part_3_a_b.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Part 3. Enhancement\n",
    "The RNN model used in Part 2 is a basic model to perform the task of sentiment classification. In\n",
    "this section, you will design strategies to improve upon the previous model you have built. You are\n",
    "required to implement the following adjustments:\n",
    "\n",
    "1. Instead of keeping the word embeddings fixed, now update the word embeddings (the same\n",
    "way as model parameters) during the training process.\n",
    "2. As discussed in Question 1(c), apply your solution in mitigating the influence of OOV words\n",
    "and train your model again.\n",
    "3. Keeping the above two adjustments, replace your simple RNN model in Part 2 with a biLSTM model and a biGRU model, incorporating recurrent computations in both directions and\n",
    "stacking multiple layers if possible.\n",
    "4. Keeping the above two adjustments, replace your simple RNN model in Part 2 with a Convolutional Neural Network (CNN) to produce sentence representations and perform sentiment\n",
    "classification.\n",
    "5. Further improve your model. You are free to use any strategy other than the above mentioned solutions. Changing hyper-parameters or stacking more layers is not counted towards\n",
    "a meaningful improvement.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim\n",
    "from common_utils import EmbeddingMatrix"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Question 3a\n",
    "\n",
    "Instead of keeping the word embeddings fixed, now update the word embeddings (the same\n",
    "way as model parameters) during the training process.\n",
    "\n",
    "### Approach\n",
    "\n",
    "We will use the same model as in part 2 notebook, but now we will also back propagate\n",
    "the loss into the word embeddings itself. This will mean that as the model learns,\n",
    "the word embeddings would also update, causing the encoding of the words to change.\n",
    "\n",
    "We keep all other variables constant such as not handling the OOV words and using\n",
    "the same hyperparameters as in part 2 notebook."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from common_utils import CustomDatasetPreparer\n",
    "\n",
    "\n",
    "dataset_preparer = CustomDatasetPreparer(\n",
    "    dataset_name=\"rotten_tomatoes\",\n",
    "    batch_size=32, # optimal batch size for the model in part 2\n",
    "    ignore_unknown=True\n",
    ")\n",
    "\n",
    "train_loader, val_loader, test_loader = dataset_preparer.get_dataloaders(ignore_unknown=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sample input: \n",
      " tensor([[ 2099,  7656,   714,  ..., 16163, 16163, 16163],\n",
      "        [ 7053,  6712,  3323,  ..., 16163, 16163, 16163],\n",
      "        [14339, 15711,  9793,  ..., 16163, 16163, 16163],\n",
      "        ...,\n",
      "        [ 7865,  9660,  2136,  ..., 16163, 16163, 16163],\n",
      "        [ 9235, 14330,   822,  ...,   732,   505,    30],\n",
      "        [14339,  5432,  8832,  ..., 16163, 16163, 16163]])\n",
      "\n",
      "Sample label size:  torch.Size([32])\n",
      "Sample label: \n",
      " tensor([0, 0, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 0, 0, 0, 1, 1, 1, 1, 1, 0, 0, 1, 0,\n",
      "        0, 0, 0, 0, 0, 0, 1, 1])\n",
      "\n",
      "Sample label size:  torch.Size([32])\n",
      "Sample label: \n",
      " tensor([16, 19, 26, 22, 15, 17, 14, 20, 22, 32, 14, 14,  6, 15, 15, 27, 37, 20,\n",
      "        12,  8, 32, 14, 24, 21, 27, 22, 17, 28, 30,  4, 37, 30])\n"
     ]
    }
   ],
   "source": [
    "# obtain one batch of training data\n",
    "dataiter = iter(train_loader)\n",
    "sample_x,_,lenghts, sample_y = next(dataiter)\n",
    "\n",
    "# print('Sample input size: ', sample_x.size()) # batch_size, seq_length\n",
    "print('Sample input: \\n', sample_x)\n",
    "print()\n",
    "print('Sample label size: ', sample_y.size()) # batch_size\n",
    "print('Sample label: \\n', sample_y)\n",
    "print()\n",
    "print('Sample label size: ', lenghts.size()) # batch_size\n",
    "print('Sample label: \\n', lenghts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The index of <PAD> is:  16163\n"
     ]
    }
   ],
   "source": [
    "class RNN(nn.Module):\n",
    "    def __init__(\n",
    "        self,\n",
    "        hidden_dim,\n",
    "        embedding_dim,\n",
    "        word_embeddings: torch.tensor,\n",
    "        pad_idx: int,\n",
    "        num_layers=1,\n",
    "        output_size=1,\n",
    "        dropout_rate=0.3,\n",
    "    ):\n",
    "        super(RNN, self).__init__()\n",
    "        # Embedding layer\n",
    "        self.embedding = nn.Embedding.from_pretrained(\n",
    "            word_embeddings, freeze=False, padding_idx=pad_idx\n",
    "        )\n",
    "        self.num_layers = num_layers\n",
    "        self.hidden_dim = hidden_dim\n",
    "        self.rnn = nn.RNN(\n",
    "            embedding_dim, hidden_dim, num_layers, batch_first=True\n",
    "        )  # this is the num rows of the input matrix\n",
    "        self.fc = nn.Linear(hidden_dim, output_size)\n",
    "        self.drop_out = nn.Dropout(dropout_rate)\n",
    "        self.sigmoid = nn.Sigmoid()\n",
    "\n",
    "    def forward(self, x, lengths):\n",
    "        # Pass input through embedding layer\n",
    "        x = self.embedding(x).float()\n",
    "        h0 = torch.zeros(\n",
    "            self.num_layers, x.size(0), self.hidden_dim, dtype=torch.float\n",
    "        ).to(x.device)\n",
    "        # Pass the embeddings through the RNN layer\n",
    "        out, _ = self.rnn(x, h0)\n",
    "        # Max pooling\n",
    "        res, _ = torch.max(out, dim=1)\n",
    "        # Only take the last output for each sequence\n",
    "        res = self.drop_out(res)\n",
    "        # Pass through the fully connected layer\n",
    "        res = self.fc(res)\n",
    "        # Apply sigmoid activation (for binary classification)\n",
    "        res = self.sigmoid(res)\n",
    "\n",
    "        return res\n",
    "\n",
    "\n",
    "class RNNBatch(nn.Module):\n",
    "    \"\"\"\n",
    "    RNN model with batch handling\n",
    "    \"\"\"\n",
    "    def __init__(\n",
    "        self,\n",
    "        hidden_dim: int,\n",
    "        embedding_dim: int,\n",
    "        word_embeddings: torch.Tensor,\n",
    "        pad_idx,\n",
    "        num_layers=1,\n",
    "        output_size=1,\n",
    "        dropout_rate=0,\n",
    "    ):\n",
    "        super().__init__()\n",
    "        self.hidden_dim = hidden_dim\n",
    "        self.num_layers = num_layers\n",
    "        self.dropout_rate = dropout_rate\n",
    "        self.embedding = nn.Embedding.from_pretrained(\n",
    "            word_embeddings, freeze=False, padding_idx=pad_idx\n",
    "        )\n",
    "        self.rnn = nn.RNN(\n",
    "            embedding_dim, hidden_dim, num_layers, batch_first=True\n",
    "        )  # this is the num rows of the input matrix\n",
    "        self.fc = nn.Linear(hidden_dim, output_size)\n",
    "        self.sigmoid = nn.Sigmoid()\n",
    "\n",
    "        self.dropout = nn.Dropout(dropout_rate)\n",
    "\n",
    "    def forward(self, x, text_lengths):\n",
    "        embedded = self.embedding(x)\n",
    "\n",
    "        h0 = torch.zeros(self.num_layers, x.size(0), self.hidden_dim).to(\n",
    "            x.device\n",
    "        )\n",
    "\n",
    "        if self.dropout_rate > 0:\n",
    "            embedded = self.dropout(self.embedding(x)).float()\n",
    "        else:\n",
    "            embedded = self.embedding(x).float()\n",
    "\n",
    "        packed_embedded = nn.utils.rnn.pack_padded_sequence(\n",
    "            embedded, text_lengths, batch_first=True, enforce_sorted=False\n",
    "        )\n",
    "        packed_output, _ = self.rnn(packed_embedded)\n",
    "        output, _ = nn.utils.rnn.pad_packed_sequence(\n",
    "            packed_output, batch_first=True\n",
    "        )\n",
    "\n",
    "        if self.dropout_rate > 0:\n",
    "            output = self.dropout(output)\n",
    "\n",
    "        output, _ = torch.max(output, 1)\n",
    "\n",
    "        out = self.fc(\n",
    "            output\n",
    "        )  # Use the last output of the RNN for classification\n",
    "\n",
    "        sig_out = self.sigmoid(out)\n",
    "        sig_out = sig_out.squeeze(1)\n",
    "        # print(\"Sig_out shape: \", sig_out.shape)\n",
    "        return sig_out\n",
    "\n",
    "\n",
    "from common_utils import HIDDEN_SIZE, EMBEDDING_DIM, LEARNING_RATE\n",
    "\n",
    "# initialize word embeddings\n",
    "word_embeddings = EmbeddingMatrix.load()\n",
    "word_embeddings.add_padding()\n",
    "\n",
    "print(\"The index of <PAD> is: \", word_embeddings.pad_idx)\n",
    "\n",
    "\n",
    "basic_RNN = RNN(\n",
    "    hidden_dim=HIDDEN_SIZE,\n",
    "    embedding_dim=EMBEDDING_DIM,\n",
    "    word_embeddings=word_embeddings.to_tensor,\n",
    "    dropout_rate=0.3,\n",
    "    pad_idx=word_embeddings.pad_idx,\n",
    "    num_layers=2,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "epochs = 100\n",
    "es_patience = 15\n",
    "\n",
    "model = basic_RNN\n",
    "optimizer = optim.Adam(model.parameters(), lr=0.0001)\n",
    "loss_fn = nn.BCELoss()\n",
    "scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=3, gamma=0.05)\n",
    "\n",
    "model = model.to(device)\n",
    "loss_fn = loss_fn.to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  1%|          | 1/100 [00:02<04:03,  2.46s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch 1, train_loss 0.681911 train_acc 0.567761, val_loss 0.655396, val_acc 0.675422\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 11%|█         | 11/100 [00:27<03:49,  2.57s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch 11, train_loss 0.485706 train_acc 0.768464, val_loss 0.530832, val_acc 0.731707\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 18%|█▊        | 18/100 [00:48<03:39,  2.68s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "early stopping after 19 epochs\n",
      "best val loss: 0.525316472439205\n",
      "best accuracy on val set: 0.7354596622889306\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "# function to predict accuracy\n",
    "def acc(pred,label):\n",
    "    pred = torch.round(pred.squeeze())\n",
    "    return torch.sum(pred == label).item()\n",
    "\n",
    "# training\n",
    "def train_loop(train_loader, model, loss_fn, optimizer, scheduler, max_norm = 5, device='cpu'):\n",
    "    train_loss = []\n",
    "    train_acc = 0.0\n",
    "    model.train()\n",
    "    for X, extra_features, lengths, Y in train_loader:\n",
    "        X, Y = X.to(device), Y.to(device)   \n",
    "\n",
    "        optimizer.zero_grad()\n",
    "        output = model(X, lengths)\n",
    "        \n",
    "        # calculate the loss and perform backprop\n",
    "        loss = loss_fn(output.squeeze(), Y.float())\n",
    "        train_loss.append(loss.item())\n",
    "        loss.backward()\n",
    "        \n",
    "        # calculating accuracy\n",
    "        accuracy = acc(output,Y)\n",
    "        train_acc += accuracy\n",
    "        \n",
    "        #`clip_grad_norm` helps prevent the exploding gradient problem in RNNs / LSTMs.\n",
    "        nn.utils.clip_grad_norm_(model.parameters(), max_norm)\n",
    "        optimizer.step()\n",
    "        \n",
    "    scheduler.step()\n",
    "    epoch_train_loss = np.mean(train_loss)\n",
    "    epoch_train_acc = train_acc/len(train_loader.dataset)\n",
    "\n",
    "    return epoch_train_loss, epoch_train_acc\n",
    "\n",
    "def test_loop(test_loader, model, loss_fn, optimizer, device='cpu'):\n",
    "    test_loss = []\n",
    "    test_acc = 0.0\n",
    "    model.eval()\n",
    "    with torch.no_grad():\n",
    "        for X,extra_features, lengths, Y in test_loader:\n",
    "            X, Y = X.to(device), Y.to(device)   \n",
    "\n",
    "            optimizer.zero_grad()\n",
    "            output = model(X, lengths)\n",
    "            \n",
    "            # calculate the loss and perform backprop\n",
    "            loss = loss_fn(output.squeeze(), Y.float())\n",
    "\n",
    "            test_loss.append(loss.item())\n",
    "            # calculating accuracy\n",
    "            accuracy = acc(output,Y)\n",
    "            test_acc += accuracy\n",
    "\n",
    "    epoch_test_loss = np.mean(test_loss)\n",
    "    epoch_test_acc = test_acc/len(test_loader.dataset)\n",
    "\n",
    "    return epoch_test_loss, epoch_test_acc\n",
    "\n",
    "def train_model(train_loader, val_loader, model, loss_fn, optimizer, scheduler, epochs, es_patience):\n",
    "    best_val_loss = np.inf\n",
    "    best_acc = 0\n",
    "    train_loss_, train_acc_, val_loss_, val_acc_ = [], [], [], []\n",
    "    from tqdm import tqdm\n",
    "    # start training\n",
    "    for epoch in tqdm(range(epochs)):\n",
    "        train_loss, train_acc = train_loop(train_loader, model, loss_fn, optimizer, scheduler)\n",
    "        val_loss, val_acc = test_loop(val_loader, model, loss_fn, optimizer)\n",
    "\n",
    "        train_loss_.append(train_loss), train_acc_.append(train_acc)\n",
    "        val_loss_.append(val_loss), val_acc_.append(val_acc)\n",
    "\n",
    "        if val_acc > best_acc:\n",
    "            best_acc = val_acc\n",
    "        # early stopping\n",
    "        if val_loss < best_val_loss:\n",
    "            best_val_loss = val_loss\n",
    "            epochs_without_improvement = 0\n",
    "            best_model = model.state_dict()\n",
    "        else:\n",
    "            epochs_without_improvement += 1\n",
    "            if epochs_without_improvement >= es_patience:\n",
    "                print(f'early stopping after {epoch+1} epochs')\n",
    "                print(f'best val loss: {best_val_loss}')\n",
    "                print(f'best accuracy on val set: {best_acc}')\n",
    "                break\n",
    "\n",
    "        if epoch % 10 == 0:\n",
    "            print(f\"epoch {epoch+1}, train_loss {train_loss:>7f} train_acc {train_acc:>4f}, val_loss {val_loss:>7f}, val_acc {val_acc:>4f}\")\n",
    "\n",
    "    return train_loss_, train_acc_, val_loss_, val_acc_\n",
    "\n",
    "train_loss_, train_acc_, val_loss_, val_acc_ = train_model(train_loader, val_loader, model, loss_fn, optimizer, scheduler, epochs, es_patience)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Question 3a Answer\n",
    "\n",
    "(a) We use the same RNN model with the same parameters and we found that by unfreezing the weights, the performance of the model went for the test set. We hypothesize this is the case as the word embeddings are\n",
    "generalized and not specific to the task at hand. By updating the word embeddings, the model can learn the specific embeddings for the small dataset that was given to us.\n",
    "\n",
    "We show our results in the following 2 cells where we obtained a much better accuracy on the validation and test set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "def plot_loss_accuracy(train_loss_, train_acc_, val_loss_, val_acc_):\n",
    "    fig = plt.figure(figsize = (20, 6))\n",
    "    plt.subplot(1, 2, 1)\n",
    "    plt.plot(train_acc_, label='Train Acc')\n",
    "    plt.plot(val_acc_, label='Validation Acc')\n",
    "    plt.title(\"Accuracy\")\n",
    "    plt.legend()\n",
    "    plt.grid()\n",
    "        \n",
    "    plt.subplot(1, 2, 2)\n",
    "    plt.plot(train_loss_, label='Train loss')\n",
    "    plt.plot(val_loss_, label='Validation loss')\n",
    "    plt.title(\"Loss\")\n",
    "    plt.legend()\n",
    "    plt.grid()\n",
    "\n",
    "    plt.show()\n",
    "\n",
    "plot_loss_accuracy(train_loss_, train_acc_, val_loss_, val_acc_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "test_loss 0.509811, test_acc 0.775797\n"
     ]
    }
   ],
   "source": [
    "test_loss, test_acc = test_loop(test_loader, model, loss_fn,optimizer)\n",
    "print(f\"test_loss {test_loss:>7f}, test_acc {test_acc:>4f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Question 3b\n",
    "\n",
    "### Approach\n",
    "\n",
    "As discussed in part 1, we have mentioned 2 approaches to handling of the\n",
    "OOV words. We will now demonstrate the first approach, which is to replace the OOV\n",
    "words with a special token. We will replace the OOV words with a special token\n",
    "`<UNK>`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading GloVe embeddings...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Repo card metadata block was not found. Setting CardData to empty.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total GloVe words loaded: 400000\n",
      "Number of missing words: 0\n",
      "The missing words are: []\n",
      "Building embedding matrix...\n",
      "Vocab size: 16164\n",
      "Embedding matrix built successfully.\n",
      "Index of <UNK> token: 226\n"
     ]
    }
   ],
   "source": [
    "from common_utils import EMBEDDING_DIM, UNK_TOKEN, load_glove_embeddings\n",
    "\n",
    "\n",
    "w2v_model = EmbeddingMatrix.load()\n",
    "extended_vocab = w2v_model.vocab\n",
    "\n",
    "glove_dict = load_glove_embeddings()\n",
    "\n",
    "# Collect words to be removed\n",
    "missing_words = []\n",
    "for word in extended_vocab:\n",
    "    if word not in glove_dict:\n",
    "        missing_words.append(word)\n",
    "\n",
    "# Remove missing words from vocab\n",
    "for word in missing_words:\n",
    "    extended_vocab.remove(word)\n",
    "        \n",
    "print(f\"Number of missing words: {len(missing_words)}\")\n",
    "print(f\"The missing words are: {missing_words}\")\n",
    "\n",
    "extended_vocab.add(UNK_TOKEN)\n",
    "\n",
    "# mapping of words to indices and vice versa\n",
    "word2idx = {word: idx for idx, word in enumerate(sorted(extended_vocab))}\n",
    "idx2word = {idx: word for word, idx in word2idx.items()}\n",
    "\n",
    "print(\"Building embedding matrix...\")\n",
    "vocab_size = len(word2idx)\n",
    "print(f\"Vocab size: {vocab_size}\")\n",
    "embedding_matrix = np.zeros((vocab_size, EMBEDDING_DIM))\n",
    "\n",
    "for word, idx in word2idx.items():\n",
    "    if word == UNK_TOKEN:\n",
    "        continue\n",
    "    embedding_matrix[idx] = glove_dict[word]\n",
    "\n",
    "# add random vector for unknown words\n",
    "embedding_matrix[word2idx[UNK_TOKEN]] = np.random.normal(scale=0.6, size=(EMBEDDING_DIM,))\n",
    "\n",
    "print(\"Embedding matrix built successfully.\")\n",
    "\n",
    "print(f\"Index of <UNK> token: {word2idx['<UNK>']}\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To aid us in the downstream task for the rest of the analysis that we will be\n",
    "doing, we will update the EmbeddingMatrix class that was earlier defined to\n",
    "also handle the case of unknown words.\n",
    "\n",
    "We define the class using an API interface which can be easily called."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "from common_utils import EMBEDDING_MATRIX_PATH, WORD2IDX_PATH, IDX2WORD_PATH\n",
    "\n",
    "class EmbeddingMatrix:\n",
    "    def __init__(self, unk_token=UNK_TOKEN, handle_unknown=True) -> None:\n",
    "        self.d = 0\n",
    "        self.v = 0\n",
    "        self.pad_idx: int\n",
    "        self.unk_idx: int\n",
    "        self.embedding_matrix: np.ndarray\n",
    "        self.word2idx: dict\n",
    "        self.idx2word: dict\n",
    "        self.unk_token = unk_token\n",
    "        self.handle_unknown = handle_unknown\n",
    "\n",
    "    @classmethod\n",
    "    def load(cls) -> \"EmbeddingMatrix\":\n",
    "        # load vectors from file\n",
    "        embedding_matrix: np.ndarray = np.load(EMBEDDING_MATRIX_PATH)\n",
    "        # set attributes\n",
    "        em = cls()\n",
    "        em.embedding_matrix = embedding_matrix\n",
    "\n",
    "        with open(WORD2IDX_PATH, \"r\", encoding=\"utf-8\") as f:\n",
    "            word2idx: dict = json.load(f)\n",
    "            em.word2idx = word2idx\n",
    "        with open(IDX2WORD_PATH, \"r\", encoding=\"utf-8\") as f:\n",
    "            idx2word: dict = json.load(f)\n",
    "            em.idx2word = idx2word\n",
    "            \n",
    "        em.v, em.d = embedding_matrix.shape\n",
    "        return em\n",
    "\n",
    "    def save(self) -> None:\n",
    "        np.save(EMBEDDING_MATRIX_PATH, self.embedding_matrix)\n",
    "        \n",
    "        with open(WORD2IDX_PATH, \"w\", encoding=\"utf-8\") as f:\n",
    "            json.dump(self.word2idx, f, ensure_ascii=False, indent=4)\n",
    "            \n",
    "        with open(IDX2WORD_PATH, \"w\", encoding=\"utf-8\") as f:\n",
    "            json.dump(self.idx2word, f, ensure_ascii=False, indent=4)\n",
    "\n",
    "    @property\n",
    "    def to_tensor(self) -> torch.Tensor:\n",
    "        return torch.tensor(self.embedding_matrix, dtype=torch.float)\n",
    "\n",
    "    def add_padding(self) -> None:\n",
    "        if \"<PAD>\" in self.word2idx:\n",
    "            return\n",
    "        padding = np.zeros((1, self.d), dtype=\"float32\")\n",
    "        self.embedding_matrix = np.vstack((self.embedding_matrix, padding))\n",
    "\n",
    "        self.v += 1\n",
    "        self.pad_idx = self.v - 1\n",
    "        self.word2idx[\"<PAD>\"] = self.pad_idx\n",
    "\n",
    "    def add_unk_token(self) -> None:\n",
    "        if self.unk_token in self.word2idx:\n",
    "            return\n",
    "        unk_vector = np.random.normal(scale=0.6, size=(EMBEDDING_DIM,))\n",
    "        self.embedding_matrix = np.vstack((self.embedding_matrix, unk_vector))\n",
    "\n",
    "        self.v += 1\n",
    "        self.unk_idx = self.v - 1\n",
    "        self.word2idx[self.unk_token] = self.unk_idx\n",
    "\n",
    "    @property\n",
    "    def dimension(self) -> int:\n",
    "        \"\"\"Dimension of the embedding matrix\n",
    "\n",
    "        :return: The dimension of the embedding matrix\n",
    "        :rtype: int\n",
    "        \"\"\"\n",
    "        return self.d\n",
    "\n",
    "    @property\n",
    "    def vocab_size(self) -> int:\n",
    "        \"\"\"Vocabulary size of the embedding matrix\n",
    "\n",
    "        :return: The vocabulary size of the embedding matrix\n",
    "        :rtype: int\n",
    "        \"\"\"\n",
    "        return self.v\n",
    "\n",
    "    @property\n",
    "    def vocab(self) -> set[str]:\n",
    "        \"\"\"Vocabulary of the embedding matrix\n",
    "\n",
    "        Set of words in the embedding matrix\n",
    "\n",
    "        :return: The vocabulary of the embedding matrix\n",
    "        :rtype: set[str]\n",
    "        \"\"\"\n",
    "        return set(self.word2idx.keys())\n",
    "\n",
    "    def __getitem__(self, word: str) -> np.ndarray:\n",
    "        return self.embedding_matrix[self.word2idx[word]]\n",
    "\n",
    "    def get_idx(self, word: str) -> int:\n",
    "        # if word not in vocab, return None\n",
    "        if self.handle_unknown:\n",
    "            return self.word2idx.get(word, self.unk_idx)\n",
    "\n",
    "        return self.word2idx.get(word, None)\n",
    "    \n",
    "    def is_in_vocab(self, word: str) -> bool:\n",
    "        return word in self.word2idx\n",
    "\n",
    "    def is_in_index(self, idx: int) -> bool:\n",
    "        return idx in self.idx2word"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Fast Text Embeddings\n",
    "For the second approach, we will use the FastText embeddings, which are trained on\n",
    "subword information. This will help in encoding the OOV words as well."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Implementation of FastText for word embedding\n",
    "import gensim\n",
    "from gensim.models import FastText\n",
    "from common_utils import EMBEDDING_DIM\n",
    "# Create a FastText model with the same dimensions as the Word2Vec model\n",
    "fasttext_model = FastText(\n",
    "    vector_size=EMBEDDING_DIM,\n",
    "    window=5, # context window size \n",
    "    min_count=1, # threshold for word frequency\n",
    "    workers=4\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The corpus has 8531 documents.\n"
     ]
    }
   ],
   "source": [
    "from datasets import load_dataset, Dataset\n",
    "from common_utils import EmbeddingMatrix, tokenize\n",
    "import nltk\n",
    "dataset = load_dataset(\"rotten_tomatoes\")\n",
    "train_dataset = dataset['train']\n",
    "\n",
    "corpus = []\n",
    "for example in train_dataset:\n",
    "    tokens = nltk.word_tokenize(example['text'])\n",
    "    corpus.append(tokens)\n",
    "    \n",
    "# add padding and unknown token\n",
    "corpus = [[\"<PAD>\", \"<UNK>\"]] + corpus\n",
    "print(\"The corpus has {} documents.\".format(len(corpus)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Length of vocabulary: 18032\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(657734, 919850)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fasttext_model.build_vocab(corpus_iterable=corpus)\n",
    "# fasttext_model.build_vocab(corpus_iterable=corpus, update=True)\n",
    "\n",
    "print(\"Length of vocabulary:\", len(fasttext_model.wv.key_to_index))\n",
    "\n",
    "\n",
    "fasttext_model.train(\n",
    "    corpus_iterable=corpus, epochs=fasttext_model.epochs,\n",
    "    total_examples=fasttext_model.corpus_count, total_words=fasttext_model.corpus_total_words,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"<UNK>\" in fasttext_model.wv.key_to_index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "fasttext_vocab = set(fasttext_model.wv.key_to_index.keys())\n",
    "w2v_model = EmbeddingMatrix.load()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "set()"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "w2v_model.vocab - fasttext_vocab"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'substance-free',\n",
       " 'affirmational',\n",
       " 'greasiest',\n",
       " 'dead-eyed',\n",
       " 'mind-destroying',\n",
       " 'twist-and-turn',\n",
       " \"'under\",\n",
       " 'decidiram',\n",
       " '179-minute',\n",
       " 'retro-refitting',\n",
       " 'stagecrafts',\n",
       " 'stop-go',\n",
       " 'homo-eroticism',\n",
       " 'excrescence',\n",
       " 'back-stabbing',\n",
       " 'brazil-like',\n",
       " 'finding-herself',\n",
       " 'unplundered',\n",
       " 'pulpiness',\n",
       " 'shrieky',\n",
       " '129-minute',\n",
       " 'snail-like',\n",
       " 'dearly-loved',\n",
       " \"'garth\",\n",
       " 'poster-boy',\n",
       " 'laugh-filled',\n",
       " 'autocritique',\n",
       " 'middle-brow',\n",
       " 'time-it-is',\n",
       " 'docu-makers',\n",
       " 'drippiness',\n",
       " 'oft-described',\n",
       " 'laser-projected',\n",
       " 'headline-fresh',\n",
       " 'bodice-ripper',\n",
       " 'self-glorification',\n",
       " 'decades-spanning',\n",
       " 'demented-funny',\n",
       " 'surfacey',\n",
       " \"'swept\",\n",
       " 'close-to-solid',\n",
       " 'skate/surf',\n",
       " 'early-on',\n",
       " 'too-conscientious',\n",
       " \"'enigma\",\n",
       " 'truck-loving',\n",
       " 'janklowicz-mann',\n",
       " 'real-live',\n",
       " \"'real\",\n",
       " 'flck',\n",
       " 'puréed',\n",
       " 'hour-and-a-half-long',\n",
       " 'dull-witted',\n",
       " 'amoses',\n",
       " \"'chick\",\n",
       " 'digital-video',\n",
       " \"'terrible\",\n",
       " 'preciosista',\n",
       " 'dialed-up',\n",
       " 'euro-film',\n",
       " 'intelectualmente',\n",
       " 'hawk-style',\n",
       " 'drama/action',\n",
       " 'mood-altering',\n",
       " 'monster/science',\n",
       " 'girls-behaving-badly',\n",
       " \"'this\",\n",
       " 'plot-wise',\n",
       " 'ultra-low-budget',\n",
       " 'rip-roaring',\n",
       " 'kickass',\n",
       " 'bang-the-drum',\n",
       " 'culture-clash',\n",
       " 'tv-cops',\n",
       " 'gorefests',\n",
       " \"'get\",\n",
       " 'large-frame',\n",
       " 'premissa',\n",
       " 'overcoming-obstacles',\n",
       " 'espectáculo',\n",
       " 'star-making',\n",
       " \"'lovely\",\n",
       " \"'just\",\n",
       " 'third-act',\n",
       " 'grand-scale',\n",
       " 'brush-up',\n",
       " \"'solaris\",\n",
       " 'estudo',\n",
       " 'schlock-filled',\n",
       " 'ankle-deep',\n",
       " '1/2-hour',\n",
       " 'video-viewing',\n",
       " 'datedness',\n",
       " 'singer/composer',\n",
       " 'petin',\n",
       " 'unfakable',\n",
       " 'over-25s',\n",
       " 'two-drink-minimum',\n",
       " 'well-wrought',\n",
       " 'surfer-girl',\n",
       " 'well-contructed',\n",
       " 'mcklusky',\n",
       " 'heart-string',\n",
       " 'one-joke',\n",
       " 'sucker-punch',\n",
       " 'opera-ish',\n",
       " 'super-',\n",
       " 'pincel',\n",
       " 'hoo-ha',\n",
       " 'travil',\n",
       " 'post-9/11',\n",
       " 'social/economic/urban',\n",
       " 'groan-inducing',\n",
       " 'brothers-style',\n",
       " 'industrial-model',\n",
       " 'self-promoter',\n",
       " 'assistir',\n",
       " 'lesser-praised',\n",
       " '-after',\n",
       " \"'credit\",\n",
       " 'car-wreck',\n",
       " 'overemphatic',\n",
       " 'double-',\n",
       " \"'reality\",\n",
       " 'still-contemporary',\n",
       " \"'classic\",\n",
       " 'well-detailed',\n",
       " 'kill-by-numbers',\n",
       " 'patriotero',\n",
       " 'foot-age',\n",
       " 'roteirista',\n",
       " 'plot-lines',\n",
       " 'wife/colleague',\n",
       " 'pretenciosas',\n",
       " 'lynch-like',\n",
       " 'near-disaster',\n",
       " 'teary-eyed',\n",
       " 'sogginess',\n",
       " 'blood-curdling',\n",
       " 'covardia',\n",
       " 'force-feed',\n",
       " 'elegiacally',\n",
       " 'something-borrowed',\n",
       " 'cameo-packed',\n",
       " \"'dog\",\n",
       " 'gay-niche',\n",
       " 'decasia',\n",
       " 'funny/gritty',\n",
       " 'good-deed/bad-deed',\n",
       " 'fever-pitched',\n",
       " 'jaglomized',\n",
       " 'playboy-mansion',\n",
       " 'pseudo-educational',\n",
       " 'talking-head',\n",
       " 'middle-agers',\n",
       " 'manipulador',\n",
       " 'japanimator',\n",
       " \"mid-'70s\",\n",
       " 'estrogen-free',\n",
       " 'aborbing',\n",
       " 'quizá',\n",
       " 'over-familiarity',\n",
       " 'delicia',\n",
       " 'chopsocky',\n",
       " 'comedy-deficient',\n",
       " 'ho-hum',\n",
       " 'mothball-y',\n",
       " 'lascivious-minded',\n",
       " 'smart-aleck',\n",
       " \"'carente\",\n",
       " \"'brazil\",\n",
       " '-stunning',\n",
       " 'ga-zillionth',\n",
       " 'snazzy-looking',\n",
       " 'non-porn',\n",
       " 'cinema-and-self',\n",
       " 'snore-fest',\n",
       " 'be-all-end-all',\n",
       " 'paint-by-numbers',\n",
       " 'lifetime-channel',\n",
       " 'eye-filling',\n",
       " 'pathos-filled',\n",
       " 'pouquíssimos',\n",
       " 'disappearing/reappearing',\n",
       " 'long-dreaded',\n",
       " 'animated-movie',\n",
       " 'ham-fisted',\n",
       " 'bar-scrapping',\n",
       " 'super-stupid',\n",
       " 'marine/legal',\n",
       " \"'if\",\n",
       " \"'compleja\",\n",
       " 'peace-and-love',\n",
       " \"ain't-\",\n",
       " 'Último',\n",
       " 'freak-out',\n",
       " 'idoosyncratic',\n",
       " 'trend-hoppy',\n",
       " \"'like\",\n",
       " 'predecible',\n",
       " 'bierbichler',\n",
       " 'bling-bling',\n",
       " 'cliche-riddled',\n",
       " 'rintarô',\n",
       " 'image-mongering',\n",
       " 'pile-ups',\n",
       " 'artnering',\n",
       " 'thrift-shop',\n",
       " 'achival',\n",
       " 'incoloro',\n",
       " 'impotentes',\n",
       " 'all-wise-guys-all-the-time',\n",
       " 'not-quite-urban',\n",
       " '3-year-olds',\n",
       " 'coriat',\n",
       " 'for-fans',\n",
       " 'satirizado',\n",
       " 'funcionar',\n",
       " 'heremakono',\n",
       " 'unentertaining',\n",
       " 'young-guns',\n",
       " 'sober-minded',\n",
       " 'techno-saturation',\n",
       " 'kibbitzes',\n",
       " 'often-funny',\n",
       " 'less-is-more',\n",
       " 'prefeminist',\n",
       " \"'issues\",\n",
       " \"'suits\",\n",
       " 'ótimo',\n",
       " 'corporate-sports',\n",
       " 'derivativeness',\n",
       " 'well-meant',\n",
       " 'cliche-ridden',\n",
       " 'hyper-time',\n",
       " 'mid-to-low',\n",
       " \"shoot-'em-up\",\n",
       " \"'top\",\n",
       " 'nohe',\n",
       " 'oh-so-important',\n",
       " 'cringe-inducing',\n",
       " 'freak-outs',\n",
       " 'wollter',\n",
       " 're-do',\n",
       " 'gilliam-esque',\n",
       " 'teen-speak',\n",
       " 'b-flick',\n",
       " 'perfervid',\n",
       " 'not-quite',\n",
       " 'hyper-artificiality',\n",
       " 'not-so-small',\n",
       " 'razor-sided',\n",
       " 'soon-to-be-forgettable',\n",
       " 'smile-button',\n",
       " 'roisterous',\n",
       " 'skyscraper-trapeze',\n",
       " 'quasi-original',\n",
       " 'american-russian',\n",
       " 'culminando',\n",
       " 'minac',\n",
       " 'hell-jaunt',\n",
       " 'lástima',\n",
       " \"'girls\",\n",
       " 'pell-mell',\n",
       " \"'evelyn\",\n",
       " 'strung-together',\n",
       " 'entretenida',\n",
       " 'art-directed',\n",
       " 'efteriades',\n",
       " \"'intro\",\n",
       " 'she-cute',\n",
       " 'hoity-toity',\n",
       " 'personagens',\n",
       " 'dictator-madman',\n",
       " 'fuddled',\n",
       " \"'blade\",\n",
       " 'well-put-together',\n",
       " 'dead-center',\n",
       " 'ming-liang',\n",
       " 'aproveitar',\n",
       " 'tear-stained',\n",
       " 'otto-sallies',\n",
       " 'adapted-',\n",
       " 'diretor',\n",
       " 'audacious-impossible',\n",
       " 'drama/character',\n",
       " 'gator-bashing',\n",
       " \"'chan\",\n",
       " 'tv-to-movie',\n",
       " 'delibrately',\n",
       " 'aceitou',\n",
       " \"'literary\",\n",
       " 'péssima',\n",
       " 'fine-looking',\n",
       " 'fato',\n",
       " 'often-deadly',\n",
       " 'zzzzzzzzz',\n",
       " 'kids-and-family-oriented',\n",
       " 'self-serious',\n",
       " 'miedos',\n",
       " 'interações',\n",
       " 'opera-to-film',\n",
       " 'amuse-bouche',\n",
       " 'estava',\n",
       " \"'abandon\",\n",
       " 'oh-so',\n",
       " \"'ick\",\n",
       " 'well-edited',\n",
       " 'not-so-hot',\n",
       " 'east-vs',\n",
       " 'of-a-sequel',\n",
       " \"'seeking\",\n",
       " 'wifty',\n",
       " 'brain-deadening',\n",
       " 'sub-sophomoric',\n",
       " 'heart-tugging',\n",
       " 'scene-stealing',\n",
       " 's1m0ne',\n",
       " 'powaqqatsi',\n",
       " 'nrelentingly',\n",
       " 'gang-infested',\n",
       " 'ball-and-chain',\n",
       " 'esfera',\n",
       " 'rat-a-tat',\n",
       " 'brûlée',\n",
       " 'too-facile',\n",
       " 'baca-asay',\n",
       " 'desfecho',\n",
       " 'movilizador',\n",
       " 'dridi',\n",
       " \"'requiem\",\n",
       " 'daytime-drama',\n",
       " \"'edgy\",\n",
       " 'makmalbaf',\n",
       " 'stomach-turning',\n",
       " 'party-hearty',\n",
       " 'clarke-williams',\n",
       " '4/5ths',\n",
       " '79-minute',\n",
       " 'navel-gazing',\n",
       " 'tone-deaf',\n",
       " 'screen-eating',\n",
       " 'character-',\n",
       " \"'cq\",\n",
       " 'phonce',\n",
       " 'nutjob',\n",
       " \"'hannibal\",\n",
       " 'desnudo',\n",
       " 'minute-by-minute',\n",
       " 'laugh-a-minute',\n",
       " 'life-embracing',\n",
       " \"'worse\",\n",
       " \"'thank\",\n",
       " 'bump-in',\n",
       " 'all-too-familiar',\n",
       " 'the-blanks',\n",
       " 'action/comedy',\n",
       " 'cheap-looking',\n",
       " 'police-oriented',\n",
       " 'alientation',\n",
       " 'message-mongering',\n",
       " \"'tap\",\n",
       " 'memória',\n",
       " 'bibbidy-bobbidi-bland',\n",
       " \"'ace\",\n",
       " \"'drama\",\n",
       " 'go-round',\n",
       " 'a-knocking',\n",
       " 'by-the-books',\n",
       " 'valley-girl',\n",
       " 'shagster',\n",
       " 'superada',\n",
       " 'hour-and-a-half',\n",
       " \"early-'80s\",\n",
       " 'sailboaters',\n",
       " 'heartbeat-like',\n",
       " 'kidlets',\n",
       " 'democracie',\n",
       " 'forgettably',\n",
       " 'punch-drunk',\n",
       " 'pseudo-hip',\n",
       " 'convencional',\n",
       " 'saímos',\n",
       " 'unlaughable',\n",
       " 'comedy/thriller',\n",
       " 'groan-to-guffaw',\n",
       " 'ultra-provincial',\n",
       " 'single-mindedly',\n",
       " 'lhe',\n",
       " 'pre-shooting',\n",
       " 'candy-coat',\n",
       " \"'rare\",\n",
       " 'disfrutable',\n",
       " 'addessi',\n",
       " 'higuchinsky',\n",
       " 'shayamalan',\n",
       " 'boom-bam',\n",
       " 'blade-thin',\n",
       " 'self-defeatingly',\n",
       " 'sytle',\n",
       " 'suspenser',\n",
       " 'nail-biter',\n",
       " 'talancón',\n",
       " 'give-a-damn',\n",
       " 'início',\n",
       " 'wishy-washy',\n",
       " 'too-tepid',\n",
       " 'true-blue',\n",
       " 'wise-cracker',\n",
       " 'sleep-inducingly',\n",
       " 'giant-screen',\n",
       " 'bait-and-switch',\n",
       " 'wind-in-the-hair',\n",
       " 'highly-praised',\n",
       " 'writer-producer-director',\n",
       " \"'significant\",\n",
       " 'push-the-limits',\n",
       " 'resultan',\n",
       " 'caper-comedy',\n",
       " 'brother-man',\n",
       " 'estafeta',\n",
       " 'kafka-inspired',\n",
       " 'video-game-based',\n",
       " 'fascinantes',\n",
       " \"'bold\",\n",
       " 'self-congratulatory',\n",
       " 'últimos',\n",
       " 'orquídeas',\n",
       " 'crime-busting',\n",
       " 'dust-caked',\n",
       " 'decent-enough',\n",
       " 'critic-proof',\n",
       " 'happily-ever',\n",
       " 'contemplarse',\n",
       " 'vidgame',\n",
       " 'camera-work',\n",
       " 'beat-the-clock',\n",
       " 'channel-style',\n",
       " 'graças',\n",
       " 'old-movie',\n",
       " 'well-trod',\n",
       " 'crime-film',\n",
       " 'light-heartedness',\n",
       " \"'jackass\",\n",
       " \"'my\",\n",
       " 'director-writer',\n",
       " 'gut-buster',\n",
       " 'killer-thrillers',\n",
       " 'sex-reassignment',\n",
       " '-west',\n",
       " '-a',\n",
       " 'style-free',\n",
       " 'raunch-fests',\n",
       " \"'swimfan\",\n",
       " \"'let\",\n",
       " 'oh-those-wacky-brits',\n",
       " 'war-movie',\n",
       " 'water-bound',\n",
       " 'ningún',\n",
       " 'out-bad-act',\n",
       " 'dogma-like',\n",
       " 'stress-reducing',\n",
       " \"'there\",\n",
       " \"'small\",\n",
       " 'consegue',\n",
       " 'super-dooper-adorability',\n",
       " '\\x91direct-to-video',\n",
       " 'neo-hitchcockianism',\n",
       " 'collage-form',\n",
       " 'eroti-comedy',\n",
       " 'soul-killing',\n",
       " 'collosum',\n",
       " 'gone-to-seed',\n",
       " 'shuck-and-jive',\n",
       " 'bigger-than-life',\n",
       " 'girl-meets-girl',\n",
       " 'out-of-kilter',\n",
       " 'meanspirited',\n",
       " 'front-loaded',\n",
       " 'right-thinking',\n",
       " 'not-exactly',\n",
       " 'no-surprise',\n",
       " 'emocionalmente',\n",
       " 'monster-in-the-',\n",
       " \"'all\",\n",
       " 'ego-destroying',\n",
       " 'live-wire',\n",
       " 'drenched-in-the-',\n",
       " \"'fatal\",\n",
       " 'triple-crosses',\n",
       " 'character-who-shall-',\n",
       " 'handsome-looking',\n",
       " 'movie-of-the-week',\n",
       " 'superlarge',\n",
       " 'cold-fish',\n",
       " \"'scooby\",\n",
       " 'unclassifiably',\n",
       " 'half-bad',\n",
       " 'remain-nameless',\n",
       " 'non-reactionary',\n",
       " 'semi-coherent',\n",
       " 'gasp-inducing',\n",
       " 'live-honed',\n",
       " \"'laughing\",\n",
       " 'shock-you-into-laughter',\n",
       " 'b-film',\n",
       " \"'do\",\n",
       " 'dreadfulness',\n",
       " 'cat-and-cat',\n",
       " 'dumbfoundingly',\n",
       " \"'frailty\",\n",
       " \"so-bad-it's-funny\",\n",
       " \"'wayne\",\n",
       " 'focus-grouped',\n",
       " 'semi-humorous',\n",
       " 'actuada',\n",
       " '3/4th',\n",
       " 'overplotted',\n",
       " 'post-saving',\n",
       " 'siuation',\n",
       " 'instante',\n",
       " \"'quit\",\n",
       " 'hat-in-hand',\n",
       " 'strip-mined',\n",
       " 'wish-fulfilling',\n",
       " 'half-sleep',\n",
       " 'nonethnic',\n",
       " 'sermonize',\n",
       " 'rah-rah',\n",
       " 'loosely-connected',\n",
       " 'snake-down-the-throat',\n",
       " 'post-camp',\n",
       " 'quasi-improvised',\n",
       " 'andamento',\n",
       " 'half-lit',\n",
       " 'projeção',\n",
       " 'headbangingly',\n",
       " 'hammily',\n",
       " 'acting-workshop',\n",
       " \"'guests\",\n",
       " 'fulford-wierzbicki',\n",
       " \"'stock\",\n",
       " 'well-honed',\n",
       " 'joylessly',\n",
       " 'ballistic-pyrotechnic',\n",
       " 'ages-old',\n",
       " 'cativante',\n",
       " 'snazziness',\n",
       " 'self-congratulation',\n",
       " \"'santa\",\n",
       " 'engaña',\n",
       " 'actory',\n",
       " 'sequel-for-the-sake-',\n",
       " 'chou-chou',\n",
       " 'down-and-dirty',\n",
       " 'hamfisted',\n",
       " 'no-bull',\n",
       " 'tear-drenched',\n",
       " 'old-hat',\n",
       " 'difficult-to-swallow',\n",
       " 'margolo',\n",
       " \"'funny\",\n",
       " 'am-radio',\n",
       " 'mind-bending',\n",
       " 'uplifter',\n",
       " 'socio-histo-political',\n",
       " \"'no\",\n",
       " 'french-produced',\n",
       " 'prejuicios',\n",
       " 'uberviolence',\n",
       " 'soft-porn',\n",
       " 'worldly-wise',\n",
       " \"'vain\",\n",
       " \"'unfaithful\",\n",
       " 'dès',\n",
       " 'food-for-thought',\n",
       " 'half-formed',\n",
       " 'loosey-goosey',\n",
       " 'edge-of-your-seat',\n",
       " 'nuttgens',\n",
       " 'ooky-spookies',\n",
       " 'so-inept-',\n",
       " 'cliché-riddled',\n",
       " 'haphazardness',\n",
       " \"'korean\",\n",
       " 'salaciously',\n",
       " 'redneck-versus-blueblood',\n",
       " 'logra',\n",
       " 'sweet-and-sour',\n",
       " 'aburrido',\n",
       " 'class-',\n",
       " 'hyper-real',\n",
       " 'shimmeringly',\n",
       " 'fustily',\n",
       " 'surehanded',\n",
       " \"'t\",\n",
       " 'romijn-stamos',\n",
       " 'anti-hollywood',\n",
       " 'head-turner',\n",
       " 'melodic/',\n",
       " 'entreter',\n",
       " 'cor-blimey-luv-a-duck',\n",
       " 'body-switching',\n",
       " 'superficiale',\n",
       " 'less-than-objective',\n",
       " 'única',\n",
       " \"'memento\",\n",
       " 'leplouff',\n",
       " 'manga-like',\n",
       " 'abandone',\n",
       " 'horror/action',\n",
       " 'narrativa',\n",
       " 'sparklingly',\n",
       " 'walking-dead',\n",
       " 'non-shakespeare',\n",
       " 'italianas',\n",
       " 'fluxing',\n",
       " 'gooeyness',\n",
       " 'strainingly',\n",
       " 'sub-aquatic',\n",
       " 'kiddie-friendly',\n",
       " 'seldahl',\n",
       " 'stadium-seat',\n",
       " 'acabamos',\n",
       " 'próprio',\n",
       " 'imponderably',\n",
       " 'all-over-the-map',\n",
       " 'wankery',\n",
       " 'age-wise',\n",
       " 'globetrotters-generals',\n",
       " 'star/producer',\n",
       " \"'it\",\n",
       " 'romething',\n",
       " 'esforço',\n",
       " 'shapable',\n",
       " 'flim-flam',\n",
       " 'montied',\n",
       " 'annie-mary',\n",
       " 'monkeyfun',\n",
       " 'watstein',\n",
       " 'qatsi',\n",
       " 'debuter',\n",
       " 'writer/director/producer',\n",
       " 'obligada',\n",
       " 'existência',\n",
       " 'kilt-wearing',\n",
       " 'gymkata',\n",
       " 'papai',\n",
       " 'warm-milk',\n",
       " 'vietnamese-born',\n",
       " 'thekids',\n",
       " 'skid-row',\n",
       " 'bolly-holly',\n",
       " 'evolução',\n",
       " 'pollyana',\n",
       " 'guilty-pleasure',\n",
       " 'less-than-magic',\n",
       " \"'lick\",\n",
       " 'well-conceived',\n",
       " 'exporing',\n",
       " 'gal-pal',\n",
       " 'often-cute',\n",
       " 'live-style',\n",
       " 'whip-crack',\n",
       " 'blown-out',\n",
       " 'natural-seeming',\n",
       " 'go-for-broke',\n",
       " 'oídos',\n",
       " 'well-observed',\n",
       " \"'have-yourself-a-happy-little-holocaust\",\n",
       " 'nietzsche-referencing',\n",
       " 'bjorkness',\n",
       " 'demencial',\n",
       " \"'videodrome\",\n",
       " 'giggle-inducing',\n",
       " 'slice-of-depression',\n",
       " 'middle-fingered',\n",
       " 'done-that',\n",
       " \"'frankly\",\n",
       " 'one-trick',\n",
       " 'guy-in-a-dress',\n",
       " 'slasher-movie',\n",
       " 'pseudo-bio',\n",
       " 'best-foreign-film',\n",
       " 'verete',\n",
       " 'denlopp',\n",
       " 'weapon-derived',\n",
       " '-the-night',\n",
       " 'age-inspired',\n",
       " 'old-world-',\n",
       " 'inhospitability',\n",
       " \"'bad\",\n",
       " 'whimsicality',\n",
       " \"'they\",\n",
       " 'travel-agency',\n",
       " 'laboriousness',\n",
       " \"'stoked\",\n",
       " 'unencouraging',\n",
       " 'out-to-change-the-world',\n",
       " 'odd-couple',\n",
       " 'over-dramatic',\n",
       " 'bottom-rung',\n",
       " 'farewell-to-innocence',\n",
       " 'once-over',\n",
       " \"'woods\",\n",
       " 'clericks',\n",
       " 'sex-as-war',\n",
       " 'out-of-field',\n",
       " 'overmanipulative',\n",
       " 'cadness',\n",
       " 'too-long',\n",
       " 'technology-of-the-moment',\n",
       " 'who-wrote-shakespeare',\n",
       " 'four-',\n",
       " 'sweaty-palmed',\n",
       " \"'topless\",\n",
       " 'big-bug',\n",
       " 'eye-opener',\n",
       " 'intentando',\n",
       " 'meet-cute',\n",
       " \"'belgium\",\n",
       " 'musclefest',\n",
       " 'bone-crushing',\n",
       " 'thrill-kill',\n",
       " 'escapa',\n",
       " 'same-old',\n",
       " 'næs',\n",
       " 'ever-watchful',\n",
       " 'chick-flick',\n",
       " 'saído',\n",
       " \"'you\",\n",
       " 'claustrophic',\n",
       " 'continuação',\n",
       " 'side-splittingly',\n",
       " 'hard-',\n",
       " 'necessidade',\n",
       " 'deadeningly',\n",
       " \"'tweener\",\n",
       " 'poke-mania',\n",
       " 'melodrama/character',\n",
       " \"'epic\",\n",
       " \"'comedian\",\n",
       " \"'aunque\",\n",
       " 'slap-happy',\n",
       " 'nonchallenging',\n",
       " \"'possession\",\n",
       " \"'anyone\",\n",
       " \"'fish\",\n",
       " 'fantasti',\n",
       " 'eyeball-to-eyeball',\n",
       " 'girl-woman',\n",
       " 'idemoto',\n",
       " 'inquestionável',\n",
       " 'complejos',\n",
       " 'rap-metal',\n",
       " 'absolutamente',\n",
       " \"'life\",\n",
       " 'goose-pimple',\n",
       " 'repellantly',\n",
       " 'stuffiest',\n",
       " 'vulakoro',\n",
       " 'schneidermeister',\n",
       " 'rubber-face',\n",
       " 'not-at-all-good',\n",
       " 'crummles',\n",
       " 'control-alt-delete',\n",
       " 'dime-store',\n",
       " 'arriesgado',\n",
       " \"'masterpiece\",\n",
       " 'lame-old',\n",
       " 'not-so-big',\n",
       " 'spiritual-uplift',\n",
       " \"'synthetic\",\n",
       " \"'christian\",\n",
       " 'none-too-original',\n",
       " 'sychowski',\n",
       " \"'children\",\n",
       " \"'qatsi\",\n",
       " 'gender-war',\n",
       " 'fire-red',\n",
       " 'super-simple',\n",
       " 're-assess',\n",
       " \"'too\",\n",
       " 'scuzbag',\n",
       " 'unreligious',\n",
       " 'grandiosa',\n",
       " \"'co-stars\",\n",
       " 'not-quite-suburban',\n",
       " 'blue-light-special',\n",
       " 'pseudo-witty',\n",
       " 'near-xenophobic',\n",
       " \"'estupendamente\",\n",
       " \"'bowling\",\n",
       " 'mcbeal-style',\n",
       " 'off-hollywood',\n",
       " 'heartwarmingly',\n",
       " 'transfigures',\n",
       " 'snap-crackle',\n",
       " 'cliff-notes',\n",
       " 'soul-stripping',\n",
       " 'materalism',\n",
       " 'frissons',\n",
       " 'spook-a-rama',\n",
       " 'star-splashed',\n",
       " 'exotic-looking',\n",
       " 'sad-sack',\n",
       " 'self-glorified',\n",
       " 'komediant',\n",
       " 'sock-you-in-the-eye',\n",
       " 'morning-glory',\n",
       " 'friday-night',\n",
       " 'anti-date',\n",
       " 'bond-inspired',\n",
       " 'media-soaked',\n",
       " 'heart-on-its-sleeve',\n",
       " 'copmovieland',\n",
       " 'witlessness',\n",
       " 'late-twenty-somethings',\n",
       " \"'what\",\n",
       " 'rápidamente',\n",
       " 'quasi-shakespearean',\n",
       " 'nebrida',\n",
       " 'cliche-bound',\n",
       " \"matrix'-style\",\n",
       " 'estranhos',\n",
       " 'handbag-clutching',\n",
       " 'outgag',\n",
       " \"'empowerment\",\n",
       " 'obviation',\n",
       " 'six-packs',\n",
       " 'oh-so-hollywood',\n",
       " 'bergmanesque',\n",
       " 'not-great',\n",
       " 'screeching-metal',\n",
       " 'hayseeds-vs',\n",
       " 'intacto',\n",
       " 'tv-style',\n",
       " 'exhilarate',\n",
       " \"'deadly\",\n",
       " \"'barbershop\",\n",
       " 'pussy-ass',\n",
       " 'little-remembered',\n",
       " 'interspliced',\n",
       " 'hard-sell',\n",
       " 'easter-egg-colored',\n",
       " 'indieflick',\n",
       " 'money-grubbing',\n",
       " \"'juvenile\",\n",
       " \"'ah\",\n",
       " 'over-romanticize',\n",
       " 'messing-about',\n",
       " 'candy-like',\n",
       " 'revigorates',\n",
       " 'kosashvili',\n",
       " 'elemento',\n",
       " 'high-wattage',\n",
       " 'not-so-divine',\n",
       " 'roteiro',\n",
       " 'dysfunctionally',\n",
       " 'mini-mod-madness',\n",
       " 'drop-dead',\n",
       " 'book-on-tape',\n",
       " 'thriller-noir',\n",
       " 'rise-and-fall',\n",
       " 'rough-trade',\n",
       " 'flame-like',\n",
       " 'independent-community',\n",
       " 'bargain-basement',\n",
       " 'cannier',\n",
       " 'felinni',\n",
       " 'unamusing',\n",
       " 'nerfs',\n",
       " 'non-exploitive',\n",
       " '112-minute',\n",
       " 'title-bout',\n",
       " 'actress-producer',\n",
       " \"'artistically\",\n",
       " 'less-than-compelling',\n",
       " 'sticky-sweet',\n",
       " \"'twist\",\n",
       " 'jean-claud',\n",
       " 'stop-and-start',\n",
       " 'mesmos',\n",
       " 'criar',\n",
       " 'un-bear-able',\n",
       " 'wewannour',\n",
       " 'silbersteins',\n",
       " 'unembarrassing',\n",
       " 'wild-and-woolly',\n",
       " 'allodi',\n",
       " '28k',\n",
       " 'whoopee-cushion',\n",
       " 'truncheoning',\n",
       " 'surface-obsession',\n",
       " 'merchant-ivory',\n",
       " 'shmear',\n",
       " 'semi-stable',\n",
       " \"'alternate\",\n",
       " \"'been\",\n",
       " 'docu-dogma',\n",
       " 'bone-dry',\n",
       " 'flatula',\n",
       " 'still-inestimable',\n",
       " 'É',\n",
       " 'super-cool',\n",
       " \"'men\",\n",
       " 'voices-from-the-other-side',\n",
       " 'costumey',\n",
       " \"'orphans\",\n",
       " 'waydowntown',\n",
       " 'played-out',\n",
       " 'expresar',\n",
       " \"cam'ron\",\n",
       " 'creature-feature',\n",
       " 're-hash',\n",
       " 'paródia',\n",
       " 'responsável',\n",
       " 'reeses',\n",
       " 'well-meaningness',\n",
       " 'real\\x97she',\n",
       " 'dead-undead',\n",
       " 'jacked-up',\n",
       " 'sub-formulaic',\n",
       " 'screwed-up',\n",
       " 'all-in-all',\n",
       " 'pseudo-rock-video',\n",
       " 'espetáculo',\n",
       " \"aren't-kids-cute\",\n",
       " 'enrapturing',\n",
       " 'spell-casting',\n",
       " 'self-amused',\n",
       " 'jar-jar',\n",
       " 'not-so-bright',\n",
       " 'perseguição',\n",
       " \"'alabama\",\n",
       " 'second-guess',\n",
       " 'colocar',\n",
       " 'splat-man',\n",
       " 'wham-bam',\n",
       " 'genial-rogue',\n",
       " 'unforgivingly',\n",
       " 'villians',\n",
       " 'ihops',\n",
       " 'marcken',\n",
       " 'sappier',\n",
       " 'extra-dry',\n",
       " 'cipherlike',\n",
       " \"'easier\",\n",
       " 'speeds/',\n",
       " 'shoplifts',\n",
       " 'flakeball',\n",
       " 'atreve',\n",
       " 'pep-talk',\n",
       " 'only-in',\n",
       " 'cyber-horror',\n",
       " 'small-budget',\n",
       " \"fun-for-fun's-sake\",\n",
       " \"'challenging\",\n",
       " 'wisegirls',\n",
       " 'suspeito',\n",
       " 'strafings',\n",
       " 'girl-',\n",
       " 'lower-wit',\n",
       " 'mass-murdering',\n",
       " 'half-dimensional',\n",
       " 'sitcom-worthy',\n",
       " 'actorish',\n",
       " 'again-courage',\n",
       " 'fresh-squeezed',\n",
       " \"'great\",\n",
       " '-of-the-week',\n",
       " 'perdona',\n",
       " 'jae-eun',\n",
       " 'direct-to-void',\n",
       " 'camareras',\n",
       " 'action-filled',\n",
       " \"so-bad-it's-good\",\n",
       " 'gore-free',\n",
       " 'têm',\n",
       " 'head-banging',\n",
       " 'pie-like',\n",
       " 'chloroform-soaked',\n",
       " 'slash-and-hack',\n",
       " 'románticas',\n",
       " 'haunted-house',\n",
       " 'preocupe',\n",
       " 'silly-looking',\n",
       " 'incompetência',\n",
       " 'stunt-hungry',\n",
       " 'shlockmeister',\n",
       " 'poor-me',\n",
       " 'untugged',\n",
       " 'get-out',\n",
       " 'code-talk',\n",
       " 'less-compelling',\n",
       " 'teeny-bopper',\n",
       " 'family-film',\n",
       " 'ill-wrought',\n",
       " 'solondzian',\n",
       " '\\x96',\n",
       " ...}"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fasttext_vocab - w2v_model.vocab"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The FastText model has 18032 words.\n",
      "The Word2Vec model has 16163 words.\n",
      "The FastText model has 1869 words that are not in the Word2Vec model.\n",
      "The Word2Vec model has 0 words that are not in the FastText model.\n"
     ]
    }
   ],
   "source": [
    "print(\"The FastText model has {} words.\".format(len(fasttext_model.wv.key_to_index)))\n",
    "print(\"The Word2Vec model has {} words.\".format(len(w2v_model.vocab)))\n",
    "print(\"The FastText model has {} words that are not in the Word2Vec model.\".format(len(fasttext_vocab - w2v_model.vocab)))\n",
    "print(\"The Word2Vec model has {} words that are not in the FastText model.\".format(len(w2v_model.vocab - fasttext_vocab)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Question 3b Answer\n",
    "\n",
    "(b) The following code shows the performance of the model with OOV handling\n",
    "where the model would have the learn the embeddings for the unknown words\n",
    "that it might occur and subsequently how to classify an `<UNK>` token that."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Testing the OOV Handling\n",
    "\n",
    "With the OOV handling defined, we now define the handing of the dataloader\n",
    "with the new embedding containing the `<UNK>` token. We will also now\n",
    "define an RNN model with the updated embedding matrix and run it on the\n",
    "data loader to test the performance gain."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "from common_utils import CustomDatasetPreparer, BATCH_SIZE\n",
    "\n",
    "dataset_preparer = CustomDatasetPreparer(\n",
    "    dataset_name=\"rotten_tomatoes\",\n",
    "    batch_size=BATCH_SIZE,\n",
    ")\n",
    "\n",
    "train_loader, val_loader, test_loader = dataset_preparer.get_dataloaders()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The index of <PAD> is:  16163\n",
      "The index of <UNK> is:  16164\n"
     ]
    }
   ],
   "source": [
    "# initialize word embeddings\n",
    "word_embeddings = EmbeddingMatrix.load()\n",
    "word_embeddings.add_padding()\n",
    "word_embeddings.add_unk_token()\n",
    "\n",
    "print(\"The index of <PAD> is: \", word_embeddings.pad_idx)\n",
    "print(\"The index of <UNK> is: \", word_embeddings.unk_idx)\n",
    "\n",
    "\n",
    "basic_RNN = RNN(\n",
    "    hidden_dim=HIDDEN_SIZE,\n",
    "    embedding_dim=EMBEDDING_DIM,\n",
    "    word_embeddings=word_embeddings.to_tensor,\n",
    "    dropout_rate=0.3,\n",
    "    pad_idx=word_embeddings.pad_idx,\n",
    "    num_layers=2,\n",
    ")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  1%|          | 1/100 [00:01<03:06,  1.89s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch 1, train_loss 0.684606 train_acc 0.556038, val_loss 0.667540, val_acc 0.666041\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 11%|█         | 11/100 [00:22<02:56,  1.98s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch 11, train_loss 0.359878 train_acc 0.847128, val_loss 0.515562, val_acc 0.761726\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 21%|██        | 21/100 [00:41<02:34,  1.95s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch 21, train_loss 0.188572 train_acc 0.930598, val_loss 0.597146, val_acc 0.769231\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 23%|██▎       | 23/100 [00:47<02:38,  2.06s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "early stopping after 24 epochs\n",
      "best val loss: 0.49421545592221344\n",
      "best accuracy on val set: 0.7711069418386491\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "epochs = 100\n",
    "es_patience = 15\n",
    "\n",
    "model = basic_RNN\n",
    "optimizer = optim.Adam(model.parameters(), lr=0.0001)\n",
    "loss_fn = nn.BCELoss()\n",
    "scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=20, gamma=0.05)\n",
    "\n",
    "model = model.to(device)\n",
    "loss_fn = loss_fn.to(device)\n",
    "\n",
    "train_loss_, train_acc_, val_loss_, val_acc_ = train_model(train_loader, val_loader, model, loss_fn, optimizer, scheduler, epochs, es_patience)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "test_loss 0.553633, test_acc 0.772045\n"
     ]
    }
   ],
   "source": [
    "plot_loss_accuracy(train_loss_, train_acc_, val_loss_, val_acc_)\n",
    "\n",
    "test_loss, test_acc = test_loop(test_loader, model, loss_fn, optimizer)\n",
    "print(f\"test_loss {test_loss:>7f}, test_acc {test_acc:>4f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Testing the FastText model\n",
    "\n",
    "Fast text model includes the handling through <UNK> and the subword information."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "from common_utils import CustomDatasetPreparer, BATCH_SIZE\n",
    "\n",
    "manual_embedding = EmbeddingMatrix()\n",
    "idx2word = {str(idx):word for idx, word in enumerate(fasttext_model.wv.index_to_key)}\n",
    "manual_embedding.load_manual(fasttext_model.wv.key_to_index, idx2word, fasttext_model.wv.vectors)\n",
    "word_embeddings = manual_embedding\n",
    "\n",
    "dataset_preparer = CustomDatasetPreparer(\n",
    "    dataset_name=\"rotten_tomatoes\",\n",
    "    batch_size=BATCH_SIZE,\n",
    "    manual_embeddings=manual_embedding,\n",
    ")\n",
    "\n",
    "train_loader, val_loader, test_loader = dataset_preparer.get_dataloaders()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The index of <PAD> is:  18031\n",
      "The index of <UNK> is:  9017\n"
     ]
    }
   ],
   "source": [
    "print(\"The index of <PAD> is: \", word_embeddings.pad_idx)\n",
    "print(\"The index of <UNK> is: \", word_embeddings.unk_idx)\n",
    "\n",
    "\n",
    "basic_RNN = RNN(\n",
    "    hidden_dim=HIDDEN_SIZE,\n",
    "    embedding_dim=EMBEDDING_DIM,\n",
    "    word_embeddings=word_embeddings.to_tensor,\n",
    "    dropout_rate=0.3,\n",
    "    pad_idx=word_embeddings.pad_idx,\n",
    "    num_layers=2,\n",
    ")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  1%|          | 1/100 [00:01<03:07,  1.90s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch 1, train_loss 0.694265 train_acc 0.505862, val_loss 0.690997, val_acc 0.548780\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 11%|█         | 11/100 [00:21<02:55,  1.97s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch 11, train_loss 0.403928 train_acc 0.828370, val_loss 0.541296, val_acc 0.727955\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 21%|██        | 21/100 [00:42<02:38,  2.01s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch 21, train_loss 0.097844 train_acc 0.973271, val_loss 0.739760, val_acc 0.746717\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 25%|██▌       | 25/100 [00:51<02:35,  2.07s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "early stopping after 26 epochs\n",
      "best val loss: 0.5412959280339155\n",
      "best accuracy on val set: 0.7598499061913696\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "epochs = 100\n",
    "es_patience = 15\n",
    "\n",
    "model = basic_RNN\n",
    "optimizer = optim.Adam(model.parameters(), lr=0.0001)\n",
    "loss_fn = nn.BCELoss()\n",
    "scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=20, gamma=0.05)\n",
    "\n",
    "model = model.to(device)\n",
    "loss_fn = loss_fn.to(device)\n",
    "\n",
    "train_loss_, train_acc_, val_loss_, val_acc_ = train_model(train_loader, val_loader, model, loss_fn, optimizer, scheduler, epochs, es_patience)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "test_loss 0.647503, test_acc 0.786116\n"
     ]
    }
   ],
   "source": [
    "plot_loss_accuracy(train_loss_, train_acc_, val_loss_, val_acc_)\n",
    "\n",
    "test_loss, test_acc = test_loop(test_loader, model, loss_fn, optimizer)\n",
    "print(f\"test_loss {test_loss:>7f}, test_acc {test_acc:>4f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".venv",
   "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": "3.12.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}