abalone

Author: mvsusp

sagemaker-python-sdk/1P_kmeans_highlevel/kmeans_mnist.ipynb

@@ -55,7 +55,9 @@
    },
    "outputs": [],
    "source": [
-    "role='<your SageMaker execution role here>'\n",
+    "from sagemaker import get_execution_role\n",
+    "\n",
+    "role = get_execution_role()\n",
     "bucket='<bucket-name>'"
    ]
   },

sagemaker-python-sdk/1P_kmeans_lowlevel/kmeans_mnist_lowlevel.ipynb

@@ -55,7 +55,9 @@
    },
    "outputs": [],
    "source": [
-    "role='<your SageMaker execution role here>'\n",
+    "from sagemaker import get_execution_role\n",
+    "\n",
+    "role = get_execution_role()\n",
     "bucket='<bucket-name>'"
    ]
   },

sagemaker-python-sdk/mxnet_gluon_cifar10/cifar10.ipynb

@@ -25,12 +25,11 @@
     "import sagemaker\n",
     "from sagemaker.mxnet import MXNet\n",
     "from mxnet import gluon\n",
+    "from sagemaker import get_execution_role\n",
     "\n",
     "sagemaker_session = sagemaker.Session()\n",
     "\n",
-    "# Replace with a role that gives SageMaker access to S3 and CloudWatch\n",
-    "# see 1-Creating_a_role_allowing_SageMaker_to_access_S3_CloudWatch_ECR.ipynb\n",
-    "role='SageMakerRole'"
+    "role = get_execution_role()"
    ]
   },
   {

sagemaker-python-sdk/mxnet_gluon_mnist/mnist_with_gluon.ipynb

@@ -21,11 +21,11 @@
     "import sagemaker\n",
     "from sagemaker.mxnet import MXNet\n",
     "from mxnet import gluon\n",
+    "from sagemaker import get_execution_role\n",
     "\n",
     "sagemaker_session = sagemaker.Session()\n",
     "\n",
-    "# Replace with a role that gives SageMaker access to S3 and CloudWatch\n",
-    "role='SageMakerRole'"
+    "role = get_execution_role()"
    ]
   },
   {

sagemaker-python-sdk/mxnet_mnist/mxnet_mnist.ipynb

@@ -18,11 +18,14 @@
    },
    "outputs": [],
    "source": [
+    "from sagemaker import get_execution_role\n",
+    "\n",
     "#Bucket location to save your custom code in tar.gz format.\n",
     "custom_code_upload_location = 's3://<bucket-name>/customcode/mxnet_mnist'\n",
     "#Bucket location where results of model training are saved.\n",
     "model_artifacts_location = 's3://<bucket-name>/artifacts'\n",
-    "role='<your SageMaker execution role here>'"
+    "\n",
+    "role = get_execution_role()"
    ]
   },
   {

sagemaker-python-sdk/tensorflow_abalone_age_predictor_using_keras/tensorflow_abalone_age_predictor_using_keras.ipynb

@@ -6,10 +6,8 @@
    "source": [
     "# Creating Estimators in tf.estimator with Keras\n",
     "\n",
-    "If you are getting started using SageMaker, check [iris-dnn-classifier-using-estimators/tutorial.ipynb](iris-dnn-classifier-using-estimators/tutorial.ipynb) first.\n",
-    "\n",
-    "This tutorial covers how to create your own `Estimator` using the building\n",
-    "blocks provided in `tf.estimator`, which will predict the ages of\n",
+    "This tutorial covers how to create your own training script using the building\n",
+    "blocks provided in `tf.keras`, which will predict the ages of\n",
     "[abalones](https://en.wikipedia.org/wiki/Abalone) based on their physical\n",
     "measurements. You'll learn how to do the following:\n",
     "\n",
@@ -54,7 +52,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Let's start by setting up the environment."
+    "### Set up the environment¶"
    ]
   },
   {
@@ -67,18 +65,18 @@
    "source": [
     "import os\n",
     "import sagemaker\n",
+    "from sagemaker import get_execution_role\n",
     "\n",
     "sagemaker_session = sagemaker.Session()\n",
     "\n",
-    "# Replace with a role (either name or full arn) that gives SageMaker access to S3 and cloudwatch\n",
-    "role='SageMakerRole'"
+    "role = get_execution_role()"
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Uploading the data"
+    "### Upload the data to a S3 bucket"
    ]
   },
   {
@@ -96,7 +94,15 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Complete source code"
+    "**sagemaker_session.upload_data** will upload the abalone dataset from your machine to a bucket named **sagemaker-{your aws account number}**, if you don't have this bucket yet, sagemaker_session will create it for you."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Complete source code\n",
+    "Here is the full code for the network model:"
    ]
   },
   {
@@ -337,11 +343,13 @@
     "  # Calculate loss using mean squared error\n",
     "  loss = tf.losses.mean_squared_error(labels, predictions)\n",
     "  ...\n",
-    "```\n",
-    "\n",
-    "See the [tf.losses$API guide](https://www.tensorflow.org/api_docs/python/tf/losses) for a\n",
-    "full list of loss functions and more details on supported arguments and usage.\n",
-    "\n",
+    "```"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
     "Supplementary metrics for evaluation can be added to an `eval_metric_ops` dict.\n",
     "The following code defines an `rmse` metric, which calculates the root mean\n",
     "squared error for the model predictions. Note that the `labels` tensor is cast\n",
@@ -375,9 +383,6 @@
     "    loss=loss, global_step=tf.train.get_global_step())\n",
     "```\n",
     "\n",
-    "For a full list of optimizers, and other details, see the\n",
-    "@{$python/train#optimizers$API guide}.\n",
-    "\n",
     "### The complete abalone `model_fn`\n",
     "\n",
     "Here's the final, complete `model_fn` for the abalone age predictor. The\n",
@@ -435,8 +440,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Submitting script for training in SageMaker\n",
-    "\n",
+    "# Submitting script for training\n",
     "\n",
     "We can use the SDK to run our local training script on SageMaker infrastructure.\n",
     "\n",
@@ -456,10 +460,12 @@
     "from sagemaker.tensorflow import TensorFlow\n",
     "\n",
     "abalone_estimator = TensorFlow(entry_point='abalone.py',\n",
-    "                                  role=role,\n",
-    "                                  hyperparameters={'training_steps': 100, 'learning_rate': 0.001},\n",
-    "                                  train_instance_count=1,\n",
-    "                                  train_instance_type='ml.c4.xlarge')\n",
+    "                               role=role,\n",
+    "                               training_steps= 100,                                  \n",
+    "                               evaluation_steps= 100,\n",
+    "                               hyperparameters={'learning_rate': 0.001},\n",
+    "                               train_instance_count=1,\n",
+    "                               train_instance_type='ml.c4.xlarge')\n",
     "\n",
     "abalone_estimator.fit(inputs)"
    ]
@@ -470,11 +476,11 @@
    "source": [
     "`estimator.fit` will deploy a script in a container for training and returs the SageMaker model name using the following arguments:\n",
     "\n",
-    "*   `framework=tensorflow`. Tells submit_training that it is a tensorflow container.\n",
-    "*   `script=\"abalone.py\"`. The relative path to the script that will be deployed to the container.\n",
-    "*   `data=\"s3://bucket_name/abalone-estimator-training\"`. The S3 location of the bucket that we uploaded earlier.\n",
-    "*   `role=\"maeve-pullrole\"`. AWS role that gives your account access to SageMaker training and hosting\n",
-    "* `hyperparameters={'training_steps' : 100}`. Model and training hyperparameters. \n",
+    "*   **`entry_point=\"abalone.py\"`** The path to the script that will be deployed to the container.\n",
+    "*   **`training_steps=100`** The number of training steps of the training job.\n",
+    "*   **`evaluation_steps=100`** The number of evaluation steps of the training job.\n",
+    "*   **`role`**. AWS role that gives your account access to SageMaker training and hosting\n",
+    "*   **`hyperparameters={'learning_rate' : 0.001}`**. Training hyperparameters. \n",
     "\n",
     "Running the code block above will do the following actions:\n",
     "* deploy your script in a container with tensorflow installed\n",
@@ -501,8 +507,7 @@
    },
    "outputs": [],
    "source": [
-    "abalone_predictor = abalone_estimator.deploy(initial_instance_count=1,\n",
-    "                                             instance_type='ml.c4.xlarge')"
+    "abalone_predictor = abalone_estimator.deploy(initial_instance_count=1, instance_type='ml.c4.xlarge')"
    ]
   },
   {