Modelling Deep Learning Part 2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "No_9ci7iEk4C"
   },
   "source": [
    "# Library"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: pandas in /usr/local/lib/python3.8/dist-packages (2.0.1)\n",
      "Requirement already satisfied: python-dateutil>=2.8.2 in /usr/local/lib/python3.8/dist-packages (from pandas) (2.8.2)\n",
      "Requirement already satisfied: numpy>=1.20.3 in /usr/local/lib/python3.8/dist-packages (from pandas) (1.23.5)\n",
      "Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.8/dist-packages (from pandas) (2023.2)\n",
      "Requirement already satisfied: tzdata>=2022.1 in /usr/local/lib/python3.8/dist-packages (from pandas) (2023.3)\n",
      "Requirement already satisfied: six>=1.5 in /usr/lib/python3/dist-packages (from python-dateutil>=2.8.2->pandas) (1.14.0)\n",
      "\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n",
      "\u001b[0m\n",
      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m A new release of pip is available: \u001b[0m\u001b[31;49m23.0.1\u001b[0m\u001b[39;49m -> \u001b[0m\u001b[32;49m23.1.2\u001b[0m\n",
      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m To update, run: \u001b[0m\u001b[32;49mpython3 -m pip install --upgrade pip\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "!pip install pandas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: matplotlib in /usr/local/lib/python3.8/dist-packages (3.7.1)\n",
      "Requirement already satisfied: importlib-resources>=3.2.0 in /usr/local/lib/python3.8/dist-packages (from matplotlib) (5.12.0)\n",
      "Requirement already satisfied: contourpy>=1.0.1 in /usr/local/lib/python3.8/dist-packages (from matplotlib) (1.0.7)\n",
      "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.8/dist-packages (from matplotlib) (0.11.0)\n",
      "Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.8/dist-packages (from matplotlib) (23.0)\n",
      "Requirement already satisfied: python-dateutil>=2.7 in /usr/local/lib/python3.8/dist-packages (from matplotlib) (2.8.2)\n",
      "Requirement already satisfied: numpy>=1.20 in /usr/local/lib/python3.8/dist-packages (from matplotlib) (1.23.5)\n",
      "Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.8/dist-packages (from matplotlib) (4.39.4)\n",
      "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.8/dist-packages (from matplotlib) (1.4.4)\n",
      "Requirement already satisfied: pyparsing>=2.3.1 in /usr/local/lib/python3.8/dist-packages (from matplotlib) (3.0.9)\n",
      "Requirement already satisfied: pillow>=6.2.0 in /usr/local/lib/python3.8/dist-packages (from matplotlib) (9.5.0)\n",
      "Requirement already satisfied: zipp>=3.1.0 in /usr/local/lib/python3.8/dist-packages (from importlib-resources>=3.2.0->matplotlib) (3.15.0)\n",
      "Requirement already satisfied: six>=1.5 in /usr/lib/python3/dist-packages (from python-dateutil>=2.7->matplotlib) (1.14.0)\n",
      "\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n",
      "\u001b[0m\n",
      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m A new release of pip is available: \u001b[0m\u001b[31;49m23.0.1\u001b[0m\u001b[39;49m -> \u001b[0m\u001b[32;49m23.1.2\u001b[0m\n",
      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m To update, run: \u001b[0m\u001b[32;49mpython3 -m pip install --upgrade pip\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "!pip install matplotlib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: opendatasets in /usr/local/lib/python3.8/dist-packages (0.1.22)\n",
      "Requirement already satisfied: click in /usr/local/lib/python3.8/dist-packages (from opendatasets) (8.1.3)\n",
      "Requirement already satisfied: tqdm in /usr/local/lib/python3.8/dist-packages (from opendatasets) (4.65.0)\n",
      "Requirement already satisfied: kaggle in /usr/local/lib/python3.8/dist-packages (from opendatasets) (1.5.13)\n",
      "Requirement already satisfied: python-slugify in /usr/local/lib/python3.8/dist-packages (from kaggle->opendatasets) (8.0.1)\n",
      "Requirement already satisfied: requests in /usr/local/lib/python3.8/dist-packages (from kaggle->opendatasets) (2.28.2)\n",
      "Requirement already satisfied: python-dateutil in /usr/local/lib/python3.8/dist-packages (from kaggle->opendatasets) (2.8.2)\n",
      "Requirement already satisfied: certifi in /usr/lib/python3/dist-packages (from kaggle->opendatasets) (2019.11.28)\n",
      "Requirement already satisfied: six>=1.10 in /usr/lib/python3/dist-packages (from kaggle->opendatasets) (1.14.0)\n",
      "Requirement already satisfied: urllib3 in /usr/lib/python3/dist-packages (from kaggle->opendatasets) (1.25.8)\n",
      "Requirement already satisfied: text-unidecode>=1.3 in /usr/local/lib/python3.8/dist-packages (from python-slugify->kaggle->opendatasets) (1.3)\n",
      "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.8/dist-packages (from requests->kaggle->opendatasets) (3.1.0)\n",
      "Requirement already satisfied: idna<4,>=2.5 in /usr/lib/python3/dist-packages (from requests->kaggle->opendatasets) (2.8)\n",
      "\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n",
      "\u001b[0m\n",
      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m A new release of pip is available: \u001b[0m\u001b[31;49m23.0.1\u001b[0m\u001b[39;49m -> \u001b[0m\u001b[32;49m23.1.2\u001b[0m\n",
      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m To update, run: \u001b[0m\u001b[32;49mpython3 -m pip install --upgrade pip\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "!pip install opendatasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: scikit-learn in /usr/local/lib/python3.8/dist-packages (1.2.2)\n",
      "Requirement already satisfied: numpy>=1.17.3 in /usr/local/lib/python3.8/dist-packages (from scikit-learn) (1.23.5)\n",
      "Requirement already satisfied: joblib>=1.1.1 in /usr/local/lib/python3.8/dist-packages (from scikit-learn) (1.2.0)\n",
      "Requirement already satisfied: threadpoolctl>=2.0.0 in /usr/local/lib/python3.8/dist-packages (from scikit-learn) (3.1.0)\n",
      "Requirement already satisfied: scipy>=1.3.2 in /usr/local/lib/python3.8/dist-packages (from scikit-learn) (1.10.1)\n",
      "\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n",
      "\u001b[0m\n",
      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m A new release of pip is available: \u001b[0m\u001b[31;49m23.0.1\u001b[0m\u001b[39;49m -> \u001b[0m\u001b[32;49m23.1.2\u001b[0m\n",
      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m To update, run: \u001b[0m\u001b[32;49mpython3 -m pip install --upgrade pip\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "!pip install scikit-learn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "id": "HqzGEMvPEk4F",
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Library to dataset\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import opendatasets as od\n",
    "import os\n",
    "os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3'\n",
    "\n",
    "# Library tensorflow\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.preprocessing.text import Tokenizer\n",
    "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
    "from tensorflow.keras.callbacks import ModelCheckpoint, EarlyStopping, ReduceLROnPlateau\n",
    "\n",
    "# Surpress warning\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "# Library sklearn\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "# pickle\n",
    "import pickle\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Skipping, found downloaded files in \"./suicide-detection-clean\" (use force=True to force download)\n"
     ]
    }
   ],
   "source": [
    "# Download dataset\n",
    "od.download(\"https://www.kaggle.com/datasets/ahmadhabibhasanzein/suicide-detection-clean\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 206
    },
    "id": "ly5YTaNrEk4G",
    "outputId": "9a0d7fcf-2d0b-4ba2-ff17-0ad1e27d4b71",
    "tags": []
   },
   "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>text</th>\n",
       "      <th>class</th>\n",
       "      <th>clean_text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Ex Wife Threatening SuicideRecently I left my ...</td>\n",
       "      <td>suicide</td>\n",
       "      <td>ex wife threatening suiciderecently left wife ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Am I weird I don't get affected by compliments...</td>\n",
       "      <td>non-suicide</td>\n",
       "      <td>weird get affected compliments coming someone ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Finally 2020 is almost over... So I can never ...</td>\n",
       "      <td>non-suicide</td>\n",
       "      <td>finally almost never hear bad year ever swear ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>i need helpjust help me im crying so hard</td>\n",
       "      <td>suicide</td>\n",
       "      <td>need helpjust help im crying hard</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>I’m so lostHello, my name is Adam (16) and I’v...</td>\n",
       "      <td>suicide</td>\n",
       "      <td>losthello name adam struggling years afraid pa...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                text        class   \n",
       "0  Ex Wife Threatening SuicideRecently I left my ...      suicide  \\\n",
       "1  Am I weird I don't get affected by compliments...  non-suicide   \n",
       "2  Finally 2020 is almost over... So I can never ...  non-suicide   \n",
       "3          i need helpjust help me im crying so hard      suicide   \n",
       "4  I’m so lostHello, my name is Adam (16) and I’v...      suicide   \n",
       "\n",
       "                                          clean_text  \n",
       "0  ex wife threatening suiciderecently left wife ...  \n",
       "1  weird get affected compliments coming someone ...  \n",
       "2  finally almost never hear bad year ever swear ...  \n",
       "3                  need helpjust help im crying hard  \n",
       "4  losthello name adam struggling years afraid pa...  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Read dataset\n",
    "data = pd.read_csv('suicide-detection-clean/clean_text.csv')\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "id": "tI4zGvPDEk4H",
    "tags": []
   },
   "outputs": [],
   "source": [
    "X = data['clean_text'].astype(str)\n",
    "y = data['class']\n",
    "\n",
    "# encode class values as integers\n",
    "# Define label encoder\n",
    "encoder = LabelEncoder()\n",
    "# fit and transform\n",
    "encoded_y = encoder.fit_transform(y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Test Train Split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "id": "m7z7S5FdEk4H",
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Test and train split\n",
    "X_train, X_test, y_train, y_test = train_test_split(X,encoded_y, test_size=.2, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "YksTB-KzEk4I"
   },
   "source": [
    "# Data Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "id": "sIEVlUS7Ek4I",
    "tags": []
   },
   "outputs": [],
   "source": [
    "vocab_size = 1000\n",
    "max_length = 1000\n",
    "trunc_type='post'\n",
    "padding_type='post'\n",
    "oov_tok = \"<OOV>\"\n",
    "\n",
    "# Initialize the Tokenizer class\n",
    "tokenizer = Tokenizer(num_words=vocab_size, oov_token=oov_tok)\n",
    "\n",
    "# Generate the word index dictionary\n",
    "tokenizer.fit_on_texts(X_train)\n",
    "word_index = tokenizer.word_index\n",
    "\n",
    "# Generate and pad the training sequences\n",
    "training_sequences = tokenizer.texts_to_sequences(X_train)\n",
    "training_padded = pad_sequences(training_sequences, maxlen=max_length, padding=padding_type, truncating=trunc_type)\n",
    "\n",
    "# Generate and pad the testing sequences\n",
    "testing_sequences = tokenizer.texts_to_sequences(X_test)\n",
    "testing_padded = pad_sequences(testing_sequences, maxlen=max_length, padding=padding_type, truncating=trunc_type)\n",
    "\n",
    "# Convert the labels lists into numpy arrays\n",
    "training_labels = np.array(y_train)\n",
    "testing_labels = np.array(y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# List Resource"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU'),\n",
       " PhysicalDevice(name='/physical_device:GPU:1', device_type='GPU')]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Lis physical devices\n",
    "devices = tf.config.experimental.list_physical_devices(\"GPU\")\n",
    "devices"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Using MirroredStrategy with devices ('/job:localhost/replica:0/task:0/device:GPU:0', '/job:localhost/replica:0/task:0/device:GPU:1')\n"
     ]
    }
   ],
   "source": [
    "# Define the GPU strategy\n",
    "mirrored_strategy = tf.distribute.MirroredStrategy(devices=[\"/gpu:0\", \"/gpu:1\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "uw8lhhWEEk4J"
   },
   "source": [
    "# Bidirectional LSTM 1 Layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "of8FCYLUEk4J",
    "outputId": "c84ae2f6-9b35-4989-e7ae-80118688de8c",
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding (Embedding)       (None, 1000, 16)          16000     \n",
      "                                                                 \n",
      " bidirectional (Bidirectiona  (None, 128)              41472     \n",
      " l)                                                              \n",
      "                                                                 \n",
      " dense (Dense)               (None, 128)               16512     \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 128)               16512     \n",
      "                                                                 \n",
      " dense_2 (Dense)             (None, 128)               16512     \n",
      "                                                                 \n",
      " dense_3 (Dense)             (None, 1)                 129       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 107,137\n",
      "Trainable params: 107,137\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "embedding_dim = 16\n",
    "with mirrored_strategy.scope():\n",
    "# Initialize the model\n",
    "    model = tf.keras.Sequential([\n",
    "        tf.keras.layers.Embedding(vocab_size, embedding_dim, input_length=max_length),\n",
    "        tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(64)),\n",
    "        tf.keras.layers.Dense(128, activation='relu'),\n",
    "        tf.keras.layers.Dense(128, activation='relu'),\n",
    "        tf.keras.layers.Dense(128, activation='relu'),\n",
    "        tf.keras.layers.Dense(1, activation='sigmoid')\n",
    "    ])\n",
    "\n",
    "    # Compile the model\n",
    "    model.compile(loss='binary_crossentropy',optimizer='adam',metrics=['accuracy'])\n",
    "    \n",
    "# Print the model summary\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Difine the callbacks\n",
    "# Save the best model\n",
    "checkpoint_callback = ModelCheckpoint(filepath=\"Bidirectional-LSTM-1-Layers.h5\", save_best_only=True, monitor=\"val_accuracy\", mode=\"max\", verbose=1)\n",
    "\n",
    "# Early stopping\n",
    "early_stopping_callback = EarlyStopping(monitor=\"val_accuracy\", mode=\"max\", patience=5, verbose=1, restore_best_weights=True)\n",
    "\n",
    "# Reduce learning rate\n",
    "reduce_lr_callback = ReduceLROnPlateau(monitor=\"val_loss\", factor=0.1, patience=5, verbose=1, mode=\"min\", min_delta=0.0001, cooldown=0, min_lr=0)\n",
    "\n",
    "# Callbacks list\n",
    "callbacks=[checkpoint_callback, early_stopping_callback, reduce_lr_callback]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "id": "wt_f2RcWEk4J",
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "INFO:tensorflow:batch_all_reduce: 14 all-reduces with algorithm = nccl, num_packs = 1\n",
      "WARNING:tensorflow:Efficient allreduce is not supported for 1 IndexedSlices\n",
      "INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:GPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:GPU:0', '/job:localhost/replica:0/task:0/device:GPU:1').\n",
      "INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).\n",
      "INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).\n",
      "INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).\n",
      "INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).\n",
      "INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).\n",
      "INFO:tensorflow:batch_all_reduce: 14 all-reduces with algorithm = nccl, num_packs = 1\n",
      "WARNING:tensorflow:Efficient allreduce is not supported for 1 IndexedSlices\n",
      "INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:GPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:GPU:0', '/job:localhost/replica:0/task:0/device:GPU:1').\n",
      "INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).\n",
      "INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).\n",
      "INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).\n",
      "INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).\n",
      "1451/1451 [==============================] - ETA: 0s - loss: 0.2437 - accuracy: 0.9066INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).\n",
      "\n",
      "Epoch 1: val_accuracy improved from -inf to 0.91871, saving model to Bidirectional-LSTM-1-Layers.h5\n",
      "1451/1451 [==============================] - 124s 80ms/step - loss: 0.2437 - accuracy: 0.9066 - val_loss: 0.2179 - val_accuracy: 0.9187 - lr: 0.0010\n",
      "Epoch 2/50\n",
      "1451/1451 [==============================] - ETA: 0s - loss: 0.2101 - accuracy: 0.9210\n",
      "Epoch 2: val_accuracy improved from 0.91871 to 0.92179, saving model to Bidirectional-LSTM-1-Layers.h5\n",
      "1451/1451 [==============================] - 94s 65ms/step - loss: 0.2101 - accuracy: 0.9210 - val_loss: 0.2081 - val_accuracy: 0.9218 - lr: 0.0010\n",
      "Epoch 3/50\n",
      "1451/1451 [==============================] - ETA: 0s - loss: 0.2016 - accuracy: 0.9233\n",
      "Epoch 3: val_accuracy improved from 0.92179 to 0.92227, saving model to Bidirectional-LSTM-1-Layers.h5\n",
      "1451/1451 [==============================] - 94s 65ms/step - loss: 0.2016 - accuracy: 0.9233 - val_loss: 0.2027 - val_accuracy: 0.9223 - lr: 0.0010\n",
      "Epoch 4/50\n",
      "1451/1451 [==============================] - ETA: 0s - loss: 0.1979 - accuracy: 0.9248\n",
      "Epoch 4: val_accuracy did not improve from 0.92227\n",
      "1451/1451 [==============================] - 94s 64ms/step - loss: 0.1979 - accuracy: 0.9248 - val_loss: 0.2053 - val_accuracy: 0.9211 - lr: 0.0010\n",
      "Epoch 5/50\n",
      "1451/1451 [==============================] - ETA: 0s - loss: 0.1910 - accuracy: 0.9269\n",
      "Epoch 5: val_accuracy improved from 0.92227 to 0.92304, saving model to Bidirectional-LSTM-1-Layers.h5\n",
      "1451/1451 [==============================] - 93s 64ms/step - loss: 0.1910 - accuracy: 0.9269 - val_loss: 0.2008 - val_accuracy: 0.9230 - lr: 0.0010\n",
      "Epoch 6/50\n",
      "1451/1451 [==============================] - ETA: 0s - loss: 0.1868 - accuracy: 0.9287\n",
      "Epoch 6: val_accuracy improved from 0.92304 to 0.92369, saving model to Bidirectional-LSTM-1-Layers.h5\n",
      "1451/1451 [==============================] - 94s 65ms/step - loss: 0.1868 - accuracy: 0.9287 - val_loss: 0.1983 - val_accuracy: 0.9237 - lr: 0.0010\n",
      "Epoch 7/50\n",
      "1451/1451 [==============================] - ETA: 0s - loss: 0.1834 - accuracy: 0.9301\n",
      "Epoch 7: val_accuracy improved from 0.92369 to 0.92393, saving model to Bidirectional-LSTM-1-Layers.h5\n",
      "1451/1451 [==============================] - 93s 64ms/step - loss: 0.1834 - accuracy: 0.9301 - val_loss: 0.2019 - val_accuracy: 0.9239 - lr: 0.0010\n",
      "Epoch 8/50\n",
      "1451/1451 [==============================] - ETA: 0s - loss: 0.1808 - accuracy: 0.9316\n",
      "Epoch 8: val_accuracy did not improve from 0.92393\n",
      "1451/1451 [==============================] - 93s 64ms/step - loss: 0.1808 - accuracy: 0.9316 - val_loss: 0.1998 - val_accuracy: 0.9234 - lr: 0.0010\n",
      "Epoch 9/50\n",
      "1451/1451 [==============================] - ETA: 0s - loss: 0.1800 - accuracy: 0.9316\n",
      "Epoch 9: val_accuracy did not improve from 0.92393\n",
      "1451/1451 [==============================] - 94s 65ms/step - loss: 0.1800 - accuracy: 0.9316 - val_loss: 0.2042 - val_accuracy: 0.9228 - lr: 0.0010\n",
      "Epoch 10/50\n",
      "1451/1451 [==============================] - ETA: 0s - loss: 0.1771 - accuracy: 0.9326\n",
      "Epoch 10: val_accuracy improved from 0.92393 to 0.92569, saving model to Bidirectional-LSTM-1-Layers.h5\n",
      "1451/1451 [==============================] - 94s 64ms/step - loss: 0.1771 - accuracy: 0.9326 - val_loss: 0.2035 - val_accuracy: 0.9257 - lr: 0.0010\n",
      "Epoch 11/50\n",
      "1451/1451 [==============================] - ETA: 0s - loss: 0.1732 - accuracy: 0.9342\n",
      "Epoch 11: val_accuracy did not improve from 0.92569\n",
      "\n",
      "Epoch 11: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
      "1451/1451 [==============================] - 94s 65ms/step - loss: 0.1732 - accuracy: 0.9342 - val_loss: 0.1991 - val_accuracy: 0.9229 - lr: 0.0010\n",
      "Epoch 12/50\n",
      "1451/1451 [==============================] - ETA: 0s - loss: 0.1601 - accuracy: 0.9396\n",
      "Epoch 12: val_accuracy did not improve from 0.92569\n",
      "1451/1451 [==============================] - 94s 64ms/step - loss: 0.1601 - accuracy: 0.9396 - val_loss: 0.1993 - val_accuracy: 0.9246 - lr: 1.0000e-04\n",
      "Epoch 13/50\n",
      "1451/1451 [==============================] - ETA: 0s - loss: 0.1585 - accuracy: 0.9398\n",
      "Epoch 13: val_accuracy did not improve from 0.92569\n",
      "1451/1451 [==============================] - 93s 64ms/step - loss: 0.1585 - accuracy: 0.9398 - val_loss: 0.1994 - val_accuracy: 0.9247 - lr: 1.0000e-04\n",
      "Epoch 14/50\n",
      "1451/1451 [==============================] - ETA: 0s - loss: 0.1575 - accuracy: 0.9403\n",
      "Epoch 14: val_accuracy did not improve from 0.92569\n",
      "1451/1451 [==============================] - 94s 65ms/step - loss: 0.1575 - accuracy: 0.9403 - val_loss: 0.1996 - val_accuracy: 0.9239 - lr: 1.0000e-04\n",
      "Epoch 15/50\n",
      "1451/1451 [==============================] - ETA: 0s - loss: 0.1566 - accuracy: 0.9408\n",
      "Epoch 15: val_accuracy did not improve from 0.92569\n",
      "Restoring model weights from the end of the best epoch: 10.\n",
      "1451/1451 [==============================] - 93s 64ms/step - loss: 0.1566 - accuracy: 0.9408 - val_loss: 0.2018 - val_accuracy: 0.9244 - lr: 1.0000e-04\n",
      "Epoch 15: early stopping\n"
     ]
    }
   ],
   "source": [
    "# Define the number of epochs\n",
    "num_epochs = 50\n",
    "\n",
    "# Train the model\n",
    "history = model.fit(training_padded, training_labels, epochs=num_epochs, validation_data=(testing_padded, testing_labels), callbacks=callbacks, batch_size=128)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1451/1451 [==============================] - 37s 25ms/step - loss: 0.2034 - accuracy: 0.9257\n",
      "Test results - Loss: 0.20344993472099304 - Accuracy: 92.56921410560608%\n"
     ]
    }
   ],
   "source": [
    "# Evaluate the model\n",
    "results_base = model.evaluate(testing_padded, testing_labels)\n",
    "\n",
    "# print results\n",
    "print(f'Test results - Loss: {results_base[0]} - Accuracy: {100*results_base[1]}%')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "id": "tscfgTUxEk4K",
    "outputId": "3c42022d-dc3a-49a8-d2e8-762226ac2682",
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot utility\n",
    "def plot_graphs(history, string):\n",
    "  plt.plot(history.history[string])\n",
    "  plt.plot(history.history['val_'+string])\n",
    "  plt.xlabel(\"Epochs\")\n",
    "  plt.ylabel(string)\n",
    "  plt.legend([string, 'val_'+string])\n",
    "  plt.show()\n",
    "  \n",
    "# Plot the accuracy and loss\n",
    "plot_graphs(history, \"accuracy\")\n",
    "plot_graphs(history, \"loss\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Testing Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load all model\n",
    "BiLSTM1L_Model = tf.keras.models.load_model('Bidirectional-LSTM-1-Layers.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Testing BiLSTM 1 Layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 1s 521ms/step\n",
      "Potential Suicide Post\n"
     ]
    }
   ],
   "source": [
    "# Define the tweet\n",
    "twt = ['I will kill myself']\n",
    "twt = tokenizer.texts_to_sequences(twt)\n",
    "twt = pad_sequences(twt, maxlen=max_length, padding=padding_type, truncating=trunc_type)\n",
    "\n",
    "# Predict the sentiment\n",
    "prediction = BiLSTM1L_Model.predict(twt)\n",
    "\n",
    "# Print the prediction\n",
    "if(np.argmax(prediction) == 0):\n",
    "    print(\"Potential Suicide Post\")\n",
    "elif (np.argmax(prediction) == 1):\n",
    "    print(\"Non Suicide Post\")"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "provenance": []
  },
  "gpuClass": "standard",
  "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": "3.8.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}