BlazingText.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 141,
   "id": "fececad9-b22c-4501-953c-823b951a2380",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package punkt to /home/sagemaker-\n",
      "[nltk_data]     user/nltk_data...\n",
      "[nltk_data]   Package punkt is already up-to-date!\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 141,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "import matplotlib.pyplot as plt\n",
    "import boto3\n",
    "import sagemaker\n",
    "from sagemaker import get_execution_role\n",
    "import nltk\n",
    "nltk.download('punkt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "44355a72-346e-4114-b1f5-43a875f33f30",
   "metadata": {},
   "outputs": [],
   "source": [
    "#getting bucket name\n",
    "s3 = boto3.client('s3')\n",
    "bucket = s3.list_buckets()\n",
    "bucket_names = [bucket['Name'] for bucket in bucket['Buckets']]\n",
    "\n",
    "for bucket_name in bucket_names:\n",
    "    bucket_url = bucket_name\n",
    "    print(f\"Bucket name: {bucket_name}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "58cb636a-c1b6-45fe-8052-6f8860e4c46c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# AWS credentials\n",
    "aws_access_key = 'xxx'\n",
    "aws_secret_access_key = 'xxx'\n",
    "aws_region = 'us-east-2'\n",
    "\n",
    "# Create a SageMaker Session\n",
    "sess = sagemaker.Session()\n",
    "# Get the SageMaker execution role\n",
    "role = get_execution_role()\n",
    "# S3 bucket and file information\n",
    "bucket_name = 'xxx'\n",
    "\n",
    "# Create an S3 client\n",
    "s3 = boto3.client('s3', aws_access_key_id=aws_access_key, aws_secret_access_key=aws_secret_access_key, region_name=aws_region)\n",
    "  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "fbfdb41a-1a76-451d-9e84-60b1c12382b2",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_blazingtext = pd.read_csv('df_clean_labeled_tokenized.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "82b4d9fa-ae70-4a9b-9db6-b7dbd364aa84",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sentiment</th>\n",
       "      <th>tokenized_review</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-1</td>\n",
       "      <td>['Carmencita', 'short', 'video', 'woman', 'dan...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-1</td>\n",
       "      <td>['Like', 'work', 'Autour', 'dune', 'cabine', '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>['1891', 'Thomas', 'Edison', 'staff', 'led', '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>['James', 'J', 'Corbett', 'better', 'known', '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>['film', 'Fred', 'Ott', 'employee', 'Thomas', ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32500</th>\n",
       "      <td>1</td>\n",
       "      <td>['watched', 'Trial', 'Catonsville', 'Nine', 'f...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32501</th>\n",
       "      <td>1</td>\n",
       "      <td>['TRILOGY', 'SWORDSMANSHIP', 'came', 'bit', 's...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32502</th>\n",
       "      <td>-1</td>\n",
       "      <td>['Trinity', 'Sartana', 'originally', 'released...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32503</th>\n",
       "      <td>0</td>\n",
       "      <td>['also', 'find', 'first', 'review', 'film', 'S...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32504</th>\n",
       "      <td>1</td>\n",
       "      <td>['good', 'movie', 'nt', 'find', 'good', 'lot',...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>32505 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       sentiment                                   tokenized_review\n",
       "0             -1  ['Carmencita', 'short', 'video', 'woman', 'dan...\n",
       "1             -1  ['Like', 'work', 'Autour', 'dune', 'cabine', '...\n",
       "2              1  ['1891', 'Thomas', 'Edison', 'staff', 'led', '...\n",
       "3              1  ['James', 'J', 'Corbett', 'better', 'known', '...\n",
       "4              0  ['film', 'Fred', 'Ott', 'employee', 'Thomas', ...\n",
       "...          ...                                                ...\n",
       "32500          1  ['watched', 'Trial', 'Catonsville', 'Nine', 'f...\n",
       "32501          1  ['TRILOGY', 'SWORDSMANSHIP', 'came', 'bit', 's...\n",
       "32502         -1  ['Trinity', 'Sartana', 'originally', 'released...\n",
       "32503          0  ['also', 'find', 'first', 'review', 'film', 'S...\n",
       "32504          1  ['good', 'movie', 'nt', 'find', 'good', 'lot',...\n",
       "\n",
       "[32505 rows x 2 columns]"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_blazingtext"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "70faa462-0623-4b09-9366-9ac7c0bcdb14",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sentiment</th>\n",
       "      <th>tokenized_review</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>__label__-1</td>\n",
       "      <td>['Carmencita', 'short', 'video', 'woman', 'dan...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>__label__-1</td>\n",
       "      <td>['Like', 'work', 'Autour', 'dune', 'cabine', '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>__label__1</td>\n",
       "      <td>['1891', 'Thomas', 'Edison', 'staff', 'led', '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>__label__1</td>\n",
       "      <td>['James', 'J', 'Corbett', 'better', 'known', '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>__label__0</td>\n",
       "      <td>['film', 'Fred', 'Ott', 'employee', 'Thomas', ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     sentiment                                   tokenized_review\n",
       "0  __label__-1  ['Carmencita', 'short', 'video', 'woman', 'dan...\n",
       "1  __label__-1  ['Like', 'work', 'Autour', 'dune', 'cabine', '...\n",
       "2   __label__1  ['1891', 'Thomas', 'Edison', 'staff', 'led', '...\n",
       "3   __label__1  ['James', 'J', 'Corbett', 'better', 'known', '...\n",
       "4   __label__0  ['film', 'Fred', 'Ott', 'employee', 'Thomas', ..."
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Changing the label type for the need of BlazingText fit function\n",
    "def prepare_data(df):\n",
    "    df = df.map(lambda sentiment : '__label__{}'.format(str(sentiment).replace('__label__', '')))\n",
    "    return df\n",
    "\n",
    "\n",
    "\n",
    "df_blazingtext['sentiment'] = prepare_data(df_blazingtext['sentiment'])\n",
    "df_blazingtext.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "f854a907-0194-4256-b5cd-9ccf773de014",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sentiment</th>\n",
       "      <th>tokenized_review</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>__label__-1</td>\n",
       "      <td>['Carmencita', 'short', 'video', 'woman', 'dan...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>__label__-1</td>\n",
       "      <td>['Like', 'work', 'Autour', 'dune', 'cabine', '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>__label__1</td>\n",
       "      <td>['1891', 'Thomas', 'Edison', 'staff', 'led', '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>__label__1</td>\n",
       "      <td>['James', 'J', 'Corbett', 'better', 'known', '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>__label__0</td>\n",
       "      <td>['film', 'Fred', 'Ott', 'employee', 'Thomas', ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32500</th>\n",
       "      <td>__label__1</td>\n",
       "      <td>['watched', 'Trial', 'Catonsville', 'Nine', 'f...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32501</th>\n",
       "      <td>__label__1</td>\n",
       "      <td>['TRILOGY', 'SWORDSMANSHIP', 'came', 'bit', 's...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32502</th>\n",
       "      <td>__label__-1</td>\n",
       "      <td>['Trinity', 'Sartana', 'originally', 'released...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32503</th>\n",
       "      <td>__label__0</td>\n",
       "      <td>['also', 'find', 'first', 'review', 'film', 'S...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32504</th>\n",
       "      <td>__label__1</td>\n",
       "      <td>['good', 'movie', 'nt', 'find', 'good', 'lot',...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>32505 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         sentiment                                   tokenized_review\n",
       "0      __label__-1  ['Carmencita', 'short', 'video', 'woman', 'dan...\n",
       "1      __label__-1  ['Like', 'work', 'Autour', 'dune', 'cabine', '...\n",
       "2       __label__1  ['1891', 'Thomas', 'Edison', 'staff', 'led', '...\n",
       "3       __label__1  ['James', 'J', 'Corbett', 'better', 'known', '...\n",
       "4       __label__0  ['film', 'Fred', 'Ott', 'employee', 'Thomas', ...\n",
       "...            ...                                                ...\n",
       "32500   __label__1  ['watched', 'Trial', 'Catonsville', 'Nine', 'f...\n",
       "32501   __label__1  ['TRILOGY', 'SWORDSMANSHIP', 'came', 'bit', 's...\n",
       "32502  __label__-1  ['Trinity', 'Sartana', 'originally', 'released...\n",
       "32503   __label__0  ['also', 'find', 'first', 'review', 'film', 'S...\n",
       "32504   __label__1  ['good', 'movie', 'nt', 'find', 'good', 'lot',...\n",
       "\n",
       "[32505 rows x 2 columns]"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_blazingtext"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "id": "fd04f0fe-7d89-4f2c-9217-7a8dbb6a2757",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train set size is: 29254\n",
      "Validation set size is: 3251\n"
     ]
    }
   ],
   "source": [
    "\n",
    "# Split all data into 90% train and 10% holdout\n",
    "df_train, df_validation = train_test_split(df_blazingtext, \n",
    "                                           test_size=0.10,\n",
    "                                           stratify=df_blazingtext['sentiment'])\n",
    "\n",
    "labels = ['train', 'validation']\n",
    "sizes = [len(df_train.index), len(df_validation.index)]\n",
    "explode = (0.1, 0)  \n",
    "\n",
    "fig1, ax1 = plt.subplots()\n",
    "\n",
    "ax1.pie(sizes, explode=explode, labels=labels, autopct='%1.1f%%', startangle=90)\n",
    "\n",
    "# Equal aspect ratio ensures that pie is drawn as a circle.\n",
    "ax1.axis('equal')  \n",
    "\n",
    "plt.show()\n",
    "print(f'Train set size is: {len(df_train)}' )\n",
    "print(f'Validation set size is: {len(df_validation)}' )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "id": "0548b425-660d-4780-b6c3-c6941ec42480",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Save result as CSV\n",
    "blazingtext_train_path = './train.csv'\n",
    "df_train[['sentiment', 'tokenized_review']].to_csv(blazingtext_train_path, index=False, header=False, sep=' ')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "id": "695da23e-9258-4c6f-ab3f-36676b6b7381",
   "metadata": {},
   "outputs": [],
   "source": [
    "blazingtext_validation_path = './validation.csv'\n",
    "df_validation[['sentiment', 'tokenized_review']].to_csv(blazingtext_validation_path, index=False, header=False, sep=' ')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "id": "b6429f3b-bf1e-4856-a785-abe49d20b54d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train and Validation sets uploaded to S3\n"
     ]
    }
   ],
   "source": [
    "#Upload train and validation set for blazingtext to s3\n",
    "train_s3_uri = sess.upload_data(bucket=bucket_url, key_prefix='blazingtext/data', path=blazingtext_train_path)\n",
    "validation_s3_uri = sess.upload_data(bucket=bucket_url, key_prefix='blazingtext/data', path=blazingtext_validation_path)\n",
    "print(\"Train and Validation sets uploaded to S3\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7e449e8a-ec23-4d3c-b17b-5e67dda4c890",
   "metadata": {},
   "source": [
    "#### Train the BlazingText model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "id": "78537fe4-c000-41bc-8dc1-eff2dd2d7c9e",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:sagemaker.image_uris:Same images used for training and inference. Defaulting to image scope: inference.\n",
      "INFO:sagemaker.image_uris:Defaulting to the only supported framework/algorithm version: 1.\n",
      "INFO:sagemaker.image_uris:Ignoring unnecessary instance type: None.\n"
     ]
    }
   ],
   "source": [
    "image_uri = sagemaker.image_uris.retrieve(\n",
    "    region=aws_region,\n",
    "    framework='blazingtext'\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "id": "9c886fad-d4b7-48f6-ae01-7e58e30a787b",
   "metadata": {},
   "outputs": [],
   "source": [
    "estimator = sagemaker.estimator.Estimator(\n",
    "    image_uri=image_uri,\n",
    "    role=role, \n",
    "    instance_count=1, \n",
    "    instance_type='ml.m5.large',\n",
    "    volume_size=30,\n",
    "    max_run=7200,\n",
    "    sagemaker_session=sess\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "id": "5feb07ce-69a7-4188-a917-02e5a69c3251",
   "metadata": {},
   "outputs": [],
   "source": [
    "estimator.set_hyperparameters(mode='supervised',   # supervised (text classification)\n",
    "                              epochs=15,           # number of complete passes through the dataset: 5 - 15\n",
    "                              learning_rate=0.001,  # step size for the  numerical optimizer: 0.005 - 0.01\n",
    "                              min_count=2,         # discard words that appear less than this number: 0 - 100                              \n",
    "                              vector_dim=300,      # number of dimensions in vector space: 32-300\n",
    "                              word_ngrams=3)       # number of words in a word n-gram: 1 - 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "id": "a292d930-7100-4a49-9348-f939f97f9857",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_data = sagemaker.inputs.TrainingInput(\n",
    "    train_s3_uri,\n",
    "    distribution='FullyReplicated', \n",
    "    content_type='text/plain', \n",
    "    s3_data_type='S3Prefix'\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "id": "11992401-f590-4ade-bd9a-684f3f5a00da",
   "metadata": {},
   "outputs": [],
   "source": [
    "validation_data = sagemaker.inputs.TrainingInput(\n",
    "    validation_s3_uri,\n",
    "    distribution='FullyReplicated', \n",
    "    content_type='text/plain', \n",
    "    s3_data_type='S3Prefix'\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "id": "8dd29b3b-b75d-47c4-bfcd-a0699d2db31a",
   "metadata": {},
   "outputs": [],
   "source": [
    "data_channels = {\n",
    "    'train': train_data, \n",
    "    'validation': validation_data \n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "id": "837f9ae7-c9a5-41ca-9477-f5e12bfa09d7",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:sagemaker:Creating training-job with name: blazingtext-2023-12-20-22-15-41-519\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Job Name:  blazingtext-2023-12-20-22-15-41-519\n"
     ]
    }
   ],
   "source": [
    "estimator.fit(\n",
    "    inputs=data_channels,\n",
    "    wait=False\n",
    ")\n",
    "\n",
    "training_job_name = estimator.latest_training_job.name\n",
    "print('Training Job Name:  {}'.format(training_job_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "id": "d6cfd100-d2e0-4fe7-ae48-95e66e0c636c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "2023-12-20 22:15:42 Starting - Starting the training job.\n",
      "2023-12-20 22:15:55 Starting - Preparing the instances for training............\n",
      "2023-12-20 22:17:02 Downloading - Downloading input data........\n",
      "2023-12-20 22:17:48 Downloading - Downloading the training image..\n",
      "2023-12-20 22:18:03 Training - Training image download completed. Training in progress..................\n",
      "2023-12-20 22:19:34 Uploading - Uploading generated training model.....................\n",
      "2023-12-20 22:21:25 Completed - Training job completed\n",
      "CPU times: user 294 ms, sys: 40.8 ms, total: 335 ms\n",
      "Wall time: 5min 43s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "estimator.latest_training_job.wait(logs=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "id": "bca28bbc-fbff-46e5-8a82-536e90aff27c",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:sagemaker.analytics:Warning: No metrics called train:mean_rho found\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>timestamp</th>\n",
       "      <th>metric_name</th>\n",
       "      <th>value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>train:accuracy</td>\n",
       "      <td>0.4739</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>validation:accuracy</td>\n",
       "      <td>0.4752</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   timestamp          metric_name   value\n",
       "0        0.0       train:accuracy  0.4739\n",
       "1        0.0  validation:accuracy  0.4752"
      ]
     },
     "execution_count": 139,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "estimator.training_job_analytics.dataframe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "id": "14d2b054-cb5e-45f3-bd75-3c32691a7047",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:sagemaker:Creating model with name: blazingtext-2023-12-20-22-25-27-217\n",
      "INFO:sagemaker:Creating endpoint-config with name blazingtext-2023-12-20-22-25-27-217\n",
      "INFO:sagemaker:Creating endpoint with name blazingtext-2023-12-20-22-25-27-217\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-----!\n",
      "Endpoint name:  blazingtext-2023-12-20-22-25-27-217\n",
      "CPU times: user 68.5 ms, sys: 11.5 ms, total: 80 ms\n",
      "Wall time: 3min 2s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "text_classifier = estimator.deploy(initial_instance_count=1,\n",
    "                                   instance_type='ml.m5.large',\n",
    "                                   serializer=sagemaker.serializers.JSONSerializer(),\n",
    "                                   deserializer=sagemaker.deserializers.JSONDeserializer())\n",
    "\n",
    "print()\n",
    "print('Endpoint name:  {}'.format(text_classifier.endpoint_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "id": "521e4292-0b90-4993-8173-b4b91881f2eb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sentiment</th>\n",
       "      <th>tokenized_review</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>23905</th>\n",
       "      <td>__label__0</td>\n",
       "      <td>['admit', 'review', 'based', '36', 'minute', '...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14533</th>\n",
       "      <td>__label__1</td>\n",
       "      <td>['TENNIS', 'CHUMPS', 'one', 'glaring', 'anomal...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32294</th>\n",
       "      <td>__label__1</td>\n",
       "      <td>['Lucio', 'Fulci', 'made', 'good', 'movie', 'k...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7369</th>\n",
       "      <td>__label__-1</td>\n",
       "      <td>['early', 'days', 'Looney', 'Tunes', 'cartoons...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>683</th>\n",
       "      <td>__label__1</td>\n",
       "      <td>['movie', 'beautiful', 'love', 'box', 'tissues...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9189</th>\n",
       "      <td>__label__0</td>\n",
       "      <td>['One', 'many', 'Universal', 'programmers', 's...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19629</th>\n",
       "      <td>__label__0</td>\n",
       "      <td>['scenario', 'rather', 'banal', 'bought', 'spl...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18048</th>\n",
       "      <td>__label__1</td>\n",
       "      <td>['young', 'remember', 'much', 'Kate', 'Smith',...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22886</th>\n",
       "      <td>__label__1</td>\n",
       "      <td>['enjoying', 'movies', 'different', 'era', 'ho...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8265</th>\n",
       "      <td>__label__0</td>\n",
       "      <td>['First', 'law', 'ever', 'film', 'Mountie', 'p...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3251 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         sentiment                                   tokenized_review\n",
       "23905   __label__0  ['admit', 'review', 'based', '36', 'minute', '...\n",
       "14533   __label__1  ['TENNIS', 'CHUMPS', 'one', 'glaring', 'anomal...\n",
       "32294   __label__1  ['Lucio', 'Fulci', 'made', 'good', 'movie', 'k...\n",
       "7369   __label__-1  ['early', 'days', 'Looney', 'Tunes', 'cartoons...\n",
       "683     __label__1  ['movie', 'beautiful', 'love', 'box', 'tissues...\n",
       "...            ...                                                ...\n",
       "9189    __label__0  ['One', 'many', 'Universal', 'programmers', 's...\n",
       "19629   __label__0  ['scenario', 'rather', 'banal', 'bought', 'spl...\n",
       "18048   __label__1  ['young', 'remember', 'much', 'Kate', 'Smith',...\n",
       "22886   __label__1  ['enjoying', 'movies', 'different', 'era', 'ho...\n",
       "8265    __label__0  ['First', 'law', 'ever', 'film', 'Mountie', 'p...\n",
       "\n",
       "[3251 rows x 2 columns]"
      ]
     },
     "execution_count": 145,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "id": "c14a1cce-03f3-4ea7-8c6c-7f82e732fb63",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "IOPub data rate exceeded.\n",
      "The Jupyter server will temporarily stop sending output\n",
      "to the client in order to avoid crashing it.\n",
      "To change this limit, set the config variable\n",
      "`--ServerApp.iopub_data_rate_limit`.\n",
      "\n",
      "Current values:\n",
      "ServerApp.iopub_data_rate_limit=1000000.0 (bytes/sec)\n",
      "ServerApp.rate_limit_window=3.0 (secs)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "reviews = df_validation['tokenized_review']\n",
    "\n",
    "tokenized_reviews = [' '.join(nltk.word_tokenize(review)) for review in reviews]\n",
    "\n",
    "payload = {\"instances\" : tokenized_reviews}\n",
    "print(payload)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "id": "f80b291d-67e5-4914-a04d-902e0fef1a22",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Done\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Predicted Label</th>\n",
       "      <th>True Label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>23905</th>\n",
       "      <td>[__label__1]</td>\n",
       "      <td>__label__0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14533</th>\n",
       "      <td>[__label__1]</td>\n",
       "      <td>__label__1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32294</th>\n",
       "      <td>[__label__0]</td>\n",
       "      <td>__label__1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7369</th>\n",
       "      <td>[__label__1]</td>\n",
       "      <td>__label__-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>683</th>\n",
       "      <td>[__label__1]</td>\n",
       "      <td>__label__1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9189</th>\n",
       "      <td>[__label__1]</td>\n",
       "      <td>__label__0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19629</th>\n",
       "      <td>[__label__0]</td>\n",
       "      <td>__label__0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18048</th>\n",
       "      <td>[__label__1]</td>\n",
       "      <td>__label__1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22886</th>\n",
       "      <td>[__label__1]</td>\n",
       "      <td>__label__1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8265</th>\n",
       "      <td>[__label__1]</td>\n",
       "      <td>__label__0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3251 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      Predicted Label   True Label\n",
       "23905    [__label__1]   __label__0\n",
       "14533    [__label__1]   __label__1\n",
       "32294    [__label__0]   __label__1\n",
       "7369     [__label__1]  __label__-1\n",
       "683      [__label__1]   __label__1\n",
       "...               ...          ...\n",
       "9189     [__label__1]   __label__0\n",
       "19629    [__label__0]   __label__0\n",
       "18048    [__label__1]   __label__1\n",
       "22886    [__label__1]   __label__1\n",
       "8265     [__label__1]   __label__0\n",
       "\n",
       "[3251 rows x 2 columns]"
      ]
     },
     "execution_count": 167,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions = text_classifier.predict(data=payload)\n",
    "pred_label = []\n",
    "\n",
    "for prediction in predictions:\n",
    "    pred_label.append(prediction['label'])  # Assuming the predicted label is stored in the 'label' key\n",
    "print('Done')\n",
    "\n",
    "\n",
    "pd.DataFrame({'Predicted Label': pred_label, 'True Label': df_validation['sentiment']})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "43961c24-e72d-4319-815c-76610157eb31",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}