final updates

Author: aarora79

0_model_training_pipeline.ipynb

@@ -75,8 +75,8 @@
    },
    "outputs": [],
    "source": [
-    "# import sys\n",
-    "# !{sys.executable} -m pip install -r requirements.txt"
+    "import sys\n",
+    "!{sys.executable} -m pip install -r requirements.txt  --upgrade-strategy only-if-needed"
    ]
   },
   {
@@ -747,7 +747,7 @@
    "source": [
     "# Start pipeline with credit data and preprocessing script\n",
     "execution = pipeline.start(\n",
-    "                execution_display_name=pipeline.name\n",
+    "                execution_display_name=pipeline.name,\n",
     "                parameters=dict(\n",
     "                AccuracyConditionThreshold=config['evaluation_step']['accuracy_condition_threshold'],\n",
     "                MaximumParallelTrainingJobs=config['tuning_step']['maximum_parallel_training_jobs'],\n",

1_batch_transform_pipeline.ipynb

@@ -30,8 +30,8 @@
    },
    "outputs": [],
    "source": [
-    "#import sys\n",
-    "#!{sys.executable} -m pip install -r requirements.txt"
+    "import sys\n",
+    "!{sys.executable} -m pip install -r requirements.txt  --upgrade-strategy only-if-needed"
    ]
   },
   {
@@ -65,9 +65,8 @@
     "from sagemaker.workflow.execution_variables import ExecutionVariables\n",
     "\n",
     "from sagemaker.model import Model\n",
-    "from sagemaker.inputs import CreateModelInput\n",
-    "from sagemaker.workflow.model_step import ModelStep\n",
-    "from sagemaker.transformer import Transformer"
+    "from sagemaker.transformer import Transformer\n",
+    "from sagemaker.workflow.model_step import ModelStep"
    ]
   },
   {

2_realtime_inference.ipynb

@@ -28,8 +28,8 @@
    },
    "outputs": [],
    "source": [
-    "#import sys\n",
-    "#!{sys.executable} -m pip install -r requirements.txt"
+    "import sys\n",
+    "!{sys.executable} -m pip install -r requirements.txt  --upgrade-strategy only-if-needed"
    ]
   },
   {
@@ -52,7 +52,6 @@
     "import sagemaker.session\n",
     "from datetime import datetime\n",
     "from typing import Dict, List\n",
-    "from utils import load_config, print_pipeline_execution_summary\n",
     "from sagemaker.workflow.pipeline_context import PipelineSession"
    ]
   },

README.md

@@ -138,11 +138,10 @@ Setup a secret in Secrets Manager for the PrestoDB username and password. Call t
 
 1. Edit the [`config`](./config.yml) as per PrestoDB connection, IAM role and other pipeline details such as instance types for various pipeline steps etc.
 
-    - [**Mandatory**] Edit the parameter values in the `presto` section.
-    - [**Mandatory**] Edit the parameter values in the `aws` section.
-    - [**Mandatory**] Edit the `query` parameter value in the `training_step` section. This is the query for retrieving the training data from the PrestoDB.
-    - [**Mandatory**] Edit the `query` parameter value in the `transform_step` section. This is the query for retrieving the data for the batch transform from the PrestoDB.
-    - [Optional] Edit the parameter values in the rest of the sections as appropriate.
+    - Edit the parameter values in the `presto` section. These parameters define the connectivity to PrestoDB.
+    - Edit the parameter values in the `aws` section. These parameters define the IAM role, bucket name, region and other AWS cloud related parameters.
+    - Edit the parameter values in the sections corresponding to the pipeline steps i.e. `training_step`, `tuning_step`, `transform_step` etc. Review all the parameters in these sections carefully and edit them as appropriate for your use-case.
+    - Review the parameters in the rest of the sections of the [`config`](./config.yml)and edit them if needed.
 
 1. Run the [`0_model_training_pipeline`](./0_model_training_pipeline.ipynb) notebook to train and tune the ML model and register it with the SageMaker model registry. All the steps in this notebook are executed as part of a training pipeline.
     - This notebook also contains an automatic model approval step that changes the state of the model registered with the model registry from `PendingForApproval` to `Approved` state. This step can be removed for prod accounts where manual or some criteria based approval would be required.

code/query.py

@@ -2,7 +2,7 @@
 aws:
   region: us-east-1
   # execution role, replace the role name below with the one you are using
-  sagemaker_execution_role_name: your-role-name
+  sagemaker_execution_role_name: your-sagemaker-execution-role
   # the execution role ARN is determined automatically by the code
   sagemaker_execution_role_arn: arn:aws:iam::{account_id}:role/{role}
   s3_bucket: sagemaker-{region}-{account_id} # region and account id are automatically replaced