{"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":"The main points of our solution are presented in the notebook.\n\nWe cleaned up some of the bad data. We prepared a separate balanced data sample for training.\n\nOne of the BERT models was used as the main model. We used PyTorch as a framework, with the Pytorch Lightning add-on for learning.\n\nWe also used a trick to replace the learning dataset in the learning process.\n\nHere's what we got.\n","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19"}},{"cell_type":"markdown","source":"![2022-07-25_23-50.png](data:image/png;base64,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)","metadata":{}},{"cell_type":"code","source":"import numpy as np\nimport re\nimport pandas as pd\nfrom sklearn.model_selection import train_test_split\nfrom tqdm.notebook import tqdm","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:15:57.503069Z","iopub.execute_input":"2022-07-28T05:15:57.503518Z","iopub.status.idle":"2022-07-28T05:15:59.126004Z","shell.execute_reply.started":"2022-07-28T05:15:57.503442Z","shell.execute_reply":"2022-07-28T05:15:59.123638Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"dataset_dir = '../input/goodreads-books-reviews-290312/'","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:15:59.133776Z","iopub.execute_input":"2022-07-28T05:15:59.136651Z","iopub.status.idle":"2022-07-28T05:15:59.146244Z","shell.execute_reply.started":"2022-07-28T05:15:59.136612Z","shell.execute_reply":"2022-07-28T05:15:59.145334Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df = pd.read_csv(dataset_dir + 'goodreads_train.csv')\ndf_test = pd.read_csv(dataset_dir + 'goodreads_test.csv')","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:15:59.151688Z","iopub.execute_input":"2022-07-28T05:15:59.153957Z","iopub.status.idle":"2022-07-28T05:16:32.848699Z","shell.execute_reply.started":"2022-07-28T05:15:59.153921Z","shell.execute_reply":"2022-07-28T05:16:32.847686Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:16:32.851439Z","iopub.execute_input":"2022-07-28T05:16:32.851801Z","iopub.status.idle":"2022-07-28T05:16:32.883162Z","shell.execute_reply.started":"2022-07-28T05:16:32.851764Z","shell.execute_reply":"2022-07-28T05:16:32.882227Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Remove needless columns\n\ndf.drop(['date_added','date_updated','read_at','started_at','n_votes','n_comments'],axis=1, inplace=True)","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:16:32.884602Z","iopub.execute_input":"2022-07-28T05:16:32.884941Z","iopub.status.idle":"2022-07-28T05:16:33.038039Z","shell.execute_reply.started":"2022-07-28T05:16:32.884907Z","shell.execute_reply":"2022-07-28T05:16:33.036688Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Remove duplicates\n\ndf.drop_duplicates(subset=['review_text'], inplace=True, keep='first')\nlen(df)","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:16:33.039919Z","iopub.execute_input":"2022-07-28T05:16:33.040394Z","iopub.status.idle":"2022-07-28T05:16:35.517089Z","shell.execute_reply.started":"2022-07-28T05:16:33.040356Z","shell.execute_reply":"2022-07-28T05:16:35.516200Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Clean the data\n\nget_mean_word_length = lambda phrase: np.mean(list(map(len, phrase.split())))\ncondition = df.review_text.apply(get_mean_word_length) > 30\ndf[condition]","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:16:35.521394Z","iopub.execute_input":"2022-07-28T05:16:35.523947Z","iopub.status.idle":"2022-07-28T05:17:27.304256Z","shell.execute_reply.started":"2022-07-28T05:16:35.523908Z","shell.execute_reply":"2022-07-28T05:17:27.303189Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"index = df.index\nunappropriate_rows_indices = index[condition].tolist()\ndf.drop(unappropriate_rows_indices,inplace=True)\nlen(df)","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:27.305584Z","iopub.execute_input":"2022-07-28T05:17:27.306394Z","iopub.status.idle":"2022-07-28T05:17:27.447082Z","shell.execute_reply.started":"2022-07-28T05:17:27.306334Z","shell.execute_reply":"2022-07-28T05:17:27.446008Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Explore whether the classes are balanced. They are unbalanced, the rarest class is 1\n\ndf.rating.value_counts()","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:27.448675Z","iopub.execute_input":"2022-07-28T05:17:27.449325Z","iopub.status.idle":"2022-07-28T05:17:27.465983Z","shell.execute_reply.started":"2022-07-28T05:17:27.449285Z","shell.execute_reply":"2022-07-28T05:17:27.465075Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Let's extract a separate part from the data with balanced classes\n# At the same time, it is important to keep the ratio of classes\n\ndf.sort_values(['book_id', 'rating', 'user_id'], inplace=True)\ndf['book_num'] = df.groupby(by=['book_id'], as_index=False, sort=False)['book_id'].transform(lambda s: np.arange(len(s))+1)\ndf.sort_values(['book_id', 'book_num'], inplace=True)\ndf['kfold'] = 0","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:27.470806Z","iopub.execute_input":"2022-07-28T05:17:27.471445Z","iopub.status.idle":"2022-07-28T05:17:42.033520Z","shell.execute_reply.started":"2022-07-28T05:17:27.471410Z","shell.execute_reply":"2022-07-28T05:17:42.032524Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Fill in kfold = 1 for all scores 1\n\ndf.loc[df['rating'] == 1,'kfold'] = 1\n\n# The volume of the minimum part will determine the size of the balanced part\n\nkfold_size = len(df[df['rating'] == 1])\nkfold_size","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:42.035090Z","iopub.execute_input":"2022-07-28T05:17:42.035464Z","iopub.status.idle":"2022-07-28T05:17:42.071622Z","shell.execute_reply.started":"2022-07-28T05:17:42.035425Z","shell.execute_reply":"2022-07-28T05:17:42.070526Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Shuffle the data and add the rest of the scores to our part\n\ndf = df.sample(frac=1)\ndf.sort_values(['book_num'], inplace=True)\ndf['rating_num'] = df.groupby(by=['rating'], as_index=False, sort=False)['rating'].transform(lambda s: np.arange(len(s))+1)\ndf.loc[df['rating_num'] <= kfold_size,'kfold'] = 1","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:42.073406Z","iopub.execute_input":"2022-07-28T05:17:42.073770Z","iopub.status.idle":"2022-07-28T05:17:42.828742Z","shell.execute_reply.started":"2022-07-28T05:17:42.073732Z","shell.execute_reply":"2022-07-28T05:17:42.827685Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Divide the balance part of the sample into train and validation\n\ndf_balance = df[df['kfold']==1]\ndf_balance_train, df_balance_valid = train_test_split(df_balance, test_size=0.2, stratify=df_balance['rating'])","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:42.829981Z","iopub.execute_input":"2022-07-28T05:17:42.830377Z","iopub.status.idle":"2022-07-28T05:17:42.967978Z","shell.execute_reply.started":"2022-07-28T05:17:42.830338Z","shell.execute_reply":"2022-07-28T05:17:42.966951Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Add the data from the balance sample to the full part\n\ndf = df[df['kfold']==0]\ndf = pd.concat([df,df_balance_train])","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:42.969645Z","iopub.execute_input":"2022-07-28T05:17:42.970007Z","iopub.status.idle":"2022-07-28T05:17:43.307161Z","shell.execute_reply.started":"2022-07-28T05:17:42.969969Z","shell.execute_reply":"2022-07-28T05:17:43.306162Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df_balance_train.rating.value_counts()","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:43.308900Z","iopub.execute_input":"2022-07-28T05:17:43.309297Z","iopub.status.idle":"2022-07-28T05:17:43.318730Z","shell.execute_reply.started":"2022-07-28T05:17:43.309258Z","shell.execute_reply":"2022-07-28T05:17:43.316761Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# For training, we use the pytorch lightning framework","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:43.320131Z","iopub.execute_input":"2022-07-28T05:17:43.321102Z","iopub.status.idle":"2022-07-28T05:17:43.326770Z","shell.execute_reply.started":"2022-07-28T05:17:43.321066Z","shell.execute_reply":"2022-07-28T05:17:43.325751Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# !pip install transformers\n# !pip install pytorch_lightning","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:43.328186Z","iopub.execute_input":"2022-07-28T05:17:43.329440Z","iopub.status.idle":"2022-07-28T05:17:43.336483Z","shell.execute_reply.started":"2022-07-28T05:17:43.329402Z","shell.execute_reply":"2022-07-28T05:17:43.335402Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from typing import Any, Dict, List\nimport importlib\nimport random\nimport torch\nfrom torch.utils.data import Dataset\nfrom torch import nn\nimport pytorch_lightning as pl\nimport torchmetrics\n\nfrom transformers import BertTokenizerFast as BertTokenizer, BertModel","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:43.338360Z","iopub.execute_input":"2022-07-28T05:17:43.338836Z","iopub.status.idle":"2022-07-28T05:17:47.434729Z","shell.execute_reply.started":"2022-07-28T05:17:43.338799Z","shell.execute_reply":"2022-07-28T05:17:47.433717Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Dataset for reading data during training\n\nclass BertDataset(Dataset):\n  def __init__(\n    self, \n    data: List[str], \n    labels: List[int],\n    tokenizer: BertTokenizer, \n    sent_len: int = 128,\n    test: bool = False\n  ):\n    self.tokenizer = tokenizer\n    self.sent_len = sent_len\n    \n    self.data = data\n    self.labels = labels\n    \n    self.test = test\n    \n  def __len__(self):\n    return len(self.data)\n\n  def getitem_index(self, index: int):\n    comment_text = self.data[index]\n    \n    if self.test:\n        labels = None\n    else:\n        labels = self.labels[index]\n\n    encoding = self.tokenizer(\n      comment_text,\n      add_special_tokens=True,\n      max_length=self.sent_len,\n      return_token_type_ids=False,\n      padding=\"max_length\",\n      truncation=True,\n      return_attention_mask=True,\n      return_tensors='pt',\n    )\n    \n    result_dict = dict(\n      input_ids=encoding[\"input_ids\"].flatten(),\n      attention_mask=encoding[\"attention_mask\"].flatten(),\n      labels=labels\n    )\n    \n    if self.test:\n        result_dict.pop('labels')\n    \n    return result_dict\n\n  def __getitem__(self, index: int):\n      return self.getitem_index(index)","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:47.436131Z","iopub.execute_input":"2022-07-28T05:17:47.436858Z","iopub.status.idle":"2022-07-28T05:17:47.448978Z","shell.execute_reply.started":"2022-07-28T05:17:47.436818Z","shell.execute_reply":"2022-07-28T05:17:47.447909Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"BERT_MODEL_NAME = 'alperiox/autonlp-user-review-classification-536415182'\nBATCH_SIZE = 64\nNUM_WORKERS = 2\nSENT_LEN = 128\nMAX_EPOCHS = 3\nLEARNING_RATE = 2.0e-05\nWEIGHT_DECAY = 0.01\nEPS = 1.0e-06\nBALANCE_EPOCHS=4","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:47.450452Z","iopub.execute_input":"2022-07-28T05:17:47.450884Z","iopub.status.idle":"2022-07-28T05:17:47.463105Z","shell.execute_reply.started":"2022-07-28T05:17:47.450839Z","shell.execute_reply":"2022-07-28T05:17:47.462225Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# The model for learning\n\nclass Net(nn.Module):\n    def __init__(self) -> None:\n        super().__init__()\n        self.encoder = BertModel.from_pretrained(pretrained_model_name_or_path=BERT_MODEL_NAME,return_dict=True) \n        self.decoder = torch.nn.Linear(self.encoder.config.hidden_size, out_features=6)\n\n    def forward(self, **kwargs):\n        out = self.encoder(**kwargs)\n        logits = self.decoder(out.pooler_output)\n        return logits","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:47.464707Z","iopub.execute_input":"2022-07-28T05:17:47.465442Z","iopub.status.idle":"2022-07-28T05:17:47.476645Z","shell.execute_reply.started":"2022-07-28T05:17:47.465405Z","shell.execute_reply":"2022-07-28T05:17:47.475723Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Pytorch Lightning Training Class Description\n\nclass LitBertNER(pl.LightningModule):\n    def __init__(self):\n        super(LitBertNER, self).__init__()\n        self.model = Net()\n        self.loss = torch.nn.CrossEntropyLoss()\n        self.metric = torchmetrics.F1Score(average = 'macro',num_classes=6)\n\n    def forward(self, input_ids, attention_mask, labels=None):\n        output = self.model(input_ids=input_ids, attention_mask=attention_mask)\n        loss = 0\n        if labels is not None:\n            loss = self.loss(output, labels)\n            \n        return loss, output\n\n    def setup(self, stage=None):\n        train_data, train_labels = df['review_text'].values, df['rating'].values\n        #valid_data, valid_labels = df_valid['review_text'].values, df_valid['rating'].values\n\n        train_balance_data, train_balance_labels = df_balance_train['review_text'].values, df_balance_train['rating'].values\n        valid_balance_data, valid_balance_labels = df_balance_valid['review_text'].values, df_balance_valid['rating'].values\n\n        tokenizer = BertTokenizer.from_pretrained(BERT_MODEL_NAME)        \n\n        sent_len = SENT_LEN\n\n        self.train_dataset = BertDataset(\n            data=train_data, labels = train_labels, tokenizer=tokenizer, \n            sent_len = sent_len\n        )\n        self.train_balance_dataset = BertDataset(\n            data=train_balance_data, labels = train_balance_labels, tokenizer=tokenizer, \n            sent_len = sent_len\n        )\n        self.valid_dataset = BertDataset(\n            data=valid_balance_data, labels = valid_balance_labels, tokenizer=tokenizer, \n            sent_len = sent_len\n        )\n\n    def train_dataloader(self):\n\n        if self.current_epoch > 1:\n          train_loader = torch.utils.data.DataLoader(\n              self.train_balance_dataset,\n              batch_size=BATCH_SIZE,\n              num_workers=NUM_WORKERS,\n              collate_fn=None,\n              shuffle=True,\n          )\n        else:\n          train_loader = torch.utils.data.DataLoader(\n              self.train_dataset,\n              batch_size=BATCH_SIZE,\n              num_workers=NUM_WORKERS,\n              collate_fn=None,\n              shuffle=True,\n          )\n        return train_loader\n\n    def val_dataloader(self):\n        valid_loader = torch.utils.data.DataLoader(\n            self.valid_dataset,\n            batch_size=BATCH_SIZE,\n            num_workers=NUM_WORKERS,\n            collate_fn=None,\n            shuffle=False,\n        )\n\n        return valid_loader\n\n    def configure_optimizers(self):\n        \n        steps_per_epoch=len(self.train_dataloader())\n        total_training_steps = steps_per_epoch * MAX_EPOCHS        \n        \n        warmup_steps = total_training_steps // 5\n        \n        no_decay = [\"bias\", \"LayerNorm.weight\"]\n        optimizer_grouped_parameters = [\n            {\n                \"params\": [p for n, p in self.model.named_parameters() if not any(nd in n for nd in no_decay)],\n                \"weight_decay\": WEIGHT_DECAY,\n            },\n            {\n                \"params\": [p for n, p in self.model.named_parameters() if any(nd in n for nd in no_decay)],\n                \"weight_decay\": 0.0,\n            },\n        ]        \n        \n        optimizer = torch.optim.AdamW(optimizer_grouped_parameters, lr = LEARNING_RATE, weight_decay = WEIGHT_DECAY, eps = EPS)\n\n        scheduler = torch.optim.lr_scheduler.OneCycleLR(optimizer, max_lr=LEARNING_RATE,\n                                                        total_steps=total_training_steps,\n                                                        pct_start=0.1)\n        return (\n            [optimizer],\n            [{'scheduler': scheduler, 'interval': 'step', 'monitor': 'valid_f1'}],\n        )\n\n    def training_step(self, batch, batch_idx):\n        input_ids = batch[\"input_ids\"]\n        attention_mask = batch[\"attention_mask\"]\n        labels = batch[\"labels\"]\n\n        loss, outputs = self(input_ids, attention_mask, labels)\n        self.log(\"train_loss\", loss, prog_bar=True, logger=True)\n        \n        score = self.metric(outputs, labels)\n        self.log(f'train_f1', score, on_step=True, on_epoch=True, prog_bar=True, logger=True)\n            \n        return loss\n\n    def validation_step(self, batch, batch_idx):\n        input_ids = batch[\"input_ids\"]\n        attention_mask = batch[\"attention_mask\"]\n        labels = batch[\"labels\"]\n\n        loss, outputs = self(input_ids, attention_mask, labels)\n        self.log(\"val_loss\", loss, prog_bar=True, logger=True)\n        \n        score = self.metric(outputs, labels)\n        self.log(f'valid_f1', score, on_step=True, on_epoch=True, prog_bar=True, logger=True)\n        \n    def predict_step(self, batch, batch_idx):\n        input_ids = batch[\"input_ids\"]\n        attention_mask = batch[\"attention_mask\"]\n        loss, outputs = self(input_ids, attention_mask)\n\n        return outputs        ","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:47.478214Z","iopub.execute_input":"2022-07-28T05:17:47.478597Z","iopub.status.idle":"2022-07-28T05:17:47.502790Z","shell.execute_reply.started":"2022-07-28T05:17:47.478560Z","shell.execute_reply":"2022-07-28T05:17:47.501784Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Training\n\npl_model = LitBertNER()","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:47.504900Z","iopub.execute_input":"2022-07-28T05:17:47.505953Z","iopub.status.idle":"2022-07-28T05:17:59.969991Z","shell.execute_reply.started":"2022-07-28T05:17:47.505919Z","shell.execute_reply":"2022-07-28T05:17:59.968980Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#Log in Google Tensorboard\nloggers = []\nloggers.append(pl.loggers.TensorBoardLogger(save_dir = 'logs'))\nloggers.append(pl.loggers.CSVLogger(save_dir = 'csv_logs'))","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:59.971397Z","iopub.execute_input":"2022-07-28T05:17:59.971845Z","iopub.status.idle":"2022-07-28T05:17:59.978390Z","shell.execute_reply.started":"2022-07-28T05:17:59.971805Z","shell.execute_reply":"2022-07-28T05:17:59.977300Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"callbacks = []\ncallbacks.append(pl.callbacks.LearningRateMonitor(logging_interval='step'))\ncallbacks.append(pl.callbacks.RichProgressBar(leave=True))\ncallbacks.append(pl.callbacks.ModelCheckpoint(monitor='valid_f1', \n                                              save_top_k=2, \n                                              save_last=True, \n                                              dirpath = 'saved_models',\n                                              filename = '{epoch}-{valid_f1:.6f}',\n                                              mode= 'max',\n                                             ))","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:59.980140Z","iopub.execute_input":"2022-07-28T05:17:59.980908Z","iopub.status.idle":"2022-07-28T05:17:59.989407Z","shell.execute_reply.started":"2022-07-28T05:17:59.980870Z","shell.execute_reply":"2022-07-28T05:17:59.988280Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"trainer = pl.Trainer(max_epochs=MAX_EPOCHS+BALANCE_EPOCHS, \n                     gpus=1, \n                     strategy='dp',\n                     reload_dataloaders_every_n_epochs = 2, \n                     log_every_n_steps=100,\n                     logger=loggers,\n                     callbacks=callbacks,\n                    )\n                     #limit_train_batches = 4, \n                     #limit_val_batches = 4, ","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:17:59.991189Z","iopub.execute_input":"2022-07-28T05:17:59.991748Z","iopub.status.idle":"2022-07-28T05:18:00.063975Z","shell.execute_reply.started":"2022-07-28T05:17:59.991714Z","shell.execute_reply":"2022-07-28T05:18:00.062963Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#Open Google Tensorboard\n#%tensorboard --logdir logs","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:18:00.066176Z","iopub.execute_input":"2022-07-28T05:18:00.066527Z","iopub.status.idle":"2022-07-28T05:18:00.074213Z","shell.execute_reply.started":"2022-07-28T05:18:00.066490Z","shell.execute_reply":"2022-07-28T05:18:00.073289Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"trainer.fit(pl_model)","metadata":{"execution":{"iopub.status.busy":"2022-07-28T05:18:00.081641Z","iopub.execute_input":"2022-07-28T05:18:00.081884Z","iopub.status.idle":"2022-07-28T13:09:05.904370Z","shell.execute_reply.started":"2022-07-28T05:18:00.081861Z","shell.execute_reply":"2022-07-28T13:09:05.903287Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#saved models dir\n!ls saved_models","metadata":{"execution":{"iopub.status.busy":"2022-07-28T13:09:05.907129Z","iopub.execute_input":"2022-07-28T13:09:05.908024Z","iopub.status.idle":"2022-07-28T13:09:06.904585Z","shell.execute_reply.started":"2022-07-28T13:09:05.907973Z","shell.execute_reply":"2022-07-28T13:09:06.903213Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#logs dir\n!ls csv_logs/lightning_logs/version_0","metadata":{"execution":{"iopub.status.busy":"2022-07-28T13:09:06.906862Z","iopub.execute_input":"2022-07-28T13:09:06.910646Z","iopub.status.idle":"2022-07-28T13:09:08.228586Z","shell.execute_reply.started":"2022-07-28T13:09:06.910563Z","shell.execute_reply":"2022-07-28T13:09:08.227366Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#read log\nlog = pd.read_csv('csv_logs/lightning_logs/version_0/metrics.csv')\nlog.head()","metadata":{"execution":{"iopub.status.busy":"2022-07-28T13:09:08.230638Z","iopub.execute_input":"2022-07-28T13:09:08.230991Z","iopub.status.idle":"2022-07-28T13:09:08.259597Z","shell.execute_reply.started":"2022-07-28T13:09:08.230960Z","shell.execute_reply":"2022-07-28T13:09:08.258500Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#predict model","metadata":{"execution":{"iopub.status.busy":"2022-07-28T13:09:08.261415Z","iopub.execute_input":"2022-07-28T13:09:08.261784Z","iopub.status.idle":"2022-07-28T13:09:08.502646Z","shell.execute_reply.started":"2022-07-28T13:09:08.261747Z","shell.execute_reply":"2022-07-28T13:09:08.501233Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#pl_model = pl_model.load_from_checkpoint('saved_models/checkpoint_filename.pkl', strict=True).","metadata":{"execution":{"iopub.status.busy":"2022-07-28T13:09:08.506474Z","iopub.execute_input":"2022-07-28T13:09:08.507269Z","iopub.status.idle":"2022-07-28T13:09:08.514166Z","shell.execute_reply.started":"2022-07-28T13:09:08.507190Z","shell.execute_reply":"2022-07-28T13:09:08.512611Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df_test = pd.read_csv(dataset_dir + 'goodreads_test.csv')","metadata":{"execution":{"iopub.status.busy":"2022-07-28T13:09:08.516048Z","iopub.execute_input":"2022-07-28T13:09:08.516732Z","iopub.status.idle":"2022-07-28T13:09:20.698198Z","shell.execute_reply.started":"2022-07-28T13:09:08.516693Z","shell.execute_reply":"2022-07-28T13:09:20.697045Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"tokenizer = BertTokenizer.from_pretrained(BERT_MODEL_NAME)\npredict_dataset = BertDataset(data=df_test['review_text'].values, labels = None, tokenizer=tokenizer, sent_len = SENT_LEN, test=True)\n\npredict_loader = torch.utils.data.DataLoader(\n            predict_dataset,\n            batch_size=BATCH_SIZE,\n            num_workers=0,\n            collate_fn=None,\n            shuffle=False,\n        )","metadata":{"execution":{"iopub.status.busy":"2022-07-28T13:09:20.699583Z","iopub.execute_input":"2022-07-28T13:09:20.700018Z","iopub.status.idle":"2022-07-28T13:09:22.841338Z","shell.execute_reply.started":"2022-07-28T13:09:20.699975Z","shell.execute_reply":"2022-07-28T13:09:22.840302Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df_test = df_test[['review_id']]","metadata":{"execution":{"iopub.status.busy":"2022-07-28T13:09:22.844150Z","iopub.execute_input":"2022-07-28T13:09:22.845109Z","iopub.status.idle":"2022-07-28T13:09:22.858681Z","shell.execute_reply.started":"2022-07-28T13:09:22.845067Z","shell.execute_reply":"2022-07-28T13:09:22.857714Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"outputs = trainer.predict(pl_model, predict_loader) ","metadata":{"execution":{"iopub.status.busy":"2022-07-28T13:09:22.860299Z","iopub.execute_input":"2022-07-28T13:09:22.860889Z","iopub.status.idle":"2022-07-28T13:45:55.323331Z","shell.execute_reply.started":"2022-07-28T13:09:22.860850Z","shell.execute_reply":"2022-07-28T13:45:55.322418Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"list_logits = []\nlist_predict = []\nfor logits in outputs:\n    soft_logits = torch.softmax(logits,axis=1)\n    list_logits.extend(soft_logits.cpu().numpy())\n\n    predict = torch.argmax(logits, axis=1)\n    list_predict.extend(predict.cpu().numpy())","metadata":{"execution":{"iopub.status.busy":"2022-07-28T13:45:55.324733Z","iopub.execute_input":"2022-07-28T13:45:55.325199Z","iopub.status.idle":"2022-07-28T13:45:55.798179Z","shell.execute_reply.started":"2022-07-28T13:45:55.325161Z","shell.execute_reply":"2022-07-28T13:45:55.797174Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df_test['rating'] = list_predict\ndf_test.to_csv('submission.csv',columns = ['review_id', 'rating'], header = ['review_id', 'rating'], index = False)","metadata":{"execution":{"iopub.status.busy":"2022-07-28T13:45:55.799516Z","iopub.execute_input":"2022-07-28T13:45:55.799883Z","iopub.status.idle":"2022-07-28T13:45:57.186611Z","shell.execute_reply.started":"2022-07-28T13:45:55.799843Z","shell.execute_reply":"2022-07-28T13:45:57.185581Z"},"trusted":true},"execution_count":null,"outputs":[]}]}