler()