colbert-train-model-notebook.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:31:03.720589Z",
     "iopub.status.busy": "2021-03-31T10:31:03.719413Z",
     "iopub.status.idle": "2021-03-31T10:31:03.778542Z",
     "shell.execute_reply": "2021-03-31T10:31:03.777730Z"
    },
    "papermill": {
     "duration": 0.095686,
     "end_time": "2021-03-31T10:31:03.778661",
     "exception": false,
     "start_time": "2021-03-31T10:31:03.682975",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# CPU\n",
      "model name\t: Intel(R) Xeon(R) CPU @ 2.00GHz\n",
      "cpu MHz\t\t: 2000.190\n",
      "cpu cores\t: 1\n",
      "# RAM\n",
      "MemTotal:       16401976 kB\n",
      "# GPU\n",
      "\n",
      "# OS\n",
      "Linux 27d52c1934f3 5.4.89+ #1 SMP Wed Feb 24 19:44:28 PST 2021 x86_64 x86_64 x86_64 GNU/Linux\n"
     ]
    }
   ],
   "source": [
    "import subprocess\n",
    "from ast import literal_eval\n",
    "\n",
    "def run(command):\n",
    "    process = subprocess.Popen(command, shell=True, stdout=subprocess.PIPE)\n",
    "    out, err = process.communicate()\n",
    "    print(out.decode('utf-8').strip())\n",
    "\n",
    "print('# CPU')\n",
    "run('cat /proc/cpuinfo | egrep -m 1 \"^model name\"')\n",
    "run('cat /proc/cpuinfo | egrep -m 1 \"^cpu MHz\"')\n",
    "run('cat /proc/cpuinfo | egrep -m 1 \"^cpu cores\"')\n",
    "\n",
    "print('# RAM')\n",
    "run('cat /proc/meminfo | egrep \"^MemTotal\"')\n",
    "\n",
    "print('# GPU')\n",
    "run('lspci | grep VGA')\n",
    "\n",
    "print('# OS')\n",
    "run('uname -a')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:31:03.831763Z",
     "iopub.status.busy": "2021-03-31T10:31:03.830862Z",
     "iopub.status.idle": "2021-03-31T10:31:16.294501Z",
     "shell.execute_reply": "2021-03-31T10:31:16.294915Z"
    },
    "papermill": {
     "duration": 12.491786,
     "end_time": "2021-03-31T10:31:16.295071",
     "exception": false,
     "start_time": "2021-03-31T10:31:03.803285",
     "status": "completed"
    },
    "scrolled": true,
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.3.1\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import sklearn\n",
    "from sklearn.model_selection import GroupKFold\n",
    "import matplotlib.pyplot as plt\n",
    "from tqdm.notebook import tqdm\n",
    "\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",
    "from tensorflow import keras \n",
    "\n",
    "import os\n",
    "from scipy.stats import spearmanr\n",
    "from math import floor, ceil\n",
    "from transformers import *\n",
    "\n",
    "import seaborn as sns\n",
    "import string\n",
    "import re    #for regex\n",
    "\n",
    "np.set_printoptions(suppress=True)\n",
    "print(tf.__version__)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.025294,
     "end_time": "2021-03-31T10:31:16.348106",
     "exception": false,
     "start_time": "2021-03-31T10:31:16.322812",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Prep / tokenizer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.025805,
     "end_time": "2021-03-31T10:31:16.399608",
     "exception": false,
     "start_time": "2021-03-31T10:31:16.373803",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### 1. Read data and tokenizer\n",
    "\n",
    "Read tokenizer and data, as well as defining the maximum sequence length that will be used for the input to Bert (maximum is usually 512 tokens)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:31:16.458619Z",
     "iopub.status.busy": "2021-03-31T10:31:16.456352Z",
     "iopub.status.idle": "2021-03-31T10:31:16.459559Z",
     "shell.execute_reply": "2021-03-31T10:31:16.460218Z"
    },
    "papermill": {
     "duration": 0.035582,
     "end_time": "2021-03-31T10:31:16.460384",
     "exception": false,
     "start_time": "2021-03-31T10:31:16.424802",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "training_sample_count = 160000 # 4000\n",
    "training_epochs = 3 # 3\n",
    "test_count = 40000\n",
    "\n",
    "# if True:\n",
    "#     training_sample_count = 500 # 4000\n",
    "#     training_epochs = 1 # 3\n",
    "#     test_count = 300\n",
    "running_folds = 1 # 2\n",
    "\n",
    "MAX_SENTENCE_LENGTH = 20\n",
    "MAX_SENTENCES = 5\n",
    "MAX_LENGTH = 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:31:16.515488Z",
     "iopub.status.busy": "2021-03-31T10:31:16.514626Z",
     "iopub.status.idle": "2021-03-31T10:31:17.230262Z",
     "shell.execute_reply": "2021-03-31T10:31:17.230822Z"
    },
    "papermill": {
     "duration": 0.746318,
     "end_time": "2021-03-31T10:31:17.230976",
     "exception": false,
     "start_time": "2021-03-31T10:31:16.484658",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dataset.csv  dev.csv  test.csv\ttrain.csv\r\n"
     ]
    }
   ],
   "source": [
    "!ls /kaggle/input/200k-short-texts-for-humor-detection"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.024228,
     "end_time": "2021-03-31T10:31:17.280464",
     "exception": false,
     "start_time": "2021-03-31T10:31:17.256236",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "### original dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:31:17.338079Z",
     "iopub.status.busy": "2021-03-31T10:31:17.337456Z",
     "iopub.status.idle": "2021-03-31T10:31:18.604693Z",
     "shell.execute_reply": "2021-03-31T10:31:18.605146Z"
    },
    "papermill": {
     "duration": 1.300574,
     "end_time": "2021-03-31T10:31:18.605289",
     "exception": false,
     "start_time": "2021-03-31T10:31:17.304715",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>text</th>\n",
       "      <th>humor</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Joe biden rules out 2020 bid: 'guys, i'm not r...</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Watch: darvish gave hitter whiplash with slow ...</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>What do you call a turtle without its shell? d...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                text  humor\n",
       "0  Joe biden rules out 2020 bid: 'guys, i'm not r...  False\n",
       "1  Watch: darvish gave hitter whiplash with slow ...  False\n",
       "2  What do you call a turtle without its shell? d...   True"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>text</th>\n",
       "      <th>humor</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>What kind of cat should you take into the  des...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Remember when people used to have to be in sha...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Pizza is always good. - everyone we'll see abo...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                text  humor\n",
       "0  What kind of cat should you take into the  des...   True\n",
       "1  Remember when people used to have to be in sha...   True\n",
       "2  Pizza is always good. - everyone we'll see abo...   True"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df = pd.read_csv('/kaggle/input/200k-short-texts-for-humor-detection/dataset.csv')\n",
    "\n",
    "df_train = pd.read_csv('/kaggle/input/200k-short-texts-for-humor-detection/train.csv')\n",
    "display(df_train.head(3))\n",
    "df_train = df_train[:training_sample_count*running_folds]\n",
    "\n",
    "df_test = pd.read_csv('/kaggle/input/200k-short-texts-for-humor-detection/dev.csv')\n",
    "display(df_test.head(3))\n",
    "df_test = df_test[:test_count]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:31:18.664719Z",
     "iopub.status.busy": "2021-03-31T10:31:18.662533Z",
     "iopub.status.idle": "2021-03-31T10:31:18.681615Z",
     "shell.execute_reply": "2021-03-31T10:31:18.682138Z"
    },
    "papermill": {
     "duration": 0.051078,
     "end_time": "2021-03-31T10:31:18.682277",
     "exception": false,
     "start_time": "2021-03-31T10:31:18.631199",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "200000 160000 40000\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>text</th>\n",
       "      <th>humor</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Joe biden rules out 2020 bid: 'guys, i'm not r...</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Watch: darvish gave hitter whiplash with slow ...</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>What do you call a turtle without its shell? d...</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5 reasons the 2016 election feels so personal</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Pasco police shot mexican migrant from behind,...</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                text  humor\n",
       "0  Joe biden rules out 2020 bid: 'guys, i'm not r...  False\n",
       "1  Watch: darvish gave hitter whiplash with slow ...  False\n",
       "2  What do you call a turtle without its shell? d...   True\n",
       "3      5 reasons the 2016 election feels so personal  False\n",
       "4  Pasco police shot mexican migrant from behind,...  False"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>What kind of cat should you take into the  des...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Remember when people used to have to be in sha...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Pizza is always good. - everyone we'll see abo...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>What's 6 inches long hard, bent, and in my pan...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Black teen's response to violence in his commu...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                text\n",
       "0  What kind of cat should you take into the  des...\n",
       "1  Remember when people used to have to be in sha...\n",
       "2  Pizza is always good. - everyone we'll see abo...\n",
       "3  What's 6 inches long hard, bent, and in my pan...\n",
       "4  Black teen's response to violence in his commu..."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_df_y = df_test.copy()\n",
    "del df_test['humor']\n",
    "\n",
    "df_sub = test_df_y.copy()\n",
    "\n",
    "print(len(df),len(df_train),len(df_test))\n",
    "display(df_train.head())\n",
    "display(df_test.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:31:18.744918Z",
     "iopub.status.busy": "2021-03-31T10:31:18.744317Z",
     "iopub.status.idle": "2021-03-31T10:31:18.749361Z",
     "shell.execute_reply": "2021-03-31T10:31:18.748866Z"
    },
    "papermill": {
     "duration": 0.038217,
     "end_time": "2021-03-31T10:31:18.749486",
     "exception": false,
     "start_time": "2021-03-31T10:31:18.711269",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['text', 'humor']\n"
     ]
    }
   ],
   "source": [
    "print(list(df_train.columns))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:31:18.817215Z",
     "iopub.status.busy": "2021-03-31T10:31:18.816559Z",
     "iopub.status.idle": "2021-03-31T10:31:18.819815Z",
     "shell.execute_reply": "2021-03-31T10:31:18.820387Z"
    },
    "papermill": {
     "duration": 0.040662,
     "end_time": "2021-03-31T10:31:18.820529",
     "exception": false,
     "start_time": "2021-03-31T10:31:18.779867",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "input categories:\n",
      "\t ['text']\n",
      "\n",
      "output TARGET_COUNT:\n",
      "\t 1\n",
      "\n",
      "output categories:\n",
      "\t ['humor']\n"
     ]
    }
   ],
   "source": [
    "output_categories = list(df_train.columns[[1]])\n",
    "input_categories = list(df_train.columns[[0]])\n",
    "\n",
    "TARGET_COUNT = len(output_categories)\n",
    "\n",
    "print('\\ninput categories:\\n\\t', input_categories)\n",
    "print('\\noutput TARGET_COUNT:\\n\\t', TARGET_COUNT)\n",
    "print('\\noutput categories:\\n\\t', output_categories)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.027981,
     "end_time": "2021-03-31T10:31:18.876428",
     "exception": false,
     "start_time": "2021-03-31T10:31:18.848447",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## 2. Preprocessing functions\n",
    "\n",
    "These are some functions that will be used to preprocess the raw text data into useable Bert inputs.<br>\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:31:18.938765Z",
     "iopub.status.busy": "2021-03-31T10:31:18.937753Z",
     "iopub.status.idle": "2021-03-31T10:31:19.401003Z",
     "shell.execute_reply": "2021-03-31T10:31:19.401469Z"
    },
    "papermill": {
     "duration": 0.496825,
     "end_time": "2021-03-31T10:31:19.401616",
     "exception": false,
     "start_time": "2021-03-31T10:31:18.904791",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "968816a6d7ee4a09a8794a75730d8132",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=231508.0, style=ProgressStyle(descripti…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "from transformers import BertTokenizer\n",
    "\n",
    "MODEL_TYPE = 'bert-base-uncased'\n",
    "tokenizer = BertTokenizer.from_pretrained(MODEL_TYPE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:31:19.464460Z",
     "iopub.status.busy": "2021-03-31T10:31:19.463772Z",
     "iopub.status.idle": "2021-03-31T10:31:20.175675Z",
     "shell.execute_reply": "2021-03-31T10:31:20.176532Z"
    },
    "papermill": {
     "duration": 0.746308,
     "end_time": "2021-03-31T10:31:20.176681",
     "exception": false,
     "start_time": "2021-03-31T10:31:19.430373",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package punkt to /usr/share/nltk_data...\n",
      "[nltk_data]   Package punkt is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "import nltk\n",
    "nltk.download('punkt')\n",
    "from nltk.tokenize import sent_tokenize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:31:20.295377Z",
     "iopub.status.busy": "2021-03-31T10:31:20.286234Z",
     "iopub.status.idle": "2021-03-31T10:31:20.298700Z",
     "shell.execute_reply": "2021-03-31T10:31:20.298206Z"
    },
    "papermill": {
     "duration": 0.051038,
     "end_time": "2021-03-31T10:31:20.298808",
     "exception": false,
     "start_time": "2021-03-31T10:31:20.247770",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def return_id(str1, str2, truncation_strategy, length):\n",
    "\n",
    "    inputs = tokenizer.encode_plus(str1, str2,\n",
    "        add_special_tokens=True,\n",
    "        max_length=length,\n",
    "        truncation_strategy=truncation_strategy)\n",
    "\n",
    "    input_ids =  inputs[\"input_ids\"]\n",
    "    input_masks = [1] * len(input_ids)\n",
    "    input_segments = inputs[\"token_type_ids\"]\n",
    "    padding_length = length - len(input_ids)\n",
    "    padding_id = tokenizer.pad_token_id\n",
    "    input_ids = input_ids + ([padding_id] * padding_length)\n",
    "    input_masks = input_masks + ([0] * padding_length)\n",
    "    input_segments = input_segments + ([0] * padding_length)\n",
    "\n",
    "    return [input_ids, input_masks, input_segments]\n",
    "\n",
    "\n",
    "def compute_input_arrays(df, columns, tokenizer):\n",
    "    model_input = []\n",
    "    for xx in range((MAX_SENTENCES*3)+3):\n",
    "        model_input.append([])\n",
    "    \n",
    "    for _, row in tqdm(df[columns].iterrows()):\n",
    "        i = 0\n",
    "        \n",
    "        # sent\n",
    "        sentences = sent_tokenize(row.text)\n",
    "        for xx in range(MAX_SENTENCES):\n",
    "            s = sentences[xx] if xx<len(sentences) else ''\n",
    "            ids_q, masks_q, segments_q = return_id(s, None, 'longest_first', MAX_SENTENCE_LENGTH)\n",
    "            model_input[i].append(ids_q)\n",
    "            i+=1\n",
    "            model_input[i].append(masks_q)\n",
    "            i+=1\n",
    "            model_input[i].append(segments_q)\n",
    "            i+=1\n",
    "        \n",
    "        # full row\n",
    "        ids_q, masks_q, segments_q = return_id(row.text, None, 'longest_first', MAX_LENGTH)\n",
    "        model_input[i].append(ids_q)\n",
    "        i+=1\n",
    "        model_input[i].append(masks_q)\n",
    "        i+=1\n",
    "        model_input[i].append(segments_q)\n",
    "        \n",
    "    for xx in range((MAX_SENTENCES*3)+3):\n",
    "        model_input[xx] = np.asarray(model_input[xx], dtype=np.int32)\n",
    "        \n",
    "    print(model_input[0].shape)\n",
    "    return model_input\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:31:20.391728Z",
     "iopub.status.busy": "2021-03-31T10:31:20.390626Z",
     "iopub.status.idle": "2021-03-31T10:38:18.189374Z",
     "shell.execute_reply": "2021-03-31T10:38:18.188356Z"
    },
    "papermill": {
     "duration": 417.859548,
     "end_time": "2021-03-31T10:38:18.189495",
     "exception": false,
     "start_time": "2021-03-31T10:31:20.329947",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b22963686f534071b666c4e084fd4d9c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=1.0, bar_style='info', max=1.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Truncation was not explicitly activated but `max_length` is provided a specific value, please use `truncation=True` to explicitly truncate examples to max length. Defaulting to 'longest_first' truncation strategy. If you encode pairs of sequences (GLUE-style) with the tokenizer you can select this strategy more precisely by providing a specific strategy to `truncation`.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "(160000, 20)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d0124c1c804e466ea5553e43c77c0076",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=1.0, bar_style='info', max=1.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "(40000, 20)\n"
     ]
    }
   ],
   "source": [
    "inputs      = compute_input_arrays(df_train, input_categories, tokenizer)\n",
    "test_inputs = compute_input_arrays(df_test, input_categories, tokenizer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:38:18.278339Z",
     "iopub.status.busy": "2021-03-31T10:38:18.276206Z",
     "iopub.status.idle": "2021-03-31T10:38:18.281789Z",
     "shell.execute_reply": "2021-03-31T10:38:18.280119Z"
    },
    "papermill": {
     "duration": 0.056925,
     "end_time": "2021-03-31T10:38:18.281923",
     "exception": false,
     "start_time": "2021-03-31T10:38:18.224998",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18 160000 20\n",
      "Why do native americans hate it when it rains in april? because it brings mayflowers.\n",
      "['Why do native americans hate it when it rains in april?', 'because it brings mayflowers.']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(array([  101,  2339,  2079,  3128,  4841,  5223,  2009,  2043,  2009,\n",
       "        15811,  1999,  2258,  1029,   102,     0,     0,     0,     0,\n",
       "            0,     0], dtype=int32),\n",
       " array([  101,  2138,  2009,  7545,  2089, 14156,  2015,  1012,   102,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0], dtype=int32),\n",
       " array([101, 102,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n",
       "          0,   0,   0,   0,   0,   0,   0], dtype=int32),\n",
       " array([  101,  2339,  2079,  3128,  4841,  5223,  2009,  2043,  2009,\n",
       "        15811,  1999,  2258,  1029,  2138,  2009,  7545,  2089, 14156,\n",
       "         2015,  1012,   102,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0,     0,     0,     0,     0,     0,     0,     0,     0,\n",
       "            0], dtype=int32))"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(len(inputs), len(inputs[0]), len(inputs[0][0]))\n",
    "\n",
    "# check out input for 7th row\n",
    "xx = 7\n",
    "print(df_train.iloc[xx,0])\n",
    "print(sent_tokenize(df_train.iloc[xx,0]))\n",
    "inputs[0][xx], inputs[3][xx], inputs[6][xx], inputs[15][xx]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:38:18.365692Z",
     "iopub.status.busy": "2021-03-31T10:38:18.364846Z",
     "iopub.status.idle": "2021-03-31T10:38:18.370508Z",
     "shell.execute_reply": "2021-03-31T10:38:18.370013Z"
    },
    "papermill": {
     "duration": 0.052895,
     "end_time": "2021-03-31T10:38:18.370613",
     "exception": false,
     "start_time": "2021-03-31T10:38:18.317718",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[False],\n",
       "       [False],\n",
       "       [ True],\n",
       "       ...,\n",
       "       [ True],\n",
       "       [ True],\n",
       "       [False]])"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def compute_output_arrays(df, columns):\n",
    "    return np.asarray(df[columns])\n",
    "\n",
    "outputs = compute_output_arrays(df_train, output_categories)\n",
    "outputs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.033568,
     "end_time": "2021-03-31T10:38:18.439346",
     "exception": false,
     "start_time": "2021-03-31T10:38:18.405778",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## 3. Create model\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:38:18.514949Z",
     "iopub.status.busy": "2021-03-31T10:38:18.514387Z",
     "iopub.status.idle": "2021-03-31T10:38:44.041101Z",
     "shell.execute_reply": "2021-03-31T10:38:44.041839Z"
    },
    "papermill": {
     "duration": 25.569068,
     "end_time": "2021-03-31T10:38:44.042048",
     "exception": false,
     "start_time": "2021-03-31T10:38:18.472980",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "fdd50281a5d2432381d1b4727c0d05da",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=536063208.0, style=ProgressStyle(descri…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Some layers from the model checkpoint at bert-base-uncased were not used when initializing TFBertModel: ['mlm___cls', 'nsp___cls']\n",
      "- This IS expected if you are initializing TFBertModel from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n",
      "- This IS NOT expected if you are initializing TFBertModel from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n",
      "All the layers of TFBertModel were initialized from the model checkpoint at bert-base-uncased.\n",
      "If your task is similar to the task the model of the checkpoint was trained on, you can already use TFBertModel for predictions without further training.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "BertConfig {\n",
       "  \"_name_or_path\": \"bert-base-uncased\",\n",
       "  \"attention_probs_dropout_prob\": 0.1,\n",
       "  \"gradient_checkpointing\": false,\n",
       "  \"hidden_act\": \"gelu\",\n",
       "  \"hidden_dropout_prob\": 0.1,\n",
       "  \"hidden_size\": 768,\n",
       "  \"initializer_range\": 0.02,\n",
       "  \"intermediate_size\": 3072,\n",
       "  \"layer_norm_eps\": 1e-12,\n",
       "  \"max_position_embeddings\": 512,\n",
       "  \"model_type\": \"bert\",\n",
       "  \"num_attention_heads\": 12,\n",
       "  \"num_hidden_layers\": 12,\n",
       "  \"pad_token_id\": 0,\n",
       "  \"type_vocab_size\": 2,\n",
       "  \"vocab_size\": 30522\n",
       "}"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "config = BertConfig() # print(config) to see settings\n",
    "config.output_hidden_states = False # Set to True to obtain hidden states\n",
    "bert_model = TFBertModel.from_pretrained('bert-base-uncased', config=config)\n",
    "config"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:38:44.178692Z",
     "iopub.status.busy": "2021-03-31T10:38:44.177733Z",
     "iopub.status.idle": "2021-03-31T10:38:56.721390Z",
     "shell.execute_reply": "2021-03-31T10:38:56.720795Z"
    },
    "papermill": {
     "duration": 12.620544,
     "end_time": "2021-03-31T10:38:56.721522",
     "exception": false,
     "start_time": "2021-03-31T10:38:44.100978",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(None, 768)\n"
     ]
    }
   ],
   "source": [
    "def create_model():\n",
    "    # model structure\n",
    "    # takes q_ids [max=20*MAX_SENTENCES] and a_ids [max=200]\n",
    "    import gc\n",
    "    \n",
    "    model_inputs = []\n",
    "    f_inputs=[]\n",
    "    for i in range(MAX_SENTENCES):\n",
    "        # bert embeddings\n",
    "        q_id = tf.keras.layers.Input((MAX_SENTENCE_LENGTH,), dtype=tf.int32)\n",
    "        q_mask = tf.keras.layers.Input((MAX_SENTENCE_LENGTH,), dtype=tf.int32)\n",
    "        q_atn = tf.keras.layers.Input((MAX_SENTENCE_LENGTH,), dtype=tf.int32)\n",
    "        q_embedding = bert_model.bert(q_id, attention_mask=q_mask, token_type_ids=q_atn)[0]\n",
    "        q = tf.keras.layers.GlobalAveragePooling1D()(q_embedding)\n",
    "        \n",
    "        # internal model\n",
    "        hidden1 = keras.layers.Dense(32, activation=\"relu\")(q)\n",
    "        hidden2 = keras.layers.Dropout(0.3)(hidden1)\n",
    "        hidden3 = keras.layers.Dense(8, activation='relu')(hidden2)\n",
    "        \n",
    "        f_inputs.append(hidden3)\n",
    "        model_inputs.extend([q_id, q_mask, q_atn])\n",
    "        \n",
    "    # whole sentence\n",
    "    a_id = tf.keras.layers.Input((MAX_LENGTH,), dtype=tf.int32)\n",
    "    a_mask = tf.keras.layers.Input((MAX_LENGTH,), dtype=tf.int32)\n",
    "    a_atn = tf.keras.layers.Input((MAX_LENGTH,), dtype=tf.int32)\n",
    "    a_embedding = bert_model.bert(a_id, attention_mask=a_mask, token_type_ids=a_atn)[0]\n",
    "    a = tf.keras.layers.GlobalAveragePooling1D()(a_embedding)\n",
    "    print(a.shape)\n",
    "    # internal model\n",
    "    hidden1 = keras.layers.Dense(256, activation=\"relu\")(a)\n",
    "    hidden2 = keras.layers.Dropout(0.2)(hidden1)\n",
    "    hidden3 = keras.layers.Dense(64, activation='relu')(hidden2)\n",
    "\n",
    "    f_inputs.append(hidden3)\n",
    "    model_inputs.extend([a_id, a_mask, a_atn])\n",
    "    \n",
    "    # final classifier\n",
    "    concat_ = keras.layers.Concatenate()(f_inputs)\n",
    "    hiddenf1 = keras.layers.Dense(512, activation='relu')(concat_)\n",
    "    hiddenf2 = keras.layers.Dropout(0.2)(hiddenf1)\n",
    "    hiddenf3 = keras.layers.Dense(256, activation='relu')(hiddenf2)\n",
    "    \n",
    "    output = keras.layers.Dense(TARGET_COUNT, activation='sigmoid')(hiddenf3) # softmax\n",
    "    model = keras.Model(inputs=model_inputs, outputs=[output] )\n",
    "    \n",
    "    gc.collect()\n",
    "    return model\n",
    "\n",
    "model = create_model()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:38:56.817715Z",
     "iopub.status.busy": "2021-03-31T10:38:56.816412Z",
     "iopub.status.idle": "2021-03-31T10:38:56.841776Z",
     "shell.execute_reply": "2021-03-31T10:38:56.841244Z"
    },
    "papermill": {
     "duration": 0.08418,
     "end_time": "2021-03-31T10:38:56.841881",
     "exception": false,
     "start_time": "2021-03-31T10:38:56.757701",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"functional_1\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_1 (InputLayer)            [(None, 20)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_2 (InputLayer)            [(None, 20)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_3 (InputLayer)            [(None, 20)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_4 (InputLayer)            [(None, 20)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_5 (InputLayer)            [(None, 20)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_6 (InputLayer)            [(None, 20)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_7 (InputLayer)            [(None, 20)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_8 (InputLayer)            [(None, 20)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_9 (InputLayer)            [(None, 20)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_10 (InputLayer)           [(None, 20)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_11 (InputLayer)           [(None, 20)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_12 (InputLayer)           [(None, 20)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_13 (InputLayer)           [(None, 20)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_14 (InputLayer)           [(None, 20)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_15 (InputLayer)           [(None, 20)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_16 (InputLayer)           [(None, 100)]        0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_17 (InputLayer)           [(None, 100)]        0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_18 (InputLayer)           [(None, 100)]        0                                            \n",
      "__________________________________________________________________________________________________\n",
      "bert (TFBertMainLayer)          multiple             109482240   input_1[0][0]                    \n",
      "                                                                 input_2[0][0]                    \n",
      "                                                                 input_3[0][0]                    \n",
      "                                                                 input_4[0][0]                    \n",
      "                                                                 input_5[0][0]                    \n",
      "                                                                 input_6[0][0]                    \n",
      "                                                                 input_7[0][0]                    \n",
      "                                                                 input_8[0][0]                    \n",
      "                                                                 input_9[0][0]                    \n",
      "                                                                 input_10[0][0]                   \n",
      "                                                                 input_11[0][0]                   \n",
      "                                                                 input_12[0][0]                   \n",
      "                                                                 input_13[0][0]                   \n",
      "                                                                 input_14[0][0]                   \n",
      "                                                                 input_15[0][0]                   \n",
      "                                                                 input_16[0][0]                   \n",
      "                                                                 input_17[0][0]                   \n",
      "                                                                 input_18[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d (Globa (None, 768)          0           bert[0][0]                       \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d_1 (Glo (None, 768)          0           bert[1][0]                       \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d_2 (Glo (None, 768)          0           bert[2][0]                       \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d_3 (Glo (None, 768)          0           bert[3][0]                       \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d_4 (Glo (None, 768)          0           bert[4][0]                       \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling1d_5 (Glo (None, 768)          0           bert[5][0]                       \n",
      "__________________________________________________________________________________________________\n",
      "dense (Dense)                   (None, 32)           24608       global_average_pooling1d[0][0]   \n",
      "__________________________________________________________________________________________________\n",
      "dense_2 (Dense)                 (None, 32)           24608       global_average_pooling1d_1[0][0] \n",
      "__________________________________________________________________________________________________\n",
      "dense_4 (Dense)                 (None, 32)           24608       global_average_pooling1d_2[0][0] \n",
      "__________________________________________________________________________________________________\n",
      "dense_6 (Dense)                 (None, 32)           24608       global_average_pooling1d_3[0][0] \n",
      "__________________________________________________________________________________________________\n",
      "dense_8 (Dense)                 (None, 32)           24608       global_average_pooling1d_4[0][0] \n",
      "__________________________________________________________________________________________________\n",
      "dense_10 (Dense)                (None, 256)          196864      global_average_pooling1d_5[0][0] \n",
      "__________________________________________________________________________________________________\n",
      "dropout_37 (Dropout)            (None, 32)           0           dense[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "dropout_38 (Dropout)            (None, 32)           0           dense_2[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "dropout_39 (Dropout)            (None, 32)           0           dense_4[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "dropout_40 (Dropout)            (None, 32)           0           dense_6[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "dropout_41 (Dropout)            (None, 32)           0           dense_8[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "dropout_42 (Dropout)            (None, 256)          0           dense_10[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense_1 (Dense)                 (None, 8)            264         dropout_37[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "dense_3 (Dense)                 (None, 8)            264         dropout_38[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "dense_5 (Dense)                 (None, 8)            264         dropout_39[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "dense_7 (Dense)                 (None, 8)            264         dropout_40[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "dense_9 (Dense)                 (None, 8)            264         dropout_41[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "dense_11 (Dense)                (None, 64)           16448       dropout_42[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "concatenate (Concatenate)       (None, 104)          0           dense_1[0][0]                    \n",
      "                                                                 dense_3[0][0]                    \n",
      "                                                                 dense_5[0][0]                    \n",
      "                                                                 dense_7[0][0]                    \n",
      "                                                                 dense_9[0][0]                    \n",
      "                                                                 dense_11[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense_12 (Dense)                (None, 512)          53760       concatenate[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "dropout_43 (Dropout)            (None, 512)          0           dense_12[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense_13 (Dense)                (None, 256)          131328      dropout_43[0][0]                 \n",
      "__________________________________________________________________________________________________\n",
      "dense_14 (Dense)                (None, 1)            257         dense_13[0][0]                   \n",
      "==================================================================================================\n",
      "Total params: 110,005,257\n",
      "Trainable params: 110,005,257\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:38:56.929637Z",
     "iopub.status.busy": "2021-03-31T10:38:56.928981Z",
     "iopub.status.idle": "2021-03-31T10:38:58.103046Z",
     "shell.execute_reply": "2021-03-31T10:38:58.103532Z"
    },
    "papermill": {
     "duration": 1.215602,
     "end_time": "2021-03-31T10:38:58.103665",
     "exception": false,
     "start_time": "2021-03-31T10:38:56.888063",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from tensorflow.keras.utils import plot_model\n",
    "plot_model(model, to_file='model_plot.png', show_shapes=True, show_layer_names=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.040872,
     "end_time": "2021-03-31T10:38:58.186341",
     "exception": false,
     "start_time": "2021-03-31T10:38:58.145469",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## 5. Training, validation and testing\n",
    "\n",
    "Loops over the folds in gkf and trains each fold for 3 epochs --- with a learning rate of 3e-5 and batch_size of 6. A simple binary crossentropy is used as the objective-/loss-function. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:38:58.287959Z",
     "iopub.status.busy": "2021-03-31T10:38:58.286954Z",
     "iopub.status.idle": "2021-03-31T10:38:58.301436Z",
     "shell.execute_reply": "2021-03-31T10:38:58.300802Z"
    },
    "papermill": {
     "duration": 0.074343,
     "end_time": "2021-03-31T10:38:58.301566",
     "exception": false,
     "start_time": "2021-03-31T10:38:58.227223",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================== \n",
      "mean_absolute_error  : 0.09999999999999999\n",
      "mean_squared_error  : 0.009999999999999998\n",
      "r2 score  : 0.96\n",
      "================== \n",
      "f1_score  : 0.6666666666666666\n",
      "[[0 1]\n",
      " [0 1]]\n",
      "Acc 0.5 Prec 0.5 Rec 1.0 F1 0.6666666666666666\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.5"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Evaluation Metrics\n",
    "import sklearn\n",
    "def print_evaluation_metrics(y_true, y_pred, label='', is_regression=True, label2=''):\n",
    "    print('==================', label2)\n",
    "    ### For regression\n",
    "    if is_regression:\n",
    "        print('mean_absolute_error',label,':', sklearn.metrics.mean_absolute_error(y_true, y_pred))\n",
    "        print('mean_squared_error',label,':', sklearn.metrics.mean_squared_error(y_true, y_pred))\n",
    "        print('r2 score',label,':', sklearn.metrics.r2_score(y_true, y_pred))\n",
    "        #     print('max_error',label,':', sklearn.metrics.max_error(y_true, y_pred))\n",
    "        return sklearn.metrics.mean_squared_error(y_true, y_pred)\n",
    "    else:\n",
    "        ### FOR Classification\n",
    "#         print('balanced_accuracy_score',label,':', sklearn.metrics.balanced_accuracy_score(y_true, y_pred))\n",
    "#         print('average_precision_score',label,':', sklearn.metrics.average_precision_score(y_true, y_pred))\n",
    "#         print('balanced_accuracy_score',label,':', sklearn.metrics.balanced_accuracy_score(y_true, y_pred))\n",
    "#         print('accuracy_score',label,':', sklearn.metrics.accuracy_score(y_true, y_pred))\n",
    "        print('f1_score',label,':', sklearn.metrics.f1_score(y_true, y_pred))\n",
    "        \n",
    "        matrix = sklearn.metrics.confusion_matrix(y_true, y_pred)\n",
    "        print(matrix)\n",
    "        TP,TN,FP,FN = matrix[1][1],matrix[0][0],matrix[0][1],matrix[1][0]\n",
    "        Accuracy = (TP+TN)/(TP+FP+FN+TN)\n",
    "        Precision = TP/(TP+FP)\n",
    "        Recall = TP/(TP+FN)\n",
    "        F1 = 2*(Recall * Precision) / (Recall + Precision)\n",
    "        print('Acc', Accuracy, 'Prec', Precision, 'Rec', Recall, 'F1',F1)\n",
    "        return sklearn.metrics.accuracy_score(y_true, y_pred)\n",
    "\n",
    "print_evaluation_metrics([1,0], [0.9,0.1], '', True)\n",
    "print_evaluation_metrics([1,0], [1,1], '', False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 0.041263,
     "end_time": "2021-03-31T10:38:58.383738",
     "exception": false,
     "start_time": "2021-03-31T10:38:58.342475",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "### Loss function selection\n",
    "Regression problem between 0 and 1, so binary_crossentropy and mean_absolute_error seem good.\n",
    "\n",
    "Here are the explanations: https://www.dlology.com/blog/how-to-choose-last-layer-activation-and-loss-function/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T10:38:58.486241Z",
     "iopub.status.busy": "2021-03-31T10:38:58.482427Z",
     "iopub.status.idle": "2021-03-31T13:56:35.906423Z",
     "shell.execute_reply": "2021-03-31T13:56:35.907193Z"
    },
    "papermill": {
     "duration": 11857.481076,
     "end_time": "2021-03-31T13:56:35.907361",
     "exception": false,
     "start_time": "2021-03-31T10:38:58.426285",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>\n",
      "LR= 1e-05\n",
      "80000 80000\n",
      "(None, 768)\n",
      "model compiled\n",
      "Epoch 1/3\n",
      "10667/10667 [==============================] - 3679s 345ms/step - loss: 0.0695 - val_loss: 0.0440\n",
      "Epoch 2/3\n",
      "10667/10667 [==============================] - 3696s 347ms/step - loss: 0.0222 - val_loss: 0.0431\n",
      "Epoch 3/3\n",
      "10667/10667 [==============================] - 3720s 349ms/step - loss: 0.0099 - val_loss: 0.0496\n",
      "================== \n",
      "mean_absolute_error on #8 : 0.016925473\n",
      "mean_squared_error on #8 : 0.012227797\n",
      "r2 score on #8 : 0.9510884324146058\n",
      "new acc >>  0.012227797\n"
     ]
    }
   ],
   "source": [
    "min_acc = 100\n",
    "min_test = []\n",
    "valid_preds = []\n",
    "test_preds = []\n",
    "best_model = False\n",
    "for BS in [6]:\n",
    "    LR = 1e-5\n",
    "    print('>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>')\n",
    "    print('LR=', LR)\n",
    "    gkf = GroupKFold(n_splits=2).split(X=df_train.text, groups=df_train.text)\n",
    "\n",
    "    for fold, (train_idx, valid_idx) in enumerate(gkf):\n",
    "        if fold not in range(running_folds):\n",
    "            continue\n",
    "        train_inputs = [(inputs[i][train_idx])[:training_sample_count] for i in range(len(inputs))]\n",
    "        train_outputs = (outputs[train_idx])[:training_sample_count]\n",
    "\n",
    "        valid_inputs = [inputs[i][valid_idx] for i in range(len(inputs))]\n",
    "        valid_outputs = outputs[valid_idx]\n",
    "\n",
    "        print(len(train_idx), len(train_outputs))\n",
    "#         print(train_idx[:10], valid_idx[:10])\n",
    "\n",
    "        model = create_model()\n",
    "        K.clear_session()\n",
    "        optimizer = tf.keras.optimizers.Adam(learning_rate=LR)\n",
    "        model.compile(loss='binary_crossentropy', optimizer=optimizer)\n",
    "        print('model compiled')\n",
    "        \n",
    "        model.fit(train_inputs, train_outputs, epochs=training_epochs, batch_size=BS,\n",
    "                 validation_split=0.2, verbose=1,\n",
    "#                  validation_data=(x_val, y_val)\n",
    "                 )\n",
    "        \n",
    "        model.save_weights(f'bert-{fold}.h5')\n",
    "\n",
    "        valid_preds.append(model.predict(valid_inputs))\n",
    "        acc = print_evaluation_metrics(np.array(valid_outputs), np.array(valid_preds[-1]), 'on #'+str(xx+1))\n",
    "        if acc < min_acc:\n",
    "            print('new acc >> ', acc)\n",
    "            min_acc = acc\n",
    "            best_model = model\n",
    "#                 min_test = model.predict(test_inputs)\n",
    "#                 test_preds.append(min_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T13:56:56.156681Z",
     "iopub.status.busy": "2021-03-31T13:56:56.155709Z",
     "iopub.status.idle": "2021-03-31T13:56:57.102722Z",
     "shell.execute_reply": "2021-03-31T13:56:57.103178Z"
    },
    "papermill": {
     "duration": 10.96577,
     "end_time": "2021-03-31T13:56:57.103323",
     "exception": false,
     "start_time": "2021-03-31T13:56:46.137553",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saved model to disk\n"
     ]
    }
   ],
   "source": [
    "# https://machinelearningmastery.com/save-load-keras-deep-learning-models/\n",
    "# serialize model to JSON\n",
    "model_json = model.to_json()\n",
    "with open(\"model.json\", \"w\") as json_file:\n",
    "    json_file.write(model_json)\n",
    "# serialize weights to HDF5\n",
    "model.save_weights(\"model.h5\")\n",
    "print(\"Saved model to disk\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T13:57:18.300631Z",
     "iopub.status.busy": "2021-03-31T13:57:18.295129Z",
     "iopub.status.idle": "2021-03-31T13:57:21.771853Z",
     "shell.execute_reply": "2021-03-31T13:57:21.773325Z"
    },
    "papermill": {
     "duration": 13.96324,
     "end_time": "2021-03-31T13:57:21.773571",
     "exception": false,
     "start_time": "2021-03-31T13:57:07.810331",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# https://www.javaer101.com/en/article/18162646.html\n",
    "model.save('my_model.h5') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T13:57:43.219438Z",
     "iopub.status.busy": "2021-03-31T13:57:43.218502Z",
     "iopub.status.idle": "2021-03-31T13:59:20.088626Z",
     "shell.execute_reply": "2021-03-31T13:59:20.090360Z"
    },
    "papermill": {
     "duration": 107.546499,
     "end_time": "2021-03-31T13:59:20.090607",
     "exception": false,
     "start_time": "2021-03-31T13:57:32.544108",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "model.save(\"colbert-trained\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T13:59:40.855142Z",
     "iopub.status.busy": "2021-03-31T13:59:40.854380Z",
     "iopub.status.idle": "2021-03-31T13:59:40.857635Z",
     "shell.execute_reply": "2021-03-31T13:59:40.858116Z"
    },
    "papermill": {
     "duration": 10.195848,
     "end_time": "2021-03-31T13:59:40.858247",
     "exception": false,
     "start_time": "2021-03-31T13:59:30.662399",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# https://keras.io/guides/serialization_and_saving/\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T14:00:01.817478Z",
     "iopub.status.busy": "2021-03-31T14:00:01.814800Z",
     "iopub.status.idle": "2021-03-31T14:00:01.820004Z",
     "shell.execute_reply": "2021-03-31T14:00:01.819422Z"
    },
    "papermill": {
     "duration": 10.16492,
     "end_time": "2021-03-31T14:00:01.820148",
     "exception": false,
     "start_time": "2021-03-31T13:59:51.655228",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "best acc >>  0.012227797\n"
     ]
    }
   ],
   "source": [
    "print('best acc >> ', acc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T14:00:22.691686Z",
     "iopub.status.busy": "2021-03-31T14:00:22.689814Z",
     "iopub.status.idle": "2021-03-31T14:00:22.694847Z",
     "shell.execute_reply": "2021-03-31T14:00:22.694366Z"
    },
    "papermill": {
     "duration": 10.704572,
     "end_time": "2021-03-31T14:00:22.694943",
     "exception": false,
     "start_time": "2021-03-31T14:00:11.990371",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "80000"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(valid_inputs[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T14:00:43.074677Z",
     "iopub.status.busy": "2021-03-31T14:00:43.072726Z",
     "iopub.status.idle": "2021-03-31T14:00:43.085886Z",
     "shell.execute_reply": "2021-03-31T14:00:43.085389Z"
    },
    "papermill": {
     "duration": 10.144396,
     "end_time": "2021-03-31T14:00:43.085995",
     "exception": false,
     "start_time": "2021-03-31T14:00:32.941599",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(80000, 1) (80000, 1)\n",
      "================== \n",
      "mean_absolute_error  : 0.016925473\n",
      "mean_squared_error  : 0.012227797\n",
      "r2 score  : 0.9510884324146058\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.012227797"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(valid_outputs.shape, valid_preds[-1].shape)\n",
    "print_evaluation_metrics(np.array(valid_outputs), np.array(valid_preds[-1]), '')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T14:01:04.251460Z",
     "iopub.status.busy": "2021-03-31T14:01:04.249857Z",
     "iopub.status.idle": "2021-03-31T14:06:13.557269Z",
     "shell.execute_reply": "2021-03-31T14:06:13.556199Z"
    },
    "papermill": {
     "duration": 320.07357,
     "end_time": "2021-03-31T14:06:13.557391",
     "exception": false,
     "start_time": "2021-03-31T14:00:53.483821",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# %%time\n",
    "min_test = best_model.predict(test_inputs)\n",
    "\n",
    "## use min_test\n",
    "# min_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T14:06:34.460792Z",
     "iopub.status.busy": "2021-03-31T14:06:34.460233Z",
     "iopub.status.idle": "2021-03-31T14:06:35.220402Z",
     "shell.execute_reply": "2021-03-31T14:06:35.218876Z"
    },
    "papermill": {
     "duration": 10.806938,
     "end_time": "2021-03-31T14:06:35.220518",
     "exception": false,
     "start_time": "2021-03-31T14:06:24.413580",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "df_test.to_csv('df_test.csv')\n",
    "test_df_y.to_csv('test_df_y.csv')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 10.649105,
     "end_time": "2021-03-31T14:06:55.940042",
     "exception": false,
     "start_time": "2021-03-31T14:06:45.290937",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## Regression submission"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T14:07:15.516448Z",
     "iopub.status.busy": "2021-03-31T14:07:15.510124Z",
     "iopub.status.idle": "2021-03-31T14:07:15.878882Z",
     "shell.execute_reply": "2021-03-31T14:07:15.879917Z"
    },
    "papermill": {
     "duration": 10.09129,
     "end_time": "2021-03-31T14:07:15.880110",
     "exception": false,
     "start_time": "2021-03-31T14:07:05.788820",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================== \n",
      "mean_absolute_error  : 0.017012216\n",
      "mean_squared_error  : 0.01243615\n",
      "r2 score  : 0.9502545684127843\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>text</th>\n",
       "      <th>humor</th>\n",
       "      <th>pred</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>What kind of cat should you take into the  des...</td>\n",
       "      <td>True</td>\n",
       "      <td>0.999954</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Remember when people used to have to be in sha...</td>\n",
       "      <td>True</td>\n",
       "      <td>0.885077</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Pizza is always good. - everyone we'll see abo...</td>\n",
       "      <td>True</td>\n",
       "      <td>0.999956</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>What's 6 inches long hard, bent, and in my pan...</td>\n",
       "      <td>True</td>\n",
       "      <td>0.999980</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Black teen's response to violence in his commu...</td>\n",
       "      <td>False</td>\n",
       "      <td>0.000116</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                text  humor      pred\n",
       "0  What kind of cat should you take into the  des...   True  0.999954\n",
       "1  Remember when people used to have to be in sha...   True  0.885077\n",
       "2  Pizza is always good. - everyone we'll see abo...   True  0.999956\n",
       "3  What's 6 inches long hard, bent, and in my pan...   True  0.999980\n",
       "4  Black teen's response to violence in his commu...  False  0.000116"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_sub = test_df_y.copy()\n",
    "# df_sub['pred'] = np.average(test_preds, axis=0) # for weighted average set weights=[...]\n",
    "df_sub['pred'] = min_test\n",
    "\n",
    "\n",
    "print_evaluation_metrics(df_sub['humor'], df_sub['pred'], '', True)\n",
    "\n",
    "\n",
    "df_sub.to_csv('sub1.csv', index=False)\n",
    "df_sub.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "papermill": {
     "duration": 10.334028,
     "end_time": "2021-03-31T14:07:36.741768",
     "exception": false,
     "start_time": "2021-03-31T14:07:26.407740",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## Binary submission"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T14:07:56.853102Z",
     "iopub.status.busy": "2021-03-31T14:07:56.852454Z",
     "iopub.status.idle": "2021-03-31T14:08:00.437874Z",
     "shell.execute_reply": "2021-03-31T14:08:00.436893Z"
    },
    "papermill": {
     "duration": 13.726902,
     "end_time": "2021-03-31T14:08:00.438005",
     "exception": false,
     "start_time": "2021-03-31T14:07:46.711103",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================== SPLIT on 0.1\n",
      "f1_score  : 0.9806121444739692\n",
      "[[19284   634]\n",
      " [  154 19928]]\n",
      "Acc 0.9803 Prec 0.9691664234996595 Rec 0.9923314410915247 F1 0.9806121444739692\n",
      "================== SPLIT on 0.2\n",
      "f1_score  : 0.983292140385566\n",
      "[[19432   486]\n",
      " [  190 19892]]\n",
      "Acc 0.9831 Prec 0.9761507508096967 Rec 0.990538790957076 F1 0.983292140385566\n",
      "================== SPLIT on 0.30000000000000004\n",
      "f1_score  : 0.9842921559883058\n",
      "[[19502   416]\n",
      " [  218 19864]]\n",
      "Acc 0.98415 Prec 0.9794871794871794 Rec 0.9891445075191714 F1 0.9842921559883058\n",
      "================== SPLIT on 0.4\n",
      "f1_score  : 0.9847612429266356\n",
      "[[19547   371]\n",
      " [  243 19839]]\n",
      "Acc 0.98465 Prec 0.9816427511133102 Rec 0.9878996115924709 F1 0.9847612429266356\n",
      "================== SPLIT on 0.5\n",
      "f1_score  : 0.9850857568978374\n",
      "[[19585   333]\n",
      " [  267 19815]]\n",
      "Acc 0.985 Prec 0.9834723049434188 Rec 0.9867045115028383 F1 0.9850857568978374\n",
      "================== SPLIT on 0.6\n",
      "f1_score  : 0.9851682261596655\n",
      "[[19610   308]\n",
      " [  288 19794]]\n",
      "Acc 0.9851 Prec 0.9846781414784599 Rec 0.9856587989244099 F1 0.9851682261596655\n",
      "================== SPLIT on 0.7000000000000001\n",
      "f1_score  : 0.9853432374495239\n",
      "[[19647   271]\n",
      " [  317 19765]]\n",
      "Acc 0.9853 Prec 0.9864743461768816 Rec 0.9842147196494373 F1 0.9853432374495239\n",
      "================== SPLIT on 0.8\n",
      "f1_score  : 0.9850157334798463\n",
      "[[19679   239]\n",
      " [  361 19721]]\n",
      "Acc 0.985 Prec 0.9880260521042085 Rec 0.9820237028184444 F1 0.9850157334798463\n",
      "================== SPLIT on 0.9\n",
      "f1_score  : 0.9841596069978445\n",
      "[[19735   183]\n",
      " [  449 19633]]\n",
      "Acc 0.9842 Prec 0.9907650383528462 Rec 0.9776416691564586 F1 0.9841596069978445\n"
     ]
    }
   ],
   "source": [
    "for split in np.arange(0.1, 0.99, 0.1).tolist():\n",
    "    df_sub['pred_bi'] = (df_sub['pred'] > split)\n",
    "\n",
    "    print_evaluation_metrics(df_sub['humor'], df_sub['pred_bi'], '', False, 'SPLIT on '+str(split))\n",
    "\n",
    "    df_sub.to_csv('sub3.csv', index=False)\n",
    "    df_sub.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T14:08:20.365062Z",
     "iopub.status.busy": "2021-03-31T14:08:20.364063Z",
     "iopub.status.idle": "2021-03-31T14:08:20.709204Z",
     "shell.execute_reply": "2021-03-31T14:08:20.708691Z"
    },
    "papermill": {
     "duration": 10.494589,
     "end_time": "2021-03-31T14:08:20.709322",
     "exception": false,
     "start_time": "2021-03-31T14:08:10.214733",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================== SPLIT on 0.9\n",
      "f1_score  : 0.9850857568978374\n",
      "[[19585   333]\n",
      " [  267 19815]]\n",
      "Acc 0.985 Prec 0.9834723049434188 Rec 0.9867045115028383 F1 0.9850857568978374\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>text</th>\n",
       "      <th>humor</th>\n",
       "      <th>pred</th>\n",
       "      <th>pred_bi</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>What kind of cat should you take into the  des...</td>\n",
       "      <td>True</td>\n",
       "      <td>0.999954</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Remember when people used to have to be in sha...</td>\n",
       "      <td>True</td>\n",
       "      <td>0.885077</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Pizza is always good. - everyone we'll see abo...</td>\n",
       "      <td>True</td>\n",
       "      <td>0.999956</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>What's 6 inches long hard, bent, and in my pan...</td>\n",
       "      <td>True</td>\n",
       "      <td>0.999980</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Black teen's response to violence in his commu...</td>\n",
       "      <td>False</td>\n",
       "      <td>0.000116</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                text  humor      pred  pred_bi\n",
       "0  What kind of cat should you take into the  des...   True  0.999954     True\n",
       "1  Remember when people used to have to be in sha...   True  0.885077     True\n",
       "2  Pizza is always good. - everyone we'll see abo...   True  0.999956     True\n",
       "3  What's 6 inches long hard, bent, and in my pan...   True  0.999980     True\n",
       "4  Black teen's response to violence in his commu...  False  0.000116    False"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_sub['pred_bi'] = (df_sub['pred'] > 0.5)\n",
    "\n",
    "print_evaluation_metrics(df_sub['humor'], df_sub['pred_bi'], '', False, 'SPLIT on '+str(split))\n",
    "\n",
    "df_sub.to_csv('sub3.csv', index=False)\n",
    "df_sub.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-31T14:08:41.367076Z",
     "iopub.status.busy": "2021-03-31T14:08:41.366433Z",
     "iopub.status.idle": "2021-03-31T14:08:41.387503Z",
     "shell.execute_reply": "2021-03-31T14:08:41.387945Z"
    },
    "papermill": {
     "duration": 10.11917,
     "end_time": "2021-03-31T14:08:41.388087",
     "exception": false,
     "start_time": "2021-03-31T14:08:31.268917",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>text</th>\n",
       "      <th>humor</th>\n",
       "      <th>pred</th>\n",
       "      <th>pred_bi</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>252</th>\n",
       "      <td>Ariana grande looks like she was designed in a...</td>\n",
       "      <td>True</td>\n",
       "      <td>0.023232</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>303</th>\n",
       "      <td>If you're happy and you know it, share your meds</td>\n",
       "      <td>True</td>\n",
       "      <td>0.003352</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>309</th>\n",
       "      <td>Friends don't let friends make harlem shake' v...</td>\n",
       "      <td>True</td>\n",
       "      <td>0.278334</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>310</th>\n",
       "      <td>The whip was especially popular in the 1800's</td>\n",
       "      <td>True</td>\n",
       "      <td>0.038513</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>345</th>\n",
       "      <td>One out of five dentists has the courage to sp...</td>\n",
       "      <td>True</td>\n",
       "      <td>0.005155</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39874</th>\n",
       "      <td>Scary but true: statistics show that everyone ...</td>\n",
       "      <td>True</td>\n",
       "      <td>0.009534</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39877</th>\n",
       "      <td>On the money! a female face on the ten dollar ...</td>\n",
       "      <td>False</td>\n",
       "      <td>0.981126</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39889</th>\n",
       "      <td>Seminar topic at the annual vampire conference...</td>\n",
       "      <td>True</td>\n",
       "      <td>0.003524</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39955</th>\n",
       "      <td>A recent study by un has found dexter to be th...</td>\n",
       "      <td>True</td>\n",
       "      <td>0.172858</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39987</th>\n",
       "      <td>Make no mistake, semi-automatic guns are very ...</td>\n",
       "      <td>False</td>\n",
       "      <td>0.886348</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>600 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                    text  humor      pred  \\\n",
       "252    Ariana grande looks like she was designed in a...   True  0.023232   \n",
       "303     If you're happy and you know it, share your meds   True  0.003352   \n",
       "309    Friends don't let friends make harlem shake' v...   True  0.278334   \n",
       "310        The whip was especially popular in the 1800's   True  0.038513   \n",
       "345    One out of five dentists has the courage to sp...   True  0.005155   \n",
       "...                                                  ...    ...       ...   \n",
       "39874  Scary but true: statistics show that everyone ...   True  0.009534   \n",
       "39877  On the money! a female face on the ten dollar ...  False  0.981126   \n",
       "39889  Seminar topic at the annual vampire conference...   True  0.003524   \n",
       "39955  A recent study by un has found dexter to be th...   True  0.172858   \n",
       "39987  Make no mistake, semi-automatic guns are very ...  False  0.886348   \n",
       "\n",
       "       pred_bi  \n",
       "252      False  \n",
       "303      False  \n",
       "309      False  \n",
       "310      False  \n",
       "345      False  \n",
       "...        ...  \n",
       "39874    False  \n",
       "39877     True  \n",
       "39889    False  \n",
       "39955    False  \n",
       "39987     True  \n",
       "\n",
       "[600 rows x 4 columns]"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_sub[df_sub['pred_bi']!=df_sub['humor']]"
   ]
  }
 ],
 "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": "3.7.6"
  },
  "papermill": {
   "duration": 13074.039543,
   "end_time": "2021-03-31T14:08:52.958573",
   "environment_variables": {},
   "exception": null,
   "input_path": "__notebook__.ipynb",
   "output_path": "__notebook__.ipynb",
   "parameters": {},
   "start_time": "2021-03-31T10:30:58.919030",
   "version": "2.1.0"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "state": {
     "048739e605074668929a36194a0ad654": {
      "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_d9a481b4fe0745bf8f434b3afdb56399",
       "placeholder": "​",
       "style": "IPY_MODEL_713b6f4ba62f43a483109268f72b574d",
       "value": " 536M/536M [00:23&lt;00:00, 23.2MB/s]"
      }
     },
     "04d0de21062449c78c00b8b7278de21b": {
      "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"
      }
     },
     "09fca212662740d9be2f1e8e7576bac3": {
      "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
      }
     },
     "11b8b846b8d9406d958f659a05d5f0b5": {
      "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
      }
     },
     "19b488c7d5b14c249661fdfcc45f8100": {
      "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_a187930500b54a119a258397ea947f0e",
       "placeholder": "​",
       "style": "IPY_MODEL_f95c140c53ff438e860f17d2902fffb1",
       "value": " 160000/? [06:03&lt;00:00, 439.65it/s]"
      }
     },
     "1fac728e8bf449648fb1b23fc2d1be7c": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "FloatProgressModel",
      "state": {
       "_dom_classes": [],
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "FloatProgressModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/controls",
       "_view_module_version": "1.5.0",
       "_view_name": "ProgressView",
       "bar_style": "success",
       "description": "Downloading: 100%",
       "description_tooltip": null,
       "layout": "IPY_MODEL_dc0ff580412e486c9255181041cf2210",
       "max": 536063208.0,
       "min": 0.0,
       "orientation": "horizontal",
       "style": "IPY_MODEL_87ff20c16ee0490f975fc17ddbf89387",
       "value": 536063208.0
      }
     },
     "45e4cf163632464485e291daebf4ea85": {
      "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
      }
     },
     "49e2f68426624249ba92a10c6b4832c4": {
      "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"
      }
     },
     "5efb74cf266843a8b2ec30d1d797f3d5": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "FloatProgressModel",
      "state": {
       "_dom_classes": [],
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "FloatProgressModel",
       "_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_c84b97b742c1486f947b06455801646e",
       "max": 1.0,
       "min": 0.0,
       "orientation": "horizontal",
       "style": "IPY_MODEL_04d0de21062449c78c00b8b7278de21b",
       "value": 1.0
      }
     },
     "679eb8020c4e475b98d57e25ef6478cb": {
      "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
      }
     },
     "6ea09fbc9fc74eb68a07d42de3f9b934": {
      "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": ""
      }
     },
     "713b6f4ba62f43a483109268f72b574d": {
      "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": ""
      }
     },
     "78dea8febd00432aa6c791108475cdeb": {
      "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"
      }
     },
     "87ff20c16ee0490f975fc17ddbf89387": {
      "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"
      }
     },
     "92515a08dd8a4c588739f0e0bd16fa37": {
      "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_a6d6887db7584116ab5043c1eae6fdb4",
       "placeholder": "​",
       "style": "IPY_MODEL_6ea09fbc9fc74eb68a07d42de3f9b934",
       "value": " 232k/232k [00:00&lt;00:00, 1.58MB/s]"
      }
     },
     "968816a6d7ee4a09a8794a75730d8132": {
      "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_e20cf09aa1904f4fb1ec7f420fee7432",
        "IPY_MODEL_92515a08dd8a4c588739f0e0bd16fa37"
       ],
       "layout": "IPY_MODEL_679eb8020c4e475b98d57e25ef6478cb"
      }
     },
     "a187930500b54a119a258397ea947f0e": {
      "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
      }
     },
     "a56cf19ff2684c709ed4019e2649dc46": {
      "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
      }
     },
     "a6d6887db7584116ab5043c1eae6fdb4": {
      "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
      }
     },
     "aa53e67a3ea54292b3c6cc404dc9bbd1": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "FloatProgressModel",
      "state": {
       "_dom_classes": [],
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "FloatProgressModel",
       "_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_f4a922c4b9424e11ab260cab0478548c",
       "max": 1.0,
       "min": 0.0,
       "orientation": "horizontal",
       "style": "IPY_MODEL_49e2f68426624249ba92a10c6b4832c4",
       "value": 1.0
      }
     },
     "b22963686f534071b666c4e084fd4d9c": {
      "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_5efb74cf266843a8b2ec30d1d797f3d5",
        "IPY_MODEL_19b488c7d5b14c249661fdfcc45f8100"
       ],
       "layout": "IPY_MODEL_09fca212662740d9be2f1e8e7576bac3"
      }
     },
     "b8c039cb85504c7e96c24a5771e078fa": {
      "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": ""
      }
     },
     "c84b97b742c1486f947b06455801646e": {
      "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
      }
     },
     "cc42e503bbf14b438ffe8f8fc2035603": {
      "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_45e4cf163632464485e291daebf4ea85",
       "placeholder": "​",
       "style": "IPY_MODEL_b8c039cb85504c7e96c24a5771e078fa",
       "value": " 40000/? [02:01&lt;00:00, 330.41it/s]"
      }
     },
     "d0124c1c804e466ea5553e43c77c0076": {
      "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_aa53e67a3ea54292b3c6cc404dc9bbd1",
        "IPY_MODEL_cc42e503bbf14b438ffe8f8fc2035603"
       ],
       "layout": "IPY_MODEL_11b8b846b8d9406d958f659a05d5f0b5"
      }
     },
     "d9a481b4fe0745bf8f434b3afdb56399": {
      "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
      }
     },
     "dc0ff580412e486c9255181041cf2210": {
      "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
      }
     },
     "e20cf09aa1904f4fb1ec7f420fee7432": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "FloatProgressModel",
      "state": {
       "_dom_classes": [],
       "_model_module": "@jupyter-widgets/controls",
       "_model_module_version": "1.5.0",
       "_model_name": "FloatProgressModel",
       "_view_count": null,
       "_view_module": "@jupyter-widgets/controls",
       "_view_module_version": "1.5.0",
       "_view_name": "ProgressView",
       "bar_style": "success",
       "description": "Downloading: 100%",
       "description_tooltip": null,
       "layout": "IPY_MODEL_a56cf19ff2684c709ed4019e2649dc46",
       "max": 231508.0,
       "min": 0.0,
       "orientation": "horizontal",
       "style": "IPY_MODEL_78dea8febd00432aa6c791108475cdeb",
       "value": 231508.0
      }
     },
     "e6bdfbb5ea4c4f93a2676145c7adb6a5": {
      "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
      }
     },
     "f4a922c4b9424e11ab260cab0478548c": {
      "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
      }
     },
     "f95c140c53ff438e860f17d2902fffb1": {
      "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": ""
      }
     },
     "fdd50281a5d2432381d1b4727c0d05da": {
      "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_1fac728e8bf449648fb1b23fc2d1be7c",
        "IPY_MODEL_048739e605074668929a36194a0ad654"
       ],
       "layout": "IPY_MODEL_e6bdfbb5ea4c4f93a2676145c7adb6a5"
      }
     }
    },
    "version_major": 2,
    "version_minor": 0
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}