DeepLearningTrafficSignals.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "8f0aefd9-48dd-4a1b-9094-9c1165432b58",
   "metadata": {},
   "source": [
    "### CONVOLUTIONAL NETWORKS: TRAFFIC SIGNALS IN SAGEMAKER"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "88120b5a-e7e0-4bd1-ac44-7ad84db247c2",
   "metadata": {},
   "source": [
    "#### Learning Objectives:\n",
    "\n",
    "1. Basics of convolutional neural networks\n",
    "2. Deep Learning\n",
    "3. Image Classification\n",
    "4. How to apply and deploy deep learning models at scale, in SageMaker\n",
    "\n",
    "#### What is an Image Classifier?\n",
    "\n",
    "Image Classifier work by predicting the class of items that are in a give image. For example, we can train a classifier to identify images of fruits, bananas, mangos, apples, etc. When we feed a trained classifier on the image of a mango for example, it will give the label to the image: \"label = mango\".\n",
    "\n",
    "Let's take a look from the Fashion MNIST Dataset:\n",
    "\n",
    "\"Context\n",
    "\n",
    "Fashion-MNIST is a dataset of Zalando's article images—consisting of a training set of 60,000 examples and a test set of 10,000 examples. Each example is a 28x28 grayscale image, associated with a label from 10 classes. Zalando intends Fashion-MNIST to serve as a direct drop-in replacement for the original MNIST dataset for benchmarking machine learning algorithms. It shares the same image size and structure of training and testing splits.\n",
    "\n",
    "The original MNIST dataset contains a lot of handwritten digits. Members of the AI/ML/Data Science community love this dataset and use it as a benchmark to validate their algorithms. In fact, MNIST is often the first dataset researchers try. \"If it doesn't work on MNIST, it won't work at all\", they said. \"Well, if it does work on MNIST, it may still fail on others.\"\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9570b29a-06c7-4461-b5d7-96f390aa9a61",
   "metadata": {},
   "source": [
    "#### TWO CLASSIFIERS: \n",
    "\n",
    "1. Binary: Between two categories\n",
    "2. Multi Class Classifier: There are more than two classes! We have 10 classes for example, mango, apple, banana, and so on.\n",
    "\n",
    "When we build these deep learning based classifiers, we take all the data and images for 80-85% for training and the rest for testing."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1d757a1e-7a42-4bfb-98ac-e9446afaa625",
   "metadata": {},
   "source": [
    "#### We will use the Multi Class Classifier Technique"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1fe56f8d-3824-4bcf-aa0f-5e3b0844e276",
   "metadata": {},
   "source": [
    "### CASE STUDY: CLASSIFYING TRAFFIC SIGNS"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b0aebcab-daac-4f2f-acc6-2a35257bc752",
   "metadata": {},
   "source": [
    "We will build a multi class classifier based on deep learning and neural network, convolutional networks and we will use this classfier to:\n",
    "\n",
    "1. Classify different images.\n",
    "\n",
    "#### USE CASES:\n",
    "\n",
    "1. Self-driving cars: To make sure automated systems are able to classify safely between traffic signs for the purpose of security, safety, and following the law.\n",
    "\n",
    "2. Making sure life is more important that AI systems\n",
    "\n",
    "3. Making sure that the automated system is following the correct dorrection.\n",
    "\n",
    "#### In this case, we have 43 different classes and this is used from the data source:\n",
    "\n",
    "DataSource Link: Given in the notebook (German Traffic Data Set)\n",
    "\n",
    "#### GOALS OF THE PROJECT:\n",
    "\n",
    "1. Create a deep network called LeNet to teach and train this network to classify these images.\n",
    "\n",
    "2. We will teach the classifier to train on red green and blue images\n",
    "\n",
    "3. We want to be able to feed the endpoints an image, and get to know what that image means.\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4c5d83f9-fce0-4000-9eb9-6c917ab654e6",
   "metadata": {},
   "source": [
    "### CONVOLUTIONAL NEURAL NETWORKS - HOW DO THESE LEARN?\n",
    "\n",
    "#### Basics:\n",
    "\n",
    "1. Based on the basic AI neural networks\n",
    "\n",
    "2. AI neural networks: Information processing models, that mimick the human intelligence\n",
    "\n",
    "3. There is a nucleus, and collects information from dendrites and so on, and we take the information into the nucleus, we process the information through the axon.\n",
    "\n",
    "4. We mimic this process mathematically, take all the inputs, mutliply the weights, and add a bias signal and generate an output - we apply an activation function 'f' to get an output 'A'\n",
    "\n",
    "5. We collect many of these neurons (inputs) and we connect these to all neurons in the subsequent layer (dense/fully connected AI network)\n",
    "\n",
    "#### Let's take a look at convulutional neural networks:\n",
    "\n",
    "1. Let's assume we have a bunch of fruits and their classes\n",
    "\n",
    "2. We obtain a dense fully connected AI network, but to deal with images, we add in additional steps:\n",
    "\n",
    "    1. We take the image, apply the convulutional image\n",
    "    2. We take each image, and extract the features from each of the image\n",
    "    3. Pooling - we compress these feature maps, and represent the information in a reduce manner.\n",
    "    4. Once we have the pool, we take the filters, flatten it, and then feed it to the dense AI network.\n",
    "    \n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0f9d1828-7dd3-4f5c-af96-66fa91117488",
   "metadata": {},
   "source": [
    "## DATA COLLECTION/VISUALIZATION"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "007d3e31-05e2-4538-b1ef-9c040a3cff09",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "## Importing pickle to convert python objects into byte level streams\n",
    "import pickle\n",
    "\n",
    "## Represents opening the data: train, valid and test data\n",
    "with open(\"test.p\", mode='rb') as testing_data:\n",
    "    test = pickle.load(testing_data)\n",
    "\n",
    "    \n",
    "## Mode is read binary\n",
    "with open(\"train.p\", mode='rb') as training_data:\n",
    "    train = pickle.load(training_data)\n",
    "\n",
    "with open(\"valid.p\", mode='rb') as validation_data:\n",
    "    valid = pickle.load(validation_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "cd5b0374-5f39-46df-bd97-b1744bff4a66",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "## Represents obtaining the features and labels within the train, test and valid\n",
    "\n",
    "## Using a supervised deep learning technique\n",
    "\n",
    "## We will use the validation dataset to perform cross validation\n",
    "X_train, y_train = train['features'], train['labels']\n",
    "X_validation, y_validation = valid['features'], valid['labels']\n",
    "X_test, y_test = test['features'], test['labels']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "7a2c2a00-32a7-4899-b60e-c2cf1d732c65",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(12630, 32, 32, 3)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "00032624-58f0-4083-9487-65f5b90b229c",
   "metadata": {},
   "source": [
    "We have 12,630 images with 32 by 32 by 3 pixels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "afe88c00-13f8-4f35-b6eb-b07e341154c1",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(34799, 32, 32, 3)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "4af22a73-069f-48c5-b2d4-4874326aef41",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4410, 32, 32, 3)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_validation.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "8532cc29-9983-4f0b-9c86-d2ee4fcd7be5",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Represents visualizing one sample image\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "## Going to select a random image\n",
    "i = np.random.randint(1, len(X_test))\n",
    "\n",
    "## Show the actual image\n",
    "plt.imshow(X_test[i])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "fc031dda-d723-4014-9838-8f1d5d8ec52f",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "label =  35\n"
     ]
    }
   ],
   "source": [
    "print('label = ', y_test[i])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ef6766b3-5d15-4212-ba6a-4690f2c3e343",
   "metadata": {},
   "source": [
    "#### SHOWING THE RANDOM SORTED TRAFFIC SIGNS WITH THEIR CORRESPODING LABELS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "bb4f7cec-8266-4616-ba75-81ea3ebc87a2",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "l_grid = 5\n",
    "w_grid = 5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "c0c6a432-f8d7-4fca-b74a-888187c2ec0c",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "fig, axes = plt.subplots(l_grid, w_grid, figsize = (10, 10))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "b266efbc-4c18-4450-a1f6-2326f62c2198",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1000x1000 with 25 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "axes = axes.ravel()\n",
    "n_training = len(X_test)\n",
    "for i in np.arange(0, l_grid * w_grid):\n",
    "    index = np.random.randint(0, len(X_test))\n",
    "    axes[i].imshow(X_test[index])\n",
    "    axes[i].set_title(y_test[index])\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec020aed-df54-4f36-955f-2656d1cb66a3",
   "metadata": {},
   "source": [
    "### SAGEMAKER INTEGRATION"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "9bbfe9c3-67b2-4751-b4ee-e4d6c35a8ad7",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "arn:aws:iam::442851008302:role/service-role/AmazonSageMaker-ExecutionRole-20230713T114638\n"
     ]
    }
   ],
   "source": [
    "import sagemaker\n",
    "import boto3\n",
    "\n",
    "## Represents creating a sagemaker session\n",
    "sagemaker_session = sagemaker.Session()\n",
    "\n",
    "prefix = 'traffic-sign-classifier'\n",
    "\n",
    "role = sagemaker.get_execution_role()\n",
    "\n",
    "print(role)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "05e69b26-a282-43d7-89d0-9c533b981f51",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "## Represents uploading the data into the bucket\n",
    "import os\n",
    "os.makedirs(\"./data\", exist_ok = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "fea82147-735d-4f57-9943-fecb9435ee84",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "## Represents saving the training and validation data in the data files\n",
    "\n",
    "np.savez('./data/training', image = X_train, label = y_train)\n",
    "np.savez('./data/validation', image = X_validation, label = y_validation)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "b30aac52-94c5-40a2-a202-02ddaabce119",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "s3://sagemaker-us-east-1-442851008302/traffic-sign-classifier/training/training.npz\n",
      "s3://sagemaker-us-east-1-442851008302/traffic-sign-classifier/validation/validation.npz\n"
     ]
    }
   ],
   "source": [
    "## Now, uploading the data that we have into amazon S3\n",
    "\n",
    "training_path = sagemaker_session.upload_data('data/training.npz', key_prefix = prefix + '/training')\n",
    "valid_path = sagemaker_session.upload_data('data/validation.npz', key_prefix = prefix + '/validation')\n",
    "\n",
    "print(training_path)\n",
    "print(valid_path)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "60a6104f-d1a5-4451-bed6-95023166d2c1",
   "metadata": {},
   "source": [
    "#### TRAINING THE CNN MODEL USING: SAGEMAKER"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8e6eecde-1f9a-4c6b-af93-7436a9b8e36a",
   "metadata": {},
   "source": [
    "We will be training the LNET Model using sagemaker:\n",
    "\n",
    "1. Step 1: Convolutional Layer \n",
    "    1. Input: 32x32x3\n",
    "    2. Output: 28x28x6\n",
    "    3. We will use the 5x5 filters\n",
    "    We will apply the RELU activation function to give the output\n",
    "    and pool layer to reduce the filters and give the output\n",
    "\n",
    "\n",
    "2. Step 2: 2nd Convolutional Layer\n",
    "    1. Input: 14x14x6\n",
    "    2. Output: 10x10x6\n",
    "    We will again, apply the RELU activation function to get the output\n",
    "\n",
    "\n",
    "3. Step 3: Flattening the Network\n",
    "    1. Here, we will flatten the output\n",
    "\n",
    "\n",
    "4. Step 4: Fully Connected Layer\n",
    "    1. We will have a fully connected layer\n",
    "\n",
    "\n",
    "5. Step 5: Another Fully Connected Layer (OUTPUT LAYER --> 43 Layers)\n",
    "    1. Output will be 43 at the end\n",
    "\n",
    "\n",
    "--> Refer to the train-cnn.py to get the code for training the CNN model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "68b1d861-4e74-4d57-8c5a-365734a9868a",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[37m## Represents the training code and define the network, model architecture and training\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "\u001b[34mimport\u001b[39;49;00m \u001b[04m\u001b[36margparse\u001b[39;49;00m, \u001b[04m\u001b[36mos\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "\u001b[34mimport\u001b[39;49;00m \u001b[04m\u001b[36mtensorflow\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "\u001b[34mimport\u001b[39;49;00m \u001b[04m\u001b[36mnumpy\u001b[39;49;00m \u001b[34mas\u001b[39;49;00m \u001b[04m\u001b[36mnp\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "\u001b[34mfrom\u001b[39;49;00m \u001b[04m\u001b[36mtensorflow\u001b[39;49;00m\u001b[04m\u001b[36m.\u001b[39;49;00m\u001b[04m\u001b[36mkeras\u001b[39;49;00m \u001b[34mimport\u001b[39;49;00m backend \u001b[34mas\u001b[39;49;00m K\u001b[37m\u001b[39;49;00m\n",
      "\u001b[34mfrom\u001b[39;49;00m \u001b[04m\u001b[36mtensorflow\u001b[39;49;00m\u001b[04m\u001b[36m.\u001b[39;49;00m\u001b[04m\u001b[36mkeras\u001b[39;49;00m\u001b[04m\u001b[36m.\u001b[39;49;00m\u001b[04m\u001b[36mmodels\u001b[39;49;00m \u001b[34mimport\u001b[39;49;00m Sequential\u001b[37m\u001b[39;49;00m\n",
      "\u001b[34mfrom\u001b[39;49;00m \u001b[04m\u001b[36mtensorflow\u001b[39;49;00m\u001b[04m\u001b[36m.\u001b[39;49;00m\u001b[04m\u001b[36mkeras\u001b[39;49;00m\u001b[04m\u001b[36m.\u001b[39;49;00m\u001b[04m\u001b[36mlayers\u001b[39;49;00m \u001b[34mimport\u001b[39;49;00m Dense, Dropout, Activation, Flatten, BatchNormalization, Conv2D, MaxPooling2D, AveragePooling2D\u001b[37m\u001b[39;49;00m\n",
      "\u001b[34mfrom\u001b[39;49;00m \u001b[04m\u001b[36mtensorflow\u001b[39;49;00m\u001b[04m\u001b[36m.\u001b[39;49;00m\u001b[04m\u001b[36mkeras\u001b[39;49;00m\u001b[04m\u001b[36m.\u001b[39;49;00m\u001b[04m\u001b[36moptimizers\u001b[39;49;00m \u001b[34mimport\u001b[39;49;00m Adam\u001b[37m\u001b[39;49;00m\n",
      "\u001b[34mfrom\u001b[39;49;00m \u001b[04m\u001b[36mtensorflow\u001b[39;49;00m\u001b[04m\u001b[36m.\u001b[39;49;00m\u001b[04m\u001b[36mkeras\u001b[39;49;00m,utils \u001b[34mimport\u001b[39;49;00m \u001b[04m\u001b[36mmulti_gpu_model\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "\u001b[37m\u001b[39;49;00m\n",
      "\u001b[37m\u001b[39;49;00m\n",
      "\u001b[37m## Training code will be run here\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "\u001b[34mif\u001b[39;49;00m \u001b[31m__name__\u001b[39;49;00m == \u001b[33m'\u001b[39;49;00m\u001b[33m__main__\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m:\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## Getting all of the arguments\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    parser = argparse.ArgumentParser()\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## Model paramters are sent as Command Line arguments\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    parser.add_argument(\u001b[33m'\u001b[39;49;00m\u001b[33m--epochs\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m, \u001b[36mtype\u001b[39;49;00m=\u001b[36mint\u001b[39;49;00m, default = \u001b[34m1\u001b[39;49;00m)\u001b[37m\u001b[39;49;00m\n",
      "    parser.add_argument(\u001b[33m'\u001b[39;49;00m\u001b[33m--learning-rate\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m, \u001b[36mtype\u001b[39;49;00m=\u001b[36mfloat\u001b[39;49;00m, default = \u001b[34m0.001\u001b[39;49;00m)\u001b[37m\u001b[39;49;00m\n",
      "    parser.add_argument(\u001b[33m'\u001b[39;49;00m\u001b[33m--batch-size\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m, \u001b[36mtype\u001b[39;49;00m=\u001b[36mint\u001b[39;49;00m, default = \u001b[34m32\u001b[39;49;00m)\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## Some more important arguments to keep in mind\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## How many gpus are available for training\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    parser.add_argument(\u001b[33m'\u001b[39;49;00m\u001b[33m--gpu-count\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m, \u001b[36mtype\u001b[39;49;00m=\u001b[36mint\u001b[39;49;00m, default = os.environ[\u001b[33m'\u001b[39;49;00m\u001b[33mSM_NUM_GPUS\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m])\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## Location of storage for model artifacts\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    parser.add_argument(\u001b[33m'\u001b[39;49;00m\u001b[33m--model_dir\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m, \u001b[36mtype\u001b[39;49;00m=\u001b[36mstr\u001b[39;49;00m, default = os.environ[\u001b[33m'\u001b[39;49;00m\u001b[33mSM_MODEL_DIR\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m])\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## Channel for training and validation\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    parser.add_argument(\u001b[33m'\u001b[39;49;00m\u001b[33m--training\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m, \u001b[36mtype\u001b[39;49;00m=\u001b[36mstr\u001b[39;49;00m, default = os.environ[\u001b[33m'\u001b[39;49;00m\u001b[33mSM_CHANNEL_TRAINING\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m])\u001b[37m\u001b[39;49;00m\n",
      "    parser.add_argument(\u001b[33m'\u001b[39;49;00m\u001b[33m--validation\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m, \u001b[36mtype\u001b[39;49;00m=\u001b[36mstr\u001b[39;49;00m, default = os.environ[\u001b[33m'\u001b[39;49;00m\u001b[33mSM_CHANNEL_VALIDATION\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m])\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## Parsing all of the arguments above\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    args, _ = parser.parse_known_args()\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## Represents initializing all of the hyperparameters\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    epochs = args.epochs\u001b[37m\u001b[39;49;00m\n",
      "    lr = args.learning_rate\u001b[37m\u001b[39;49;00m\n",
      "    batch_size = args.batch_size\u001b[37m\u001b[39;49;00m\n",
      "    gpu_count = args.gpu_count\u001b[37m\u001b[39;49;00m\n",
      "    model_dir = args.model_dir\u001b[37m\u001b[39;49;00m\n",
      "    training_dir = args.training\u001b[37m\u001b[39;49;00m\n",
      "    valitation_dr = args.validation\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## Now, we will load the trianing and validation bucket from the S3 bucket\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    train_image = np.load(os.path.join(training_dir, \u001b[33m'\u001b[39;49;00m\u001b[33mtraining.npz\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m))[\u001b[33m'\u001b[39;49;00m\u001b[33mimage\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m]\u001b[37m\u001b[39;49;00m\n",
      "    train_label = np.load(os.path.join(training_dir, \u001b[33m'\u001b[39;49;00m\u001b[33mtraining.npz\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m))[\u001b[33m'\u001b[39;49;00m\u001b[33mlabel\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m]\u001b[37m\u001b[39;49;00m\n",
      "    test_image = np.load(os.path.join(validation_dir, \u001b[33m'\u001b[39;49;00m\u001b[33mvalidation.npz\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m))[\u001b[33m'\u001b[39;49;00m\u001b[33mimage\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m]\u001b[37m\u001b[39;49;00m\n",
      "    test_label = np.load(os.path.join(validation_dir, \u001b[33m'\u001b[39;49;00m\u001b[33mvalidation.npz\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m))[\u001b[33m'\u001b[39;49;00m\u001b[33mlabel\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m]\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    K.set_image_data_format(\u001b[33m'\u001b[39;49;00m\u001b[33mchannels_last\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m)\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## We have used all the code above, because we are running sagemaker.\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## Now, we will add batch dimension to our input (re image it)\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    train_image = train_image.reshape(train_image.shape[\u001b[34m0\u001b[39;49;00m], \u001b[34m32\u001b[39;49;00m, \u001b[34m32\u001b[39;49;00m, \u001b[34m3\u001b[39;49;00m)\u001b[37m\u001b[39;49;00m\n",
      "    test_image = test_image.reshape(test_image.shape[\u001b[34m0\u001b[39;49;00m], \u001b[34m32\u001b[39;49;00m, \u001b[34m32\u001b[39;49;00m, \u001b[34m3\u001b[39;49;00m)\u001b[37m\u001b[39;49;00m\n",
      "    input_shape = (\u001b[34m32\u001b[39;49;00m, \u001b[34m32\u001b[39;49;00m, \u001b[34m3\u001b[39;49;00m)\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## Now, we will normalize all of the data that we have:\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    train_image = train_image.astype(\u001b[33m'\u001b[39;49;00m\u001b[33mfloat32\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m)\u001b[37m\u001b[39;49;00m\n",
      "    test_image = test_image.astype(\u001b[33m'\u001b[39;49;00m\u001b[33mfloat32\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m)\u001b[37m\u001b[39;49;00m\n",
      "    train_image /= \u001b[34m255\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    test_image /= \u001b[34m255\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## Using tensorflow to provide the data in terms of categorical data\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    train_label = tensorflow.keras.utils.to_categorical(train_label, \u001b[34m43\u001b[39;49;00m)\u001b[37m\u001b[39;49;00m\n",
      "    test_label = tensorflow.keras.utils.to_categorical(test_label, \u001b[34m43\u001b[39;49;00m)\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## Represents the LeNet Network architecture below\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "        model = Sequential() \u001b[37m## from left to right\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## Adding the convolution 2D, specifying the input_shape as 32, 32, 3\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    model.add(Conv2D(filters=\u001b[34m6\u001b[39;49;00m, kernel_size=(\u001b[34m5\u001b[39;49;00m, \u001b[34m5\u001b[39;49;00m), activation=\u001b[33m'\u001b[39;49;00m\u001b[33mrelu\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m, input_shape= input_shape))\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## have the average pooling to reduce the size of feature maps\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    model.add(AveragePooling2D())\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## Another convolution layer, filters are 16, and we apply the RELU activation funciton again\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    model.add(Conv2D(filters=\u001b[34m16\u001b[39;49;00m, kernel_size=(\u001b[34m5\u001b[39;49;00m, \u001b[34m5\u001b[39;49;00m), activation=\u001b[33m'\u001b[39;49;00m\u001b[33mrelu\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m))   \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    model.add(AveragePooling2D())\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    model.add(Flatten())\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## Add dense layers below until we get the output, and activate it using softmax\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## to get the saturated value in the output to know which neuron is fired to get\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## the true class of the given image\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    model.add(Dense(units=\u001b[34m120\u001b[39;49;00m, activation=\u001b[33m'\u001b[39;49;00m\u001b[33mrelu\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m))\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    model.add(Dense(units=\u001b[34m84\u001b[39;49;00m, activation=\u001b[33m'\u001b[39;49;00m\u001b[33mrelu\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m))\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    model.add(Dense(units=\u001b[34m43\u001b[39;49;00m, activation = \u001b[33m'\u001b[39;49;00m\u001b[33msoftmax\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m))\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[36mprint\u001b[39;49;00m(model.summary())\u001b[37m\u001b[39;49;00m\n",
      "\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m# If more than one GPU is available, convert the model to multi-gpu model\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[34mif\u001b[39;49;00m gpu_count > \u001b[34m1\u001b[39;49;00m:\u001b[37m\u001b[39;49;00m\n",
      "        \u001b[37m\u001b[39;49;00m\n",
      "        model = multi_gpu_model(model, gpus=gpu_count)\u001b[37m\u001b[39;49;00m\n",
      "\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m# Compile and train the model\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    model.compile(loss=tensorflow.keras.losses.categorical_crossentropy,\u001b[37m\u001b[39;49;00m\n",
      "                  optimizer=Adam(lr=lr),\u001b[37m\u001b[39;49;00m\n",
      "                  metrics=[\u001b[33m'\u001b[39;49;00m\u001b[33maccuracy\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m])\u001b[37m\u001b[39;49;00m\n",
      "\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## fit the model, specify the data and validation, to get the epochs and verbose\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m## for more information\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    model.fit(train_image, train_label, batch_size=batch_size,\u001b[37m\u001b[39;49;00m\n",
      "                  validation_data=(test_image, test_label),\u001b[37m\u001b[39;49;00m\n",
      "                  epochs=epochs,\u001b[37m\u001b[39;49;00m\n",
      "                  verbose=\u001b[34m2\u001b[39;49;00m)\u001b[37m\u001b[39;49;00m\n",
      "\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m# Evaluating the model\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    score = model.evaluate(test_image, test_label, verbose=\u001b[34m0\u001b[39;49;00m)\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[36mprint\u001b[39;49;00m(\u001b[33m'\u001b[39;49;00m\u001b[33mValidation loss    :\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m, score[\u001b[34m0\u001b[39;49;00m])\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[36mprint\u001b[39;49;00m(\u001b[33m'\u001b[39;49;00m\u001b[33mValidation accuracy:\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m, score[\u001b[34m1\u001b[39;49;00m])\u001b[37m\u001b[39;49;00m\n",
      "\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m# save trained CNN Keras model to \"model_dir\" (path specificied earlier)\u001b[39;49;00m\u001b[37m\u001b[39;49;00m\n",
      "    sess = K.get_session()\u001b[37m\u001b[39;49;00m\n",
      "    tensorflow.saved_model.simple_save(\u001b[37m\u001b[39;49;00m\n",
      "        sess,\u001b[37m\u001b[39;49;00m\n",
      "        os.path.join(model_dir, \u001b[33m'\u001b[39;49;00m\u001b[33mmodel/1\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m),\u001b[37m\u001b[39;49;00m\n",
      "        inputs={\u001b[33m'\u001b[39;49;00m\u001b[33minputs\u001b[39;49;00m\u001b[33m'\u001b[39;49;00m: model.input},\u001b[37m\u001b[39;49;00m\n",
      "        outputs={t.name: t \u001b[34mfor\u001b[39;49;00m t \u001b[35min\u001b[39;49;00m model.outputs})\u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n",
      "    \u001b[37m\u001b[39;49;00m\n"
     ]
    }
   ],
   "source": [
    "## Now, we will run the trainCNN.py file to train the CNN network\n",
    "!pygmentize trainCNN.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "a82278fc-fac5-4b48-b186-36640463532d",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "train_instance_type has been renamed in sagemaker>=2.\n",
      "See: https://sagemaker.readthedocs.io/en/stable/v2.html for details.\n",
      "train_instance_count has been renamed in sagemaker>=2.\n",
      "See: https://sagemaker.readthedocs.io/en/stable/v2.html for details.\n",
      "train_instance_type has been renamed in sagemaker>=2.\n",
      "See: https://sagemaker.readthedocs.io/en/stable/v2.html for details.\n"
     ]
    }
   ],
   "source": [
    "from sagemaker.tensorflow import TensorFlow\n",
    "\n",
    "## To train the tensorflow model, we will use the estimator from the SM SDK\n",
    "\n",
    "tf_estimator = TensorFlow(entry_point='trainCNN.py',\n",
    "                         role=role, \n",
    "                         train_instance_count = 1,\n",
    "                         train_instance_type = 'ml.c4.2xlarge',\n",
    "                         framework_version = '1.12',\n",
    "                         py_version = 'py3',\n",
    "                         script_mode = True,\n",
    "                         hyperparameters = {'epochs': 13, \n",
    "                                           'batch-size': 32, \n",
    "                                           'learning-rate': 0.001}\n",
    "                         )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "9c58919d-2672-4825-8a21-a229784b0674",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:sagemaker.image_uris:image_uri is not presented, retrieving image_uri based on instance_type, framework etc.\n",
      "INFO:sagemaker:Creating training-job with name: sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using provided s3_resource\n",
      "2023-07-21 14:06:28 Starting - Starting the training job...\n",
      "2023-07-21 14:06:45 Starting - Preparing the instances for training.........\n",
      "2023-07-21 14:08:28 Downloading - Downloading input data\n",
      "2023-07-21 14:08:28 Training - Training image download completed. Training in progress...\u001b[34m2023-07-21 14:08:33,894 sagemaker-containers INFO     Imported framework sagemaker_tensorflow_container.training\u001b[0m\n",
      "\u001b[34m2023-07-21 14:08:33,900 sagemaker-containers INFO     No GPUs detected (normal if no gpus installed)\u001b[0m\n",
      "\u001b[34m2023-07-21 14:08:34,095 sagemaker-containers INFO     No GPUs detected (normal if no gpus installed)\u001b[0m\n",
      "\u001b[34m2023-07-21 14:08:34,111 sagemaker-containers INFO     No GPUs detected (normal if no gpus installed)\u001b[0m\n",
      "\u001b[34m2023-07-21 14:08:34,122 sagemaker-containers INFO     Invoking user script\u001b[0m\n",
      "\u001b[34mTraining Env:\u001b[0m\n",
      "\u001b[34m{\n",
      "    \"additional_framework_parameters\": {},\n",
      "    \"channel_input_dirs\": {\n",
      "        \"training\": \"/opt/ml/input/data/training\",\n",
      "        \"validation\": \"/opt/ml/input/data/validation\"\n",
      "    },\n",
      "    \"current_host\": \"algo-1\",\n",
      "    \"framework_module\": \"sagemaker_tensorflow_container.training:main\",\n",
      "    \"hosts\": [\n",
      "        \"algo-1\"\n",
      "    ],\n",
      "    \"hyperparameters\": {\n",
      "        \"batch-size\": 32,\n",
      "        \"epochs\": 13,\n",
      "        \"learning-rate\": 0.001,\n",
      "        \"model_dir\": \"s3://sagemaker-us-east-1-442851008302/sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405/model\"\n",
      "    },\n",
      "    \"input_config_dir\": \"/opt/ml/input/config\",\n",
      "    \"input_data_config\": {\n",
      "        \"training\": {\n",
      "            \"TrainingInputMode\": \"File\",\n",
      "            \"S3DistributionType\": \"FullyReplicated\",\n",
      "            \"RecordWrapperType\": \"None\"\n",
      "        },\n",
      "        \"validation\": {\n",
      "            \"TrainingInputMode\": \"File\",\n",
      "            \"S3DistributionType\": \"FullyReplicated\",\n",
      "            \"RecordWrapperType\": \"None\"\n",
      "        }\n",
      "    },\n",
      "    \"input_dir\": \"/opt/ml/input\",\n",
      "    \"is_master\": true,\n",
      "    \"job_name\": \"sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405\",\n",
      "    \"log_level\": 20,\n",
      "    \"master_hostname\": \"algo-1\",\n",
      "    \"model_dir\": \"/opt/ml/model\",\n",
      "    \"module_dir\": \"s3://sagemaker-us-east-1-442851008302/sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405/source/sourcedir.tar.gz\",\n",
      "    \"module_name\": \"trainCNN\",\n",
      "    \"network_interface_name\": \"eth0\",\n",
      "    \"num_cpus\": 8,\n",
      "    \"num_gpus\": 0,\n",
      "    \"output_data_dir\": \"/opt/ml/output/data\",\n",
      "    \"output_dir\": \"/opt/ml/output\",\n",
      "    \"output_intermediate_dir\": \"/opt/ml/output/intermediate\",\n",
      "    \"resource_config\": {\n",
      "        \"current_host\": \"algo-1\",\n",
      "        \"current_instance_type\": \"ml.c4.2xlarge\",\n",
      "        \"current_group_name\": \"homogeneousCluster\",\n",
      "        \"hosts\": [\n",
      "            \"algo-1\"\n",
      "        ],\n",
      "        \"instance_groups\": [\n",
      "            {\n",
      "                \"instance_group_name\": \"homogeneousCluster\",\n",
      "                \"instance_type\": \"ml.c4.2xlarge\",\n",
      "                \"hosts\": [\n",
      "                    \"algo-1\"\n",
      "                ]\n",
      "            }\n",
      "        ],\n",
      "        \"network_interface_name\": \"eth0\"\n",
      "    },\n",
      "    \"user_entry_point\": \"trainCNN.py\"\u001b[0m\n",
      "\u001b[34m}\u001b[0m\n",
      "\u001b[34mEnvironment variables:\u001b[0m\n",
      "\u001b[34mSM_HOSTS=[\"algo-1\"]\u001b[0m\n",
      "\u001b[34mSM_NETWORK_INTERFACE_NAME=eth0\u001b[0m\n",
      "\u001b[34mSM_HPS={\"batch-size\":32,\"epochs\":13,\"learning-rate\":0.001,\"model_dir\":\"s3://sagemaker-us-east-1-442851008302/sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405/model\"}\u001b[0m\n",
      "\u001b[34mSM_USER_ENTRY_POINT=trainCNN.py\u001b[0m\n",
      "\u001b[34mSM_FRAMEWORK_PARAMS={}\u001b[0m\n",
      "\u001b[34mSM_RESOURCE_CONFIG={\"current_group_name\":\"homogeneousCluster\",\"current_host\":\"algo-1\",\"current_instance_type\":\"ml.c4.2xlarge\",\"hosts\":[\"algo-1\"],\"instance_groups\":[{\"hosts\":[\"algo-1\"],\"instance_group_name\":\"homogeneousCluster\",\"instance_type\":\"ml.c4.2xlarge\"}],\"network_interface_name\":\"eth0\"}\u001b[0m\n",
      "\u001b[34mSM_INPUT_DATA_CONFIG={\"training\":{\"RecordWrapperType\":\"None\",\"S3DistributionType\":\"FullyReplicated\",\"TrainingInputMode\":\"File\"},\"validation\":{\"RecordWrapperType\":\"None\",\"S3DistributionType\":\"FullyReplicated\",\"TrainingInputMode\":\"File\"}}\u001b[0m\n",
      "\u001b[34mSM_OUTPUT_DATA_DIR=/opt/ml/output/data\u001b[0m\n",
      "\u001b[34mSM_CHANNELS=[\"training\",\"validation\"]\u001b[0m\n",
      "\u001b[34mSM_CURRENT_HOST=algo-1\u001b[0m\n",
      "\u001b[34mSM_MODULE_NAME=trainCNN\u001b[0m\n",
      "\u001b[34mSM_LOG_LEVEL=20\u001b[0m\n",
      "\u001b[34mSM_FRAMEWORK_MODULE=sagemaker_tensorflow_container.training:main\u001b[0m\n",
      "\u001b[34mSM_INPUT_DIR=/opt/ml/input\u001b[0m\n",
      "\u001b[34mSM_INPUT_CONFIG_DIR=/opt/ml/input/config\u001b[0m\n",
      "\u001b[34mSM_OUTPUT_DIR=/opt/ml/output\u001b[0m\n",
      "\u001b[34mSM_NUM_CPUS=8\u001b[0m\n",
      "\u001b[34mSM_NUM_GPUS=0\u001b[0m\n",
      "\u001b[34mSM_MODEL_DIR=/opt/ml/model\u001b[0m\n",
      "\u001b[34mSM_MODULE_DIR=s3://sagemaker-us-east-1-442851008302/sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405/source/sourcedir.tar.gz\u001b[0m\n",
      "\u001b[34mSM_TRAINING_ENV={\"additional_framework_parameters\":{},\"channel_input_dirs\":{\"training\":\"/opt/ml/input/data/training\",\"validation\":\"/opt/ml/input/data/validation\"},\"current_host\":\"algo-1\",\"framework_module\":\"sagemaker_tensorflow_container.training:main\",\"hosts\":[\"algo-1\"],\"hyperparameters\":{\"batch-size\":32,\"epochs\":13,\"learning-rate\":0.001,\"model_dir\":\"s3://sagemaker-us-east-1-442851008302/sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405/model\"},\"input_config_dir\":\"/opt/ml/input/config\",\"input_data_config\":{\"training\":{\"RecordWrapperType\":\"None\",\"S3DistributionType\":\"FullyReplicated\",\"TrainingInputMode\":\"File\"},\"validation\":{\"RecordWrapperType\":\"None\",\"S3DistributionType\":\"FullyReplicated\",\"TrainingInputMode\":\"File\"}},\"input_dir\":\"/opt/ml/input\",\"is_master\":true,\"job_name\":\"sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405\",\"log_level\":20,\"master_hostname\":\"algo-1\",\"model_dir\":\"/opt/ml/model\",\"module_dir\":\"s3://sagemaker-us-east-1-442851008302/sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405/source/sourcedir.tar.gz\",\"module_name\":\"trainCNN\",\"network_interface_name\":\"eth0\",\"num_cpus\":8,\"num_gpus\":0,\"output_data_dir\":\"/opt/ml/output/data\",\"output_dir\":\"/opt/ml/output\",\"output_intermediate_dir\":\"/opt/ml/output/intermediate\",\"resource_config\":{\"current_group_name\":\"homogeneousCluster\",\"current_host\":\"algo-1\",\"current_instance_type\":\"ml.c4.2xlarge\",\"hosts\":[\"algo-1\"],\"instance_groups\":[{\"hosts\":[\"algo-1\"],\"instance_group_name\":\"homogeneousCluster\",\"instance_type\":\"ml.c4.2xlarge\"}],\"network_interface_name\":\"eth0\"},\"user_entry_point\":\"trainCNN.py\"}\u001b[0m\n",
      "\u001b[34mSM_USER_ARGS=[\"--batch-size\",\"32\",\"--epochs\",\"13\",\"--learning-rate\",\"0.001\",\"--model_dir\",\"s3://sagemaker-us-east-1-442851008302/sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405/model\"]\u001b[0m\n",
      "\u001b[34mSM_OUTPUT_INTERMEDIATE_DIR=/opt/ml/output/intermediate\u001b[0m\n",
      "\u001b[34mSM_CHANNEL_TRAINING=/opt/ml/input/data/training\u001b[0m\n",
      "\u001b[34mSM_CHANNEL_VALIDATION=/opt/ml/input/data/validation\u001b[0m\n",
      "\u001b[34mSM_HP_BATCH-SIZE=32\u001b[0m\n",
      "\u001b[34mSM_HP_EPOCHS=13\u001b[0m\n",
      "\u001b[34mSM_HP_LEARNING-RATE=0.001\u001b[0m\n",
      "\u001b[34mSM_HP_MODEL_DIR=s3://sagemaker-us-east-1-442851008302/sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405/model\u001b[0m\n",
      "\u001b[34mPYTHONPATH=/opt/ml/code:/usr/local/bin:/usr/lib/python36.zip:/usr/lib/python3.6:/usr/lib/python3.6/lib-dynload:/usr/local/lib/python3.6/dist-packages:/usr/lib/python3/dist-packages\u001b[0m\n",
      "\u001b[34mInvoking script with the following command:\u001b[0m\n",
      "\u001b[34m/usr/bin/python trainCNN.py --batch-size 32 --epochs 13 --learning-rate 0.001 --model_dir s3://sagemaker-us-east-1-442851008302/sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405/model\n",
      "  File \"trainCNN.py\", line 9\n",
      "    from tensorflow.keras,utils import multi_gpu_model\n",
      "                         ^\u001b[0m\n",
      "\u001b[34mSyntaxError: invalid syntax\u001b[0m\n",
      "\u001b[34m2023-07-21 14:08:34,150 sagemaker-containers ERROR    ExecuteUserScriptError:\u001b[0m\n",
      "\u001b[34mCommand \"/usr/bin/python trainCNN.py --batch-size 32 --epochs 13 --learning-rate 0.001 --model_dir s3://sagemaker-us-east-1-442851008302/sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405/model\"\u001b[0m\n",
      "\n",
      "2023-07-21 14:08:45 Uploading - Uploading generated training model\n",
      "2023-07-21 14:08:45 Failed - Training job failed\n"
     ]
    },
    {
     "ename": "UnexpectedStatusException",
     "evalue": "Error for Training job sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405: Failed. Reason: AlgorithmError: ExecuteUserScriptError:\nCommand \"/usr/bin/python trainCNN.py --batch-size 32 --epochs 13 --learning-rate 0.001 --model_dir s3://sagemaker-us-east-1-442851008302/sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405/model\", exit code: 1",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mUnexpectedStatusException\u001b[0m                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-20-f93828a3e509>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mtf_estimator\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m{\u001b[0m\u001b[0;34m'training'\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mtraining_path\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'validation'\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mvalid_path\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/opt/conda/lib/python3.7/site-packages/sagemaker/workflow/pipeline_context.py\u001b[0m in \u001b[0;36mwrapper\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    298\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0m_StepArguments\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mretrieve_caller_name\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself_instance\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrun_func\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    299\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 300\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mrun_func\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    301\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    302\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0mwrapper\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.7/site-packages/sagemaker/estimator.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, inputs, wait, logs, job_name, experiment_config)\u001b[0m\n\u001b[1;32m   1250\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mjobs\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlatest_training_job\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1251\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mwait\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1252\u001b[0;31m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlatest_training_job\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwait\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlogs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mlogs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1253\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1254\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_compilation_job_name\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.7/site-packages/sagemaker/estimator.py\u001b[0m in \u001b[0;36mwait\u001b[0;34m(self, logs)\u001b[0m\n\u001b[1;32m   2396\u001b[0m         \u001b[0;31m# If logs are requested, call logs_for_jobs.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2397\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mlogs\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0;34m\"None\"\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2398\u001b[0;31m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msagemaker_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlogs_for_job\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mjob_name\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mwait\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlog_type\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mlogs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2399\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2400\u001b[0m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msagemaker_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwait_for_job\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mjob_name\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.7/site-packages/sagemaker/session.py\u001b[0m in \u001b[0;36mlogs_for_job\u001b[0;34m(self, job_name, wait, poll, log_type, timeout)\u001b[0m\n\u001b[1;32m   4759\u001b[0m             \u001b[0mexceptions\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mUnexpectedStatusException\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mIf\u001b[0m \u001b[0mwaiting\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mthe\u001b[0m \u001b[0mtraining\u001b[0m \u001b[0mjob\u001b[0m \u001b[0mfails\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   4760\u001b[0m         \"\"\"\n\u001b[0;32m-> 4761\u001b[0;31m         \u001b[0m_logs_for_job\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mboto_session\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mjob_name\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mwait\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpoll\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlog_type\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtimeout\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   4762\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   4763\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mlogs_for_processing_job\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mjob_name\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mwait\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpoll\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m10\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.7/site-packages/sagemaker/session.py\u001b[0m in \u001b[0;36m_logs_for_job\u001b[0;34m(boto_session, job_name, wait, poll, log_type, timeout)\u001b[0m\n\u001b[1;32m   6628\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   6629\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mwait\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 6630\u001b[0;31m         \u001b[0m_check_job_status\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mjob_name\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdescription\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"TrainingJobStatus\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   6631\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mdot\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   6632\u001b[0m             \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.7/site-packages/sagemaker/session.py\u001b[0m in \u001b[0;36m_check_job_status\u001b[0;34m(job, desc, status_key_name)\u001b[0m\n\u001b[1;32m   6684\u001b[0m             \u001b[0mmessage\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mmessage\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   6685\u001b[0m             \u001b[0mallowed_statuses\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m\"Completed\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"Stopped\"\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 6686\u001b[0;31m             \u001b[0mactual_status\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mstatus\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   6687\u001b[0m         )\n\u001b[1;32m   6688\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mUnexpectedStatusException\u001b[0m: Error for Training job sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405: Failed. Reason: AlgorithmError: ExecuteUserScriptError:\nCommand \"/usr/bin/python trainCNN.py --batch-size 32 --epochs 13 --learning-rate 0.001 --model_dir s3://sagemaker-us-east-1-442851008302/sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405/model\", exit code: 1"
     ]
    }
   ],
   "source": [
    "tf_estimator.fit({'training': training_path, 'validation': valid_path})"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0eecf4aa-899c-4056-95c5-983cf6d8b02d",
   "metadata": {},
   "source": [
    "## DEPLOYING THE TRAINED MODEL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "296fb822-6c52-4904-ac2f-d9ca19fa3043",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:sagemaker.tensorflow.model:image_uri is not presented, retrieving image_uri based on instance_type, framework etc.\n",
      "INFO:sagemaker:Creating model with name: sagemaker-tensorflow-scriptmode-2023-07-21-14-13-23-593\n"
     ]
    },
    {
     "ename": "ClientError",
     "evalue": "An error occurred (ValidationException) when calling the CreateModel operation: Could not find model data at s3://sagemaker-us-east-1-442851008302/sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405/output/model.tar.gz.",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mClientError\u001b[0m                               Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-21-27a21dca063a>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      5\u001b[0m tf_predictor = tf_estimator.deploy(initial_instance_count = 1,\n\u001b[1;32m      6\u001b[0m                                    \u001b[0minstance_type\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m'ml.t2.medium'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 7\u001b[0;31m                                    endpoint_name = tf_endpoint_name)\n\u001b[0m",
      "\u001b[0;32m/opt/conda/lib/python3.7/site-packages/sagemaker/estimator.py\u001b[0m in \u001b[0;36mdeploy\u001b[0;34m(self, initial_instance_count, instance_type, serializer, deserializer, accelerator_type, endpoint_name, use_compiled_model, wait, model_name, kms_key, data_capture_config, tags, serverless_inference_config, async_inference_config, volume_size, model_data_download_timeout, container_startup_health_check_timeout, inference_recommendation_id, explainer_config, **kwargs)\u001b[0m\n\u001b[1;32m   1578\u001b[0m             \u001b[0mmodel_data_download_timeout\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mmodel_data_download_timeout\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1579\u001b[0m             \u001b[0mcontainer_startup_health_check_timeout\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mcontainer_startup_health_check_timeout\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1580\u001b[0;31m             \u001b[0minference_recommendation_id\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0minference_recommendation_id\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1581\u001b[0m         )\n\u001b[1;32m   1582\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.7/site-packages/sagemaker/tensorflow/model.py\u001b[0m in \u001b[0;36mdeploy\u001b[0;34m(self, initial_instance_count, instance_type, serializer, deserializer, accelerator_type, endpoint_name, tags, kms_key, wait, data_capture_config, async_inference_config, serverless_inference_config, volume_size, model_data_download_timeout, container_startup_health_check_timeout, inference_recommendation_id, explainer_config, **kwargs)\u001b[0m\n\u001b[1;32m    361\u001b[0m             \u001b[0minference_recommendation_id\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0minference_recommendation_id\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    362\u001b[0m             \u001b[0mexplainer_config\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mexplainer_config\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 363\u001b[0;31m             \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    364\u001b[0m         )\n\u001b[1;32m    365\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.7/site-packages/sagemaker/model.py\u001b[0m in \u001b[0;36mdeploy\u001b[0;34m(self, initial_instance_count, instance_type, serializer, deserializer, accelerator_type, endpoint_name, tags, kms_key, wait, data_capture_config, async_inference_config, serverless_inference_config, volume_size, model_data_download_timeout, container_startup_health_check_timeout, inference_recommendation_id, explainer_config, **kwargs)\u001b[0m\n\u001b[1;32m   1299\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1300\u001b[0m         self._create_sagemaker_model(\n\u001b[0;32m-> 1301\u001b[0;31m             \u001b[0minstance_type\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maccelerator_type\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtags\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mserverless_inference_config\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1302\u001b[0m         )\n\u001b[1;32m   1303\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.7/site-packages/sagemaker/model.py\u001b[0m in \u001b[0;36m_create_sagemaker_model\u001b[0;34m(self, instance_type, accelerator_type, tags, serverless_inference_config)\u001b[0m\n\u001b[1;32m    759\u001b[0m             \u001b[0mtags\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mtags\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    760\u001b[0m         )\n\u001b[0;32m--> 761\u001b[0;31m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msagemaker_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcreate_model\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m**\u001b[0m\u001b[0mcreate_model_args\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    762\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    763\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_ensure_base_name_if_needed\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mimage_uri\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mscript_uri\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel_uri\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.7/site-packages/sagemaker/session.py\u001b[0m in \u001b[0;36mcreate_model\u001b[0;34m(self, name, role, container_defs, vpc_config, enable_network_isolation, primary_container, tags)\u001b[0m\n\u001b[1;32m   3489\u001b[0m                     \u001b[0;32mraise\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   3490\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 3491\u001b[0;31m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_intercept_create_request\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcreate_model_request\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msubmit\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcreate_model\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__name__\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   3492\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   3493\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.7/site-packages/sagemaker/session.py\u001b[0m in \u001b[0;36m_intercept_create_request\u001b[0;34m(self, request, create, func_name)\u001b[0m\n\u001b[1;32m   5368\u001b[0m             \u001b[0mfunc_name\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mstr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mthe\u001b[0m \u001b[0mname\u001b[0m \u001b[0mof\u001b[0m \u001b[0mthe\u001b[0m \u001b[0mfunction\u001b[0m \u001b[0mneeded\u001b[0m \u001b[0mintercepting\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   5369\u001b[0m         \"\"\"\n\u001b[0;32m-> 5370\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mcreate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrequest\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   5371\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   5372\u001b[0m     def _create_inference_recommendations_job_request(\n",
      "\u001b[0;32m/opt/conda/lib/python3.7/site-packages/sagemaker/session.py\u001b[0m in \u001b[0;36msubmit\u001b[0;34m(request)\u001b[0m\n\u001b[1;32m   3477\u001b[0m             \u001b[0mLOGGER\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdebug\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"CreateModel request: %s\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mjson\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdumps\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrequest\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mindent\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m4\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   3478\u001b[0m             \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 3479\u001b[0;31m                 \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msagemaker_client\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcreate_model\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m**\u001b[0m\u001b[0mrequest\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   3480\u001b[0m             \u001b[0;32mexcept\u001b[0m \u001b[0mClientError\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   3481\u001b[0m                 \u001b[0merror_code\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mresponse\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m\"Error\"\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m\"Code\"\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.7/site-packages/botocore/client.py\u001b[0m in \u001b[0;36m_api_call\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m    528\u001b[0m                 )\n\u001b[1;32m    529\u001b[0m             \u001b[0;31m# The \"self\" in this scope is referring to the BaseClient.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 530\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_make_api_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0moperation_name\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    531\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    532\u001b[0m         \u001b[0m_api_call\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__name__\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpy_operation_name\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.7/site-packages/botocore/client.py\u001b[0m in \u001b[0;36m_make_api_call\u001b[0;34m(self, operation_name, api_params)\u001b[0m\n\u001b[1;32m    962\u001b[0m             \u001b[0merror_code\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mparsed_response\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Error\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Code\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    963\u001b[0m             \u001b[0merror_class\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexceptions\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfrom_code\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0merror_code\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 964\u001b[0;31m             \u001b[0;32mraise\u001b[0m \u001b[0merror_class\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mparsed_response\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moperation_name\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    965\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    966\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0mparsed_response\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mClientError\u001b[0m: An error occurred (ValidationException) when calling the CreateModel operation: Could not find model data at s3://sagemaker-us-east-1-442851008302/sagemaker-tensorflow-scriptmode-2023-07-21-14-06-28-405/output/model.tar.gz."
     ]
    }
   ],
   "source": [
    "import time\n",
    "\n",
    "tf_endpoint_name = 'trafficsign-' + time.strftime(\"%Y-%m-%d-%H-%M-%S\", time.gmtime())\n",
    "\n",
    "tf_predictor = tf_estimator.deploy(initial_instance_count = 1,\n",
    "                                   instance_type = 'ml.t2.medium',\n",
    "                                   endpoint_name = tf_endpoint_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "793398b1-78ea-4bd6-82e4-9d8b20c2d11f",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "availableInstances": [
   {
    "_defaultOrder": 0,
    "_isFastLaunch": true,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 4,
    "name": "ml.t3.medium",
    "vcpuNum": 2
   },
   {
    "_defaultOrder": 1,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 8,
    "name": "ml.t3.large",
    "vcpuNum": 2
   },
   {
    "_defaultOrder": 2,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 16,
    "name": "ml.t3.xlarge",
    "vcpuNum": 4
   },
   {
    "_defaultOrder": 3,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 32,
    "name": "ml.t3.2xlarge",
    "vcpuNum": 8
   },
   {
    "_defaultOrder": 4,
    "_isFastLaunch": true,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 8,
    "name": "ml.m5.large",
    "vcpuNum": 2
   },
   {
    "_defaultOrder": 5,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 16,
    "name": "ml.m5.xlarge",
    "vcpuNum": 4
   },
   {
    "_defaultOrder": 6,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 32,
    "name": "ml.m5.2xlarge",
    "vcpuNum": 8
   },
   {
    "_defaultOrder": 7,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 64,
    "name": "ml.m5.4xlarge",
    "vcpuNum": 16
   },
   {
    "_defaultOrder": 8,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 128,
    "name": "ml.m5.8xlarge",
    "vcpuNum": 32
   },
   {
    "_defaultOrder": 9,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 192,
    "name": "ml.m5.12xlarge",
    "vcpuNum": 48
   },
   {
    "_defaultOrder": 10,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 256,
    "name": "ml.m5.16xlarge",
    "vcpuNum": 64
   },
   {
    "_defaultOrder": 11,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 384,
    "name": "ml.m5.24xlarge",
    "vcpuNum": 96
   },
   {
    "_defaultOrder": 12,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 8,
    "name": "ml.m5d.large",
    "vcpuNum": 2
   },
   {
    "_defaultOrder": 13,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 16,
    "name": "ml.m5d.xlarge",
    "vcpuNum": 4
   },
   {
    "_defaultOrder": 14,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 32,
    "name": "ml.m5d.2xlarge",
    "vcpuNum": 8
   },
   {
    "_defaultOrder": 15,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 64,
    "name": "ml.m5d.4xlarge",
    "vcpuNum": 16
   },
   {
    "_defaultOrder": 16,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 128,
    "name": "ml.m5d.8xlarge",
    "vcpuNum": 32
   },
   {
    "_defaultOrder": 17,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 192,
    "name": "ml.m5d.12xlarge",
    "vcpuNum": 48
   },
   {
    "_defaultOrder": 18,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 256,
    "name": "ml.m5d.16xlarge",
    "vcpuNum": 64
   },
   {
    "_defaultOrder": 19,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 384,
    "name": "ml.m5d.24xlarge",
    "vcpuNum": 96
   },
   {
    "_defaultOrder": 20,
    "_isFastLaunch": false,
    "category": "General purpose",
    "gpuNum": 0,
    "hideHardwareSpecs": true,
    "memoryGiB": 0,
    "name": "ml.geospatial.interactive",
    "supportedImageNames": [
     "sagemaker-geospatial-v1-0"
    ],
    "vcpuNum": 0
   },
   {
    "_defaultOrder": 21,
    "_isFastLaunch": true,
    "category": "Compute optimized",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 4,
    "name": "ml.c5.large",
    "vcpuNum": 2
   },
   {
    "_defaultOrder": 22,
    "_isFastLaunch": false,
    "category": "Compute optimized",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 8,
    "name": "ml.c5.xlarge",
    "vcpuNum": 4
   },
   {
    "_defaultOrder": 23,
    "_isFastLaunch": false,
    "category": "Compute optimized",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 16,
    "name": "ml.c5.2xlarge",
    "vcpuNum": 8
   },
   {
    "_defaultOrder": 24,
    "_isFastLaunch": false,
    "category": "Compute optimized",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 32,
    "name": "ml.c5.4xlarge",
    "vcpuNum": 16
   },
   {
    "_defaultOrder": 25,
    "_isFastLaunch": false,
    "category": "Compute optimized",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 72,
    "name": "ml.c5.9xlarge",
    "vcpuNum": 36
   },
   {
    "_defaultOrder": 26,
    "_isFastLaunch": false,
    "category": "Compute optimized",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 96,
    "name": "ml.c5.12xlarge",
    "vcpuNum": 48
   },
   {
    "_defaultOrder": 27,
    "_isFastLaunch": false,
    "category": "Compute optimized",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 144,
    "name": "ml.c5.18xlarge",
    "vcpuNum": 72
   },
   {
    "_defaultOrder": 28,
    "_isFastLaunch": false,
    "category": "Compute optimized",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 192,
    "name": "ml.c5.24xlarge",
    "vcpuNum": 96
   },
   {
    "_defaultOrder": 29,
    "_isFastLaunch": true,
    "category": "Accelerated computing",
    "gpuNum": 1,
    "hideHardwareSpecs": false,
    "memoryGiB": 16,
    "name": "ml.g4dn.xlarge",
    "vcpuNum": 4
   },
   {
    "_defaultOrder": 30,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 1,
    "hideHardwareSpecs": false,
    "memoryGiB": 32,
    "name": "ml.g4dn.2xlarge",
    "vcpuNum": 8
   },
   {
    "_defaultOrder": 31,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 1,
    "hideHardwareSpecs": false,
    "memoryGiB": 64,
    "name": "ml.g4dn.4xlarge",
    "vcpuNum": 16
   },
   {
    "_defaultOrder": 32,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 1,
    "hideHardwareSpecs": false,
    "memoryGiB": 128,
    "name": "ml.g4dn.8xlarge",
    "vcpuNum": 32
   },
   {
    "_defaultOrder": 33,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 4,
    "hideHardwareSpecs": false,
    "memoryGiB": 192,
    "name": "ml.g4dn.12xlarge",
    "vcpuNum": 48
   },
   {
    "_defaultOrder": 34,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 1,
    "hideHardwareSpecs": false,
    "memoryGiB": 256,
    "name": "ml.g4dn.16xlarge",
    "vcpuNum": 64
   },
   {
    "_defaultOrder": 35,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 1,
    "hideHardwareSpecs": false,
    "memoryGiB": 61,
    "name": "ml.p3.2xlarge",
    "vcpuNum": 8
   },
   {
    "_defaultOrder": 36,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 4,
    "hideHardwareSpecs": false,
    "memoryGiB": 244,
    "name": "ml.p3.8xlarge",
    "vcpuNum": 32
   },
   {
    "_defaultOrder": 37,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 8,
    "hideHardwareSpecs": false,
    "memoryGiB": 488,
    "name": "ml.p3.16xlarge",
    "vcpuNum": 64
   },
   {
    "_defaultOrder": 38,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 8,
    "hideHardwareSpecs": false,
    "memoryGiB": 768,
    "name": "ml.p3dn.24xlarge",
    "vcpuNum": 96
   },
   {
    "_defaultOrder": 39,
    "_isFastLaunch": false,
    "category": "Memory Optimized",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 16,
    "name": "ml.r5.large",
    "vcpuNum": 2
   },
   {
    "_defaultOrder": 40,
    "_isFastLaunch": false,
    "category": "Memory Optimized",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 32,
    "name": "ml.r5.xlarge",
    "vcpuNum": 4
   },
   {
    "_defaultOrder": 41,
    "_isFastLaunch": false,
    "category": "Memory Optimized",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 64,
    "name": "ml.r5.2xlarge",
    "vcpuNum": 8
   },
   {
    "_defaultOrder": 42,
    "_isFastLaunch": false,
    "category": "Memory Optimized",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 128,
    "name": "ml.r5.4xlarge",
    "vcpuNum": 16
   },
   {
    "_defaultOrder": 43,
    "_isFastLaunch": false,
    "category": "Memory Optimized",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 256,
    "name": "ml.r5.8xlarge",
    "vcpuNum": 32
   },
   {
    "_defaultOrder": 44,
    "_isFastLaunch": false,
    "category": "Memory Optimized",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 384,
    "name": "ml.r5.12xlarge",
    "vcpuNum": 48
   },
   {
    "_defaultOrder": 45,
    "_isFastLaunch": false,
    "category": "Memory Optimized",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 512,
    "name": "ml.r5.16xlarge",
    "vcpuNum": 64
   },
   {
    "_defaultOrder": 46,
    "_isFastLaunch": false,
    "category": "Memory Optimized",
    "gpuNum": 0,
    "hideHardwareSpecs": false,
    "memoryGiB": 768,
    "name": "ml.r5.24xlarge",
    "vcpuNum": 96
   },
   {
    "_defaultOrder": 47,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 1,
    "hideHardwareSpecs": false,
    "memoryGiB": 16,
    "name": "ml.g5.xlarge",
    "vcpuNum": 4
   },
   {
    "_defaultOrder": 48,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 1,
    "hideHardwareSpecs": false,
    "memoryGiB": 32,
    "name": "ml.g5.2xlarge",
    "vcpuNum": 8
   },
   {
    "_defaultOrder": 49,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 1,
    "hideHardwareSpecs": false,
    "memoryGiB": 64,
    "name": "ml.g5.4xlarge",
    "vcpuNum": 16
   },
   {
    "_defaultOrder": 50,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 1,
    "hideHardwareSpecs": false,
    "memoryGiB": 128,
    "name": "ml.g5.8xlarge",
    "vcpuNum": 32
   },
   {
    "_defaultOrder": 51,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 1,
    "hideHardwareSpecs": false,
    "memoryGiB": 256,
    "name": "ml.g5.16xlarge",
    "vcpuNum": 64
   },
   {
    "_defaultOrder": 52,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 4,
    "hideHardwareSpecs": false,
    "memoryGiB": 192,
    "name": "ml.g5.12xlarge",
    "vcpuNum": 48
   },
   {
    "_defaultOrder": 53,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 4,
    "hideHardwareSpecs": false,
    "memoryGiB": 384,
    "name": "ml.g5.24xlarge",
    "vcpuNum": 96
   },
   {
    "_defaultOrder": 54,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 8,
    "hideHardwareSpecs": false,
    "memoryGiB": 768,
    "name": "ml.g5.48xlarge",
    "vcpuNum": 192
   },
   {
    "_defaultOrder": 55,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 8,
    "hideHardwareSpecs": false,
    "memoryGiB": 1152,
    "name": "ml.p4d.24xlarge",
    "vcpuNum": 96
   },
   {
    "_defaultOrder": 56,
    "_isFastLaunch": false,
    "category": "Accelerated computing",
    "gpuNum": 8,
    "hideHardwareSpecs": false,
    "memoryGiB": 1152,
    "name": "ml.p4de.24xlarge",
    "vcpuNum": 96
   }
  ],
  "instance_type": "ml.t3.medium",
  "kernelspec": {
   "display_name": "Python 3 (Data Science)",
   "language": "python",
   "name": "python3__SAGEMAKER_INTERNAL__arn:aws:sagemaker:us-east-1:081325390199:image/datascience-1.0"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}