{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"**Thư viện**","metadata":{}},{"cell_type":"markdown","source":"# Phần 1","metadata":{}},{"cell_type":"code","source":"import numpy as np\nimport pandas as pd\nimport os\nimport random\nimport time\n\nimport re\nimport string\nimport nltk\nfrom nltk.corpus import stopwords\n\nimport matplotlib.pyplot as plt\nimport seaborn as sns\nsns.set(style=\"ticks\", context=\"talk\")\nplt.style.use('dark_background')\n\nfrom tqdm import tqdm\n\nimport torch\nimport torch.nn as nn\nimport torch.nn.functional as func\nfrom torch.utils.data import DataLoader, Dataset\n\nimport transformers\nfrom transformers import AdamW, get_linear_schedule_with_warmup\n\nimport tokenizers\nfrom sklearn.metrics import mean_squared_error, roc_auc_score, roc_curve, auc\n\nimport warnings\nwarnings.simplefilter('ignore')","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:28.294506Z","iopub.execute_input":"2022-01-06T15:53:28.294786Z","iopub.status.idle":"2022-01-06T15:53:28.308424Z","shell.execute_reply.started":"2022-01-06T15:53:28.294756Z","shell.execute_reply":"2022-01-06T15:53:28.306863Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Data","metadata":{}},{"cell_type":"code","source":"train = pd.read_csv('../input/quora-insincere-questions-classification/train.csv')\ntrain.head()","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:28.311331Z","iopub.execute_input":"2022-01-06T15:53:28.311540Z","iopub.status.idle":"2022-01-06T15:53:30.786526Z","shell.execute_reply.started":"2022-01-06T15:53:28.311515Z","shell.execute_reply":"2022-01-06T15:53:30.785774Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"temp = train[train['target'] == 1]\ntemp.head()","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:30.787788Z","iopub.execute_input":"2022-01-06T15:53:30.788203Z","iopub.status.idle":"2022-01-06T15:53:30.816326Z","shell.execute_reply.started":"2022-01-06T15:53:30.788154Z","shell.execute_reply":"2022-01-06T15:53:30.815520Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test = pd.read_csv('../input/quora-insincere-questions-classification/test.csv')\ntest.head()","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:30.818414Z","iopub.execute_input":"2022-01-06T15:53:30.818672Z","iopub.status.idle":"2022-01-06T15:53:32.144986Z","shell.execute_reply.started":"2022-01-06T15:53:30.818636Z","shell.execute_reply":"2022-01-06T15:53:32.144259Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"submission = pd.read_csv('../input/quora-insincere-questions-classification/sample_submission.csv')\nsubmission.head()","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:32.148820Z","iopub.execute_input":"2022-01-06T15:53:32.150698Z","iopub.status.idle":"2022-01-06T15:53:32.440162Z","shell.execute_reply.started":"2022-01-06T15:53:32.150658Z","shell.execute_reply":"2022-01-06T15:53:32.439501Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Tổng hợp và khảo sát data**","metadata":{}},{"cell_type":"markdown","source":"Check thông tin không xác định","metadata":{}},{"cell_type":"code","source":"train.isnull().sum()","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:32.441500Z","iopub.execute_input":"2022-01-06T15:53:32.441757Z","iopub.status.idle":"2022-01-06T15:53:32.702100Z","shell.execute_reply.started":"2022-01-06T15:53:32.441722Z","shell.execute_reply":"2022-01-06T15:53:32.701280Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test.isnull().sum()","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:32.703598Z","iopub.execute_input":"2022-01-06T15:53:32.703881Z","iopub.status.idle":"2022-01-06T15:53:32.783687Z","shell.execute_reply.started":"2022-01-06T15:53:32.703845Z","shell.execute_reply":"2022-01-06T15:53:32.782944Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train.shape, test.shape","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:32.785099Z","iopub.execute_input":"2022-01-06T15:53:32.785680Z","iopub.status.idle":"2022-01-06T15:53:32.791855Z","shell.execute_reply.started":"2022-01-06T15:53:32.785594Z","shell.execute_reply":"2022-01-06T15:53:32.791146Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Kiểm tra số lượng target = 1","metadata":{}},{"cell_type":"code","source":"df_train = train.drop(['qid', 'question_text'], axis = 1)\nlabel_counts = df_train.sum()\ndf_counts = pd.DataFrame(label_counts)\ndf_counts.rename(columns = {0:'counts'}, inplace = True)\ndf_counts = df_counts.sort_values('counts', ascending = False)\ndf_counts","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:32.795059Z","iopub.execute_input":"2022-01-06T15:53:32.795748Z","iopub.status.idle":"2022-01-06T15:53:32.814871Z","shell.execute_reply.started":"2022-01-06T15:53:32.795718Z","shell.execute_reply":"2022-01-06T15:53:32.814258Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"tỉ lệ trực quan","metadata":{}},{"cell_type":"code","source":"labels = np.round(df_train.sum()/len(df_train)*100, 1)\nlabels","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:32.817397Z","iopub.execute_input":"2022-01-06T15:53:32.817780Z","iopub.status.idle":"2022-01-06T15:53:32.827712Z","shell.execute_reply.started":"2022-01-06T15:53:32.817753Z","shell.execute_reply":"2022-01-06T15:53:32.826799Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Kết quả khảo sát:**\n*     - Không có dữ liệu không xác định (null).\n*     - Tập train có 1306122 giá trị, tập test có 375806 giá trị.\n*     - Dữ liệu target = 1 có 80810 giá trị, chiếm khoảng 6.2% tập train.","metadata":{}},{"cell_type":"markdown","source":"**Chuẩn hóa dữ liệu**","metadata":{}},{"cell_type":"markdown","source":"*Dọn dẹp dữ liệu:*\n* Xóa siêu liên kết, dấu chấm câu và số.\n* Mã hóa.\n* Bỏ qua việc thay đổi tất cả thành chữ thường và giữ cho các chữ cái phù hợp với từng trường hợp. \"BAD !!\" >> \"bad !!\".\n* Không xóa các từ dừng, Vì với các mô hình ngữ cảnh như BERT và ROBERTA, tốt hơn hết là (hầu như) không xử lý các văn bản xóa từ dừng. Các mô hình này được đào tạo trước với các từ dừng: nếu loại bỏ các từ dừng, mô hình có thể mất ngữ cảnh. Điều này cũng đúng với các kỹ thuật tiền xử lý gốc và lemmatization. Vì vậy, ta cũng bỏ qua chúng!\n","metadata":{}},{"cell_type":"code","source":"def clean_text(text):\n    '''Make text lowercase, remove text in square brackets,remove links,remove punctuation\n    and remove words containing numbers.'''\n    #text = text.lower()\n    \n    #pattern = [zero or more character]\n    text = re.sub('\\[.*?\\]', '', text)\n    \n    #pattern = with or without(http),://, one or more non-white space character, OR www, .,one or more non-white space character\n    text = re.sub('https?://\\S+|www\\.\\S+', '', text)\n    \n    #pattern = <, zero or more characters, >, (one or more occurance of >)\n    text = re.sub('<.*?>+', '', text)\n    \n    #pattern = any punctionation\n    text = re.sub('[%s]' % re.escape(string.punctuation), '', text)\n    \n    #pattern = any new line\n    text = re.sub('\\n', '', text)\n    \n    #pattern = any from[a-zA-Z0-9_], any from[0-9], any from [a-zA-Z0-9_]\n    text = re.sub('\\w*\\d\\w*', '', text)\n    return text","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:32.829388Z","iopub.execute_input":"2022-01-06T15:53:32.829641Z","iopub.status.idle":"2022-01-06T15:53:32.835832Z","shell.execute_reply.started":"2022-01-06T15:53:32.829609Z","shell.execute_reply":"2022-01-06T15:53:32.835040Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Để giảm thời gian train, ta chỉ sử dụng 5000 giá trị","metadata":{}},{"cell_type":"code","source":"%%time\ntrain = pd.read_csv('../input/quora-insincere-questions-classification/train.csv', nrows = 50000)\ntest = pd.read_csv('../input/quora-insincere-questions-classification/test.csv', nrows = 100)\ntrain['clean_text'] = train['question_text'].apply(str).apply(lambda x: clean_text(x))\ntest['clean_text'] = test['question_text'].apply(str).apply(lambda x: clean_text(x))","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:32.837373Z","iopub.execute_input":"2022-01-06T15:53:32.837695Z","iopub.status.idle":"2022-01-06T15:53:32.994521Z","shell.execute_reply.started":"2022-01-06T15:53:32.837662Z","shell.execute_reply":"2022-01-06T15:53:32.993805Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Sử dụng 20% dữ liệu training làm bộ validation.","metadata":{}},{"cell_type":"code","source":"kfold = 5\ntrain['kfold'] = train.index % kfold\ntrain.index % kfold","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:32.995732Z","iopub.execute_input":"2022-01-06T15:53:32.995983Z","iopub.status.idle":"2022-01-06T15:53:33.007486Z","shell.execute_reply.started":"2022-01-06T15:53:32.995946Z","shell.execute_reply":"2022-01-06T15:53:33.006811Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"p_train = train[train[\"kfold\"] != 0].reset_index(drop = True)\np_valid = train[train[\"kfold\"] == 0].reset_index(drop = True)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:33.009109Z","iopub.execute_input":"2022-01-06T15:53:33.009990Z","iopub.status.idle":"2022-01-06T15:53:33.019033Z","shell.execute_reply.started":"2022-01-06T15:53:33.009852Z","shell.execute_reply":"2022-01-06T15:53:33.018199Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"p_train.head()","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:33.020372Z","iopub.execute_input":"2022-01-06T15:53:33.020739Z","iopub.status.idle":"2022-01-06T15:53:33.038118Z","shell.execute_reply.started":"2022-01-06T15:53:33.020701Z","shell.execute_reply":"2022-01-06T15:53:33.037332Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"tokenizer = transformers.BertTokenizer.from_pretrained('bert-base-cased')","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:33.039569Z","iopub.execute_input":"2022-01-06T15:53:33.039838Z","iopub.status.idle":"2022-01-06T15:53:35.453320Z","shell.execute_reply.started":"2022-01-06T15:53:33.039803Z","shell.execute_reply":"2022-01-06T15:53:35.452587Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"%%time\nsenten_len = []\nfor sentence in tqdm(p_train['clean_text']):\n    token_words = tokenizer.encode_plus(sentence)['input_ids']\n    senten_len.append(len(token_words))","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:35.454556Z","iopub.execute_input":"2022-01-06T15:53:35.454814Z","iopub.status.idle":"2022-01-06T15:53:37.085884Z","shell.execute_reply.started":"2022-01-06T15:53:35.454780Z","shell.execute_reply":"2022-01-06T15:53:37.085237Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Xem xét sự phân bố độ dài cho các question_text được mã hóa để lựa chọn max_length cho các dữ liệu được mã hóa của chúng ta.","metadata":{}},{"cell_type":"code","source":"sns.distplot(senten_len)\nplt.xlim([0, 50])\nplt.xlabel('Tocken count')","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:37.086988Z","iopub.execute_input":"2022-01-06T15:53:37.087652Z","iopub.status.idle":"2022-01-06T15:53:37.378628Z","shell.execute_reply.started":"2022-01-06T15:53:37.087612Z","shell.execute_reply":"2022-01-06T15:53:37.377936Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"max_len = 30","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:37.379798Z","iopub.execute_input":"2022-01-06T15:53:37.380110Z","iopub.status.idle":"2022-01-06T15:53:37.385158Z","shell.execute_reply.started":"2022-01-06T15:53:37.380070Z","shell.execute_reply":"2022-01-06T15:53:37.384278Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Định nghĩa lớp BertDataSet vói Dataset là siêu lớp và ghi đè hàm init, len và getitem. Nó sẽ lấy danh sách question và toxic label rồi tạo token ids và attention mask để phân biệt các câu hỏi với zero padding.\n\ntorch.tensor vs np.ndarray: Nếu chỉ quan tâm đến cách hiệu quả và dễ dàng để thực hiện các phép toán trên ma trận, np.ndarray hoặc torch.tensor có thể được sử dụng thay thế cho nhau.\n\nTuy nhiên, torch.tensors được thiết kế để sử dụng trong bối cảnh tối ưu hóa độ dốc gradient, và do đó chúng không chỉ giữ một tensor với các giá trị số, mà (quan trọng hơn) biểu đồ tính toán dẫn đến các giá trị này. Sau đó, đồ thị tính toán này được sử dụng (sử dụng quy tắc chuỗi của đạo hàm) để tính đạo hàm của hàm tổn thất w.r.t từng biến độc lập được sử dụng để tính toán tổn thất.\n\nĐối tượng np.ndarray không có thêm lớp \"đồ thị tính toán\" này và do đó, khi chuyển đổi torch.tensor thành np.ndarray, ta phải xóa rõ ràng đồ thị tính toán của tensor bằng lệnh detach ().","metadata":{}},{"cell_type":"code","source":"class BertDataSet(Dataset):\n    \n    def __init__(self, sentences, toxic_labels):\n        self.sentences = sentences\n        self.targets = toxic_labels.to_numpy()\n    \n    def __len__(self):\n        return len(self.sentences)\n    \n    \n    def __getitem__(self, idx):\n        sentence = self.sentences[idx]\n        bert_senten = tokenizer.encode_plus(sentence, \n                                            add_special_tokens = True, # [CLS],[SEP]\n                                            max_length = max_len,\n                                            pad_to_max_length = True,\n                                            truncation = True,\n                                            return_attention_mask = True\n                                             )\n        ids = torch.tensor(bert_senten['input_ids'], dtype = torch.long)\n        mask = torch.tensor(bert_senten['attention_mask'], dtype = torch.long)\n        toxic_label = torch.tensor(self.targets[idx], dtype = torch.float)\n        \n        \n        return {\n            'ids' : ids,\n            'mask' : mask,\n            'toxic_label':toxic_label\n        }","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:37.386354Z","iopub.execute_input":"2022-01-06T15:53:37.387132Z","iopub.status.idle":"2022-01-06T15:53:37.395937Z","shell.execute_reply.started":"2022-01-06T15:53:37.387095Z","shell.execute_reply":"2022-01-06T15:53:37.395260Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_dataset = BertDataSet(p_train['clean_text'], p_train['target'])\nvalid_dataset = BertDataSet(p_valid['clean_text'], p_valid['target'])","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:37.397023Z","iopub.execute_input":"2022-01-06T15:53:37.397413Z","iopub.status.idle":"2022-01-06T15:53:37.407688Z","shell.execute_reply.started":"2022-01-06T15:53:37.397377Z","shell.execute_reply":"2022-01-06T15:53:37.406959Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Đặt train_batch và valid_batch là 32. Kích thước batch lớn hơn giúp tăng tốc độ tính toán. Tuy nhiên, kết quả xác nhận rằng việc sử dụng kích thước batch nhỏ đạt được hiệu suất tổng quát hóa tốt nhất, với một chi phí tính toán nhất định. Trong mọi trường hợp, kết quả tốt nhất đã thu được với kích thước batch là 32 hoặc nhỏ hơn. Thường kích thước bacth nhỏ như 2 hoặc 4 mang lại kết quả tối ưu.","metadata":{}},{"cell_type":"code","source":"train_batch = 32\nvalid_batch = 32","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:37.408875Z","iopub.execute_input":"2022-01-06T15:53:37.409559Z","iopub.status.idle":"2022-01-06T15:53:37.415694Z","shell.execute_reply.started":"2022-01-06T15:53:37.409516Z","shell.execute_reply":"2022-01-06T15:53:37.414940Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"pin_memory = True trong DataLoader sẽ tự động đặt Tensors dữ liệu đã tìm nạp vào bộ nhớ được ghim, và do đó cho phép truyền dữ liệu nhanh hơn đến các GPU hỗ trợ CUDA. Điều này được giải thích rõ nhất trong bài đăng trên blog của NVIDIA.\n![download.png](attachment:1ebb3def-cfa9-4f82-9745-605372320c87.png)","metadata":{},"attachments":{"1ebb3def-cfa9-4f82-9745-605372320c87.png":{"image/png":"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"}}},{"cell_type":"code","source":"train_dataloader = DataLoader(train_dataset, batch_size = train_batch, pin_memory = True, num_workers = 4, shuffle = True)\nvalid_dataloader = DataLoader(valid_dataset, batch_size = valid_batch, pin_memory = True, num_workers = 4, shuffle = False)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:37.416795Z","iopub.execute_input":"2022-01-06T15:53:37.417243Z","iopub.status.idle":"2022-01-06T15:53:37.424746Z","shell.execute_reply.started":"2022-01-06T15:53:37.417090Z","shell.execute_reply":"2022-01-06T15:53:37.423984Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Mô hình BERT**","metadata":{}},{"cell_type":"code","source":"device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\ndevice","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:37.429753Z","iopub.execute_input":"2022-01-06T15:53:37.430423Z","iopub.status.idle":"2022-01-06T15:53:37.436293Z","shell.execute_reply.started":"2022-01-06T15:53:37.430291Z","shell.execute_reply":"2022-01-06T15:53:37.435591Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"%%time\nmodel = transformers.BertForSequenceClassification.from_pretrained('bert-base-cased', num_labels = 1)\nmodel.to(device)\nmodel.train()","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:37.437610Z","iopub.execute_input":"2022-01-06T15:53:37.438676Z","iopub.status.idle":"2022-01-06T15:53:39.862828Z","shell.execute_reply.started":"2022-01-06T15:53:37.438641Z","shell.execute_reply":"2022-01-06T15:53:39.862088Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Mô hình sử dụng id và mask từ token encoding trong tập train.","metadata":{}},{"cell_type":"code","source":"%%time\nfor a in train_dataloader:\n    ids = a['ids'].to(device)\n    mask = a['mask'].to(device)\n    output = model(ids, mask)\n    break","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:39.864281Z","iopub.execute_input":"2022-01-06T15:53:39.864537Z","iopub.status.idle":"2022-01-06T15:53:41.300303Z","shell.execute_reply.started":"2022-01-06T15:53:39.864503Z","shell.execute_reply":"2022-01-06T15:53:41.299490Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"output","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:41.301757Z","iopub.execute_input":"2022-01-06T15:53:41.302009Z","iopub.status.idle":"2022-01-06T15:53:41.338610Z","shell.execute_reply.started":"2022-01-06T15:53:41.301971Z","shell.execute_reply":"2022-01-06T15:53:41.337894Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"func.softmax(output['logits'], dim = 1)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:41.339800Z","iopub.execute_input":"2022-01-06T15:53:41.340117Z","iopub.status.idle":"2022-01-06T15:53:41.352988Z","shell.execute_reply.started":"2022-01-06T15:53:41.340078Z","shell.execute_reply":"2022-01-06T15:53:41.352363Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"output_probs = func.softmax(output['logits'], dim = 1)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:41.355112Z","iopub.execute_input":"2022-01-06T15:53:41.355391Z","iopub.status.idle":"2022-01-06T15:53:41.359114Z","shell.execute_reply.started":"2022-01-06T15:53:41.355356Z","shell.execute_reply":"2022-01-06T15:53:41.358352Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"torch.max(output_probs, dim = 1)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:41.360820Z","iopub.execute_input":"2022-01-06T15:53:41.361588Z","iopub.status.idle":"2022-01-06T15:53:41.378812Z","shell.execute_reply.started":"2022-01-06T15:53:41.361551Z","shell.execute_reply":"2022-01-06T15:53:41.378177Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Chuẩn bị mô hình training và validating**","metadata":{}},{"cell_type":"markdown","source":"Cài đặt tham số","metadata":{}},{"cell_type":"code","source":"epochs = 5\nLR = 2e-5 #Learning rate\noptimizer = AdamW(model.parameters(), LR, betas = (0.9, 0.999), weight_decay = 1e-2, correct_bias = False)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:41.379786Z","iopub.execute_input":"2022-01-06T15:53:41.382348Z","iopub.status.idle":"2022-01-06T15:53:41.391011Z","shell.execute_reply.started":"2022-01-06T15:53:41.382309Z","shell.execute_reply":"2022-01-06T15:53:41.390269Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Ta sử dụng tham số learning rate như đã thiết lập ở trên cho 10% tổng thời gian training. Sau đó, giảm dần learning rate về không.","metadata":{}},{"cell_type":"code","source":"train_steps = int((len(train) * epochs)/train_batch)\nnum_steps = int(train_steps * 0.1)\nscheduler = get_linear_schedule_with_warmup(optimizer, num_steps, train_steps)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:41.392376Z","iopub.execute_input":"2022-01-06T15:53:41.392889Z","iopub.status.idle":"2022-01-06T15:53:41.397857Z","shell.execute_reply.started":"2022-01-06T15:53:41.392854Z","shell.execute_reply":"2022-01-06T15:53:41.397047Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"le = []\nfor b in tqdm(range(epochs)):\n    for a in train_dataloader:\n        le.append(scheduler.get_last_lr())\n        scheduler.step()\nplt.plot(np.arange(len(le)), le)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:41.399081Z","iopub.execute_input":"2022-01-06T15:53:41.399418Z","iopub.status.idle":"2022-01-06T15:53:52.871823Z","shell.execute_reply.started":"2022-01-06T15:53:41.399385Z","shell.execute_reply":"2022-01-06T15:53:52.871081Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"loss_fn = nn.BCEWithLogitsLoss()\nloss_fn.to(device)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:52.873308Z","iopub.execute_input":"2022-01-06T15:53:52.873797Z","iopub.status.idle":"2022-01-06T15:53:52.880676Z","shell.execute_reply.started":"2022-01-06T15:53:52.873742Z","shell.execute_reply":"2022-01-06T15:53:52.879640Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"scaler = torch.cuda.amp.GradScaler()","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:52.882280Z","iopub.execute_input":"2022-01-06T15:53:52.882769Z","iopub.status.idle":"2022-01-06T15:53:52.889824Z","shell.execute_reply.started":"2022-01-06T15:53:52.882729Z","shell.execute_reply":"2022-01-06T15:53:52.888853Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Hàm training**","metadata":{}},{"cell_type":"code","source":"def training(train_dataloader, model, optimizer, scheduler):\n    model.train()\n    torch.backends.cudnn.benchmark = True\n    correct_predictions = 0\n    \n    for a in train_dataloader:\n        losses = []\n        optimizer.zero_grad()\n        \n        #allpreds = []\n        #alltargets = []\n        \n        with torch.cuda.amp.autocast():\n            \n            ids = a['ids'].to(device, non_blocking = True)\n            mask = a['mask'].to(device, non_blocking = True) \n\n            output = model(ids, mask) #This gives model as output, however we want the values at the output\n            output = output['logits'].squeeze(-1).to(torch.float32)\n\n            output_probs = torch.sigmoid(output)\n            preds = torch.where(output_probs > 0.5, 1, 0)\n            \n            toxic_label = a['toxic_label'].to(device, non_blocking = True) \n            loss = loss_fn(output, toxic_label)            \n            \n            losses.append(loss.item())\n            #allpreds.append(output.detach().cpu().numpy())\n            #alltargets.append(toxic.detach().squeeze(-1).cpu().numpy())\n            correct_predictions += torch.sum(preds == toxic_label)\n        \n        scaler.scale(loss).backward() #Multiplies (‘scales’) a tensor or list of tensors by the scale factor.\n                                      #Returns scaled outputs. If this instance of GradScaler is not enabled, outputs are returned unmodified.\n        scaler.step(optimizer) #Returns the return value of optimizer.step(*args, **kwargs).\n        scaler.update() #Updates the scale factor.If any optimizer steps were skipped the scale is multiplied by backoff_factor to reduce it. \n                        #If growth_interval unskipped iterations occurred consecutively, the scale is multiplied by growth_factor to increase it\n        scheduler.step() # Update learning rate schedule\n    \n    losses = np.mean(losses)\n    corr_preds = correct_predictions.detach().cpu().numpy()\n    accuracy = corr_preds/len(p_train)\n    \n    return losses, accuracy","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:52.891408Z","iopub.execute_input":"2022-01-06T15:53:52.891696Z","iopub.status.idle":"2022-01-06T15:53:52.904232Z","shell.execute_reply.started":"2022-01-06T15:53:52.891659Z","shell.execute_reply":"2022-01-06T15:53:52.903415Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"amp: automatic mixed precision.\n\ncâu lệnh with được sử dụng trong xử lý ngoại lệ để làm cho mã sạch hơn và dễ đọc hơn nhiều. Nó đơn giản hóa việc quản lý các tài nguyên chung như các luồng tệp.\n\nOptimizer.step(): thực hiện cập nhật tham số dựa trên gradient hiện tại (được lưu trữ trong thuộc tính .grad của tham số) và quy tắc cập nhật.\n\nLoss.backward() gọi .backward() nhiều lần tích lũy gradient (bằng cách cộng) cho mỗi tham số. Đây là lý do tại sao ta nên gọi Optimizer.zero_grad() sau mỗi lần gọi .step().","metadata":{}},{"cell_type":"markdown","source":"**Hàm validating**","metadata":{}},{"cell_type":"markdown","source":"Hàm validating khá giống với hàm training. Sự khác biệt là không có sự lan truyền ngược và tối ưu hóa cho các tham số trong đó.","metadata":{}},{"cell_type":"code","source":"def validating(valid_dataloader, model):\n    \n    model.eval()\n    correct_predictions = 0\n    all_output_probs = []\n    \n    for a in valid_dataloader:\n        losses = []\n        ids = a['ids'].to(device, non_blocking = True)\n        mask = a['mask'].to(device, non_blocking = True)\n        output = model(ids, mask)\n        output = output['logits'].squeeze(-1).to(torch.float32)\n        output_probs = torch.sigmoid(output)\n        preds = torch.where(output_probs > 0.5, 1, 0)\n            \n        toxic_label = a['toxic_label'].to(device, non_blocking = True)\n        loss = loss_fn(output, toxic_label)\n        losses.append(loss.item())\n        all_output_probs.extend(output_probs.detach().cpu().numpy())\n        \n        correct_predictions += torch.sum(preds == toxic_label)\n        corr_preds = correct_predictions.detach().cpu().numpy()\n    \n    losses = np.mean(losses)\n    corr_preds = correct_predictions.detach().cpu().numpy()\n    accuracy = corr_preds/len(p_valid)\n    \n    return losses, accuracy, all_output_probs","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:52.905240Z","iopub.execute_input":"2022-01-06T15:53:52.905478Z","iopub.status.idle":"2022-01-06T15:53:52.917965Z","shell.execute_reply.started":"2022-01-06T15:53:52.905449Z","shell.execute_reply":"2022-01-06T15:53:52.917156Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Train mô hình**","metadata":{}},{"cell_type":"code","source":"%%time\n\nbest_score = 1000\ntrain_accs = []\nvalid_accs = []\ntrain_losses = []\nvalid_losses = []\n\nfor eboch in tqdm(range(epochs)):\n    \n    train_loss, train_acc = training(train_dataloader, model, optimizer, scheduler)\n    valid_loss, valid_acc, valid_probs = validating(valid_dataloader, model)\n    \n    print('train losses: %.4f' % train_loss, 'train accuracy: %.3f' % train_acc)\n    print('valid losses: %.4f' % valid_loss, 'valid accuracy: %.3f' % valid_acc)\n    train_losses.append(train_loss)\n    valid_losses.append(valid_loss)\n    train_accs.append(train_acc)\n    valid_accs.append(valid_acc)\n    \n    \n    if valid_loss < best_score:\n        best_score = valid_loss\n        print('Found a good model!')\n        state = {\n            'state_dict': model.state_dict(),\n            'optimizer_dict': optimizer.state_dict(),\n            'best_score': best_score\n        }\n        torch.save(state, 'best_model.pth')\n    else:\n        pass","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:53:52.919161Z","iopub.execute_input":"2022-01-06T15:53:52.919509Z","iopub.status.idle":"2022-01-06T15:55:25.757774Z","shell.execute_reply.started":"2022-01-06T15:53:52.919469Z","shell.execute_reply":"2022-01-06T15:55:25.757027Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"x = np.arange(epochs)\nfig, ax = plt.subplots(1, 2, figsize = (15,4))\nax[0].plot(x, train_losses)\nax[0].plot(x, valid_losses)\nax[0].set_ylabel('Losses', weight = 'bold')\nax[0].set_xlabel('Epochs')\nax[0].grid(alpha = 0.3)\nax[0].legend(labels = ['train losses', 'valid losses'])\n\nax[1].plot(x, train_accs)\nax[1].plot(x, valid_accs)\nax[1].set_ylabel('Accuracy', weight = 'bold')\nax[1].set_xlabel('Epochs')\nax[1].legend(labels = ['train acc', 'valid acc'])\n\nax[1].grid(alpha = 0.3)\nfig.suptitle('Fold = 0', weight = 'bold') ","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:55:25.759201Z","iopub.execute_input":"2022-01-06T15:55:25.760054Z","iopub.status.idle":"2022-01-06T15:55:26.133737Z","shell.execute_reply.started":"2022-01-06T15:55:25.760010Z","shell.execute_reply":"2022-01-06T15:55:26.133094Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Đánh giá mô hình","metadata":{}},{"cell_type":"markdown","source":"Ta sử dụng auc để đánh giá","metadata":{}},{"cell_type":"code","source":"valid_loss, valid_acc, valid_probs = validating(valid_dataloader, model)\nvalid_probs = np.asarray(valid_probs).flatten()\ny_valid = p_valid['target'].to_numpy().flatten()\nfpr, tpr, _ = roc_curve(y_valid, valid_probs)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:55:26.134986Z","iopub.execute_input":"2022-01-06T15:55:26.135759Z","iopub.status.idle":"2022-01-06T15:55:27.607826Z","shell.execute_reply.started":"2022-01-06T15:55:26.135719Z","shell.execute_reply":"2022-01-06T15:55:27.606953Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"fig, ax = plt.subplots()\nax.plot(fpr, tpr)\nax.set_title('ROC Curv')\nax.set_xlabel('FPR')\nax.set_ylabel('TPR')\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:55:27.609423Z","iopub.execute_input":"2022-01-06T15:55:27.609682Z","iopub.status.idle":"2022-01-06T15:55:27.787755Z","shell.execute_reply.started":"2022-01-06T15:55:27.609634Z","shell.execute_reply":"2022-01-06T15:55:27.787081Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"auc(fpr, tpr)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:55:27.788975Z","iopub.execute_input":"2022-01-06T15:55:27.789703Z","iopub.status.idle":"2022-01-06T15:55:27.796621Z","shell.execute_reply.started":"2022-01-06T15:55:27.789664Z","shell.execute_reply":"2022-01-06T15:55:27.795775Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Phần 2","metadata":{}},{"cell_type":"code","source":"%%time\nimport numpy as np\nimport pandas as pd\nimport os\nimport random\nimport time\n\nimport re\nimport string\nimport nltk\nfrom nltk.corpus import stopwords\n\nimport matplotlib.pyplot as plt\nimport seaborn as sns\nsns.set(style=\"ticks\", context=\"talk\")\nplt.style.use('dark_background')\n\nfrom tqdm import tqdm\n\nimport torch\nimport torch.nn as nn\nimport torch.nn.functional as func\nfrom torch.utils.data import DataLoader, Dataset\n\nimport transformers\nfrom transformers import AdamW, get_linear_schedule_with_warmup\n\nimport tokenizers\nfrom sklearn.metrics import mean_squared_error, roc_auc_score, roc_curve, auc\n\nimport warnings\nwarnings.simplefilter('ignore')\n\ntrain = pd.read_csv('../input/quora-insincere-questions-classification/train.csv', nrows = 50000)\ntest = pd.read_csv('../input/quora-insincere-questions-classification/test.csv', nrows = 100)\nsubmission = pd.read_csv('../input/quora-insincere-questions-classification/sample_submission.csv')\n\n\ndef clean_text(text):\n\n    text = re.sub('\\[.*?\\]', '', text)\n    text = re.sub('https?://\\S+|www\\.\\S+', '', text)\n    text = re.sub('<.*?>+', '', text)\n    text = re.sub('[%s]' % re.escape(string.punctuation), '', text)\n    text = re.sub('\\n', '', text)\n    text = re.sub('\\w*\\d\\w*', '', text)\n    return text\n\n\ntrain['clean_text'] = train['question_text'].apply(str).apply(lambda x: clean_text(x))\ntest['clean_text'] = test['question_text'].apply(str).apply(lambda x: clean_text(x))\n\nkfold = 5\ntrain['kfold'] = train.index % kfold\n\ntokenizer = transformers.BertTokenizer.from_pretrained('bert-base-cased')\nmax_len = 30\n\nclass BertDataSet(Dataset):\n    \n    def __init__(self, sentences, toxic_labels):\n        self.sentences = sentences\n        self.targets = toxic_labels.to_numpy()\n    \n    def __len__(self):\n        return len(self.sentences)\n    \n    \n    def __getitem__(self, idx):\n        sentence = self.sentences[idx]\n        bert_senten = tokenizer.encode_plus(sentence, \n                                            add_special_tokens = True, # [CLS],[SEP]\n                                            max_length = max_len,\n                                            pad_to_max_length = True,\n                                            truncation = True,\n                                            return_attention_mask = True\n                                             )\n        ids = torch.tensor(bert_senten['input_ids'], dtype = torch.long)\n        mask = torch.tensor(bert_senten['attention_mask'], dtype = torch.long)\n        toxic_label = torch.tensor(self.targets[idx], dtype = torch.float)\n        \n        \n        return {\n            'ids' : ids,\n            'mask' : mask,\n            'toxic_label':toxic_label\n        }\n\nepochs = 5\ntrain_batch = 32\nvalid_batch = 32\ndevice = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n\nloss_fn = nn.BCEWithLogitsLoss()\nloss_fn.to(device)\nscaler = torch.cuda.amp.GradScaler()\n\ndef training(train_dataloader, model, optimizer, scheduler):\n    model.train()\n    torch.backends.cudnn.benchmark = True\n    correct_predictions = 0\n    \n    for a in train_dataloader:\n        losses = []\n        optimizer.zero_grad()\n        \n        with torch.cuda.amp.autocast():\n            \n            ids = a['ids'].to(device, non_blocking = True)\n            mask = a['mask'].to(device, non_blocking = True) \n\n            output = model(ids, mask) #This gives model as output, however we want the values at the output\n            output = output['logits'].squeeze(-1).to(torch.float32)\n\n            output_probs = torch.sigmoid(output)\n            preds = torch.where(output_probs > 0.5, 1, 0)\n            \n            toxic_label = a['toxic_label'].to(device, non_blocking = True) \n            loss = loss_fn(output, toxic_label)            \n            \n            losses.append(loss.item())\n            correct_predictions += torch.sum(preds == toxic_label)\n        \n        scaler.scale(loss).backward()\n        scaler.step(optimizer)\n        scaler.update()\n        scheduler.step()\n    \n    losses = np.mean(losses)\n    corr_preds = correct_predictions.detach().cpu().numpy()\n    accuracy = corr_preds/len(p_train)\n    \n    return losses, accuracy\n\ndef validating(valid_dataloader, model):\n    \n    model.eval()\n    correct_predictions = 0\n    all_output_probs = []\n    \n    for a in valid_dataloader:\n        losses = []\n        ids = a['ids'].to(device, non_blocking = True)\n        mask = a['mask'].to(device, non_blocking = True)\n        output = model(ids, mask)\n        output = output['logits'].squeeze(-1).to(torch.float32)\n        output_probs = torch.sigmoid(output)\n        preds = torch.where(output_probs > 0.5, 1, 0)\n            \n        toxic_label = a['toxic_label'].to(device, non_blocking = True)\n        loss = loss_fn(output, toxic_label)\n        losses.append(loss.item())\n        all_output_probs.extend(output_probs.detach().cpu().numpy())\n        \n        correct_predictions += torch.sum(preds == toxic_label)\n        corr_preds = correct_predictions.detach().cpu().numpy()\n    \n    losses = np.mean(losses)\n    corr_preds = correct_predictions.detach().cpu().numpy()\n    accuracy = corr_preds/len(p_valid)\n    \n    return losses, accuracy, all_output_probs","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:55:27.798167Z","iopub.execute_input":"2022-01-06T15:55:27.798698Z","iopub.status.idle":"2022-01-06T15:55:30.650394Z","shell.execute_reply.started":"2022-01-06T15:55:27.798629Z","shell.execute_reply":"2022-01-06T15:55:30.649590Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Để cải thiện mô hình, ta lặp lại quy trình training tương tự cho mỗi lần gấp k-folds","metadata":{}},{"cell_type":"markdown","source":"**Luyện tập lặp lại cho K-fold**","metadata":{}},{"cell_type":"code","source":"%%time\n\nbest_scores = []\nfor fold in tqdm(range(0,5)):\n\n    # initializing the data\n    p_train = train[train['kfold'] != fold].reset_index(drop = True)\n    p_valid = train[train['kfold'] == fold].reset_index(drop = True)\n\n    train_dataset = BertDataSet(p_train['clean_text'], p_train['target'])\n    valid_dataset = BertDataSet(p_valid['clean_text'], p_valid['target'])\n\n    train_dataloader = DataLoader(train_dataset, batch_size = train_batch, shuffle = True, num_workers = 4, pin_memory = True)\n    valid_dataloader = DataLoader(valid_dataset, batch_size = valid_batch, shuffle = False, num_workers = 4, pin_memory = True)\n\n    model = transformers.BertForSequenceClassification.from_pretrained(\"bert-base-cased\", num_labels = 1)\n    model.to(device)\n    \n    LR = 2e-5\n    optimizer = AdamW(model.parameters(), LR,betas = (0.9, 0.999), weight_decay = 1e-2) # AdamW optimizer\n\n    train_steps = int(len(p_train)/train_batch * epochs)\n    num_steps = int(train_steps * 0.1)\n\n    scheduler = get_linear_schedule_with_warmup(optimizer, num_steps, train_steps)\n    \n    best_score = 1000\n    train_accs = []\n    valid_accs = []\n    train_losses = []\n    valid_losses = []\n    best_valid_probs = []\n    \n    print(\"-------------- Fold = \" + str(fold) + \"-------------\")\n    \n    for epoch in tqdm(range(epochs)):\n        print(\"Epoch = \" + str(epoch))\n\n        train_loss, train_acc = training(train_dataloader, model, optimizer, scheduler)\n        valid_loss, valid_acc, valid_probs = validating(valid_dataloader, model)\n\n        train_losses.append(train_loss)\n        train_accs.append(train_acc)\n        valid_losses.append(valid_loss)\n        valid_accs.append(valid_acc)\n        \n        print('train losses: %.4f' %(train_loss), 'train accuracy: %.3f' %(train_acc))\n        print('valid losses: %.4f' %(valid_loss), 'valid accuracy: %.3f' %(valid_acc))\n\n        if (valid_loss < best_score):\n\n            best_score = valid_loss\n            print(\"Found an improved model! :)\")\n\n            state = {'state_dict': model.state_dict(),\n                     'optimizer_dict': optimizer.state_dict(),\n                     'best_score':best_score\n                    }\n\n            torch.save(state, \"model\" + str(fold) + \".pth\")\n            best_valid_prob = valid_probs\n            torch.cuda.memory_summary(device = None, abbreviated = False)\n        else:\n            pass\n\n\n    best_scores.append(best_score)\n    best_valid_probs.append(best_valid_prob)\n    \n    ##Plotting the result for each fold\n    x = np.arange(epochs)\n    fig, ax = plt.subplots(1, 2, figsize = (15,4))\n    ax[0].plot(x, train_losses)\n    ax[0].plot(x, valid_losses)\n    ax[0].set_ylabel('Losses', weight = 'bold')\n    ax[0].set_xlabel('Epochs')\n    ax[0].grid(alpha = 0.3)\n    ax[0].legend(labels = ['train losses', 'valid losses'])\n\n    ax[1].plot(x, train_accs)\n    ax[1].plot(x, valid_accs)\n    ax[1].set_ylabel('Accuracy', weight = 'bold')\n    ax[1].set_xlabel('Epochs')\n    ax[1].legend(labels = ['train acc', 'valid acc'])\n\n    ax[1].grid(alpha = 0.3)\n    fig.suptitle('Fold = '+str(fold), weight = 'bold') ","metadata":{"execution":{"iopub.status.busy":"2022-01-06T15:55:30.651964Z","iopub.execute_input":"2022-01-06T15:55:30.652461Z","iopub.status.idle":"2022-01-06T16:03:55.759597Z","shell.execute_reply.started":"2022-01-06T15:55:30.652419Z","shell.execute_reply":"2022-01-06T16:03:55.758828Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"best_scores","metadata":{"execution":{"iopub.status.busy":"2022-01-06T16:03:55.761062Z","iopub.execute_input":"2022-01-06T16:03:55.761500Z","iopub.status.idle":"2022-01-06T16:03:55.769019Z","shell.execute_reply.started":"2022-01-06T16:03:55.761449Z","shell.execute_reply":"2022-01-06T16:03:55.768186Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print('Mean of',kfold, 'folds for best loss in', epochs, 'epochs cross-validation folds is %.4f.' %(np.mean(best_scores)))","metadata":{"execution":{"iopub.status.busy":"2022-01-06T16:03:55.770685Z","iopub.execute_input":"2022-01-06T16:03:55.770964Z","iopub.status.idle":"2022-01-06T16:03:55.780685Z","shell.execute_reply.started":"2022-01-06T16:03:55.770925Z","shell.execute_reply":"2022-01-06T16:03:55.779928Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Đánh giá cho k mô hình**","metadata":{}},{"cell_type":"markdown","source":"Ta sử dụng tập hợp để đánh giá k mô hình trong validation set","metadata":{}},{"cell_type":"code","source":"def predicting(test_dataloader, model, pthes):\n    allpreds = []\n    \n    for pth in pthes:\n        state = torch.load(pth)\n        model.load_state_dict(state['state_dict'])\n        model.to(device)\n        model.eval()\n        preds = []\n        with torch.no_grad():\n            for a in test_dataloader:\n                ids = a['ids'].to(device)\n                mask = a['mask'].to(device)\n                output = model(ids, mask)\n                output = output['logits'].squeeze(-1)\n                output_probs = torch.sigmoid(output)\n                preds.append(output_probs.cpu().numpy())\n            preds = np.concatenate(preds)\n            allpreds.append(preds)\n      \n    return allpreds","metadata":{"execution":{"iopub.status.busy":"2022-01-06T16:03:55.782040Z","iopub.execute_input":"2022-01-06T16:03:55.782344Z","iopub.status.idle":"2022-01-06T16:03:55.790452Z","shell.execute_reply.started":"2022-01-06T16:03:55.782286Z","shell.execute_reply":"2022-01-06T16:03:55.789448Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"pthes = [os.path.join(\"./\",s) for s in os.listdir(\"./\") if \".pth\" in s]\npthes","metadata":{"execution":{"iopub.status.busy":"2022-01-06T16:03:55.791817Z","iopub.execute_input":"2022-01-06T16:03:55.792029Z","iopub.status.idle":"2022-01-06T16:03:55.803184Z","shell.execute_reply.started":"2022-01-06T16:03:55.792004Z","shell.execute_reply":"2022-01-06T16:03:55.802302Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"allpreds = predicting(valid_dataloader, model, pthes)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T16:03:55.804488Z","iopub.execute_input":"2022-01-06T16:03:55.805120Z","iopub.status.idle":"2022-01-06T16:05:21.914187Z","shell.execute_reply.started":"2022-01-06T16:03:55.805070Z","shell.execute_reply":"2022-01-06T16:05:21.913242Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Lấy trung bình của xác suất từ k-mô hình cho mỗi mẫu trong tập validation.","metadata":{}},{"cell_type":"code","source":"valid_probs = np.zeros(len(p_valid))\nfor i in range(kfold):\n    valid_probs += allpreds[i]\nvalid_probs = valid_probs / kfold","metadata":{"execution":{"iopub.status.busy":"2022-01-06T16:05:21.919762Z","iopub.execute_input":"2022-01-06T16:05:21.926400Z","iopub.status.idle":"2022-01-06T16:05:21.932192Z","shell.execute_reply.started":"2022-01-06T16:05:21.926348Z","shell.execute_reply":"2022-01-06T16:05:21.931473Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"valid_probs = np.asarray(valid_probs).flatten()","metadata":{"execution":{"iopub.status.busy":"2022-01-06T16:05:21.936390Z","iopub.execute_input":"2022-01-06T16:05:21.937966Z","iopub.status.idle":"2022-01-06T16:05:21.947736Z","shell.execute_reply.started":"2022-01-06T16:05:21.937925Z","shell.execute_reply":"2022-01-06T16:05:21.946673Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"y_valid = p_valid['target'].to_numpy().flatten()","metadata":{"execution":{"iopub.status.busy":"2022-01-06T16:05:21.948976Z","iopub.execute_input":"2022-01-06T16:05:21.949257Z","iopub.status.idle":"2022-01-06T16:05:21.962617Z","shell.execute_reply.started":"2022-01-06T16:05:21.949199Z","shell.execute_reply":"2022-01-06T16:05:21.961683Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"fpr, tpr, _ = roc_curve(y_valid, valid_probs)\nprint('auc score for kfold =', kfold, 'models is: %.2f' %(auc(fpr, tpr)*100))","metadata":{"execution":{"iopub.status.busy":"2022-01-06T16:05:21.964047Z","iopub.execute_input":"2022-01-06T16:05:21.964380Z","iopub.status.idle":"2022-01-06T16:05:21.980005Z","shell.execute_reply.started":"2022-01-06T16:05:21.964343Z","shell.execute_reply":"2022-01-06T16:05:21.978619Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"fig, ax = plt.subplots()\nax.plot(fpr, tpr)\nax.set_title('ROC Curv')\nax.set_xlabel('FPR')\nax.set_ylabel('TPR')\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2022-01-06T16:05:21.985076Z","iopub.execute_input":"2022-01-06T16:05:21.985522Z","iopub.status.idle":"2022-01-06T16:05:22.217581Z","shell.execute_reply.started":"2022-01-06T16:05:21.985488Z","shell.execute_reply":"2022-01-06T16:05:22.216931Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"class BERTinferenceDataSet(Dataset):\n    \n    def __init__(self, sentences):\n        self.sentences = sentences\n\n    def __len__(self):\n        return len(self.sentences)\n\n    def __getitem__(self, idx):\n        sentence = self.sentences[idx]\n        bert_sent = tokenizer.encode_plus(sentence, \n                                         add_special_tokens = True, #[SEP][PAD]\n                                         max_length = max_len,\n                                         pad_to_max_length = True,\n                                         truncation = True)\n\n        ids = torch.tensor(bert_sent['input_ids'], dtype = torch.long)\n        mask = torch.tensor(bert_sent['attention_mask'], dtype = torch.long)\n\n        return{\n            'ids' : ids,\n            'mask' : mask\n             }","metadata":{"execution":{"iopub.status.busy":"2022-01-06T16:05:22.221376Z","iopub.execute_input":"2022-01-06T16:05:22.223253Z","iopub.status.idle":"2022-01-06T16:05:22.234164Z","shell.execute_reply.started":"2022-01-06T16:05:22.223201Z","shell.execute_reply":"2022-01-06T16:05:22.233412Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test_batch = 32\ntest_dataset = BERTinferenceDataSet(test['clean_text'])\ntest_dataloader = DataLoader(test_dataset, batch_size = test_batch, shuffle = False, num_workers = 4, pin_memory = True)\npthes","metadata":{"execution":{"iopub.status.busy":"2022-01-06T16:05:22.235532Z","iopub.execute_input":"2022-01-06T16:05:22.236002Z","iopub.status.idle":"2022-01-06T16:05:22.251796Z","shell.execute_reply.started":"2022-01-06T16:05:22.235967Z","shell.execute_reply":"2022-01-06T16:05:22.250834Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"allpreds = predicting(test_dataloader, model, pthes)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T16:05:22.253612Z","iopub.execute_input":"2022-01-06T16:05:22.254389Z","iopub.status.idle":"2022-01-06T16:06:50.322311Z","shell.execute_reply.started":"2022-01-06T16:05:22.254350Z","shell.execute_reply":"2022-01-06T16:06:50.321451Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"allpreds[0][0]","metadata":{"execution":{"iopub.status.busy":"2022-01-06T16:06:50.324396Z","iopub.execute_input":"2022-01-06T16:06:50.325916Z","iopub.status.idle":"2022-01-06T16:06:50.333178Z","shell.execute_reply.started":"2022-01-06T16:06:50.325870Z","shell.execute_reply":"2022-01-06T16:06:50.332233Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"preds = np.zeros(len(test_dataset))\nfor i in range(kfold):\n    preds += allpreds[i]\npreds = preds / kfold","metadata":{"execution":{"iopub.status.busy":"2022-01-06T16:06:50.334597Z","iopub.execute_input":"2022-01-06T16:06:50.334859Z","iopub.status.idle":"2022-01-06T16:06:50.343062Z","shell.execute_reply.started":"2022-01-06T16:06:50.334825Z","shell.execute_reply":"2022-01-06T16:06:50.342332Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"results = pd.DataFrame(preds)\nsubmission = pd.concat([test,results], axis = 1).drop(['question_text', 'clean_text'], axis = 1)\nsubmission.rename(columns = { 0:'target'}, inplace = True)\nsubmission.to_csv(\"submission.csv\", index = False)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T16:14:17.024509Z","iopub.execute_input":"2022-01-06T16:14:17.025416Z","iopub.status.idle":"2022-01-06T16:14:17.036577Z","shell.execute_reply.started":"2022-01-06T16:14:17.025360Z","shell.execute_reply":"2022-01-06T16:14:17.035869Z"},"trusted":true},"execution_count":null,"outputs":[]}]}