pytorch-sentiment-analysis-on-amazon-reviews.ipynb

{"metadata":{"colab":{"provenance":[],"collapsed_sections":["QkUdpVi4Zun8","EFDG8Q8W6KVc","5-Hiad5HtfZp","x5Q57ZIiAzNZ","igbS2x6LZXq6"],"gpuType":"T4"},"kernelspec":{"name":"python3","display_name":"Python 3","language":"python"},"language_info":{"name":"python","version":"3.10.10","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"gpuClass":"standard","accelerator":"GPU"},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import torch\nfrom torch import nn\nimport torchtext\nimport numpy as np\nimport random, time, json\nfrom tqdm import tqdm\n\nprint(torch.__version__)\nprint(torchtext.__version__)","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"wDY3qCk_K-Uf","outputId":"f3a7d7b1-e037-4647-e940-a64d3cefc02f"},"execution_count":2,"outputs":[{"output_type":"stream","name":"stdout","text":"2.0.1+cu118\n\n0.15.2+cpu\n"}]},{"cell_type":"code","source":"DEVICE = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\nDEVICE","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"g0La2XDvZW5q","outputId":"bf71546b-689f-4e6b-eb5a-0267c648f121"},"execution_count":3,"outputs":[{"output_type":"execute_result","execution_count":3,"data":{"text/plain":["device(type='cuda')"]},"metadata":{}}]},{"cell_type":"markdown","source":"# Dataset","metadata":{"id":"xP1Jo2JqtZUR"}},{"cell_type":"markdown","source":"## Loading Data with Filter","metadata":{"id":"PMM1QQs0tatY"}},{"cell_type":"code","source":"import os\nfor dirname, _, filenames in os.walk('/kaggle/input'):\n    for filename in filenames:\n        print(os.path.join(dirname, filename))","metadata":{"execution":{"iopub.status.busy":"2023-05-21T16:05:42.454087Z","iopub.execute_input":"2023-05-21T16:05:42.454497Z","iopub.status.idle":"2023-05-21T16:05:42.485333Z","shell.execute_reply.started":"2023-05-21T16:05:42.454444Z","shell.execute_reply":"2023-05-21T16:05:42.484217Z"},"trusted":true},"execution_count":1,"outputs":[{"name":"stdout","text":"/kaggle/input/amazonreviews/test.ft.txt.bz2\n/kaggle/input/amazonreviews/train.ft.txt.bz2\n","output_type":"stream"}]},{"cell_type":"code","source":"TRAIN_PATH = \"/kaggle/input/amazonreviews/train.ft.txt.bz2\"\nTEST_PATH = \"/kaggle/input/amazonreviews/test.ft.txt.bz2\"","metadata":{"id":"8AtLhrWfF60t"},"execution_count":7,"outputs":[]},{"cell_type":"code","source":"MIN_LEN_ALLOWED = 3     # select review that has more than MIN_LEN_ALLOWED words\nMAX_LEN_ALLOWED = 60    # select review that has less than MAX_LEN_ALLOWED words\nN_TRAIN = 1_000_000     # maximum number of training samples to select\nN_TEST = 250_000        # maximum number of test samples to select","metadata":{"id":"nwAy7TVq16vm"},"execution_count":8,"outputs":[]},{"cell_type":"code","source":"# since we have large dataset of 3.6 million reviews, we will filter only a few using this function\ndef text_filter(text):\n    alpha = \"abcdefghijklnopqrstuvwxyz \"\n    text = \"\".join([i for i in text.lower().strip() if i in alpha])\n    text_len = len(text.split())\n    return text_len >= MIN_LEN_ALLOWED - 1 and text_len <= MAX_LEN_ALLOWED","metadata":{"id":"rxYb9lWt16vq"},"execution_count":9,"outputs":[]},{"cell_type":"code","source":"import bz2\n\n# thsi function reads data from the file and selects the line that passes through the filter function\ndef extract_content(file_path, N=None):\n    data = []\n    with bz2.open(file_path, 'rt') as f:\n        with tqdm(total=N) as pbar:\n            while (line := f.readline()):\n                line = f.readline()\n                label = int(line[9])\n                text = line[11:]\n                if text_filter(text):\n                    data.append(\n                        [label, text]\n                    )\n                    pbar.update(1)\n                if pbar.n == N: break\n\n    return data","metadata":{"id":"Rq1As9J1AqUW"},"execution_count":10,"outputs":[]},{"cell_type":"code","source":"train = extract_content(TRAIN_PATH, N=N_TRAIN)","metadata":{"id":"pWsnARzrGBG-","colab":{"base_uri":"https://localhost:8080/"},"outputId":"0f701e77-3393-40f2-e862-365058664a40"},"execution_count":11,"outputs":[{"output_type":"stream","name":"stderr","text":" 77%|███████▋  | 767319/1000000 [02:42<00:49, 4717.34it/s]\n"}]},{"cell_type":"code","source":"test = extract_content(TEST_PATH, N=N_TEST)","metadata":{"id":"x_dZtOWQGBU9","colab":{"base_uri":"https://localhost:8080/"},"outputId":"000f896e-cb23-4e50-beed-8f78714ee17c"},"execution_count":12,"outputs":[{"output_type":"stream","name":"stderr","text":" 34%|███▍      | 85135/250000 [00:22<00:42, 3861.90it/s]\n"}]},{"cell_type":"code","source":"# this function balances the dataset such that each of the 2 classes has equal number of samples in train dataset\n\ndef equalize_labels(data):\n    # Count the number of tuples with label 1 and label 2\n    count_1 = sum(1 for item in data if item[0] == 1)\n    count_2 = sum(1 for item in data if item[0] == 2)\n\n\n    # Determine the minimum count between label 1 and label 2\n    min_count = min(count_1, count_2)\n\n    # Collect tuples with label 1 and label 2 up to the minimum count\n    new_data = [item for item in data if item[0] == 1][:min_count]\n    new_data += [item for item in data if item[0] == 2][:min_count]\n\n    # shuffling the data\n    random.shuffle(new_data)\n\n    return new_data","metadata":{"id":"dN2w5DOlRrpn"},"execution_count":13,"outputs":[]},{"cell_type":"code","source":"train = equalize_labels(train)","metadata":{"id":"y4oNMAqFRzWG"},"execution_count":14,"outputs":[]},{"cell_type":"code","source":"N_TRAIN = len(train)\nN_TEST = len(test)\n\nprint(\"Total Train Samples:\", N_TRAIN)\nprint(\"Total TEST Samples:\", N_TEST)","metadata":{"id":"BirtrKsF7wBa","colab":{"base_uri":"https://localhost:8080/"},"outputId":"7951ca6f-70fc-4abd-a299-61751f8b2a98"},"execution_count":15,"outputs":[{"output_type":"stream","name":"stdout","text":"Total Train Samples: 704848\n\nTotal TEST Samples: 85135\n"}]},{"cell_type":"code","source":"train[0], test[0]","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"2Z6RidQX1Q6Z","outputId":"2f8249db-1035-4f51-aad5-2685fc58f9ad"},"execution_count":16,"outputs":[{"output_type":"execute_result","execution_count":16,"data":{"text/plain":["([2,\n","  'One of our 5 Favorite Movies!: This is an excellent movie that can be watched by an entire family - from young children to grandparents.The story and acting are superb! We watch the movie at least once a year and it has become one of our five favorites.\\n'],\n"," [2,\n","  \"works fine, but Maha Energy is better: Check out Maha Energy's website. Their Powerex MH-C204F charger works in 100 minutes for rapid charge, with option for slower charge (better for batteries). And they have 2200 mAh batteries.\\n\"])"]},"metadata":{}}]},{"cell_type":"markdown","source":"## Description of Data\n- Distribution of Number of words in each sentence\n- Label Distribution\n    - 1: Negative Sentiment\n    - 2: Positive Sentiment","metadata":{"id":"EFDG8Q8W6KVc"}},{"cell_type":"code","source":"import pandas as pd\nimport seaborn as sns\nimport matplotlib.pyplot as plt","metadata":{"id":"zsoG3hwD69Et"},"execution_count":17,"outputs":[]},{"cell_type":"code","source":"def plot_length_distribution(data, name, ax):\n    data = [len(i[1].split(\" \")) for i in data]\n    sns.histplot(data, bins=MAX_LEN_ALLOWED-MIN_LEN_ALLOWED, color='skyblue', edgecolor='white', ax=ax)\n    ax.set_xlabel('Length of Sentences')\n    ax.set_ylabel('Frequency')\n    ax.set_title(f'Distribution of Length of Sentences in {name} Data')\n\ndef plot_label_distribution(data, name, ax):\n    data = pd.Series([i[0] for i in data])\n    data = data.value_counts()\n    data.plot(kind='bar', title=f\"Label Distribution in {name} Data\", color=\"skyblue\", ax=ax)","metadata":{"id":"XRSqum4VS4oc"},"execution_count":18,"outputs":[]},{"cell_type":"code","source":"fig, axes = plt.subplots(2, 2, figsize=(12, 12))\n\nplot_length_distribution(train, name=\"Train\", ax=axes[0][0])\nplot_length_distribution(test, name=\"Test\", ax=axes[0][1])\nplot_label_distribution(train, name=\"Train\", ax=axes[1][0])\nplot_label_distribution(test, name=\"Test\", ax=axes[1][1])","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":1000},"id":"OZ7Fdr7dTycV","outputId":"33da88b9-0fbb-4dea-cd0e-77f0c7fb184e"},"execution_count":19,"outputs":[{"output_type":"display_data","data":{"text/plain":["<Figure size 1200x1200 with 4 Axes>"],"image/png":"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\n"},"metadata":{}}]},{"cell_type":"markdown","source":"## Data Pipeline\n- Label Pipeline\n    - converting labels from [1, 2] to [0, 1], negative and positive respectively\n- Text Pipeline\n    - cleaning\n    - tokenizing\n    - vocab","metadata":{"id":"5-Hiad5HtfZp"}},{"cell_type":"markdown","source":"### Text Pipeline","metadata":{"id":"iEPVcCsc2pQY"}},{"cell_type":"code","source":"example_text = \"This is, an example text\"","metadata":{"id":"K7BNHfEJ3oeK"},"execution_count":20,"outputs":[]},{"cell_type":"markdown","source":"#### 1. Cleaning","metadata":{"id":"KtxGpwsy3DzS"}},{"cell_type":"code","source":"import re\ndef clean_text(text):\n    # Remove extra spaces\n    text = text.strip()\n    # Convert multiple spaces to single spaces\n    text = re.sub('\\s+', ' ', text)\n    # Lowercase the text\n    text = text.lower()\n    # Remove punctuation marks\n    text = re.sub('[^\\w\\s]', '', text)\n    return text","metadata":{"id":"QK0RBjt92q48"},"execution_count":21,"outputs":[]},{"cell_type":"code","source":"clean_text(example_text)","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":36},"id":"6qdD17Td2q7b","outputId":"7e6651fb-2090-4d3f-b19b-188f24f54ca4"},"execution_count":22,"outputs":[{"output_type":"execute_result","execution_count":22,"data":{"text/plain":["'this is an example text'"],"application/vnd.google.colaboratory.intrinsic+json":{"type":"string"}},"metadata":{}}]},{"cell_type":"markdown","source":"#### 2. Tokenizing","metadata":{"id":"KFNHRpha3vEL"}},{"cell_type":"code","source":"tokenizer = torchtext.data.utils.get_tokenizer('basic_english')","metadata":{"id":"umBscShV2q-a"},"execution_count":23,"outputs":[]},{"cell_type":"code","source":"tokenizer(\n    clean_text(example_text)\n)","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"RC1mkpgV3-zf","outputId":"9b345346-b24f-4bdb-9f14-d6eae3074ea9"},"execution_count":24,"outputs":[{"output_type":"execute_result","execution_count":24,"data":{"text/plain":["['this', 'is', 'an', 'example', 'text']"]},"metadata":{}}]},{"cell_type":"markdown","source":"#### 3. Vocab","metadata":{"id":"lPcl8zk14ECF"}},{"cell_type":"code","source":"VOCAB_SIZE = 8_000","metadata":{"id":"hxISHNhNuD1C"},"execution_count":25,"outputs":[]},{"cell_type":"code","source":"def vocab_iter(data):\n    for row in data:\n        text = row[1]\n        yield tokenizer(clean_text(text))","metadata":{"id":"h_tOyRG24cZb"},"execution_count":26,"outputs":[]},{"cell_type":"code","source":"print(next(vocab_iter(test)))","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"v7V5O4KTXpNm","outputId":"acc1cfa9-82cb-452e-87c2-bd457318b372"},"execution_count":27,"outputs":[{"output_type":"stream","name":"stdout","text":"['works', 'fine', 'but', 'maha', 'energy', 'is', 'better', 'check', 'out', 'maha', 'energys', 'website', 'their', 'powerex', 'mhc204f', 'charger', 'works', 'in', '100', 'minutes', 'for', 'rapid', 'charge', 'with', 'option', 'for', 'slower', 'charge', 'better', 'for', 'batteries', 'and', 'they', 'have', '2200', 'mah', 'batteries']\n"}]},{"cell_type":"code","source":"vocab = torchtext.vocab.build_vocab_from_iterator(\n    vocab_iter(train),\n    specials=[\"<OOV>\"],\n    max_tokens = VOCAB_SIZE,\n)\nvocab.set_default_index(vocab[\"<OOV>\"])","metadata":{"id":"zl9DpIGsVtnJ"},"execution_count":28,"outputs":[]},{"cell_type":"code","source":"print(vocab.get_itos()[:20])","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"--6TN7ZaNpw6","outputId":"070568bc-22f3-4d52-c2f6-c8f2102b8ffa"},"execution_count":29,"outputs":[{"output_type":"stream","name":"stdout","text":"['<OOV>', 'the', 'i', 'and', 'a', 'it', 'this', 'to', 'is', 'of', 'for', 'not', 'in', 'was', 'my', 'book', 'but', 'you', 'that', 'with']\n"}]},{"cell_type":"code","source":"vocab(\n    tokenizer(\n        clean_text(example_text)\n    )\n)","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"tarnHAmB4R_j","outputId":"3db71099-c9fb-4cae-887a-d63ee91135c0"},"execution_count":30,"outputs":[{"output_type":"execute_result","execution_count":30,"data":{"text/plain":["[6, 8, 44, 1146, 727]"]},"metadata":{}}]},{"cell_type":"markdown","source":"#### 4. Creating Pipeline","metadata":{"id":"VoYS2f0N5Zgu"}},{"cell_type":"code","source":"def TEXT_PIPELINE(text):\n    return vocab(tokenizer(clean_text(text)))\n\nTEXT_PIPELINE(example_text)","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"7NIR6H8H5cqs","outputId":"62b934c5-f314-448f-ee92-a89abb6b2f7e"},"execution_count":31,"outputs":[{"output_type":"execute_result","execution_count":31,"data":{"text/plain":["[6, 8, 44, 1146, 727]"]},"metadata":{}}]},{"cell_type":"markdown","source":"### Label Pipeline","metadata":{"id":"KqlyJKSf52_G"}},{"cell_type":"code","source":"def LABEL_PIPELINE(label):\n    return label - 1\n\nLABEL_PIPELINE(2)","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"UUvC-US-5ctP","outputId":"be6e8458-76ae-4c61-cc58-d5e957a3d236"},"execution_count":32,"outputs":[{"output_type":"execute_result","execution_count":32,"data":{"text/plain":["1"]},"metadata":{}}]},{"cell_type":"markdown","source":"### Pipeline","metadata":{"id":"x5Q57ZIiAzNZ"}},{"cell_type":"code","source":"def PIPELINE(sample):\n    return (LABEL_PIPELINE(sample[0]), TEXT_PIPELINE(sample[1]))\n\nprint(PIPELINE(train[0]))","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"7M8JwEHv5cve","outputId":"abb4ec27-cd6d-4086-a3d3-3643f958cc64"},"execution_count":33,"outputs":[{"output_type":"stream","name":"stdout","text":"(1, [27, 9, 179, 224, 244, 253, 6, 8, 44, 118, 39, 18, 65, 28, 576, 70, 44, 624, 331, 38, 571, 433, 7, 0, 105, 3, 460, 25, 1126, 81, 163, 1, 39, 36, 416, 330, 4, 186, 3, 5, 47, 916, 27, 9, 179, 708, 1024])\n"}]},{"cell_type":"code","source":"new_train = []\nfor item in tqdm(train):\n    new_train.append(PIPELINE(item))\ntrain = new_train\n\nnew_test = []\nfor item in tqdm(test):\n    new_test.append(PIPELINE(item))\ntest = new_test","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"mIa7XgS1WClu","outputId":"4ddf130f-f2be-4941-8377-b9c69642753a"},"execution_count":34,"outputs":[{"output_type":"stream","name":"stderr","text":"100%|██████████| 704848/704848 [01:02<00:00, 11315.60it/s]\n\n100%|██████████| 85135/85135 [00:06<00:00, 12192.95it/s]\n"}]},{"cell_type":"markdown","source":"## Dataloader","metadata":{"id":"ZzzEAceCleFE"}},{"cell_type":"code","source":"PADDING_IDX = vocab['<OOV>']","metadata":{"id":"TqYuUrOKXez7"},"execution_count":35,"outputs":[]},{"cell_type":"code","source":"# this function is called when a batch is generated in from DataLoader\ndef collate_batch(batch):\n    time_i = time.time()\n    label_list, text_list = [], []\n    for (_label, _text) in batch:\n        label_list.append(_label)\n        text_list.append(torch.tensor(_text))\n            \n    label_list = torch.tensor(label_list, dtype=torch.float32).to(DEVICE)\n    sequences =  torch.nn.utils.rnn.pad_sequence(\n        text_list,\n        batch_first = True,\n        padding_value = PADDING_IDX\n    ).to(DEVICE)\n\n    return sequences, label_list\n\n# this function decides how to split the dataset into batches, here sentences of similar lengths are kept in a batch together\nclass SortedSampler(torch.utils.data.Sampler):\n    def __init__(self, data_source, batch_size, shuffle_times):\n        self.batch_size = batch_size\n        self.shuffle_times = shuffle_times\n\n        # getting sorted index\n        self.idx = [i[0] for i in sorted(enumerate(data_source), key=lambda x: len(x[1]))]\n\n    def shuffle_idx(self):\n        # starting points for each batch\n        offsets = list(range(0, len(self.idx), self.batch_size))[:-1]\n        # shuffling inter-batch\n        for i in range(self.shuffle_times):\n            i1, i2 = random.choices(offsets, k=2)\n            self.idx[i1:i1+self.batch_size], self.idx[i2:i2+self.batch_size] = self.idx[i2:i2+self.batch_size], self.idx[i1:i1+self.batch_size]\n\n    def __iter__(self):\n        return iter(self.idx)\n\n    def __len__(self):\n        return len(self.idx)","metadata":{"id":"35XhPtzqlh2B"},"execution_count":36,"outputs":[]},{"cell_type":"markdown","source":"# Modelling","metadata":{"id":"iQ7fcO0niUOI"}},{"cell_type":"markdown","source":"## Custom Embeddings","metadata":{"id":"igbS2x6LZXq6"}},{"cell_type":"code","source":"from gensim.models import Word2Vec\n\n# Initializing\nmodel = Word2Vec(\n    vector_size = 128,\n    window = 5, \n    min_count = 1\n)\n\n# Vocabulary\nmodel.build_vocab_from_freq({i: VOCAB_SIZE - i + 1 for i in range(VOCAB_SIZE)})\nfor k, v in model.wv.key_to_index.items():\n    assert k == v\n\n# Training\nmodel.train(\n    [i[1] for i in train],\n    total_examples=len(train),\n    epochs=3\n)\n\n# Extracting Embeddings\npretrained_embeddings = model.wv.vectors\npretrained_embeddings = torch.tensor(pretrained_embeddings)\n\npretrained_embeddings.shape","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"XjBL1Y39ZZu6","outputId":"f634141c-279e-483d-c23f-62c8ffe1b5bb"},"execution_count":37,"outputs":[{"output_type":"execute_result","execution_count":37,"data":{"text/plain":["torch.Size([8000, 128])"]},"metadata":{}}]},{"cell_type":"code","source":"word = \"good\"\nprint(f\"Similar Words for \\\"{word}\\\":\")\nfor i in model.wv.most_similar(vocab([word])[0]):\n    print(\"->\", vocab.get_itos()[i[0]])","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"aNkE5lTwd8Yi","outputId":"44d8847c-bf4d-464f-8b4f-78caa3996bf2"},"execution_count":38,"outputs":[{"output_type":"stream","name":"stdout","text":"Similar Words for \"good\":\n\n-> great\n\n-> decent\n\n-> nice\n\n-> bad\n\n-> cool\n\n-> excellent\n\n-> neat\n\n-> ok\n\n-> okay\n\n-> terrific\n"}]},{"cell_type":"markdown","source":"## Utilities","metadata":{"id":"Ryxo8IY3GnCi"}},{"cell_type":"code","source":"def accuracy(preds, labels):\n    bools = preds.round() == labels\n    return bools.sum(0) / len(bools)\n\ndef predict_texts(texts, _print=True):\n\n    x, y = collate_batch(\n        [[0, TEXT_PIPELINE(text)] for text in texts]\n    )\n    model.eval()\n    preds = model(x)\n    preds = preds.tolist()\n\n    if _print:\n        for pred, orig_text in zip(preds, texts):\n            print(f\"{int(round(pred))} ({round(pred, 3)}) - {orig_text}\")\n    else:\n        return preds\n\ndef plot_history(history, skip_keys=[], splits=1, figsize=None):\n\n    if figsize == None: figsize = (15, 15//len(history))\n    fig, axs = plt.subplots(1, len(history), figsize=figsize)\n\n    for i, (_key, _val) in enumerate(history.items()):\n        for key, val in _val.items():\n            if key in skip_keys: continue\n            # creating splits\n            val = torch.tensor(val)\n            val = val[:(len(val) // splits) * splits].view(splits, -1).mean(axis=1)\n            axs[i].plot(val, label=key)\n        axs[i].legend()\n        axs[i].set_title(_key)\n\n    plt.show()","metadata":{"id":"f5DOd15ziV4I"},"execution_count":39,"outputs":[]},{"cell_type":"markdown","source":"## Model Architechture\n- Input `(batch_size, seq_len)`\n- Embedding Layer `(batch_size, seq_len, emb_dim)`\n- RNN Layer `(batch_size, seq_len, hidden_size)`\n- Attention `(batch_size, seq_len, hidden_size)`\n- RNN Layer `(batch_size, hidden_size)`\n- Neural Network `(batch_size, inner_hidden_size)`\n- Output `(batch_size, 1)`","metadata":{"id":"2v61O646HZIT"}},{"cell_type":"code","source":"def self_attention(Q, K, V):\t\n    d = K.shape[-1]\n    QK = Q @ K.transpose(-2, -1)\n    QK_d = QK / (d ** 0.5)\n    weights = torch.softmax(QK_d, axis=-1)\n    outputs = weights @ V\n    return outputs\n\nclass Attention(torch.nn.Module):\n    def __init__(self, emb_dim, n_heads):\n        super(Attention, self).__init__()\n\n        self.emb_dim = emb_dim\n        self.n_heads = n_heads\n\n    def forward(self, X):\n\n        batch_size, seq_len, emb_dim = X.size() # (batch_size, seq_len, emb_dim)\n        n_heads = self.n_heads\n        emb_dim_per_head = emb_dim // n_heads\n\n        assert emb_dim == self.emb_dim\n        assert emb_dim_per_head * n_heads == emb_dim\n\n        X = X.transpose(1, 2)\n        output = self_attention(X, X, X)       # (batch_size, n_heads, seq_len, emb_dim_per_head)\n        output = output.transpose(1, 2)   # (batch_size, seq_len, n_heads, emb_dim_per_head)\n        output = output.contiguous().view(batch_size, seq_len, emb_dim)    # (batch_size, seq_len, emb_dim)\n\n        return output\n\nclass ClassifierAttention(nn.Module):\n    def __init__(self, vocab_size, emb_dim, padding_idx, hidden_size, n_layers, attention_heads, hidden_layer_units, dropout):\n        super(ClassifierAttention, self).__init__()\n\n        self.embedding = nn.Embedding(\n            num_embeddings = vocab_size,\n            embedding_dim = emb_dim,\n            padding_idx = padding_idx\n        )\n\n        self.rnn_1 = nn.LSTM(\n            emb_dim,\n            hidden_size,\n            n_layers,\n            bidirectional = False,\n            batch_first = True,\n        )\n\n        self.attention = Attention(hidden_size, attention_heads)\n\n        self.rnn_2 = nn.LSTM(\n            hidden_size,\n            hidden_size,\n            n_layers,\n            bidirectional = False,\n            batch_first = True,\n        )\n\n        self.dropout = nn.Dropout(dropout)\n\n        hidden_layer_units = [hidden_size, *hidden_layer_units]\n        self.hidden_layers = nn.ModuleList([])\n        for in_unit, out_unit in zip(hidden_layer_units[:-1], hidden_layer_units[1:]):\n            self.hidden_layers.append(nn.Linear(in_unit, out_unit))\n            self.hidden_layers.append(nn.ReLU())\n            self.hidden_layers.append(self.dropout)\n        self.hidden_layers.append(nn.Linear(hidden_layer_units[-1], 1))\n\n        self.sigmoid = nn.Sigmoid()\n\n        self.add_custom_embeddings()\n\n    def add_custom_embeddings(self):\n        self.embedding.weight.data.copy_(pretrained_embeddings)\n        self.embedding.weight.requires_grad = False\n\n    def forward(self, x):\n        # x: (batch_size, seq_len)\n\n        out = self.embedding(x)     # (batch_size, seq_len, emb_dim)\n        out, (hidden_state, cell_state) = self.rnn_1(out)\n        out = self.attention(out)    # (batch_size, seq_len, hidden_size)\n        out = self.dropout(out)\n        output, (hidden_state, cell_state) = self.rnn_2(out)\n        out = hidden_state[-1]      # (batch_size, hidden_size)\n        out = self.dropout(out)\n        # (batch_size, seq_len, hidden_dim)\n        # (n_layers*n_direction, batch_size, hidden_size)\n        # (n_layers*n_direction, batch_size, hidden_size)\n\n        for layer in self.hidden_layers:\n            out = layer(out)\n\n        out = self.sigmoid(out) # (batch_size, 1)\n        out = out.squeeze(-1)   # (batch_size)\n\n        return out  \n","metadata":{"id":"Kmoxe3DieS0B"},"execution_count":50,"outputs":[]},{"cell_type":"markdown","source":"## Training","metadata":{"id":"t6AFhT_NHghE"}},{"cell_type":"code","source":"# DATA PARAMETERS\nBATCH_SIZE = 8\nTEST_BATCH_SIZE = 64\nSHUFFLE_TIMES = 500_000\nTRAIN_ITERATIONS_PER_EPOCH = N_TRAIN // BATCH_SIZE\nTEST_ITERATIONS_PER_EPOCH = N_TEST // TEST_BATCH_SIZE\n# MODEL PARAMETERS\nEMB_DIM = 128\nLSTM_HIDDEN_SIZE = 64\nLSTM_N_LAYERS = 1\nATTENTION_HEADS = 2\nHIDDEN_LAYER_UNITS = [64, 64]\nDROPOUT = 0.4\n# TRAINGING PARAMETERS\nEPOCHS = 3\nLR = 0.12\nMOMENTUM = 0\nSCHEDULER_STEP_SIZE = 1\nSCHEDULER_FACTOR = 0.8\n# TESTING PARAMETERS\nINFO_SPLIT = 5\nTEST_SENTENCES = [\n    # negative sentences\n    \"I'm extremely disappointed with the quality of this item; it broke within a week.\",\n    \"The customer support for this product is terrible; they were unhelpful and rude.\",#\n    \"I had high expectations for this item, but it turned out to be a total letdown.\",\n    \"The shipping was delayed, and the packaging was damaged when the product arrived.\",\n    \"I encountered multiple issues with this item, and the manufacturer provided no solutions.\",\n    \"The features advertised for this product are misleading; it doesn't perform as claimed.\",#\n    \"I found this item to be overpriced for its quality; it doesn't justify the cost.\",\n    \"The user interface is confusing and difficult to navigate; it's not user-friendly at all.\",#\n\n    # possitive sentences\n    \"The product exceeded my expectations and I am extremely satisfied with it.\",#\n    \"This is the best product I've ever purchased, and I can't recommend it enough.\",\n    \"The quality of the product is outstanding, and it's worth every penny.\",#\n    \"I've had this item for a week now, and I'm impressed with its performance.\",#\n    \"I was skeptical at first, but this product turned out to be a pleasant surprise.\",\n    \"The product arrived in a damaged box, but fortunately, the item itself was fine.\",#\n    \"I've tried many similar products, but this one stands out from the rest.\",#\n    \"The features of this item are impressive, and it offers great value for money.\",#\n]","metadata":{"id":"XwvZxZMMpO3t"},"execution_count":98,"outputs":[]},{"cell_type":"code","source":"print(\"TRAIN_ITERATIONS_PER_EPOCH\", TRAIN_ITERATIONS_PER_EPOCH)\nprint(\"TEST_ITERATIONS_PER_EPOCH\", TEST_ITERATIONS_PER_EPOCH)\n\nlr = LR\nsteps = 0\nfor i in range(EPOCHS):\n    print(f\"Epoch: {i+1}\\tLR: {lr}\")\n    steps += 1\n    if steps == SCHEDULER_STEP_SIZE:\n        steps = 0\n        lr *= SCHEDULER_FACTOR","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"h_QqtjiYXeae","outputId":"0fd56ab8-f993-4f12-f4ca-d9566eeddfb8"},"execution_count":99,"outputs":[{"output_type":"stream","name":"stdout","text":"TRAIN_ITERATIONS_PER_EPOCH 88106\n\nTEST_ITERATIONS_PER_EPOCH 1330\n\nEpoch: 1\tLR: 0.12\n\nEpoch: 2\tLR: 0.096\n\nEpoch: 3\tLR: 0.07680000000000001\n"}]},{"cell_type":"code","source":"# ===> TRAIN DATALOADER\ntrain_dataloader = torch.utils.data.DataLoader(\n    train,\n    batch_size = BATCH_SIZE,\n    collate_fn = collate_batch,\n    sampler = SortedSampler(train, BATCH_SIZE, SHUFFLE_TIMES)\n)\n\n# ===> TEST DATALOADER\ntest_dataloader = torch.utils.data.DataLoader(\n    test,\n    batch_size = TEST_BATCH_SIZE,\n    collate_fn = collate_batch,\n    sampler = SortedSampler(test, TEST_BATCH_SIZE, 0)\n)","metadata":{"id":"025uAlpy8a8E"},"execution_count":100,"outputs":[]},{"cell_type":"code","source":"# ===> MODEL\nmodel = ClassifierAttention(VOCAB_SIZE, EMB_DIM, PADDING_IDX, LSTM_HIDDEN_SIZE, LSTM_N_LAYERS, ATTENTION_HEADS, HIDDEN_LAYER_UNITS, DROPOUT).to(DEVICE)\n\n# ===> LOSS\ncriterion = nn.BCELoss()\n\n# ===> OPTIMIZER\noptimizer = torch.optim.SGD(model.parameters(), lr=LR, momentum=MOMENTUM)\n\n# ===> SCHEDULER\nscheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=SCHEDULER_STEP_SIZE, gamma=SCHEDULER_FACTOR)","metadata":{"id":"2psT8kQXpUxF"},"execution_count":101,"outputs":[]},{"cell_type":"code","source":"model","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"Svpf2jWqN5B7","outputId":"5bc90a4d-c4ba-4451-9a6d-e81a37e2401e"},"execution_count":102,"outputs":[{"output_type":"execute_result","execution_count":102,"data":{"text/plain":["ClassifierAttention(\n","  (embedding): Embedding(8000, 128, padding_idx=0)\n","  (rnn_1): LSTM(128, 64, batch_first=True)\n","  (attention): Attention()\n","  (rnn_2): LSTM(64, 64, batch_first=True)\n","  (dropout): Dropout(p=0.4, inplace=False)\n","  (hidden_layers): ModuleList(\n","    (0): Linear(in_features=64, out_features=64, bias=True)\n","    (1): ReLU()\n","    (2): Dropout(p=0.4, inplace=False)\n","    (3): Linear(in_features=64, out_features=64, bias=True)\n","    (4): ReLU()\n","    (5): Dropout(p=0.4, inplace=False)\n","    (6): Linear(in_features=64, out_features=1, bias=True)\n","  )\n","  (sigmoid): Sigmoid()\n",")"]},"metadata":{}}]},{"cell_type":"code","source":"predict_texts(TEST_SENTENCES)","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"5_J--QZG9QSF","outputId":"265bdb61-b05b-449b-aa0d-cc5a1ea59881"},"execution_count":103,"outputs":[{"output_type":"stream","name":"stdout","text":"1 (0.502) - I'm extremely disappointed with the quality of this item; it broke within a week.\n\n1 (0.502) - The customer support for this product is terrible; they were unhelpful and rude.\n\n1 (0.502) - I had high expectations for this item, but it turned out to be a total letdown.\n\n1 (0.502) - The shipping was delayed, and the packaging was damaged when the product arrived.\n\n1 (0.502) - I encountered multiple issues with this item, and the manufacturer provided no solutions.\n\n1 (0.502) - The features advertised for this product are misleading; it doesn't perform as claimed.\n\n1 (0.502) - I found this item to be overpriced for its quality; it doesn't justify the cost.\n\n1 (0.502) - The user interface is confusing and difficult to navigate; it's not user-friendly at all.\n\n1 (0.502) - The product exceeded my expectations and I am extremely satisfied with it.\n\n1 (0.502) - This is the best product I've ever purchased, and I can't recommend it enough.\n\n1 (0.502) - The quality of the product is outstanding, and it's worth every penny.\n\n1 (0.502) - I've had this item for a week now, and I'm impressed with its performance.\n\n1 (0.502) - I was skeptical at first, but this product turned out to be a pleasant surprise.\n\n1 (0.502) - The product arrived in a damaged box, but fortunately, the item itself was fine.\n\n1 (0.502) - I've tried many similar products, but this one stands out from the rest.\n\n1 (0.502) - The features of this item are impressive, and it offers great value for money.\n"}]},{"cell_type":"code","source":"losses = {\n    \"train\": [],\n    \"test\": [],\n}\nmetrics = {\n    \"train\": [],\n    \"test\": [],\n}\n\nfor epoch in range(EPOCHS):\n\n    # TRAINING\n    with tqdm(total=TRAIN_ITERATIONS_PER_EPOCH) as pbar:\n        model.train()\n        temp_acc = 0\n        temp_loss = 0\n        i = 0\n        for x, y in train_dataloader:\n            optimizer.zero_grad()\n            preds = model(x)\n            loss = criterion(preds, y)\n            loss.backward()\n            optimizer.step()\n            metric = accuracy(preds, y)\n\n            losses['train'].append(loss.item())\n            metrics['train'].append(metric.item())\n\n            pbar.update(1)\n\n            if pbar.n % (pbar.total // INFO_SPLIT) == 0:\n                # interrupting the progress bar so last metrics and losses are visible\n                # writing to progress bar\n                round_to = 4\n                temp_loss = str(round(np.mean(losses['train'][-pbar.total // INFO_SPLIT:]), round_to))\n                temp_acc = str(round(np.mean(metrics['train'][-pbar.total // INFO_SPLIT:]), round_to))\n                temp_loss += \" \" * (round_to + 2 - len(temp_loss))\n                temp_acc += \" \" * (round_to + 2 - len(temp_acc))\n\n                pbar.set_description(f\"Epoch {epoch+1}/{EPOCHS} ----> [ Loss: {temp_loss} Accuracy: {temp_acc} ]\")\n                print()\n\n    # TESTING\n    with torch.no_grad():\n        model.eval()\n        for x, y in test_dataloader:\n            preds = model(x)\n            loss = criterion(preds, y)\n            metric = accuracy(preds, y)\n\n            losses['test'].append(loss.item())\n            metrics['test'].append(metric.item())\n        print(f\"Epoch {epoch+1}/{EPOCHS} ----> [ Test Loss: {np.mean(losses['test'])} Test Accuracy: {np.mean(metrics['test'])} ]\")\n        print()\n\n    # END OF EPOCH\n    scheduler.step()\n\n    predict_texts(TEST_SENTENCES)\n    print()\n\n    if epoch > 0:\n        plot_history(\n            splits = epoch+1,\n            skip_keys = [],\n            figsize = (10, 10//3),\n            history = {\n                \"loss\": losses,\n                \"accuracy\": metrics,\n            }\n        )\n","metadata":{"id":"kcVV3oeQiV67","colab":{"base_uri":"https://localhost:8080/","height":1000},"outputId":"90607aa7-c3ff-44ee-e9b5-83283370f7f8"},"execution_count":104,"outputs":[{"output_type":"stream","name":"stderr","text":"Epoch 1/3 ----> [ Loss: 0.415  Accuracy: 0.775  ]:  20%|██        | 17667/88106 [01:09<04:18, 272.65it/s]"},{"output_type":"stream","name":"stdout","text":"\n"},{"output_type":"stream","name":"stderr","text":"Epoch 1/3 ----> [ Loss: 0.227  Accuracy: 0.9131 ]:  40%|████      | 35284/88106 [02:19<03:25, 256.90it/s]"},{"output_type":"stream","name":"stdout","text":"\n"},{"output_type":"stream","name":"stderr","text":"Epoch 1/3 ----> [ Loss: 0.2092 Accuracy: 0.9212 ]:  60%|██████    | 52910/88106 [03:31<02:19, 253.17it/s]"},{"output_type":"stream","name":"stdout","text":"\n"},{"output_type":"stream","name":"stderr","text":"Epoch 1/3 ----> [ Loss: 0.2009 Accuracy: 0.9248 ]:  80%|████████  | 70536/88106 [04:42<01:05, 269.96it/s]"},{"output_type":"stream","name":"stdout","text":"\n"},{"output_type":"stream","name":"stderr","text":"Epoch 1/3 ----> [ Loss: 0.1921 Accuracy: 0.9275 ]: 100%|██████████| 88106/88106 [05:52<00:00, 249.70it/s]\n"},{"output_type":"stream","name":"stdout","text":"\n\nEpoch 1/3 ----> [ Test Loss: 0.18662463441619875 Test Accuracy: 0.9306293827174392 ]\n\n\n\n0 (0.0) - I'm extremely disappointed with the quality of this item; it broke within a week.\n\n0 (0.0) - The customer support for this product is terrible; they were unhelpful and rude.\n\n0 (0.031) - I had high expectations for this item, but it turned out to be a total letdown.\n\n0 (0.084) - The shipping was delayed, and the packaging was damaged when the product arrived.\n\n1 (0.541) - I encountered multiple issues with this item, and the manufacturer provided no solutions.\n\n0 (0.01) - The features advertised for this product are misleading; it doesn't perform as claimed.\n\n0 (0.005) - I found this item to be overpriced for its quality; it doesn't justify the cost.\n\n0 (0.001) - The user interface is confusing and difficult to navigate; it's not user-friendly at all.\n\n1 (0.998) - The product exceeded my expectations and I am extremely satisfied with it.\n\n1 (0.999) - This is the best product I've ever purchased, and I can't recommend it enough.\n\n1 (0.999) - The quality of the product is outstanding, and it's worth every penny.\n\n1 (0.98) - I've had this item for a week now, and I'm impressed with its performance.\n\n1 (0.874) - I was skeptical at first, but this product turned out to be a pleasant surprise.\n\n0 (0.028) - The product arrived in a damaged box, but fortunately, the item itself was fine.\n\n0 (0.34) - I've tried many similar products, but this one stands out from the rest.\n\n1 (0.999) - The features of this item are impressive, and it offers great value for money.\n\n\n"},{"output_type":"stream","name":"stderr","text":"Epoch 2/3 ----> [ Loss: 0.1853 Accuracy: 0.9313 ]:  20%|██        | 17667/88106 [01:11<04:24, 266.79it/s]"},{"output_type":"stream","name":"stdout","text":"\n"},{"output_type":"stream","name":"stderr","text":"Epoch 2/3 ----> [ Loss: 0.1778 Accuracy: 0.933  ]:  40%|████      | 35279/88106 [02:21<03:19, 264.97it/s]"},{"output_type":"stream","name":"stdout","text":"\n"},{"output_type":"stream","name":"stderr","text":"Epoch 2/3 ----> [ Loss: 0.1778 Accuracy: 0.9335 ]:  60%|██████    | 52911/88106 [03:32<02:15, 260.12it/s]"},{"output_type":"stream","name":"stdout","text":"\n"},{"output_type":"stream","name":"stderr","text":"Epoch 2/3 ----> [ Loss: 0.1772 Accuracy: 0.9348 ]:  80%|████████  | 70504/88106 [04:43<01:32, 190.34it/s]"},{"output_type":"stream","name":"stdout","text":"\n"},{"output_type":"stream","name":"stderr","text":"Epoch 2/3 ----> [ Loss: 0.1735 Accuracy: 0.9363 ]: 100%|██████████| 88106/88106 [05:54<00:00, 248.62it/s]\n"},{"output_type":"stream","name":"stdout","text":"\n\nEpoch 2/3 ----> [ Test Loss: 0.1825205061429374 Test Accuracy: 0.9331181129954257 ]\n\n\n\n0 (0.0) - I'm extremely disappointed with the quality of this item; it broke within a week.\n\n0 (0.0) - The customer support for this product is terrible; they were unhelpful and rude.\n\n0 (0.018) - I had high expectations for this item, but it turned out to be a total letdown.\n\n0 (0.165) - The shipping was delayed, and the packaging was damaged when the product arrived.\n\n1 (0.874) - I encountered multiple issues with this item, and the manufacturer provided no solutions.\n\n0 (0.015) - The features advertised for this product are misleading; it doesn't perform as claimed.\n\n0 (0.003) - I found this item to be overpriced for its quality; it doesn't justify the cost.\n\n0 (0.0) - The user interface is confusing and difficult to navigate; it's not user-friendly at all.\n\n1 (0.999) - The product exceeded my expectations and I am extremely satisfied with it.\n\n1 (0.998) - This is the best product I've ever purchased, and I can't recommend it enough.\n\n1 (0.999) - The quality of the product is outstanding, and it's worth every penny.\n\n1 (0.99) - I've had this item for a week now, and I'm impressed with its performance.\n\n1 (0.946) - I was skeptical at first, but this product turned out to be a pleasant surprise.\n\n0 (0.166) - The product arrived in a damaged box, but fortunately, the item itself was fine.\n\n0 (0.065) - I've tried many similar products, but this one stands out from the rest.\n\n1 (0.999) - The features of this item are impressive, and it offers great value for money.\n\n\n"},{"output_type":"display_data","data":{"text/plain":["<Figure size 1000x300 with 2 Axes>"],"image/png":"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\n"},"metadata":{}},{"output_type":"stream","name":"stderr","text":"Epoch 3/3 ----> [ Loss: 0.1699 Accuracy: 0.9369 ]:  20%|██        | 17665/88106 [01:11<04:28, 262.54it/s]"},{"output_type":"stream","name":"stdout","text":"\n"},{"output_type":"stream","name":"stderr","text":"Epoch 3/3 ----> [ Loss: 0.1653 Accuracy: 0.9385 ]:  40%|████      | 35282/88106 [02:23<04:16, 205.90it/s]"},{"output_type":"stream","name":"stdout","text":"\n"},{"output_type":"stream","name":"stderr","text":"Epoch 3/3 ----> [ Loss: 0.1636 Accuracy: 0.9395 ]:  60%|██████    | 52917/88106 [03:33<02:09, 272.43it/s]"},{"output_type":"stream","name":"stdout","text":"\n"},{"output_type":"stream","name":"stderr","text":"Epoch 3/3 ----> [ Loss: 0.1654 Accuracy: 0.9392 ]:  80%|████████  | 70512/88106 [04:44<01:07, 261.60it/s]"},{"output_type":"stream","name":"stdout","text":"\n"},{"output_type":"stream","name":"stderr","text":"Epoch 3/3 ----> [ Loss: 0.1618 Accuracy: 0.9403 ]: 100%|██████████| 88106/88106 [05:54<00:00, 248.50it/s]\n"},{"output_type":"stream","name":"stdout","text":"\n\nEpoch 3/3 ----> [ Test Loss: 0.17608294603450986 Test Accuracy: 0.9357174534920669 ]\n\n\n\n0 (0.0) - I'm extremely disappointed with the quality of this item; it broke within a week.\n\n0 (0.0) - The customer support for this product is terrible; they were unhelpful and rude.\n\n0 (0.012) - I had high expectations for this item, but it turned out to be a total letdown.\n\n0 (0.446) - The shipping was delayed, and the packaging was damaged when the product arrived.\n\n1 (0.745) - I encountered multiple issues with this item, and the manufacturer provided no solutions.\n\n0 (0.015) - The features advertised for this product are misleading; it doesn't perform as claimed.\n\n0 (0.015) - I found this item to be overpriced for its quality; it doesn't justify the cost.\n\n0 (0.0) - The user interface is confusing and difficult to navigate; it's not user-friendly at all.\n\n1 (0.999) - The product exceeded my expectations and I am extremely satisfied with it.\n\n1 (0.998) - This is the best product I've ever purchased, and I can't recommend it enough.\n\n1 (0.999) - The quality of the product is outstanding, and it's worth every penny.\n\n1 (0.996) - I've had this item for a week now, and I'm impressed with its performance.\n\n1 (0.873) - I was skeptical at first, but this product turned out to be a pleasant surprise.\n\n0 (0.197) - The product arrived in a damaged box, but fortunately, the item itself was fine.\n\n0 (0.155) - I've tried many similar products, but this one stands out from the rest.\n\n1 (0.999) - The features of this item are impressive, and it offers great value for money.\n\n\n"},{"output_type":"display_data","data":{"text/plain":["<Figure size 1000x300 with 2 Axes>"],"image/png":"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\n"},"metadata":{}}]},{"cell_type":"code","source":"plot_history(\n    splits = 200,\n    skip_keys = [\n        # \"train\",\n        \"test\"\n    ],\n    figsize = (14, 14//3),\n    history = {\n        \"loss\": losses,\n        \"accuracy\": metrics,\n    }\n)","metadata":{"id":"HwOMeqlbtolr","colab":{"base_uri":"https://localhost:8080/","height":391},"outputId":"76234581-3b7a-44a5-8600-7a26ace16a40"},"execution_count":111,"outputs":[{"output_type":"display_data","data":{"text/plain":["<Figure size 1400x400 with 2 Axes>"],"image/png":"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\n"},"metadata":{}}]},{"cell_type":"markdown","source":"## Validating","metadata":{"id":"tTYm2BN_lfDo"}},{"cell_type":"code","source":"# GETTING THE PREDICTIONS\n\nvalid_preds = []\nvalid_labels = []\nwith torch.no_grad():\n    model.eval()\n    for x, y in test_dataloader:\n        preds = model(x)\n        valid_preds.extend(preds.round().type(torch.int32).tolist())\n        valid_labels.extend(y.type(torch.int32).tolist())","metadata":{"id":"m5rJEY89hrd_"},"execution_count":106,"outputs":[]},{"cell_type":"code","source":"import numpy as np\nimport matplotlib.pyplot as plt\nfrom sklearn.metrics import accuracy_score, confusion_matrix, precision_score, recall_score, f1_score, classification_report\n\n# Calculate accuracy\nacc = accuracy_score(valid_labels, valid_preds)\nprint(f\"Accuracy: {acc:.4f}\")\n\n# Calculate confusion matrix\ncm = confusion_matrix(valid_labels, valid_preds)\nprint()\nprint(\"Confusion Matrix:\")\nprint(cm)\nprint()\n\n# Plot heatmap of confusion matrix\nplt.imshow(cm, cmap='Blues')\nplt.title('Confusion Matrix')\nplt.colorbar()\nplt.xlabel('Predicted Label')\nplt.ylabel('True Label')\nplt.xticks([0, 1], ['0', '1'])\nplt.yticks([0, 1], ['0', '1'])\nplt.show()\n\n# Calculate precision, recall, and F1 score\nprecision = precision_score(valid_labels, valid_preds)\nrecall = recall_score(valid_labels, valid_preds)\nf1 = f1_score(valid_labels, valid_preds)\n\nprint()\nprint(f\"Precision: {precision:.4f}\")\nprint(f\"Recall: {recall:.4f}\")\nprint(f\"F1 Score: {f1:.4f}\")\n\n# Print classification report\nclassification_rep = classification_report(valid_labels, valid_preds)\nprint()\nprint(\"Classification Report:\")\nprint(classification_rep)","metadata":{"id":"_yUrWx6rkryr","colab":{"base_uri":"https://localhost:8080/","height":836},"outputId":"e02064a8-82df-445f-d3bc-3e71e4827053"},"execution_count":107,"outputs":[{"output_type":"stream","name":"stdout","text":"Accuracy: 0.9409\n\n\n\nConfusion Matrix:\n\n[[36074  3042]\n\n [ 1991 44028]]\n\n\n"},{"output_type":"display_data","data":{"text/plain":["<Figure size 640x480 with 2 Axes>"],"image/png":"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\n"},"metadata":{}},{"output_type":"stream","name":"stdout","text":"\n\nPrecision: 0.9354\n\nRecall: 0.9567\n\nF1 Score: 0.9459\n\n\n\nClassification Report:\n\n              precision    recall  f1-score   support\n\n\n\n           0       0.95      0.92      0.93     39116\n\n           1       0.94      0.96      0.95     46019\n\n\n\n    accuracy                           0.94     85135\n\n   macro avg       0.94      0.94      0.94     85135\n\nweighted avg       0.94      0.94      0.94     85135\n\n\n"}]},{"cell_type":"code","source":"predict_texts([\"that was not a bad experience of my life\"])\npredict_texts([\"that was a bad experience of my life\"])","metadata":{"id":"0-cpWP_Onchh","colab":{"base_uri":"https://localhost:8080/"},"outputId":"52a9fa19-80c6-4d6a-9937-8a3b608c7bf3"},"execution_count":108,"outputs":[{"output_type":"stream","name":"stdout","text":"1 (0.953) - that was not a bad experience of my life\n\n1 (0.601) - that was a bad experience of my life\n"}]},{"cell_type":"code","source":"x, y = collate_batch(\n    [[0, TEXT_PIPELINE(text)] for text in TEST_SENTENCES]\n)\nmodel.eval()\npreds = model(x)\n# preds = preds.round()\npreds = preds.tolist()\n\nfor pred, orig_text in zip(preds, TEST_SENTENCES):\n    print(f\"Text: {orig_text}\")\n    print(f\"Prediction: {['Negative Sentiment', 'Positive Sentiment'][int(round(pred))]} ( Confidence in Prediction: {abs(round(( pred - (1 - round(pred)) ) * 100 , 2))}% )\")\n    print()","metadata":{"id":"zoGDB4W4Zc3p","colab":{"base_uri":"https://localhost:8080/"},"outputId":"ce1a7b02-bfd7-4bbf-d839-7adb3da4ed4a"},"execution_count":109,"outputs":[{"output_type":"stream","name":"stdout","text":"Text: I'm extremely disappointed with the quality of this item; it broke within a week.\n\nPrediction: Negative Sentiment ( Confidence in Prediction: 99.99% )\n\n\n\nText: The customer support for this product is terrible; they were unhelpful and rude.\n\nPrediction: Negative Sentiment ( Confidence in Prediction: 99.95% )\n\n\n\nText: I had high expectations for this item, but it turned out to be a total letdown.\n\nPrediction: Negative Sentiment ( Confidence in Prediction: 98.76% )\n\n\n\nText: The shipping was delayed, and the packaging was damaged when the product arrived.\n\nPrediction: Negative Sentiment ( Confidence in Prediction: 55.43% )\n\n\n\nText: I encountered multiple issues with this item, and the manufacturer provided no solutions.\n\nPrediction: Positive Sentiment ( Confidence in Prediction: 74.52% )\n\n\n\nText: The features advertised for this product are misleading; it doesn't perform as claimed.\n\nPrediction: Negative Sentiment ( Confidence in Prediction: 98.47% )\n\n\n\nText: I found this item to be overpriced for its quality; it doesn't justify the cost.\n\nPrediction: Negative Sentiment ( Confidence in Prediction: 98.53% )\n\n\n\nText: The user interface is confusing and difficult to navigate; it's not user-friendly at all.\n\nPrediction: Negative Sentiment ( Confidence in Prediction: 99.98% )\n\n\n\nText: The product exceeded my expectations and I am extremely satisfied with it.\n\nPrediction: Positive Sentiment ( Confidence in Prediction: 99.86% )\n\n\n\nText: This is the best product I've ever purchased, and I can't recommend it enough.\n\nPrediction: Positive Sentiment ( Confidence in Prediction: 99.85% )\n\n\n\nText: The quality of the product is outstanding, and it's worth every penny.\n\nPrediction: Positive Sentiment ( Confidence in Prediction: 99.86% )\n\n\n\nText: I've had this item for a week now, and I'm impressed with its performance.\n\nPrediction: Positive Sentiment ( Confidence in Prediction: 99.57% )\n\n\n\nText: I was skeptical at first, but this product turned out to be a pleasant surprise.\n\nPrediction: Positive Sentiment ( Confidence in Prediction: 87.31% )\n\n\n\nText: The product arrived in a damaged box, but fortunately, the item itself was fine.\n\nPrediction: Negative Sentiment ( Confidence in Prediction: 80.27% )\n\n\n\nText: I've tried many similar products, but this one stands out from the rest.\n\nPrediction: Negative Sentiment ( Confidence in Prediction: 84.53% )\n\n\n\nText: The features of this item are impressive, and it offers great value for money.\n\nPrediction: Positive Sentiment ( Confidence in Prediction: 99.87% )\n\n\n"}]},{"cell_type":"code","source":"","metadata":{"id":"VMws7p5R97aS"},"execution_count":109,"outputs":[]}]}