{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.7.10","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\nfrom tqdm import tqdm\nfrom sklearn.model_selection import train_test_split\nimport tensorflow as tf\nfrom keras.models import Sequential\nfrom keras.layers.recurrent import LSTM, GRU,SimpleRNN\nfrom keras.layers.core import Dense, Activation, Dropout\nfrom keras.layers.embeddings import Embedding\nfrom keras.layers.normalization import BatchNormalization\nfrom keras.utils import np_utils\nfrom sklearn import preprocessing, decomposition, model_selection, metrics, pipeline\nfrom keras.layers import GlobalMaxPooling1D, Conv1D, MaxPooling1D, Flatten, Bidirectional, SpatialDropout1D\nfrom keras.preprocessing import sequence, text\nfrom keras.callbacks import EarlyStopping\n\n\nimport matplotlib.pyplot as plt\nimport seaborn as sns\n%matplotlib inline\nfrom plotly import graph_objs as go\nimport plotly.express as px\nimport plotly.figure_factory as ff","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","execution":{"iopub.status.busy":"2022-11-27T18:22:11.088279Z","iopub.execute_input":"2022-11-27T18:22:11.088875Z","iopub.status.idle":"2022-11-27T18:22:20.081805Z","shell.execute_reply.started":"2022-11-27T18:22:11.088751Z","shell.execute_reply":"2022-11-27T18:22:20.08078Z"},"trusted":true},"execution_count":1,"outputs":[]},{"cell_type":"markdown","source":"# Configuring TPU's\n\nFor this version of Notebook we will be using TPU's as we have to built a BERT Model","metadata":{}},{"cell_type":"code","source":"# Detect hardware, return appropriate distribution strategy\ntry:\n    # TPU detection. No parameters necessary if TPU_NAME environment variable is\n    # set: this is always the case on Kaggle.\n    tpu = tf.distribute.cluster_resolver.TPUClusterResolver()\n    print('Running on TPU ', tpu.master())\nexcept ValueError:\n    tpu = None\n\nif tpu:\n    tf.config.experimental_connect_to_cluster(tpu)\n    tf.tpu.experimental.initialize_tpu_system(tpu)\n    strategy = tf.distribute.experimental.TPUStrategy(tpu)\nelse:\n    # Default distribution strategy in Tensorflow. Works on CPU and single GPU.\n    strategy = tf.distribute.get_strategy()\n\nprint(\"REPLICAS: \", strategy.num_replicas_in_sync)","metadata":{"execution":{"iopub.status.busy":"2022-11-27T18:22:33.534591Z","iopub.execute_input":"2022-11-27T18:22:33.535189Z","iopub.status.idle":"2022-11-27T18:22:33.550298Z","shell.execute_reply.started":"2022-11-27T18:22:33.535133Z","shell.execute_reply":"2022-11-27T18:22:33.549407Z"},"trusted":true},"execution_count":2,"outputs":[{"name":"stdout","text":"REPLICAS:  1\n","output_type":"stream"}]},{"cell_type":"code","source":"train = pd.read_csv('../input/main-dataset-amc/one_col_train.csv')\n# validation = pd.read_csv('/kaggle/input/jigsaw-multilingual-toxic-comment-classification/validation.csv')\ntest = pd.read_csv('../input/main-dataset-amc/one_col_test.csv')","metadata":{"_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","execution":{"iopub.status.busy":"2022-11-27T18:22:37.13911Z","iopub.execute_input":"2022-11-27T18:22:37.139615Z","iopub.status.idle":"2022-11-27T18:23:36.404857Z","shell.execute_reply.started":"2022-11-27T18:22:37.13958Z","shell.execute_reply":"2022-11-27T18:23:36.403991Z"},"trusted":true},"execution_count":3,"outputs":[]},{"cell_type":"code","source":"def roc_auc(predictions,target):\n    '''\n    This methods returns the AUC Score when given the Predictions\n    and Labels\n    '''\n\n    fpr, tpr, thresholds = metrics.roc_curve(target, predictions)\n    roc_auc = metrics.auc(fpr, tpr)\n    return roc_auc","metadata":{"execution":{"iopub.status.busy":"2022-11-27T18:23:46.816097Z","iopub.execute_input":"2022-11-27T18:23:46.81652Z","iopub.status.idle":"2022-11-27T18:23:46.822317Z","shell.execute_reply.started":"2022-11-27T18:23:46.816484Z","shell.execute_reply":"2022-11-27T18:23:46.821062Z"},"trusted":true},"execution_count":4,"outputs":[]},{"cell_type":"code","source":"# random_indexes = train[train.BROWSE_NODE_ID == 1045].index\n# import random\n# rand_indexes = random.choices(random_indexes, k=1e5)\n# rand_indexes\n# random_indexes[1]","metadata":{"execution":{"iopub.status.busy":"2022-11-27T18:23:55.903145Z","iopub.execute_input":"2022-11-27T18:23:55.903924Z","iopub.status.idle":"2022-11-27T18:23:55.90811Z","shell.execute_reply.started":"2022-11-27T18:23:55.903875Z","shell.execute_reply":"2022-11-27T18:23:55.907206Z"},"trusted":true},"execution_count":6,"outputs":[]},{"cell_type":"markdown","source":"### Data Preparation","metadata":{}},{"cell_type":"code","source":"# xtrain, xvalid, ytrain, yvalid = train_test_split(train.TDBB.values, train.BROWSE_NODE_ID.values, \n#                                                   random_state=42, \n#                                     test_size=0.2, shuffle=True)","metadata":{"execution":{"iopub.status.busy":"2021-07-31T10:03:34.389846Z","iopub.execute_input":"2021-07-31T10:03:34.390471Z","iopub.status.idle":"2021-07-31T10:03:35.78512Z","shell.execute_reply.started":"2021-07-31T10:03:34.390425Z","shell.execute_reply":"2021-07-31T10:03:35.784126Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# del train","metadata":{"execution":{"iopub.status.busy":"2021-07-31T10:03:46.962082Z","iopub.execute_input":"2021-07-31T10:03:46.962452Z","iopub.status.idle":"2021-07-31T10:03:46.966823Z","shell.execute_reply.started":"2021-07-31T10:03:46.962423Z","shell.execute_reply":"2021-07-31T10:03:46.965725Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Loading Dependencies\nimport os\nimport tensorflow as tf\nfrom tensorflow.keras.layers import Dense, Input\nfrom tensorflow.keras.optimizers import Adam\nfrom tensorflow.keras.models import Model\nfrom tensorflow.keras.callbacks import ModelCheckpoint\nfrom kaggle_datasets import KaggleDatasets\nimport transformers\n\nfrom tokenizers import BertWordPieceTokenizer","metadata":{"execution":{"iopub.status.busy":"2022-11-27T18:24:00.605296Z","iopub.execute_input":"2022-11-27T18:24:00.605917Z","iopub.status.idle":"2022-11-27T18:24:01.827517Z","shell.execute_reply.started":"2022-11-27T18:24:00.605878Z","shell.execute_reply":"2022-11-27T18:24:01.826378Z"},"trusted":true},"execution_count":7,"outputs":[]},{"cell_type":"code","source":"# # LOADING THE DATA\n\n# train1 = pd.read_csv(\"/kaggle/input/jigsaw-multilingual-toxic-comment-classification/jigsaw-toxic-comment-train.csv\")\n# valid = pd.read_csv('/kaggle/input/jigsaw-multilingual-toxic-comment-classification/validation.csv')\n# test = pd.read_csv('/kaggle/input/jigsaw-multilingual-toxic-comment-classification/test.csv')\n# sub = pd.read_csv('/kaggle/input/jigsaw-multilingual-toxic-comment-classification/sample_submission.csv')","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Encoder FOr DATA for understanding waht encode batch does read documentation of hugging face tokenizer :\nhttps://huggingface.co/transformers/main_classes/tokenizer.html here","metadata":{}},{"cell_type":"code","source":"def fast_encode(texts, tokenizer, chunk_size=256, maxlen=512):\n    \"\"\"\n    Encoder for encoding the text into sequence of integers for BERT Input\n    \"\"\"\n    tokenizer.enable_truncation(max_length=maxlen)\n    tokenizer.enable_padding(length=maxlen)\n    all_ids = []\n    \n    for i in tqdm(range(0, len(texts), chunk_size)):\n        text_chunk = texts[i:i+chunk_size].tolist()\n        encs = tokenizer.encode_batch(text_chunk)\n        all_ids.extend([enc.ids for enc in encs])\n    \n    return np.array(all_ids)","metadata":{"execution":{"iopub.status.busy":"2021-07-31T10:28:13.014262Z","iopub.execute_input":"2021-07-31T10:28:13.014672Z","iopub.status.idle":"2021-07-31T10:28:13.022692Z","shell.execute_reply.started":"2021-07-31T10:28:13.014636Z","shell.execute_reply":"2021-07-31T10:28:13.021577Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"#IMP DATA FOR CONFIG\n\nAUTO = tf.data.experimental.AUTOTUNE\n\n# Configuration\nEPOCHS = 2\nBATCH_SIZE = 16 * strategy.num_replicas_in_sync\nMAX_LEN = 192","metadata":{"execution":{"iopub.status.busy":"2021-07-31T10:28:22.203036Z","iopub.execute_input":"2021-07-31T10:28:22.203579Z","iopub.status.idle":"2021-07-31T10:28:22.207487Z","shell.execute_reply.started":"2021-07-31T10:28:22.203544Z","shell.execute_reply":"2021-07-31T10:28:22.206718Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Tokenization\n\n","metadata":{}},{"cell_type":"code","source":"# First load the real tokenizer\ntokenizer = transformers.DistilBertTokenizer.from_pretrained('distilbert-base-multilingual-cased')\n# Save the loaded tokenizer locally\ntokenizer.save_pretrained('.')\n# Reload it with the huggingface tokenizers library\nfast_tokenizer = BertWordPieceTokenizer('vocab.txt', lowercase=False)\nfast_tokenizer","metadata":{"execution":{"iopub.status.busy":"2021-07-31T10:28:26.882834Z","iopub.execute_input":"2021-07-31T10:28:26.883497Z","iopub.status.idle":"2021-07-31T10:28:29.468765Z","shell.execute_reply.started":"2021-07-31T10:28:26.883458Z","shell.execute_reply":"2021-07-31T10:28:29.467413Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"x_train = fast_encode(train.TDBB.astype(str), fast_tokenizer, maxlen=MAX_LEN)\n# x_valid = fast_encode(valid.TDBB.astype(str), fast_tokenizer, maxlen=MAX_LEN)\nx_test = fast_encode(test.content.astype(str), fast_tokenizer, maxlen=MAX_LEN)\n\n# y_train = train.BROWSE_NODE_ID.values\n# y_valid = valid.BROWSE_NODE_ID.values","metadata":{"execution":{"iopub.status.busy":"2021-07-31T10:28:58.833789Z","iopub.execute_input":"2021-07-31T10:28:58.834166Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import pickle\noutfile1 = open('x_train.pkl','wb')\npickle.dump(x_train,outfile1)\noutfile1.close()\n\noutfile2 = open('x_test.pkl','wb')\npickle.dump(x_test,outfile2)\noutfile2.close()\n\n# x_train.to_csv('x_train', index=False)\n# x_test.to_csv('x_test', index=False)","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# train_dataset = (\n#     tf.data.Dataset\n#     .from_tensor_slices((x_train, y_train))\n#     .repeat()\n#     .shuffle(2048)\n#     .batch(BATCH_SIZE)\n#     .prefetch(AUTO)\n# )\n\n# valid_dataset = (\n#     tf.data.Dataset\n#     .from_tensor_slices((x_valid, y_valid))\n#     .batch(BATCH_SIZE)\n#     .cache()\n#     .prefetch(AUTO)\n# )\n\n# test_dataset = (\n#     tf.data.Dataset\n#     .from_tensor_slices(x_test)\n#     .batch(BATCH_SIZE)\n# )","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# def build_model(transformer, max_len=512):\n#     \"\"\"\n#     function for training the BERT model\n#     \"\"\"\n#     input_word_ids = Input(shape=(max_len,), dtype=tf.int32, name=\"input_word_ids\")\n#     sequence_output = transformer(input_word_ids)[0]\n#     cls_token = sequence_output[:, 0, :]\n#     out = Dense(9919, activation='sigmoid')(cls_token)\n    \n#     model = Model(inputs=input_word_ids, outputs=out)\n#     model.compile(Adam(lr=1e-5), loss='binary_crossentropy', metrics=['accuracy'])\n    \n#     return model","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Starting Training\n\n","metadata":{}},{"cell_type":"code","source":"# %%time\n# with strategy.scope():\n#     transformer_layer = (\n#         transformers.TFDistilBertModel\n#         .from_pretrained('distilbert-base-multilingual-cased')\n#     )\n#     model = build_model(transformer_layer, max_len=MAX_LEN)\n# model.summary()","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# n_steps = x_train.shape[0] // BATCH_SIZE\n# train_history = model.fit(\n#     train_dataset,\n#     steps_per_epoch=n_steps,\n#     validation_data=valid_dataset,\n#     epochs=EPOCHS\n# )","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# n_steps = x_valid.shape[0] // BATCH_SIZE\n# train_history_2 = model.fit(\n#     valid_dataset.repeat(),\n#     steps_per_epoch=n_steps,\n#     epochs=EPOCHS*2\n# )","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# sub['toxic'] = model.predict(test_dataset, verbose=1)\n# sub.to_csv('submission.csv', index=False)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# using pytorch \n\nimport pandas as pd\nimport numpy as np\nfrom sklearn.model_selection import train_test_split\n\nimport torch\nfrom tqdm.notebook import tqdm\n\nfrom transformers import BertTokenizer\nfrom torch.utils.data import TensorDataset\n\nfrom transformers import BertForSequenceClassification\n\n\n\n## Loading data\n\ntrain = pd.read_csv('../input/main-dataset-amc/one_col_train.csv')\ntest = pd.read_csv('../input/main-dataset-amc/one_col_test.csv')\n\n## Splitting the data\n\nX_train, X_val, y_train, y_val = train_test_split(train.index.values, \n                                                  train.BROWSE_NODE_ID.values, \n                                                  test_size=0.3, \n                                                  random_state=2001)\n\ntrain['data_type'] = ['not_set']*train.shape[0]\n\ntrain.loc[X_train, 'data_type'] = 'train'\ntrain.loc[X_val, 'data_type'] = 'val'\n\ntokenizer = BertTokenizer.from_pretrained('bert-base-uncased', \n                                          do_lower_case=True)\n                                          \nencoded_data_train = tokenizer.batch_encode_plus(\n    train[train.data_type=='train'].TDBB.values, \n    add_special_tokens=True, \n    return_attention_mask=True, \n    pad_to_max_length=True, \n    max_length=256, \n    return_tensors='pt'\n)\n\nencoded_data_val = tokenizer.batch_encode_plus(\n    train[train.data_type=='val'].TDBB.values, \n    add_special_tokens=True, \n    return_attention_mask=True, \n    pad_to_max_length=True, \n    max_length=256, \n    return_tensors='pt'\n)\n\n\ninput_ids_train = encoded_data_train['input_ids']\nattention_masks_train = encoded_data_train['attention_mask']\nlabels_train = torch.tensor(train[train.data_type=='train'].BROWSE_NODE_ID.values)\n\ninput_ids_val = encoded_data_val['input_ids']\nattention_masks_val = encoded_data_val['attention_mask']\nlabels_val = torch.tensor(train[train.data_type=='val'].BROWSE_NODE_ID.values)\n\ndataset_train = TensorDataset(input_ids_train, attention_masks_train, labels_train)\ndataset_val = TensorDataset(input_ids_val, attention_masks_val, labels_val)","metadata":{"execution":{"iopub.status.busy":"2021-08-01T07:38:53.29094Z","iopub.execute_input":"2021-08-01T07:38:53.291595Z"},"trusted":true},"execution_count":null,"outputs":[]}]}