Created
February 19, 2020 14:00
-
-
Save jshirius/f57e5c3044d4e56d50d7c4dc5553d185 to your computer and use it in GitHub Desktop.
Google QUEST Q&A Labelingコンペ-銅メダル検証
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| { | |
| "cells": [ | |
| { | |
| "cell_type": "code", | |
| "execution_count": 1, | |
| "metadata": { | |
| "scrolled": true | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "import pandas as pd\n", | |
| "import numpy as np\n", | |
| "from sklearn.model_selection import GroupKFold\n", | |
| "import matplotlib.pyplot as plt\n", | |
| "from tqdm.notebook import tqdm\n", | |
| "import tensorflow_hub as hub\n", | |
| "import tensorflow as tf\n", | |
| "import bert_tokenization as tokenization\n", | |
| "import tensorflow.keras.backend as K\n", | |
| "import gc\n", | |
| "import os\n", | |
| "from scipy.stats import spearmanr\n", | |
| "from math import floor, ceil\n", | |
| "from tensorflow.keras.models import load_model\n", | |
| "\n", | |
| "np.set_printoptions(suppress=True)" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": {}, | |
| "source": [ | |
| "## Abuout this kernel\n", | |
| "\n", | |
| "In this kernel, an example of adding features other than character strings is described with reference to \"l Bert-base TF2.0 (minimalistic) III\".\n", | |
| "\n", | |
| "Referenced kernel Bert-base TF2.0 (minimalistic) III<br>\n", | |
| "https://www.kaggle.com/bibek777/bert-base-tf2-0-minimalistic-iii\n", | |
| " " | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 3, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "name": "stdout", | |
| "output_type": "stream", | |
| "text": [ | |
| "train shape = (6079, 41)\n", | |
| "test shape = (476, 11)\n", | |
| "\n", | |
| "output categories:\n", | |
| "\t ['question_asker_intent_understanding', 'question_body_critical', 'question_conversational', 'question_expect_short_answer', 'question_fact_seeking', 'question_has_commonly_accepted_answer', 'question_interestingness_others', 'question_interestingness_self', 'question_multi_intent', 'question_not_really_a_question', 'question_opinion_seeking', 'question_type_choice', 'question_type_compare', 'question_type_consequence', 'question_type_definition', 'question_type_entity', 'question_type_instructions', 'question_type_procedure', 'question_type_reason_explanation', 'question_type_spelling', 'question_well_written', 'answer_helpful', 'answer_level_of_information', 'answer_plausible', 'answer_relevance', 'answer_satisfaction', 'answer_type_instructions', 'answer_type_procedure', 'answer_type_reason_explanation', 'answer_well_written']\n", | |
| "\n", | |
| "input categories:\n", | |
| "\t ['question_title', 'question_body', 'answer']\n" | |
| ] | |
| } | |
| ], | |
| "source": [ | |
| "PATH = '../input/google-quest-challenge/'\n", | |
| "\n", | |
| "\n", | |
| "#事前学習のモデルをLargeに変更する\n", | |
| "BERT_PATH = '../input/bert-base-from-tfhub/bert_en_uncased_L-12_H-768_A-12'\n", | |
| "\n", | |
| "#現状、Largeだとメモリーエラーになる\n", | |
| "#BERT_PATH = '../input/bert-base-from-tfhub-l24-h1024-a16/tf_bert_en_cased_L-24_H-1024_A-16'\n", | |
| "\n", | |
| "\n", | |
| "tokenizer = tokenization.FullTokenizer(BERT_PATH+'/assets/vocab.txt', True)\n", | |
| "MAX_SEQUENCE_LENGTH = 512\n", | |
| "\n", | |
| "df_train = pd.read_csv(PATH+'train.csv')\n", | |
| "df_test = pd.read_csv(PATH+'test.csv')\n", | |
| "df_sub = pd.read_csv(PATH+'sample_submission.csv')\n", | |
| "\n", | |
| "TARGET_COLUMNS = df_sub.columns.values[1:].tolist()\n", | |
| "\n", | |
| "print('train shape =', df_train.shape)\n", | |
| "print('test shape =', df_test.shape)\n", | |
| "\n", | |
| "output_categories = list(df_train.columns[11:])\n", | |
| "input_categories = list(df_train.columns[[1,2,5]])\n", | |
| "print('\\noutput categories:\\n\\t', output_categories)\n", | |
| "print('\\ninput categories:\\n\\t', input_categories)" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 4, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "text/plain": [ | |
| "'\\nprint(\"データ追加前\"), print(df_train.shape)\\n\\n#データを追加する\\n#/kaggle/input/to-google-quest-qa-labeling/train_append.csv\\ndf_train_append = pd.read_csv(\\'../input/to-google-quest-qa-labeling/train_append_id800up.csv\\')\\ndf_train_append = df_train_append.drop(\"index\", axis=1)\\nprint(\"追加するデータ量\"), print(df_train_append.shape)\\n\\n#追加後の行数\\ndf_train = df_train.append(df_train_append)\\ndf_train = df_train.reset_index(drop=True)\\nprint(\"データ追加後\"), print(df_train.shape)\\n'" | |
| ] | |
| }, | |
| "execution_count": 4, | |
| "metadata": {}, | |
| "output_type": "execute_result" | |
| } | |
| ], | |
| "source": [ | |
| "#データ追加前の形状\n", | |
| "#結局、やくたたずだった\n", | |
| "\"\"\"\n", | |
| "print(\"データ追加前\"), print(df_train.shape)\n", | |
| "\n", | |
| "#データを追加する\n", | |
| "#/kaggle/input/to-google-quest-qa-labeling/train_append.csv\n", | |
| "df_train_append = pd.read_csv('../input/to-google-quest-qa-labeling/train_append_id800up.csv')\n", | |
| "df_train_append = df_train_append.drop(\"index\", axis=1)\n", | |
| "print(\"追加するデータ量\"), print(df_train_append.shape)\n", | |
| "\n", | |
| "#追加後の行数\n", | |
| "df_train = df_train.append(df_train_append)\n", | |
| "df_train = df_train.reset_index(drop=True)\n", | |
| "print(\"データ追加後\"), print(df_train.shape)\n", | |
| "\"\"\"\n" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 5, | |
| "metadata": {}, | |
| "outputs": [], | |
| "source": [ | |
| "#トレーニングデータ、テストデータをを少なく(100個くらい)する\n", | |
| "#コードチェック用の\n", | |
| "\n", | |
| "#df_train = pd.read_csv(PATH+'train.csv')\n", | |
| "#df_test = pd.read_csv(PATH+'test.csv')\n" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 6, | |
| "metadata": {}, | |
| "outputs": [], | |
| "source": [ | |
| "def _get_masks(tokens, max_seq_length):\n", | |
| " \"\"\"Mask for padding\"\"\"\n", | |
| " if len(tokens)>max_seq_length:\n", | |
| " raise IndexError(\"Token length more than max seq length!\")\n", | |
| " return [1]*len(tokens) + [0] * (max_seq_length - len(tokens))\n", | |
| "\n", | |
| "def _get_segments(tokens, max_seq_length):\n", | |
| " \"\"\"Segments: 0 for the first sequence, 1 for the second\"\"\"\n", | |
| " if len(tokens)>max_seq_length:\n", | |
| " raise IndexError(\"Token length more than max seq length!\")\n", | |
| " segments = []\n", | |
| " first_sep = True\n", | |
| " current_segment_id = 0\n", | |
| " for token in tokens:\n", | |
| " segments.append(current_segment_id)\n", | |
| " if token == \"[SEP]\":\n", | |
| " if first_sep:\n", | |
| " first_sep = False \n", | |
| " else:\n", | |
| " current_segment_id = 1\n", | |
| " return segments + [0] * (max_seq_length - len(tokens))\n", | |
| "\n", | |
| "def _get_ids(tokens, tokenizer, max_seq_length):\n", | |
| " \"\"\"Token ids from Tokenizer vocab\"\"\"\n", | |
| " token_ids = tokenizer.convert_tokens_to_ids(tokens)\n", | |
| " input_ids = token_ids + [0] * (max_seq_length-len(token_ids))\n", | |
| " return input_ids\n", | |
| "\n", | |
| "def _trim_input(title, question, answer, max_sequence_length, \n", | |
| " t_max_len=30, q_max_len=239, a_max_len=239):\n", | |
| "\n", | |
| " t = tokenizer.tokenize(title)\n", | |
| " q = tokenizer.tokenize(question)\n", | |
| " a = tokenizer.tokenize(answer)\n", | |
| " \n", | |
| " t_len = len(t)\n", | |
| " q_len = len(q)\n", | |
| " a_len = len(a)\n", | |
| "\n", | |
| " if (t_len+q_len+a_len+4) > max_sequence_length:\n", | |
| " \n", | |
| " if t_max_len > t_len:\n", | |
| " t_new_len = t_len\n", | |
| " a_max_len = a_max_len + floor((t_max_len - t_len)/2)\n", | |
| " q_max_len = q_max_len + ceil((t_max_len - t_len)/2)\n", | |
| " else:\n", | |
| " t_new_len = t_max_len\n", | |
| " \n", | |
| " if a_max_len > a_len:\n", | |
| " a_new_len = a_len \n", | |
| " q_new_len = q_max_len + (a_max_len - a_len)\n", | |
| " elif q_max_len > q_len:\n", | |
| " a_new_len = a_max_len + (q_max_len - q_len)\n", | |
| " q_new_len = q_len\n", | |
| " else:\n", | |
| " a_new_len = a_max_len\n", | |
| " q_new_len = q_max_len\n", | |
| " \n", | |
| " \n", | |
| " if t_new_len+a_new_len+q_new_len+4 != max_sequence_length:\n", | |
| " raise ValueError(\"New sequence length should be %d, but is %d\" \n", | |
| " % (max_sequence_length, (t_new_len+a_new_len+q_new_len+4)))\n", | |
| " \n", | |
| " t = t[:t_new_len]\n", | |
| " q = q[:q_new_len]\n", | |
| " a = a[:a_new_len]\n", | |
| " \n", | |
| " return t, q, a\n", | |
| "\n", | |
| "def _convert_to_bert_inputs(title, question, answer, tokenizer, max_sequence_length):\n", | |
| " \"\"\"Converts tokenized input to ids, masks and segments for BERT\"\"\"\n", | |
| " \n", | |
| " stoken = [\"[CLS]\"] + title + [\"[SEP]\"] + question + [\"[SEP]\"] + answer + [\"[SEP]\"]\n", | |
| "\n", | |
| " input_ids = _get_ids(stoken, tokenizer, max_sequence_length)\n", | |
| " input_masks = _get_masks(stoken, max_sequence_length)\n", | |
| " input_segments = _get_segments(stoken, max_sequence_length)\n", | |
| "\n", | |
| " return [input_ids, input_masks, input_segments]\n", | |
| "\n", | |
| "def compute_input_arays(df, columns, tokenizer, max_sequence_length):\n", | |
| " input_ids, input_masks, input_segments = [], [], []\n", | |
| " for _, instance in tqdm(df[columns].iterrows()):\n", | |
| " t, q, a = instance.question_title, instance.question_body, instance.answer\n", | |
| "\n", | |
| " t, q, a = _trim_input(t, q, a, max_sequence_length)\n", | |
| "\n", | |
| " ids, masks, segments = _convert_to_bert_inputs(t, q, a, tokenizer, max_sequence_length)\n", | |
| " input_ids.append(ids)\n", | |
| " input_masks.append(masks)\n", | |
| " input_segments.append(segments)\n", | |
| " \n", | |
| " return [np.asarray(input_ids, dtype=np.int32), \n", | |
| " np.asarray(input_masks, dtype=np.int32), \n", | |
| " np.asarray(input_segments, dtype=np.int32)]\n", | |
| "\n", | |
| "\n", | |
| "def compute_output_arrays(df, columns):\n", | |
| " return np.asarray(df[columns])" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 7, | |
| "metadata": {}, | |
| "outputs": [], | |
| "source": [ | |
| "def compute_spearmanr(trues, preds):\n", | |
| " rhos = []\n", | |
| " for col_trues, col_pred in zip(trues.T, preds.T):\n", | |
| " rhos.append(\n", | |
| " spearmanr(col_trues, col_pred + np.random.normal(0, 1e-7, col_pred.shape[0])).correlation)\n", | |
| " return np.mean(rhos)\n", | |
| "\n", | |
| "\n", | |
| "class CustomCallback(tf.keras.callbacks.Callback):\n", | |
| " \n", | |
| " def __init__(self, valid_data, test_data, batch_size=16, fold=None):\n", | |
| "\n", | |
| " self.valid_inputs = valid_data[0]\n", | |
| " self.valid_outputs = valid_data[1]\n", | |
| " self.test_inputs = test_data\n", | |
| " \n", | |
| " self.batch_size = batch_size\n", | |
| " self.fold = fold\n", | |
| " \n", | |
| " def on_train_begin(self, logs={}):\n", | |
| " self.valid_predictions = []\n", | |
| " self.test_predictions = []\n", | |
| " \n", | |
| " def on_epoch_end(self, epoch, logs={}):\n", | |
| " self.valid_predictions.append(\n", | |
| " self.model.predict(self.valid_inputs, batch_size=self.batch_size))\n", | |
| " \n", | |
| " rho_val = compute_spearmanr(\n", | |
| " self.valid_outputs, np.average(self.valid_predictions, axis=0))\n", | |
| " \n", | |
| " print(\"\\nvalidation rho: %.4f\" % rho_val)\n", | |
| " \n", | |
| " if self.fold is not None:\n", | |
| " self.model.save_weights(f'bert-base-{fold}-{epoch}.h5py')\n", | |
| " \n", | |
| " self.test_predictions.append(\n", | |
| " self.model.predict(self.test_inputs, batch_size=self.batch_size)\n", | |
| " )\n", | |
| "\n", | |
| "def bert_model():\n", | |
| " \n", | |
| " input_word_ids = tf.keras.layers.Input(\n", | |
| " (MAX_SEQUENCE_LENGTH,), dtype=tf.int32, name='input_word_ids')\n", | |
| " input_masks = tf.keras.layers.Input(\n", | |
| " (MAX_SEQUENCE_LENGTH,), dtype=tf.int32, name='input_masks')\n", | |
| " input_segments = tf.keras.layers.Input(\n", | |
| " (MAX_SEQUENCE_LENGTH,), dtype=tf.int32, name='input_segments')\n", | |
| " \n", | |
| " bert_layer = hub.KerasLayer(BERT_PATH, trainable=True)\n", | |
| " \n", | |
| " _, x = bert_layer([input_word_ids, input_masks, input_segments])\n", | |
| " \n", | |
| " #x = tf.keras.layers.GlobalAveragePooling1D()(sequence_output)\n", | |
| " x = tf.keras.layers.Lambda(lambda x: x[:, 0])(x)\n", | |
| " x = tf.keras.layers.Dropout(0.2)(x)\n", | |
| " out = tf.keras.layers.Dense(30, activation=\"sigmoid\", name=\"dense_output\")(x)\n", | |
| "\n", | |
| " model = tf.keras.models.Model(\n", | |
| " inputs=[input_word_ids, input_masks, input_segments], outputs=out)\n", | |
| " \n", | |
| " return model \n", | |
| " \n", | |
| "def train_and_predict(model, train_data, valid_data, test_data, \n", | |
| " learning_rate, epochs, batch_size, loss_function, fold):\n", | |
| " \n", | |
| " custom_callback = CustomCallback(\n", | |
| " valid_data=(valid_data[0], valid_data[1]), \n", | |
| " test_data=test_data,\n", | |
| " batch_size=batch_size,\n", | |
| " fold=None)\n", | |
| "\n", | |
| " optimizer = tf.keras.optimizers.Adam(learning_rate=learning_rate)\n", | |
| " model.compile(loss=loss_function, optimizer=optimizer)\n", | |
| " model.fit(train_data[0], train_data[1], epochs=epochs, \n", | |
| " batch_size=batch_size, callbacks=[custom_callback])\n", | |
| " \n", | |
| " return custom_callback\n" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 8, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "data": { | |
| "application/vnd.jupyter.widget-view+json": { | |
| "model_id": "59527e1231e1400795cea7f4c477ab21", | |
| "version_major": 2, | |
| "version_minor": 0 | |
| }, | |
| "text/plain": [ | |
| "HBox(children=(IntProgress(value=1, bar_style='info', max=1), HTML(value='')))" | |
| ] | |
| }, | |
| "metadata": {}, | |
| "output_type": "display_data" | |
| }, | |
| { | |
| "name": "stdout", | |
| "output_type": "stream", | |
| "text": [ | |
| "\n" | |
| ] | |
| }, | |
| { | |
| "data": { | |
| "application/vnd.jupyter.widget-view+json": { | |
| "model_id": "9e5ed41366914a33a6b1fcbd992e2df0", | |
| "version_major": 2, | |
| "version_minor": 0 | |
| }, | |
| "text/plain": [ | |
| "HBox(children=(IntProgress(value=1, bar_style='info', max=1), HTML(value='')))" | |
| ] | |
| }, | |
| "metadata": {}, | |
| "output_type": "display_data" | |
| }, | |
| { | |
| "name": "stdout", | |
| "output_type": "stream", | |
| "text": [ | |
| "\n" | |
| ] | |
| } | |
| ], | |
| "source": [ | |
| "gkf = GroupKFold(n_splits=10).split(X=df_train.question_body, groups=df_train.question_body) ############## originaln_splits=5\n", | |
| "\n", | |
| "outputs = compute_output_arrays(df_train, output_categories)\n", | |
| "inputs = compute_input_arays(df_train, input_categories, tokenizer, MAX_SEQUENCE_LENGTH)\n", | |
| "test_inputs = compute_input_arays(df_test, input_categories, tokenizer, MAX_SEQUENCE_LENGTH)" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 9, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "name": "stdout", | |
| "output_type": "stream", | |
| "text": [ | |
| "inputの形状は？\n", | |
| "<class 'list'>\n" | |
| ] | |
| }, | |
| { | |
| "data": { | |
| "text/plain": [ | |
| "3" | |
| ] | |
| }, | |
| "execution_count": 9, | |
| "metadata": {}, | |
| "output_type": "execute_result" | |
| } | |
| ], | |
| "source": [ | |
| "print(\"inputの形状は？\")\n", | |
| "print(type(inputs))\n", | |
| "len(inputs)" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 10, | |
| "metadata": {}, | |
| "outputs": [ | |
| { | |
| "name": "stdout", | |
| "output_type": "stream", | |
| "text": [ | |
| "1回目のループ\n", | |
| "[ 0 3 4 ... 6076 6077 6078]\n", | |
| "train_inputsの形状\n", | |
| "3\n", | |
| "5471\n", | |
| "Model: \"model\"\n", | |
| "__________________________________________________________________________________________________\n", | |
| "Layer (type) Output Shape Param # Connected to \n", | |
| "==================================================================================================\n", | |
| "input_word_ids (InputLayer) [(None, 512)] 0 \n", | |
| "__________________________________________________________________________________________________\n", | |
| "input_masks (InputLayer) [(None, 512)] 0 \n", | |
| "__________________________________________________________________________________________________\n", | |
| "input_segments (InputLayer) [(None, 512)] 0 \n", | |
| "__________________________________________________________________________________________________\n", | |
| "keras_layer (KerasLayer) [(None, 768), (None, 109482241 input_word_ids[0][0] \n", | |
| " input_masks[0][0] \n", | |
| " input_segments[0][0] \n", | |
| "__________________________________________________________________________________________________\n", | |
| "lambda (Lambda) (None, 768) 0 keras_layer[0][1] \n", | |
| "__________________________________________________________________________________________________\n", | |
| "dropout (Dropout) (None, 768) 0 lambda[0][0] \n", | |
| "__________________________________________________________________________________________________\n", | |
| "dense_output (Dense) (None, 30) 23070 dropout[0][0] \n", | |
| "==================================================================================================\n", | |
| "Total params: 109,505,311\n", | |
| "Trainable params: 109,505,310\n", | |
| "Non-trainable params: 1\n", | |
| "__________________________________________________________________________________________________\n", | |
| "Train on 5471 samples\n", | |
| "Epoch 1/5\n", | |
| "5464/5471 [============================>.] - ETA: 0s - loss: 0.3947" | |
| ] | |
| }, | |
| { | |
| "name": "stderr", | |
| "output_type": "stream", | |
| "text": [ | |
| "/opt/conda/lib/python3.6/site-packages/numpy/lib/function_base.py:2530: RuntimeWarning: invalid value encountered in true_divide\n", | |
| " c /= stddev[:, None]\n", | |
| "/opt/conda/lib/python3.6/site-packages/numpy/lib/function_base.py:2531: RuntimeWarning: invalid value encountered in true_divide\n", | |
| " c /= stddev[None, :]\n", | |
| "/opt/conda/lib/python3.6/site-packages/scipy/stats/_distn_infrastructure.py:877: RuntimeWarning: invalid value encountered in greater\n", | |
| " return (self.a < x) & (x < self.b)\n", | |
| "/opt/conda/lib/python3.6/site-packages/scipy/stats/_distn_infrastructure.py:877: RuntimeWarning: invalid value encountered in less\n", | |
| " return (self.a < x) & (x < self.b)\n", | |
| "/opt/conda/lib/python3.6/site-packages/scipy/stats/_distn_infrastructure.py:1831: RuntimeWarning: invalid value encountered in less_equal\n", | |
| " cond2 = cond0 & (x <= self.a)\n" | |
| ] | |
| }, | |
| { | |
| "name": "stdout", | |
| "output_type": "stream", | |
| "text": [ | |
| "\n", | |
| "validation rho: nan\n", | |
| "5471/5471 [==============================] - 412s 75ms/sample - loss: 0.3947\n", | |
| "Epoch 2/5\n", | |
| "5464/5471 [============================>.] - ETA: 0s - loss: 0.3638\n", | |
| "validation rho: nan\n", | |
| "5471/5471 [==============================] - 379s 69ms/sample - loss: 0.3638\n", | |
| "Epoch 3/5\n", | |
| "5464/5471 [============================>.] - ETA: 0s - loss: 0.3426\n", | |
| "validation rho: nan\n", | |
| "5471/5471 [==============================] - 379s 69ms/sample - loss: 0.3426\n", | |
| "Epoch 4/5\n", | |
| "5464/5471 [============================>.] - ETA: 0s - loss: 0.3175\n", | |
| "validation rho: nan\n", | |
| "5471/5471 [==============================] - 380s 69ms/sample - loss: 0.3175\n", | |
| "Epoch 5/5\n", | |
| "5464/5471 [============================>.] - ETA: 0s - loss: 0.2930\n", | |
| "validation rho: nan\n", | |
| "5471/5471 [==============================] - 379s 69ms/sample - loss: 0.2930\n", | |
| "2回目のループ\n", | |
| "[ 1 2 3 ... 6076 6077 6078]\n", | |
| "train_inputsの形状\n", | |
| "3\n", | |
| "5471\n", | |
| "Model: \"model\"\n", | |
| "__________________________________________________________________________________________________\n", | |
| "Layer (type) Output Shape Param # Connected to \n", | |
| "==================================================================================================\n", | |
| "input_word_ids (InputLayer) [(None, 512)] 0 \n", | |
| "__________________________________________________________________________________________________\n", | |
| "input_masks (InputLayer) [(None, 512)] 0 \n", | |
| "__________________________________________________________________________________________________\n", | |
| "input_segments (InputLayer) [(None, 512)] 0 \n", | |
| "__________________________________________________________________________________________________\n", | |
| "keras_layer (KerasLayer) [(None, 768), (None, 109482241 input_word_ids[0][0] \n", | |
| " input_masks[0][0] \n", | |
| " input_segments[0][0] \n", | |
| "__________________________________________________________________________________________________\n", | |
| "lambda (Lambda) (None, 768) 0 keras_layer[0][1] \n", | |
| "__________________________________________________________________________________________________\n", | |
| "dropout (Dropout) (None, 768) 0 lambda[0][0] \n", | |
| "__________________________________________________________________________________________________\n", | |
| "dense_output (Dense) (None, 30) 23070 dropout[0][0] \n", | |
| "==================================================================================================\n", | |
| "Total params: 109,505,311\n", | |
| "Trainable params: 109,505,310\n", | |
| "Non-trainable params: 1\n", | |
| "__________________________________________________________________________________________________\n", | |
| "Train on 5471 samples\n", | |
| "Epoch 1/5\n", | |
| "5464/5471 [============================>.] - ETA: 0s - loss: 0.3919\n", | |
| "validation rho: 0.3691\n", | |
| "5471/5471 [==============================] - 413s 75ms/sample - loss: 0.3920\n", | |
| "Epoch 2/5\n", | |
| "5464/5471 [============================>.] - ETA: 0s - loss: 0.3622\n", | |
| "validation rho: 0.3899\n", | |
| "5471/5471 [==============================] - 380s 69ms/sample - loss: 0.3622\n", | |
| "Epoch 3/5\n", | |
| "5464/5471 [============================>.] - ETA: 0s - loss: 0.3414\n", | |
| "validation rho: 0.3952\n", | |
| "5471/5471 [==============================] - 380s 69ms/sample - loss: 0.3413\n", | |
| "Epoch 4/5\n", | |
| "5464/5471 [============================>.] - ETA: 0s - loss: 0.3160\n", | |
| "validation rho: 0.3957\n", | |
| "5471/5471 [==============================] - 380s 69ms/sample - loss: 0.3160\n", | |
| "Epoch 5/5\n", | |
| "5464/5471 [============================>.] - ETA: 0s - loss: 0.2928\n", | |
| "validation rho: 0.3970\n", | |
| "5471/5471 [==============================] - 380s 69ms/sample - loss: 0.2928\n", | |
| "3回目のループ\n", | |
| "[ 0 1 2 ... 6076 6077 6078]\n", | |
| "train_inputsの形状\n", | |
| "3\n", | |
| "5471\n", | |
| "Model: \"model\"\n", | |
| "__________________________________________________________________________________________________\n", | |
| "Layer (type) Output Shape Param # Connected to \n", | |
| "==================================================================================================\n", | |
| "input_word_ids (InputLayer) [(None, 512)] 0 \n", | |
| "__________________________________________________________________________________________________\n", | |
| "input_masks (InputLayer) [(None, 512)] 0 \n", | |
| "__________________________________________________________________________________________________\n", | |
| "input_segments (InputLayer) [(None, 512)] 0 \n", | |
| "__________________________________________________________________________________________________\n", | |
| "keras_layer (KerasLayer) [(None, 768), (None, 109482241 input_word_ids[0][0] \n", | |
| " input_masks[0][0] \n", | |
| " input_segments[0][0] \n", | |
| "__________________________________________________________________________________________________\n", | |
| "lambda (Lambda) (None, 768) 0 keras_layer[0][1] \n", | |
| "__________________________________________________________________________________________________\n", | |
| "dropout (Dropout) (None, 768) 0 lambda[0][0] \n", | |
| "__________________________________________________________________________________________________\n", | |
| "dense_output (Dense) (None, 30) 23070 dropout[0][0] \n", | |
| "==================================================================================================\n", | |
| "Total params: 109,505,311\n", | |
| "Trainable params: 109,505,310\n", | |
| "Non-trainable params: 1\n", | |
| "__________________________________________________________________________________________________\n", | |
| "Train on 5471 samples\n", | |
| "Epoch 1/5\n", | |
| "5464/5471 [============================>.] - ETA: 0s - loss: 0.3921" | |
| ] | |
| }, | |
| { | |
| "name": "stderr", | |
| "output_type": "stream", | |
| "text": [ | |
| "/opt/conda/lib/python3.6/site-packages/numpy/lib/function_base.py:2530: RuntimeWarning: invalid value encountered in true_divide\n", | |
| " c /= stddev[:, None]\n", | |
| "/opt/conda/lib/python3.6/site-packages/numpy/lib/function_base.py:2531: RuntimeWarning: invalid value encountered in true_divide\n", | |
| " c /= stddev[None, :]\n", | |
| "/opt/conda/lib/python3.6/site-packages/scipy/stats/_distn_infrastructure.py:877: RuntimeWarning: invalid value encountered in greater\n", | |
| " return (self.a < x) & (x < self.b)\n", | |
| "/opt/conda/lib/python3.6/site-packages/scipy/stats/_distn_infrastructure.py:877: RuntimeWarning: invalid value encountered in less\n", | |
| " return (self.a < x) & (x < self.b)\n", | |
| "/opt/conda/lib/python3.6/site-packages/scipy/stats/_distn_infrastructure.py:1831: RuntimeWarning: invalid value encountered in less_equal\n", | |
| " cond2 = cond0 & (x <= self.a)\n" | |
| ] | |
| }, | |
| { | |
| "name": "stdout", | |
| "output_type": "stream", | |
| "text": [ | |
| "\n", | |
| "validation rho: nan\n", | |
| "5471/5471 [==============================] - 412s 75ms/sample - loss: 0.3921\n", | |
| "Epoch 2/5\n", | |
| "5464/5471 [============================>.] - ETA: 0s - loss: 0.3618\n", | |
| "validation rho: nan\n", | |
| "5471/5471 [==============================] - 380s 69ms/sample - loss: 0.3619\n", | |
| "Epoch 3/5\n", | |
| "5464/5471 [============================>.] - ETA: 0s - loss: 0.3412\n", | |
| "validation rho: nan\n", | |
| "5471/5471 [==============================] - 380s 69ms/sample - loss: 0.3412\n", | |
| "Epoch 4/5\n", | |
| "5464/5471 [============================>.] - ETA: 0s - loss: 0.3169\n", | |
| "validation rho: nan\n", | |
| "5471/5471 [==============================] - 380s 69ms/sample - loss: 0.3169\n", | |
| "Epoch 5/5\n", | |
| "5464/5471 [============================>.] - ETA: 0s - loss: 0.2938\n", | |
| "validation rho: nan\n", | |
| "5471/5471 [==============================] - 380s 69ms/sample - loss: 0.2939\n", | |
| "4回目のループ\n", | |
| "[ 0 1 2 ... 6076 6077 6078]\n", | |
| "4回目のループ\n", | |
| "[ 0 1 2 ... 6076 6077 6078]\n", | |
| "4回目のループ\n", | |
| "[ 0 1 2 ... 6076 6077 6078]\n", | |
| "4回目のループ\n", | |
| "[ 0 1 2 ... 6076 6077 6078]\n", | |
| "4回目のループ\n", | |
| "[ 0 1 2 ... 6074 6076 6078]\n", | |
| "4回目のループ\n", | |
| "[ 0 1 2 ... 6076 6077 6078]\n", | |
| "4回目のループ\n", | |
| "[ 0 1 2 ... 6074 6075 6077]\n" | |
| ] | |
| } | |
| ], | |
| "source": [ | |
| "histories = []\n", | |
| "count = 1\n", | |
| "for fold, (train_idx, valid_idx) in enumerate(gkf):\n", | |
| " print(str(count) + \"回目のループ\")\n", | |
| " print(train_idx)\n", | |
| " \n", | |
| " # will actually only do 3 folds (out of 5) to manage < 2h\n", | |
| " if fold < 3:\n", | |
| " K.clear_session()\n", | |
| " model = bert_model()\n", | |
| "\n", | |
| "\n", | |
| " \n", | |
| " train_inputs = [inputs[i][train_idx] for i in range(3)]\n", | |
| " train_outputs = outputs[train_idx]\n", | |
| "\n", | |
| " valid_inputs = [inputs[i][valid_idx] for i in range(3)]\n", | |
| " valid_outputs = outputs[valid_idx]\n", | |
| "\n", | |
| " \n", | |
| " \n", | |
| " print(\"train_inputsの形状\")\n", | |
| " print(len(train_inputs)) #一周目の答えは\n", | |
| " print(len(train_inputs[0])) #\n", | |
| " \n", | |
| " \n", | |
| " model.summary()\n", | |
| " name = \"model_\" + str(count) + \".png\"\n", | |
| " tf.keras.utils.plot_model(model, to_file=name)\n", | |
| " count +=1\n", | |
| " \n", | |
| " \n", | |
| " # history contains two lists of valid and test preds respectively:\n", | |
| " # [valid_predictions_{fold}, test_predictions_{fold}]\n", | |
| " # Largeでbatch_sizeを8だとメモリ不足になるため6に調整 epochsを5から３にした→元の戻した\n", | |
| " history = train_and_predict(model, \n", | |
| " train_data=(train_inputs, train_outputs), \n", | |
| " valid_data=(valid_inputs, valid_outputs),\n", | |
| " test_data=test_inputs, \n", | |
| " #learning_rateから「1e-5」→「 3e-5」\n", | |
| " learning_rate=3e-5, epochs=5, batch_size=8,\n", | |
| " loss_function='binary_crossentropy', fold=fold)\n", | |
| " \n", | |
| " \n", | |
| "\n", | |
| " histories.append(history)" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 11, | |
| "metadata": {}, | |
| "outputs": [], | |
| "source": [ | |
| "from sklearn.preprocessing import MinMaxScaler\n", | |
| "\n", | |
| "\n", | |
| "\n", | |
| "def postprocessing(oof_df):\n", | |
| " \n", | |
| " scaler = MinMaxScaler()\n", | |
| "\n", | |
| " type_one_column_list = [\n", | |
| " 'question_conversational', \\\n", | |
| " 'question_has_commonly_accepted_answer', \\\n", | |
| " 'question_not_really_a_question', \\\n", | |
| " 'question_type_choice', \\\n", | |
| " 'question_type_compare', \\\n", | |
| " 'question_type_consequence', \\\n", | |
| " 'question_type_definition', \\\n", | |
| " 'question_type_entity', \\\n", | |
| " 'question_type_instructions', \n", | |
| " ]\n", | |
| " \n", | |
| " oof_df[type_one_column_list] = scaler.fit_transform(oof_df[type_one_column_list])\n", | |
| " \n", | |
| " tmp = oof_df.copy(deep=True)\n", | |
| " \n", | |
| " for column in type_one_column_list:\n", | |
| " \n", | |
| " oof_df.loc[tmp[column] <= 0.16667, column] = 0\n", | |
| " oof_df.loc[(tmp[column] > 0.16667) & (tmp[column] <= 0.41667), column] = 0.333333\n", | |
| " oof_df.loc[(tmp[column] > 0.41667) & (tmp[column] <= 0.58333), column] = 0.500000\n", | |
| " oof_df.loc[(tmp[column] > 0.58333) & (tmp[column] <= 0.73333), column] = 0.666667\n", | |
| " oof_df.loc[(tmp[column] > 0.73333), column] = 1\n", | |
| " \n", | |
| " \n", | |
| " \n", | |
| " \n", | |
| " \n", | |
| " ################################################# handle type 3 columns \n", | |
| " type_three_column_list = [\n", | |
| " 'question_interestingness_self', \n", | |
| " ]\n", | |
| " scaler = MinMaxScaler(feature_range=(0, 1))\n", | |
| " oof_df[type_three_column_list] = scaler.fit_transform(oof_df[type_three_column_list])\n", | |
| " tmp[type_three_column_list] = scaler.fit_transform(tmp[type_three_column_list])\n", | |
| " \n", | |
| " for column in type_three_column_list:\n", | |
| " oof_df.loc[tmp[column] <= 0.385, column] = 0.333333\n", | |
| " oof_df.loc[(tmp[column] > 0.385) & (tmp[column] <= 0.47), column] = 0.444444\n", | |
| " oof_df.loc[(tmp[column] > 0.47) & (tmp[column] <= 0.525), column] = 0.5\n", | |
| " oof_df.loc[(tmp[column] > 0.525) & (tmp[column] <= 0.605), column] = 0.555556\n", | |
| " oof_df.loc[(tmp[column] > 0.605) & (tmp[column] <= 0.715), column] = 0.666667\n", | |
| " oof_df.loc[(tmp[column] > 0.715) & (tmp[column] <= 0.8), column] = 0.833333\n", | |
| " oof_df.loc[(tmp[column] > 0.8) & (tmp[column] <= 0.94), column] = 0.888889\n", | |
| " oof_df.loc[(tmp[column] > 0.94), column] = 1\n", | |
| " \n", | |
| " \n", | |
| " \n", | |
| " type_four_column_list = [\n", | |
| " 'answer_satisfaction'\n", | |
| " ]\n", | |
| " scaler = MinMaxScaler(feature_range=(0.2, 1))\n", | |
| " oof_df[type_four_column_list] = scaler.fit_transform(oof_df[type_four_column_list])\n", | |
| " tmp[type_four_column_list] = scaler.fit_transform(tmp[type_four_column_list])\n", | |
| " \n", | |
| " for column in type_four_column_list:\n", | |
| " \n", | |
| " oof_df.loc[tmp[column] <= 0.233, column] = 0.200000\n", | |
| " oof_df.loc[(tmp[column] > 0.233) & (tmp[column] <= 0.283), column] = 0.266667\n", | |
| " oof_df.loc[(tmp[column] > 0.283) & (tmp[column] <= 0.315), column] = 0.300000\n", | |
| " oof_df.loc[(tmp[column] > 0.315) & (tmp[column] <= 0.365), column] = 0.333333\n", | |
| " oof_df.loc[(tmp[column] > 0.365) & (tmp[column] <= 0.433), column] = 0.400000\n", | |
| " oof_df.loc[(tmp[column] > 0.433) & (tmp[column] <= 0.483), column] = 0.466667\n", | |
| " oof_df.loc[(tmp[column] > 0.483) & (tmp[column] <= 0.517), column] = 0.500000\n", | |
| " oof_df.loc[(tmp[column] > 0.517) & (tmp[column] <= 0.567), column] = 0.533333\n", | |
| " oof_df.loc[(tmp[column] > 0.567) & (tmp[column] <= 0.633), column] = 0.600000\n", | |
| " oof_df.loc[(tmp[column] > 0.633) & (tmp[column] <= 0.683), column] = 0.666667\n", | |
| " oof_df.loc[(tmp[column] > 0.683) & (tmp[column] <= 0.715), column] = 0.700000\n", | |
| " oof_df.loc[(tmp[column] > 0.715) & (tmp[column] <= 0.767), column] = 0.733333\n", | |
| " oof_df.loc[(tmp[column] > 0.767) & (tmp[column] <= 0.833), column] = 0.800000\n", | |
| " oof_df.loc[(tmp[column] > 0.883) & (tmp[column] <= 0.915), column] = 0.900000\n", | |
| " oof_df.loc[(tmp[column] > 0.915) & (tmp[column] <= 0.967), column] = 0.933333\n", | |
| " oof_df.loc[(tmp[column] > 0.967), column] = 1\n", | |
| " \n", | |
| " \n", | |
| " oof_values = oof_df[TARGET_COLUMNS].values\n", | |
| " DEGREE = len(oof_df)//45*9\n", | |
| " oof_values = np.around(oof_values * DEGREE) / DEGREE \n", | |
| " oof_df[TARGET_COLUMNS] = oof_values\n", | |
| " \n", | |
| " return oof_df" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 12, | |
| "metadata": {}, | |
| "outputs": [], | |
| "source": [ | |
| "test_predictions = [histories[i].test_predictions for i in range(len(histories))]\n", | |
| "test_predictions = [np.average(test_predictions[i], axis=0) for i in range(len(test_predictions))]\n", | |
| "test_predictions = np.mean(test_predictions, axis=0)\n", | |
| "\n", | |
| "#post process \n", | |
| "\n", | |
| "df_sub.iloc[:, 1:] = test_predictions\n", | |
| "\n", | |
| "#提出用に加工\n", | |
| "df_sub = postprocessing(df_sub) \n", | |
| "\n", | |
| "df_sub.to_csv('submission.csv', index=False)" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": 13, | |
| "metadata": {}, | |
| "outputs": [], | |
| "source": [ | |
| "#df_sub = postprocessing(df_sub) " | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": {}, | |
| "outputs": [], | |
| "source": [] | |
| } | |
| ], | |
| "metadata": { | |
| "kernelspec": { | |
| "display_name": "Python 3", | |
| "language": "python", | |
| "name": "python3" | |
| }, | |
| "language_info": { | |
| "codemirror_mode": { | |
| "name": "ipython", | |
| "version": 3 | |
| }, | |
| "file_extension": ".py", | |
| "mimetype": "text/x-python", | |
| "name": "python", | |
| "nbconvert_exporter": "python", | |
| "pygments_lexer": "ipython3", | |
| "version": "3.6.8" | |
| }, | |
| "widgets": { | |
| "application/vnd.jupyter.widget-state+json": { | |
| "state": { | |
| "310411ea7eb0474284a449db736aa803": { | |
| "model_module": "@jupyter-widgets/controls", | |
| "model_module_version": "1.5.0", | |
| "model_name": "DescriptionStyleModel", | |
| "state": { | |
| "_model_module": "@jupyter-widgets/controls", | |
| "_model_module_version": "1.5.0", | |
| "_model_name": "DescriptionStyleModel", | |
| "_view_count": null, | |
| "_view_module": "@jupyter-widgets/base", | |
| "_view_module_version": "1.2.0", | |
| "_view_name": "StyleView", | |
| "description_width": "" | |
| } | |
| }, | |
| "425a789d044140ac9e5b492882e2a5f7": { | |
| "model_module": "@jupyter-widgets/base", | |
| "model_module_version": "1.2.0", | |
| "model_name": "LayoutModel", | |
| "state": { | |
| "_model_module": "@jupyter-widgets/base", | |
| "_model_module_version": "1.2.0", | |
| "_model_name": "LayoutModel", | |
| "_view_count": null, | |
| "_view_module": "@jupyter-widgets/base", | |
| "_view_module_version": "1.2.0", | |
| "_view_name": "LayoutView", | |
| "align_content": null, | |
| "align_items": null, | |
| "align_self": null, | |
| "border": null, | |
| "bottom": null, | |
| "display": null, | |
| "flex": null, | |
| "flex_flow": null, | |
| "grid_area": null, | |
| "grid_auto_columns": null, | |
| "grid_auto_flow": null, | |
| "grid_auto_rows": null, | |
| "grid_column": null, | |
| "grid_gap": null, | |
| "grid_row": null, | |
| "grid_template_areas": null, | |
| "grid_template_columns": null, | |
| "grid_template_rows": null, | |
| "height": null, | |
| "justify_content": null, | |
| "justify_items": null, | |
| "left": null, | |
| "margin": null, | |
| "max_height": null, | |
| "max_width": null, | |
| "min_height": null, | |
| "min_width": null, | |
| "object_fit": null, | |
| "object_position": null, | |
| "order": null, | |
| "overflow": null, | |
| "overflow_x": null, | |
| "overflow_y": null, | |
| "padding": null, | |
| "right": null, | |
| "top": null, | |
| "visibility": null, | |
| "width": null | |
| } | |
| }, | |
| "51573fdd88b646119b4224ea99aea29b": { | |
| "model_module": "@jupyter-widgets/base", | |
| "model_module_version": "1.2.0", | |
| "model_name": "LayoutModel", | |
| "state": { | |
| "_model_module": "@jupyter-widgets/base", | |
| "_model_module_version": "1.2.0", | |
| "_model_name": "LayoutModel", | |
| "_view_count": null, | |
| "_view_module": "@jupyter-widgets/base", | |
| "_view_module_version": "1.2.0", | |
| "_view_name": "LayoutView", | |
| "align_content": null, | |
| "align_items": null, | |
| "align_self": null, | |
| "border": null, | |
| "bottom": null, | |
| "display": null, | |
| "flex": null, | |
| "flex_flow": null, | |
| "grid_area": null, | |
| "grid_auto_columns": null, | |
| "grid_auto_flow": null, | |
| "grid_auto_rows": null, | |
| "grid_column": null, | |
| "grid_gap": null, | |
| "grid_row": null, | |
| "grid_template_areas": null, | |
| "grid_template_columns": null, | |
| "grid_template_rows": null, | |
| "height": null, | |
| "justify_content": null, | |
| "justify_items": null, | |
| "left": null, | |
| "margin": null, | |
| "max_height": null, | |
| "max_width": null, | |
| "min_height": null, | |
| "min_width": null, | |
| "object_fit": null, | |
| "object_position": null, | |
| "order": null, | |
| "overflow": null, | |
| "overflow_x": null, | |
| "overflow_y": null, | |
| "padding": null, | |
| "right": null, | |
| "top": null, | |
| "visibility": null, | |
| "width": null | |
| } | |
| }, | |
| "59527e1231e1400795cea7f4c477ab21": { | |
| "model_module": "@jupyter-widgets/controls", | |
| "model_module_version": "1.5.0", | |
| "model_name": "HBoxModel", | |
| "state": { | |
| "_dom_classes": [], | |
| "_model_module": "@jupyter-widgets/controls", | |
| "_model_module_version": "1.5.0", | |
| "_model_name": "HBoxModel", | |
| "_view_count": null, | |
| "_view_module": "@jupyter-widgets/controls", | |
| "_view_module_version": "1.5.0", | |
| "_view_name": "HBoxView", | |
| "box_style": "", | |
| "children": [ | |
| "IPY_MODEL_f5e02903db024c87b7793931efcd2770", | |
| "IPY_MODEL_f573014b23344654a6c3447294e5bead" | |
| ], | |
| "layout": "IPY_MODEL_e7b563eb2e584dca8bd046b327040b68" | |
| } | |
| }, | |
| "6ce4a4332bea45ccbf6b6c3df20cb4ac": { | |
| "model_module": "@jupyter-widgets/controls", | |
| "model_module_version": "1.5.0", | |
| "model_name": "ProgressStyleModel", | |
| "state": { | |
| "_model_module": "@jupyter-widgets/controls", | |
| "_model_module_version": "1.5.0", | |
| "_model_name": "ProgressStyleModel", | |
| "_view_count": null, | |
| "_view_module": "@jupyter-widgets/base", | |
| "_view_module_version": "1.2.0", | |
| "_view_name": "StyleView", | |
| "bar_color": null, | |
| "description_width": "initial" | |
| } | |
| }, | |
| "90cb37ef2c834e2aabc9ae268bd77cc2": { | |
| "model_module": "@jupyter-widgets/controls", | |
| "model_module_version": "1.5.0", | |
| "model_name": "HTMLModel", | |
| "state": { | |
| "_dom_classes": [], | |
| "_model_module": "@jupyter-widgets/controls", | |
| "_model_module_version": "1.5.0", | |
| "_model_name": "HTMLModel", | |
| "_view_count": null, | |
| "_view_module": "@jupyter-widgets/controls", | |
| "_view_module_version": "1.5.0", | |
| "_view_name": "HTMLView", | |
| "description": "", | |
| "description_tooltip": null, | |
| "layout": "IPY_MODEL_425a789d044140ac9e5b492882e2a5f7", | |
| "placeholder": "", | |
| "style": "IPY_MODEL_cc7781cb82db4bb294428f0483ebb176", | |
| "value": " 476/? [00:08<00:00, 57.12it/s]" | |
| } | |
| }, | |
| "9d6ad2a884fd43fb98121b15077d15ef": { | |
| "model_module": "@jupyter-widgets/base", | |
| "model_module_version": "1.2.0", | |
| "model_name": "LayoutModel", | |
| "state": { | |
| "_model_module": "@jupyter-widgets/base", | |
| "_model_module_version": "1.2.0", | |
| "_model_name": "LayoutModel", | |
| "_view_count": null, | |
| "_view_module": "@jupyter-widgets/base", | |
| "_view_module_version": "1.2.0", | |
| "_view_name": "LayoutView", | |
| "align_content": null, | |
| "align_items": null, | |
| "align_self": null, | |
| "border": null, | |
| "bottom": null, | |
| "display": null, | |
| "flex": null, | |
| "flex_flow": null, | |
| "grid_area": null, | |
| "grid_auto_columns": null, | |
| "grid_auto_flow": null, | |
| "grid_auto_rows": null, | |
| "grid_column": null, | |
| "grid_gap": null, | |
| "grid_row": null, | |
| "grid_template_areas": null, | |
| "grid_template_columns": null, | |
| "grid_template_rows": null, | |
| "height": null, | |
| "justify_content": null, | |
| "justify_items": null, | |
| "left": null, | |
| "margin": null, | |
| "max_height": null, | |
| "max_width": null, | |
| "min_height": null, | |
| "min_width": null, | |
| "object_fit": null, | |
| "object_position": null, | |
| "order": null, | |
| "overflow": null, | |
| "overflow_x": null, | |
| "overflow_y": null, | |
| "padding": null, | |
| "right": null, | |
| "top": null, | |
| "visibility": null, | |
| "width": null | |
| } | |
| }, | |
| "9e5ed41366914a33a6b1fcbd992e2df0": { | |
| "model_module": "@jupyter-widgets/controls", | |
| "model_module_version": "1.5.0", | |
| "model_name": "HBoxModel", | |
| "state": { | |
| "_dom_classes": [], | |
| "_model_module": "@jupyter-widgets/controls", | |
| "_model_module_version": "1.5.0", | |
| "_model_name": "HBoxModel", | |
| "_view_count": null, | |
| "_view_module": "@jupyter-widgets/controls", | |
| "_view_module_version": "1.5.0", | |
| "_view_name": "HBoxView", | |
| "box_style": "", | |
| "children": [ | |
| "IPY_MODEL_d4286c52a36e4821974598b32b5d87ab", | |
| "IPY_MODEL_90cb37ef2c834e2aabc9ae268bd77cc2" | |
| ], | |
| "layout": "IPY_MODEL_9d6ad2a884fd43fb98121b15077d15ef" | |
| } | |
| }, | |
| "afa0da908a964e8eabbf673a669275b9": { | |
| "model_module": "@jupyter-widgets/controls", | |
| "model_module_version": "1.5.0", | |
| "model_name": "ProgressStyleModel", | |
| "state": { | |
| "_model_module": "@jupyter-widgets/controls", | |
| "_model_module_version": "1.5.0", | |
| "_model_name": "ProgressStyleModel", | |
| "_view_count": null, | |
| "_view_module": "@jupyter-widgets/base", | |
| "_view_module_version": "1.2.0", | |
| "_view_name": "StyleView", | |
| "bar_color": null, | |
| "description_width": "initial" | |
| } | |
| }, | |
| "bba2fc23f48b42208ac796e85148a3e3": { | |
| "model_module": "@jupyter-widgets/base", | |
| "model_module_version": "1.2.0", | |
| "model_name": "LayoutModel", | |
| "state": { | |
| "_model_module": "@jupyter-widgets/base", | |
| "_model_module_version": "1.2.0", | |
| "_model_name": "LayoutModel", | |
| "_view_count": null, | |
| "_view_module": "@jupyter-widgets/base", | |
| "_view_module_version": "1.2.0", | |
| "_view_name": "LayoutView", | |
| "align_content": null, | |
| "align_items": null, | |
| "align_self": null, | |
| "border": null, | |
| "bottom": null, | |
| "display": null, | |
| "flex": null, | |
| "flex_flow": null, | |
| "grid_area": null, | |
| "grid_auto_columns": null, | |
| "grid_auto_flow": null, | |
| "grid_auto_rows": null, | |
| "grid_column": null, | |
| "grid_gap": null, | |
| "grid_row": null, | |
| "grid_template_areas": null, | |
| "grid_template_columns": null, | |
| "grid_template_rows": null, | |
| "height": null, | |
| "justify_content": null, | |
| "justify_items": null, | |
| "left": null, | |
| "margin": null, | |
| "max_height": null, | |
| "max_width": null, | |
| "min_height": null, | |
| "min_width": null, | |
| "object_fit": null, | |
| "object_position": null, | |
| "order": null, | |
| "overflow": null, | |
| "overflow_x": null, | |
| "overflow_y": null, | |
| "padding": null, | |
| "right": null, | |
| "top": null, | |
| "visibility": null, | |
| "width": null | |
| } | |
| }, | |
| "bf184a4683f14d21aa21f905350a4bd2": { | |
| "model_module": "@jupyter-widgets/base", | |
| "model_module_version": "1.2.0", | |
| "model_name": "LayoutModel", | |
| "state": { | |
| "_model_module": "@jupyter-widgets/base", | |
| "_model_module_version": "1.2.0", | |
| "_model_name": "LayoutModel", | |
| "_view_count": null, | |
| "_view_module": "@jupyter-widgets/base", | |
| "_view_module_version": "1.2.0", | |
| "_view_name": "LayoutView", | |
| "align_content": null, | |
| "align_items": null, | |
| "align_self": null, | |
| "border": null, | |
| "bottom": null, | |
| "display": null, | |
| "flex": null, | |
| "flex_flow": null, | |
| "grid_area": null, | |
| "grid_auto_columns": null, | |
| "grid_auto_flow": null, | |
| "grid_auto_rows": null, | |
| "grid_column": null, | |
| "grid_gap": null, | |
| "grid_row": null, | |
| "grid_template_areas": null, | |
| "grid_template_columns": null, | |
| "grid_template_rows": null, | |
| "height": null, | |
| "justify_content": null, | |
| "justify_items": null, | |
| "left": null, | |
| "margin": null, | |
| "max_height": null, | |
| "max_width": null, | |
| "min_height": null, | |
| "min_width": null, | |
| "object_fit": null, | |
| "object_position": null, | |
| "order": null, | |
| "overflow": null, | |
| "overflow_x": null, | |
| "overflow_y": null, | |
| "padding": null, | |
| "right": null, | |
| "top": null, | |
| "visibility": null, | |
| "width": null | |
| } | |
| }, | |
| "cc7781cb82db4bb294428f0483ebb176": { | |
| "model_module": "@jupyter-widgets/controls", | |
| "model_module_version": "1.5.0", | |
| "model_name": "DescriptionStyleModel", | |
| "state": { | |
| "_model_module": "@jupyter-widgets/controls", | |
| "_model_module_version": "1.5.0", | |
| "_model_name": "DescriptionStyleModel", | |
| "_view_count": null, | |
| "_view_module": "@jupyter-widgets/base", | |
| "_view_module_version": "1.2.0", | |
| "_view_name": "StyleView", | |
| "description_width": "" | |
| } | |
| }, | |
| "d4286c52a36e4821974598b32b5d87ab": { | |
| "model_module": "@jupyter-widgets/controls", | |
| "model_module_version": "1.5.0", | |
| "model_name": "IntProgressModel", | |
| "state": { | |
| "_dom_classes": [], | |
| "_model_module": "@jupyter-widgets/controls", | |
| "_model_module_version": "1.5.0", | |
| "_model_name": "IntProgressModel", | |
| "_view_count": null, | |
| "_view_module": "@jupyter-widgets/controls", | |
| "_view_module_version": "1.5.0", | |
| "_view_name": "ProgressView", | |
| "bar_style": "success", | |
| "description": "", | |
| "description_tooltip": null, | |
| "layout": "IPY_MODEL_bf184a4683f14d21aa21f905350a4bd2", | |
| "max": 1, | |
| "min": 0, | |
| "orientation": "horizontal", | |
| "style": "IPY_MODEL_6ce4a4332bea45ccbf6b6c3df20cb4ac", | |
| "value": 1 | |
| } | |
| }, | |
| "e7b563eb2e584dca8bd046b327040b68": { | |
| "model_module": "@jupyter-widgets/base", | |
| "model_module_version": "1.2.0", | |
| "model_name": "LayoutModel", | |
| "state": { | |
| "_model_module": "@jupyter-widgets/base", | |
| "_model_module_version": "1.2.0", | |
| "_model_name": "LayoutModel", | |
| "_view_count": null, | |
| "_view_module": "@jupyter-widgets/base", | |
| "_view_module_version": "1.2.0", | |
| "_view_name": "LayoutView", | |
| "align_content": null, | |
| "align_items": null, | |
| "align_self": null, | |
| "border": null, | |
| "bottom": null, | |
| "display": null, | |
| "flex": null, | |
| "flex_flow": null, | |
| "grid_area": null, | |
| "grid_auto_columns": null, | |
| "grid_auto_flow": null, | |
| "grid_auto_rows": null, | |
| "grid_column": null, | |
| "grid_gap": null, | |
| "grid_row": null, | |
| "grid_template_areas": null, | |
| "grid_template_columns": null, | |
| "grid_template_rows": null, | |
| "height": null, | |
| "justify_content": null, | |
| "justify_items": null, | |
| "left": null, | |
| "margin": null, | |
| "max_height": null, | |
| "max_width": null, | |
| "min_height": null, | |
| "min_width": null, | |
| "object_fit": null, | |
| "object_position": null, | |
| "order": null, | |
| "overflow": null, | |
| "overflow_x": null, | |
| "overflow_y": null, | |
| "padding": null, | |
| "right": null, | |
| "top": null, | |
| "visibility": null, | |
| "width": null | |
| } | |
| }, | |
| "f573014b23344654a6c3447294e5bead": { | |
| "model_module": "@jupyter-widgets/controls", | |
| "model_module_version": "1.5.0", | |
| "model_name": "HTMLModel", | |
| "state": { | |
| "_dom_classes": [], | |
| "_model_module": "@jupyter-widgets/controls", | |
| "_model_module_version": "1.5.0", | |
| "_model_name": "HTMLModel", | |
| "_view_count": null, | |
| "_view_module": "@jupyter-widgets/controls", | |
| "_view_module_version": "1.5.0", | |
| "_view_name": "HTMLView", | |
| "description": "", | |
| "description_tooltip": null, | |
| "layout": "IPY_MODEL_bba2fc23f48b42208ac796e85148a3e3", | |
| "placeholder": "", | |
| "style": "IPY_MODEL_310411ea7eb0474284a449db736aa803", | |
| "value": " 6079/? [01:15<00:00, 81.02it/s]" | |
| } | |
| }, | |
| "f5e02903db024c87b7793931efcd2770": { | |
| "model_module": "@jupyter-widgets/controls", | |
| "model_module_version": "1.5.0", | |
| "model_name": "IntProgressModel", | |
| "state": { | |
| "_dom_classes": [], | |
| "_model_module": "@jupyter-widgets/controls", | |
| "_model_module_version": "1.5.0", | |
| "_model_name": "IntProgressModel", | |
| "_view_count": null, | |
| "_view_module": "@jupyter-widgets/controls", | |
| "_view_module_version": "1.5.0", | |
| "_view_name": "ProgressView", | |
| "bar_style": "success", | |
| "description": "", | |
| "description_tooltip": null, | |
| "layout": "IPY_MODEL_51573fdd88b646119b4224ea99aea29b", | |
| "max": 1, | |
| "min": 0, | |
| "orientation": "horizontal", | |
| "style": "IPY_MODEL_afa0da908a964e8eabbf673a669275b9", | |
| "value": 1 | |
| } | |
| } | |
| }, | |
| "version_major": 2, | |
| "version_minor": 0 | |
| } | |
| } | |
| }, | |
| "nbformat": 4, | |
| "nbformat_minor": 1 | |
| } |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment