[homework]summarization.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "tWJTCBu4Tveb"
   },
   "source": [
    "<h3 style=\"text-align: center;\"><b>Физтех-Школа Прикладной математики и информатики (ФПМИ) МФТИ</b></h3>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Kmb8UhIzOnfK"
   },
   "source": [
    "# Text Summarization\n",
    "\n",
    "Суммаризация текста на примере генерации \"сжатых\" новостей. Рассмотрим некоторые базовые решения и познакомимся с архитектурами нейросетей для решения задачи.\n",
    "Датасет: gazeta.ru\n",
    "\n",
    "Загрузим датасет и необходимые библиотеки"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "id": "SXS1sdYZCluU"
   },
   "outputs": [],
   "source": [
    "from IPython.display import clear_output\n",
    "!pip -q install razdel networkx pymorphy2 nltk rouge==0.3.1 summa\n",
    "!pip install youtokentome\n",
    "clear_output()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "id": "FwO834E6WdQI"
   },
   "outputs": [],
   "source": [
    "import json\n",
    "import random\n",
    "import youtokentome as yttm\n",
    "from nltk.translate.bleu_score import corpus_bleu\n",
    "from rouge import Rouge\n",
    "import razdel\n",
    "from tqdm.notebook import tqdm\n",
    "import copy\n",
    "import math\n",
    "import numpy as np\n",
    "import os\n",
    "\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "import torch.optim as optim\n",
    "from torch.autograd import Variable\n",
    "from torch.nn.utils.rnn import pack_padded_sequence as pack\n",
    "from torch.nn.utils.rnn import pad_packed_sequence as unpack\n",
    "\n",
    "from sklearn.preprocessing import minmax_scale\n",
    "\n",
    "import matplotlib\n",
    "import time\n",
    "import re\n",
    "matplotlib.rcParams.update({'figure.figsize': (16, 12), 'font.size': 14})\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "id": "3QVCmnFGT0k4"
   },
   "outputs": [],
   "source": [
    "# фиксация SEED\n",
    "SEED = 42\n",
    "\n",
    "random.seed(SEED)\n",
    "np.random.seed(SEED)\n",
    "torch.manual_seed(SEED)\n",
    "torch.cuda.manual_seed(SEED)\n",
    "torch.backends.cudnn.deterministic = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "id": "OqkLTkFRfXvA"
   },
   "outputs": [],
   "source": [
    "!wget -q https://www.dropbox.com/s/43l702z5a5i2w8j/gazeta_train.txt\n",
    "!wget -q https://www.dropbox.com/s/k2egt3sug0hb185/gazeta_val.txt\n",
    "!wget -q https://www.dropbox.com/s/3gki5n5djs9w0v6/gazeta_test.txt\n",
    "clear_output()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "OyUAtZA9WyL-",
    "outputId": "f9b767bc-1eb2-4389-d671-c90ab2abadb2"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CUDA is available!  Training on GPU ...\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "device(type='cuda', index=0)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_on_gpu = torch.cuda.is_available()\n",
    "\n",
    "if not train_on_gpu:\n",
    "    print('CUDA is not available.  Training on CPU ...')\n",
    "else:\n",
    "    print('CUDA is available!  Training on GPU ...')\n",
    "\n",
    "# Определяем на чем у нас будут производиться вычисления (cuda/cpu)\n",
    "device = torch.device('cuda:0') if torch.cuda.is_available() else torch.device('cpu')  \n",
    "device"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "16pgJGsjXBdI",
    "outputId": "260b41a3-b15a-4af6-d07d-98c6701f19f2"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sun Jun 12 09:58:11 2022       \n",
      "+-----------------------------------------------------------------------------+\n",
      "| NVIDIA-SMI 460.32.03    Driver Version: 460.32.03    CUDA Version: 11.2     |\n",
      "|-------------------------------+----------------------+----------------------+\n",
      "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
      "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n",
      "|                               |                      |               MIG M. |\n",
      "|===============================+======================+======================|\n",
      "|   0  Tesla T4            Off  | 00000000:00:04.0 Off |                    0 |\n",
      "| N/A   41C    P8     9W /  70W |      3MiB / 15109MiB |      0%      Default |\n",
      "|                               |                      |                  N/A |\n",
      "+-------------------------------+----------------------+----------------------+\n",
      "                                                                               \n",
      "+-----------------------------------------------------------------------------+\n",
      "| Processes:                                                                  |\n",
      "|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\n",
      "|        ID   ID                                                   Usage      |\n",
      "|=============================================================================|\n",
      "|  No running processes found                                                 |\n",
      "+-----------------------------------------------------------------------------+\n"
     ]
    }
   ],
   "source": [
    "!nvidia-smi"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Wa0NfryxbPUP"
   },
   "source": [
    "## Dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "eesnclfDDV3F"
   },
   "source": [
    "Посмотрим на то, как устроен датасет"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "Mz6CZYKQhnd-",
    "outputId": "a6106045-d14c-4f82-bb42-36518dacd520"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{\"url\": \"https://www.gazeta.ru/financial/2011/11/30/3852658.shtml\", \"text\": \"«По итогам 2011 года чистый отток может составить примерно $80 млрд, в следующем году — около $20 млрд. При этом мы ожидаем, что со второго полугодия 2012 года начнется приток капитала», — заявил «Интерфаксу» замминистра экономического развития Андрей Клепач. Официальные прогнозы по выводу капитала из России становятся все пессимистичными: еще летом власти полагали, что из страны уйдет не более $35 млрд, в сентябре Минэкономразвития назвал цифру $50 млрд, в начале ноября Центробанк пересмотрел оценку до $70 млрд. Очередное изменение прогноза было ожидаемо: по расчетам Центробанка , за январь — октябрь чистый отток капитала достиг $64 млрд, причем в последние месяцы он ускорился: в сентябре он составил $14 млрд, в октябре — $13 млрд против среднего ежемесячного оттока в $6—8 млрд в первом полугодии. «После октябрьских данных Минэкономразвития вынуждено было изменить оценку, настаивать на $70 млрд означало ожидать серьезного замедления оттока капитала на непонятно каких причинах», — говорит главный экономист BNP Paribas Юлия Цепляева. «В последние два месяца отток капитала ускорится, на декабрь приходится значительная часть выплат по внешним долгам, что приводит к усилению оттока, особенно если они не рефинансируются новыми кредитами», — соглашается главный экономист ФК «Открытие» Владимир Тихомиров. Прогнозируемый Минэкономразвития отток капитала — один из самых высоких за последние 20 лет. Больше ушло лишь в 2008 году на фоне разрастания финансового кризиса и российско-грузинской войны — $133,7 млрд. В кризисный 2009 год из России утекло $56,1 млрд. Главный фактор ускорения оттока капитала в 2011 году — нестабильность на внешних финансовых рынках и рост опасений относительно второй волны рецессии. «Это реакция на неуверенность, которую генерирует Европа с долговыми проблемами. В случае новой волны глобальной турбулентности Россия — одна из самых уязвимых стран», — говорит Цепляева. Еще одна причина — ослабление рубля. «Привлекательность вложений снижается на фоне того, что рубль перестал укрепляться, а ставки по депозитам достаточно низкие. В результате экспортеры не полностью возвращают экспортную выручку», — говорит Тихомиров. Внутри страны эксперты не видят особых причин для бегства капитала. «Ситуация выглядит достаточно позитивно, очень хорошие макроэкономические результаты за год, особенно на фоне других стран. С политической точки зрения все достаточно понятно и предсказуемо, итог выборов очевиден», — говорит экономист ИК «Тройка Диалог» Антон Струченевский. Тем не менее политический фактор играет роль. «Бизнесу важно не только, кто будет президентом, он ждет ясности с перестановками в правительстве. В наших условиях административный ресурс важнее всего для успешности бизнеса», — говорит Цепляева, добавляя, что отток капитала продолжится до завершения президентских выборов.\", \"title\": \"Прогноз не успевает за оттоком\", \"summary\": \"В 2011 году из России уйдет $80 млрд, считают в Минэкономразвития. Менее месяца назад Центробанк давал оценку $70 млрд, повысив первоначальный прогноз вдвое. Отток капитала из страны усиливается из-за кризиса в Европе, а в декабре российским компаниям выплачивать внешние долги. На движение капитала повлияли и выборы: несмотря на их предсказуемость, бизнес хочет ясности с перестановками в правительстве.\", \"date\": \"2011-11-30 18:33:39\"}\n",
      "52400\n",
      "5265\n",
      "5770\n"
     ]
    }
   ],
   "source": [
    "!head -n 1 gazeta_train.txt\n",
    "!cat gazeta_train.txt | wc -l\n",
    "!cat gazeta_val.txt | wc -l\n",
    "!cat gazeta_test.txt | wc -l"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "id": "5pZ2UGS2DGjH"
   },
   "outputs": [],
   "source": [
    "def read_gazeta_records(file_name, shuffle=True, sort_by_date=False):\n",
    "    assert shuffle != sort_by_date\n",
    "    records = []\n",
    "    with open(file_name, \"r\") as r:\n",
    "        for line in r:\n",
    "            records.append(json.loads(line))\n",
    "    if sort_by_date:\n",
    "        records.sort(key=lambda x: x[\"date\"])\n",
    "    if shuffle:\n",
    "        random.shuffle(records)\n",
    "    return records"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "id": "GNDp-BunEA91"
   },
   "outputs": [],
   "source": [
    "train_records = read_gazeta_records(\"gazeta_train.txt\")\n",
    "val_records = read_gazeta_records(\"gazeta_val.txt\")\n",
    "test_records = read_gazeta_records(\"gazeta_test.txt\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "id": "397gjsNfFBZ_"
   },
   "outputs": [],
   "source": [
    "def calc_scores(references, predictions, metric=\"all\"):\n",
    "    print(\"Count:\", len(predictions))\n",
    "    print(\"Ref:\", references[-1])\n",
    "    print(\"Hyp:\", predictions[-1])\n",
    "\n",
    "    if metric in (\"bleu\", \"all\"):\n",
    "        print(\"BLEU: \", corpus_bleu([[r] for r in references], predictions))\n",
    "    if metric in (\"rouge\", \"all\"):\n",
    "        rouge = Rouge()\n",
    "        scores = rouge.get_scores(predictions, references, avg=True)\n",
    "        print(\"ROUGE: \", scores)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "aaJKNsUGFBaA"
   },
   "source": [
    "## Extractive RNN"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "l3izlm8HFBaC"
   },
   "source": [
    "### BPE\n",
    "Для начала сделаем BPE токенизацию"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "id": "Yg9T6q0wFBaF"
   },
   "outputs": [],
   "source": [
    "def train_bpe(records, model_path, model_type=\"bpe\", vocab_size=10000, lower=True):\n",
    "    temp_file_name = \"temp.txt\"\n",
    "    with open(temp_file_name, \"w\") as temp:\n",
    "        for record in records:\n",
    "            text, summary = record['text'], record['summary']\n",
    "            if lower:\n",
    "                summary = summary.lower()\n",
    "                text = text.lower()\n",
    "            if not text or not summary:\n",
    "                continue\n",
    "            temp.write(text + \"\\n\")\n",
    "            temp.write(summary + \"\\n\")\n",
    "    yttm.BPE.train(data=temp_file_name, vocab_size=vocab_size, model=model_path)\n",
    "\n",
    "train_bpe(train_records, \"BPE_model.bin\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "jJFAJHTtFBaF"
   },
   "source": [
    "### Словарь\n",
    "Составим словарь для индексации токенов"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "id": "MueXtatmFBaG"
   },
   "outputs": [],
   "source": [
    "bpe_processor = yttm.BPE('BPE_model.bin')\n",
    "vocabulary = bpe_processor.vocab()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Z_C_p7tHFBaH"
   },
   "source": [
    "### Кэш oracle summary\n",
    "Закэшируем oracle summary, чтобы не пересчитывать их каждый раз"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 119,
     "referenced_widgets": [
      "f685af33f30641c3a7748055889735b9",
      "ce408e42ec60499ab70b08419f4ba4a8",
      "a1c7e56bfbe04fcebf4bc4630a8dc246",
      "9f9b84b6da494f21bf9355fc2f06b833",
      "5c3fe8bfd6ba46d583aa840568f43e86",
      "8e3b02911605486a9728833cfcde4181",
      "dd07dd4af55040c9af1d3b092234c60e",
      "5a9036fe34b44f5f86232dbe71b30cd4",
      "4342980193114cceb3cc871eaeb6f651",
      "8357ebe3c260419eb58bdcae946f776a",
      "78a3f4154ee340acb57ba40ed14f6f93",
      "e2d49cdfe422414a873fcc36d13cb0ec",
      "85c35253f55b4c608caddc371b3d1494",
      "4b29e97fabd74fea9c2e45a1668cf4ed",
      "3abe49bef0b44456b86ff5c8ecaec6ae",
      "a0eea8faa4234c95b053bbd473a7d526",
      "64bccb899a7348649b9f9fb1ff169507",
      "200c9d6fa70648b2917dcd85001740d1",
      "56bc5f6da9734bcc8d6df691a64ec3c8",
      "3f4692e0c3814364a2e1dba7947a53cd",
      "49d8eed23498449ca5a6e5fd491361b5",
      "7d4ef9892dc144899d2a00555aa75349",
      "e36e0519cb484752a0616c192903c392",
      "5ab446dfdb9f41bebb504f0f44e6fe0a",
      "915270cc25724dcabb18779e4b1f8fe8",
      "a9044743cd094564a29fa4020df1ba7f",
      "32f724270b614855b691cb7338c3ee6a",
      "43c237a6422b4bdeb32b65838fd033de",
      "a8d4eff3041e4b2f9ca0fe608b629fc8",
      "963c405588f8448f984edfb693ad2c7e",
      "65bad71fdd534716bcf2598e0cedadce",
      "32e5625802744f3f96fada6cbb429cac",
      "aaa3d9e464be467096fcbf5a722fe8d6"
     ]
    },
    "id": "Fp23tuPbFBaH",
    "outputId": "0abfe30f-417e-4ae4-84f6-a2fb75d0d7c9"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f685af33f30641c3a7748055889735b9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "0it [00:00, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e2d49cdfe422414a873fcc36d13cb0ec",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "0it [00:00, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e36e0519cb484752a0616c192903c392",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "0it [00:00, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def build_oracle_summary_greedy(text, gold_summary, calc_score, lower=True, max_sentences=30):\n",
    "    '''\n",
    "    Жадное построение oracle summary\n",
    "    '''\n",
    "    gold_summary = gold_summary.lower() if lower else gold_summary\n",
    "    # Делим текст на предложения\n",
    "    sentences = [sentence.text.lower() if lower else sentence.text for sentence in razdel.sentenize(text)][:max_sentences]\n",
    "    n_sentences = len(sentences)\n",
    "    oracle_summary_sentences = set()\n",
    "    \n",
    "    score = -1.0\n",
    "    summaries = []\n",
    "    for _ in range(n_sentences):\n",
    "        for i in range(n_sentences):\n",
    "            if i in oracle_summary_sentences:\n",
    "                continue\n",
    "            current_summary_sentences = copy.copy(oracle_summary_sentences)\n",
    "            # Добавляем какое-то предложения к уже существующему summary\n",
    "            current_summary_sentences.add(i)\n",
    "            current_summary = \" \".join([sentences[index] for index in sorted(list(current_summary_sentences))])\n",
    "            # Считаем метрики\n",
    "            current_score = calc_score(current_summary, gold_summary)\n",
    "            summaries.append((current_score, current_summary_sentences))\n",
    "        # Если получилось улучшить метрики с добавлением какого-либо предложения, то пробуем добавить ещё\n",
    "        # Иначе на этом заканчиваем\n",
    "        best_summary_score, best_summary_sentences = max(summaries)\n",
    "        if best_summary_score <= score:\n",
    "            break\n",
    "        oracle_summary_sentences = best_summary_sentences\n",
    "        score = best_summary_score\n",
    "    oracle_summary = \" \".join([sentences[index] for index in sorted(list(oracle_summary_sentences))])\n",
    "    return oracle_summary, oracle_summary_sentences\n",
    "\n",
    "def calc_single_score(pred_summary, gold_summary, rouge):\n",
    "    return rouge.get_scores([pred_summary], [gold_summary], avg=True)['rouge-2']['f']\n",
    "\n",
    "def add_oracle_summary_to_records(records, max_sentences=30, lower=True, nrows=1000):\n",
    "    rouge = Rouge()\n",
    "    for i, record in tqdm(enumerate(records)):\n",
    "        if i >= nrows:\n",
    "            break\n",
    "        text = record[\"text\"]\n",
    "        summary = record[\"summary\"]\n",
    "\n",
    "        summary = summary.lower() if lower else summary\n",
    "        sentences = [sentence.text.lower() if lower else sentence.text for sentence in razdel.sentenize(text)][:max_sentences]\n",
    "        oracle_summary, sentences_indicies = build_oracle_summary_greedy(text, summary, calc_score=lambda x, y: calc_single_score(x, y, rouge),\n",
    "                                                                         lower=lower, max_sentences=max_sentences)\n",
    "        record[\"sentences\"] = sentences\n",
    "        record[\"oracle_sentences\"] = list(sentences_indicies)\n",
    "        record[\"oracle_summary\"] = oracle_summary\n",
    "\n",
    "    return records[:nrows]\n",
    "\n",
    "ext_train_records = add_oracle_summary_to_records(train_records, nrows=2048)\n",
    "ext_val_records = add_oracle_summary_to_records(val_records, nrows=256)\n",
    "ext_test_records = add_oracle_summary_to_records(test_records, nrows=256)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "UlXXc8qUHC5m"
   },
   "source": [
    "### Составление батчей"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "id": "MNyxstTChK3C"
   },
   "outputs": [],
   "source": [
    "class BatchIterator():\n",
    "    def __init__(self, records, vocabulary, batch_size, bpe_processor, shuffle=True, lower=True, max_sentences=30, max_sentence_length=50, device=torch.device('cpu')):\n",
    "        self.records = records\n",
    "        self.num_samples = len(records)\n",
    "        self.batch_size = batch_size\n",
    "        self.bpe_processor = bpe_processor\n",
    "        self.shuffle = shuffle\n",
    "        self.batches_count = int(math.ceil(self.num_samples / batch_size))\n",
    "        self.lower = lower\n",
    "        self.rouge = Rouge()\n",
    "        self.vocabulary = vocabulary\n",
    "        self.max_sentences = max_sentences\n",
    "        self.max_sentence_length = max_sentence_length\n",
    "        self.device = device\n",
    "        \n",
    "    def __len__(self):\n",
    "        return self.batches_count\n",
    "    \n",
    "    def __iter__(self):\n",
    "        indices = np.arange(self.num_samples)\n",
    "        if self.shuffle:\n",
    "            np.random.shuffle(indices)\n",
    "\n",
    "        for start in range(0, self.num_samples, self.batch_size):\n",
    "            end = min(start + self.batch_size, self.num_samples)\n",
    "            batch_indices = indices[start:end]\n",
    "\n",
    "            batch_inputs = []\n",
    "            batch_outputs = []\n",
    "            max_sentence_length = 0\n",
    "            max_sentences = 0\n",
    "            batch_records = []\n",
    "\n",
    "            for data_ind in batch_indices:\n",
    "                \n",
    "                record = self.records[data_ind]\n",
    "                batch_records.append(record)\n",
    "                text = record[\"text\"]\n",
    "                summary = record[\"summary\"]\n",
    "                summary = summary.lower() if self.lower else summary\n",
    "\n",
    "                if \"sentences\" not in record:\n",
    "                    sentences = [sentence.text.lower() if self.lower else sentence.text for sentence in razdel.sentenize(text)][:self.max_sentences]\n",
    "                else:\n",
    "                    sentences = record[\"sentences\"]\n",
    "                max_sentences = max(len(sentences), max_sentences)\n",
    "                \n",
    "                # номера предложений, которые в нашем саммари\n",
    "                if \"oracle_sentences\" not in record:\n",
    "                    calc_score = lambda x, y: calc_single_score(x, y, self.rouge)\n",
    "                    sentences_indicies = build_oracle_summary_greedy(text, summary, calc_score=calc_score, lower=self.lower, max_sentences=self.max_sentences)[1]\n",
    "                else:   \n",
    "                    sentences_indicies = record[\"oracle_sentences\"]\n",
    "                \n",
    "                # inputs - индексы слов в предложении\n",
    "                inputs = [bpe_processor.encode(sentence)[:self.max_sentence_length] for sentence in sentences]\n",
    "                max_sentence_length = max(max_sentence_length, max([len(tokens) for tokens in inputs]))\n",
    "                \n",
    "                # получение метки класса предложения\n",
    "                outputs = [int(i in sentences_indicies) for i in range(len(sentences))]\n",
    "                batch_inputs.append(inputs)\n",
    "                batch_outputs.append(outputs)\n",
    "\n",
    "            tensor_inputs = torch.zeros((self.batch_size, max_sentences, max_sentence_length), dtype=torch.long, device=self.device)\n",
    "            # we add index 2 for padding\n",
    "            # YOUR CODE\n",
    "            tensor_outputs = torch.ones((self.batch_size, max_sentences), dtype=torch.float32, device=self.device) * 2\n",
    "\n",
    "            for i, inputs in enumerate(batch_inputs):\n",
    "                for j, sentence_tokens in enumerate(inputs):\n",
    "                    tensor_inputs[i][j][:len(sentence_tokens)] = torch.tensor(sentence_tokens, dtype=torch.int64)\n",
    "\n",
    "            for i, outputs in enumerate(batch_outputs):\n",
    "                tensor_outputs[i][:len(outputs)] = torch.LongTensor(outputs)\n",
    "\n",
    "            tensor_outputs = tensor_outputs.long()\n",
    "            yield {\n",
    "                'inputs': tensor_inputs,\n",
    "                'outputs': tensor_outputs,\n",
    "                'records': batch_records\n",
    "            }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "id": "5ug9MIObdi03"
   },
   "outputs": [],
   "source": [
    "train_iterator = BatchIterator(ext_train_records, vocabulary, 32, bpe_processor, device=device)\n",
    "val_iterator = BatchIterator(ext_val_records, vocabulary, 32, bpe_processor, device=device)\n",
    "test_iterator = BatchIterator(ext_test_records, vocabulary, 32, bpe_processor, device=device)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "yPlJMg0_dQM-"
   },
   "source": [
    "## Extractor -  SummaRuNNer\n",
    " https://arxiv.org/pdf/1611.04230.pdf\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "4BSsnfe4t1uK"
   },
   "source": [
    "### Homework\n",
    "\n",
    "* В данной реализации в `outputs` в качестве padding используется индекс 0. Измените в функции \\_\\_iter__ индекс padding, чтобы он не совпадал с классом 0 или 1, например, 2.\n",
    "* В качестве criterion используйте `CrossEntropyLoss`вместо `BCEWithLogitsLoss`\n",
    "* Из-за смены criterion, вы уже должны подавать на вход criterion ни одно число, а logits для каждого класса. Перед подачей logits вы можете отфильтровать предсказания для класса padding. В этом пункте вам придется изменять файл `train_model.py`, а именно функциии `train` и `evaluate`.\n",
    "* Используйте два варианта обучения: c весами в `CrossEntropyLoss` и без\n",
    "* Также сравните `inference`, когда вы ранжируете logits, и когды вы выбирате предложения, у которых logits > 0, в двух вариантах обучения. \n",
    "* Реализуйте дополнительно характеристику предложения `novelty`. Как влияет добавление `novelty` на качество summary?\n",
    "* Постарайтесь улучшить качество модели, полученной на семинаре: $BLEU \\approx 0.45$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "id": "MAnENgEIemim"
   },
   "outputs": [],
   "source": [
    "PAD_IDX = 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "id": "iW7iS76KeEdO"
   },
   "outputs": [],
   "source": [
    "class SentenceEncoderRNN(nn.Module):\n",
    "    def __init__(self, input_size, embedding_dim, hidden_size, n_layers=3, dropout=0.3, bidirectional=True):\n",
    "        super().__init__()\n",
    "\n",
    "        num_directions = 2 if bidirectional else 1\n",
    "        assert hidden_size % num_directions == 0\n",
    "        hidden_size = hidden_size // num_directions\n",
    "\n",
    "        self.embedding_dim = embedding_dim\n",
    "        self.input_size = input_size\n",
    "        self.hidden_size = hidden_size\n",
    "        self.n_layers = n_layers\n",
    "        self.dropout = dropout\n",
    "        self.bidirectional = bidirectional\n",
    "\n",
    "        self.embedding_layer = nn.Embedding(input_size, embedding_dim)\n",
    "        self.rnn_layer = nn.LSTM(embedding_dim, hidden_size, n_layers, dropout=dropout, bidirectional=bidirectional, batch_first=True)\n",
    "        self.dropout_layer = nn.Dropout(dropout)\n",
    "\n",
    "    def forward(self, inputs, hidden=None):\n",
    "        embedded = self.dropout_layer(self.embedding_layer(inputs))\n",
    "        outputs, _ = self.rnn_layer(embedded, hidden)\n",
    "        sentences_embeddings = torch.mean(outputs, 1)\n",
    "        # [batch_size, hidden_size]\n",
    "        return sentences_embeddings\n",
    "\n",
    "class SentenceTaggerRNN(nn.Module):\n",
    "    def __init__(self,\n",
    "                 vocabulary_size,\n",
    "                 use_content=True,\n",
    "                 use_salience=True,\n",
    "                 use_novelty=True,\n",
    "                 token_embedding_dim = 128,\n",
    "                 sentence_encoder_hidden_size = 256,\n",
    "                 hidden_size = 256,\n",
    "                 bidirectional=True,\n",
    "                 sentence_encoder_n_layers = 2,\n",
    "                 sentence_encoder_dropout=0.3,\n",
    "                 sentence_encoder_bidirectional=True,\n",
    "                 n_layers = 2,\n",
    "                 dropout = 0.3):\n",
    "        \n",
    "        super().__init__()\n",
    "\n",
    "        num_directions = 2 if bidirectional else 1\n",
    "        assert hidden_size % num_directions == 0\n",
    "        hidden_size = hidden_size // num_directions\n",
    "\n",
    "        self.hidden_size = hidden_size\n",
    "        self.n_layers = n_layers\n",
    "        self.dropout = dropout\n",
    "        self.bidirectional = bidirectional\n",
    "\n",
    "        self.sentence_encoder = SentenceEncoderRNN(vocabulary_size, token_embedding_dim,\n",
    "                                                   sentence_encoder_hidden_size, sentence_encoder_n_layers, \n",
    "                                                   sentence_encoder_dropout, sentence_encoder_bidirectional)\n",
    "        \n",
    "        self.rnn_layer = nn.LSTM(sentence_encoder_hidden_size, hidden_size, n_layers, dropout=dropout,\n",
    "                           bidirectional=bidirectional, batch_first=True)\n",
    "        \n",
    "        self.dropout_layer = nn.Dropout(dropout)\n",
    "        self.content_linear_layer = nn.Linear(hidden_size * 2, 1)\n",
    "        self.document_linear_layer = nn.Linear(hidden_size * 2, hidden_size * 2)\n",
    "        self.salience_linear_layer = nn.Linear(hidden_size * 2, hidden_size * 2)\n",
    "        self.novelty_linear_layer = nn.Linear(hidden_size * 2, hidden_size * 2)\n",
    "        self.tanh_layer = nn.Tanh()\n",
    "\n",
    "        self.use_content = use_content\n",
    "        self.use_salience = use_salience\n",
    "        self.use_novelty = use_novelty\n",
    "\n",
    "    def forward(self, inputs, hidden=None):\n",
    "        # parameters of the probability\n",
    "        content = 0\n",
    "        salience = 0\n",
    "        novelty = 0\n",
    "\n",
    "        # [batch_size, seq num, seq_len]\n",
    "        batch_size = inputs.size(0)\n",
    "        sentences_count = inputs.size(1)\n",
    "        tokens_count = inputs.size(2)\n",
    "        inputs = inputs.reshape(-1, tokens_count)\n",
    "        # [batch_size * seq num, seq_len]\n",
    "\n",
    "        embedded_sentences = self.sentence_encoder(inputs)\n",
    "        embedded_sentences = self.dropout_layer(embedded_sentences.reshape(batch_size, sentences_count, -1))\n",
    "        # [batch_size *  seq num, seq_len, hidden_size] -> [batch_size, seq num, hidden_size]\n",
    "\n",
    "        outputs, _ = self.rnn_layer(embedded_sentences, hidden)\n",
    "        # [batch_size, seq num, hidden_size]\n",
    "\n",
    "        document_embedding = self.tanh_layer(self.document_linear_layer(torch.mean(outputs, 1)))\n",
    "        # [batch_size, hidden_size]\n",
    "\n",
    "        # W * h^T\n",
    "        if self.use_content:\n",
    "            content = self.content_linear_layer(outputs).squeeze(2) # 1-representation\n",
    "        # [batch_size, seq num]\n",
    "\n",
    "        # h^T * W * d\n",
    "        if self.use_salience:\n",
    "            salience = torch.bmm(outputs, self.salience_linear_layer(document_embedding).unsqueeze(2)).squeeze(2) # 2-representation\n",
    "        # [batch_size, seq num, hidden_size] * [batch_size, hidden_size, 1] = [batch_size, seq num, ]\n",
    "\n",
    "        if self.use_novelty:\n",
    "            # at every step add novelty to prediction of the sentence\n",
    "            predictions = content + salience\n",
    "            \n",
    "            # 0) initialize summary_representation and novelty by zeros\n",
    "            # https://github.com/kedz/nnsum/blob/master/nnsum/module/sentence_extractor/summarunner_sentence_extractor.py\n",
    "            novelty = torch.zeros((batch_size, sentences_count)).to(device)\n",
    "            summary_representation = torch.zeros((batch_size, self.hidden_size*2)).to(device)\n",
    "\n",
    "            for sentence_num in range(sentences_count):\n",
    "\n",
    "                # 1) take sentence_num_state from outputs(representation of the sentence with number sentence_num)\n",
    "                sentence_num_state = outputs[:, sentence_num, :]\n",
    "                # [seq num, hidden_size]\n",
    "\n",
    "                # 2) calculate novelty for current sentence\n",
    "                novelty[:, sentence_num] = -torch.bmm(sentence_num_state.unsqueeze(1), \n",
    "                                                      self.novelty_linear_layer(torch.tanh(summary_representation)).unsqueeze(2)).squeeze(2).squeeze(1)\n",
    "                # [seq num]\n",
    "                \n",
    "                # 3) add novelty to predictions\n",
    "                predictions[:, sentence_num] += novelty[:, sentence_num]\n",
    "\n",
    "                # 4) calculcate probability for current sentence\n",
    "                probabilities = torch.sigmoid(predictions[:, sentence_num])\n",
    "\n",
    "                # 5) add sentence_num_state with the weight which is equal to probability to summary_representation\n",
    "                summary_representation += torch.mv(sentence_num_state.transpose(0, 1), probabilities)\n",
    "\n",
    "            return predictions\n",
    "\n",
    "\n",
    "        return content + salience"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "QxpL3AtOrctD"
   },
   "source": [
    "## Model\n",
    "$P\\left(y_{j} = 1 \\mid \\mathbf{h}_{j}, \\mathbf{s}_{j}, \\mathbf{d}\\right)=\\sigma\\left(W_{c} \\mathbf{h}_{j} + \\mathbf{h}_{j}^{T} W_{s} \\mathbf{d}\\right)$\n",
    "--------------------"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "-QC1ZmuQfB7f",
    "outputId": "789c5941-88df-41ac-9667-cd102eec7a25"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Trainable params: 2927617\n"
     ]
    }
   ],
   "source": [
    "vocab_size = len(vocabulary)\n",
    "model = SentenceTaggerRNN(vocab_size).to(device)\n",
    "\n",
    "params_count = np.sum([p.numel() for p in model.parameters() if p.requires_grad])\n",
    "print(\"Trainable params: {}\".format(params_count))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "f9Q3aFHhgsB4",
    "outputId": "446b1acf-9994-4dc2-ecb5-135b2421690f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "sentence_encoder.embedding_layer.weight: 1280000\n",
      "sentence_encoder.rnn_layer.weight_ih_l0: 65536\n",
      "sentence_encoder.rnn_layer.weight_hh_l0: 65536\n",
      "sentence_encoder.rnn_layer.bias_ih_l0: 512\n",
      "sentence_encoder.rnn_layer.bias_hh_l0: 512\n",
      "sentence_encoder.rnn_layer.weight_ih_l0_reverse: 65536\n",
      "sentence_encoder.rnn_layer.weight_hh_l0_reverse: 65536\n",
      "sentence_encoder.rnn_layer.bias_ih_l0_reverse: 512\n",
      "sentence_encoder.rnn_layer.bias_hh_l0_reverse: 512\n",
      "sentence_encoder.rnn_layer.weight_ih_l1: 131072\n",
      "sentence_encoder.rnn_layer.weight_hh_l1: 65536\n",
      "sentence_encoder.rnn_layer.bias_ih_l1: 512\n",
      "sentence_encoder.rnn_layer.bias_hh_l1: 512\n",
      "sentence_encoder.rnn_layer.weight_ih_l1_reverse: 131072\n",
      "sentence_encoder.rnn_layer.weight_hh_l1_reverse: 65536\n",
      "sentence_encoder.rnn_layer.bias_ih_l1_reverse: 512\n",
      "sentence_encoder.rnn_layer.bias_hh_l1_reverse: 512\n",
      "rnn_layer.weight_ih_l0: 131072\n",
      "rnn_layer.weight_hh_l0: 65536\n",
      "rnn_layer.bias_ih_l0: 512\n",
      "rnn_layer.bias_hh_l0: 512\n",
      "rnn_layer.weight_ih_l0_reverse: 131072\n",
      "rnn_layer.weight_hh_l0_reverse: 65536\n",
      "rnn_layer.bias_ih_l0_reverse: 512\n",
      "rnn_layer.bias_hh_l0_reverse: 512\n",
      "rnn_layer.weight_ih_l1: 131072\n",
      "rnn_layer.weight_hh_l1: 65536\n",
      "rnn_layer.bias_ih_l1: 512\n",
      "rnn_layer.bias_hh_l1: 512\n",
      "rnn_layer.weight_ih_l1_reverse: 131072\n",
      "rnn_layer.weight_hh_l1_reverse: 65536\n",
      "rnn_layer.bias_ih_l1_reverse: 512\n",
      "rnn_layer.bias_hh_l1_reverse: 512\n",
      "content_linear_layer.weight: 256\n",
      "content_linear_layer.bias: 1\n",
      "document_linear_layer.weight: 65536\n",
      "document_linear_layer.bias: 256\n",
      "salience_linear_layer.weight: 65536\n",
      "salience_linear_layer.bias: 256\n",
      "novelty_linear_layer.weight: 65536\n",
      "novelty_linear_layer.bias: 256\n"
     ]
    }
   ],
   "source": [
    "for name, param in model.named_parameters():\n",
    "    print(f\"{name}: {param.numel()}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "id": "cTFOSjCYYeTU"
   },
   "outputs": [],
   "source": [
    "def train(model, iterator, optimizer, criterion, clip, train_history=None, valid_history=None):\n",
    "    model.train()\n",
    "    \n",
    "    epoch_loss = 0\n",
    "    history = []\n",
    "    for i, batch in enumerate(iterator):\n",
    "        \n",
    "        # YOUR CODE\n",
    "        model.train()\n",
    "        logits = model(batch[\"inputs\"])\n",
    "        target = batch[\"outputs\"]\n",
    "        prob = torch.sigmoid(logits)\n",
    "        input = torch.cat([(1-prob).unsqueeze(-1), prob.unsqueeze(-1)], dim=2)\n",
    "        input = input.view(-1,2)\n",
    "        target = target.view(-1,1)[:,0]\n",
    "        loss = criterion(input, target)\n",
    "\n",
    "        #choice_2\n",
    "        #output_1d = model(batch[\"inputs\"].to(device)).view(-1)\n",
    "        #batch_outputs_1d = batch[\"outputs\"].view(-1).to(device)\n",
    "        #output_1d = output_1d.unsqueeze(-1)\n",
    "        #output = torch.cat((-output_1d, output_1d),-1).to(device)\n",
    "        #loss = criterion(output, batch_outputs_1d)\n",
    "\n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        torch.nn.utils.clip_grad_norm_(model.parameters(), clip)\n",
    "        optimizer.step()\n",
    "        \n",
    "        epoch_loss += loss.item()\n",
    "        history.append(loss.item())\n",
    "        if (i+1)%10==0:\n",
    "            fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(12, 8))\n",
    "\n",
    "            clear_output(True)\n",
    "            ax[0].plot(history, label='train loss')\n",
    "            ax[0].set_xlabel('Batch')\n",
    "            ax[0].set_title('Train loss')\n",
    "            ax[0].grid()\n",
    "            if train_history is not None:\n",
    "                ax[1].plot(train_history, label='general train history')\n",
    "                ax[1].set_xlabel('Epoch')\n",
    "            if valid_history is not None:\n",
    "                ax[1].plot(valid_history, label='general valid history')\n",
    "            ax[1].grid()\n",
    "            plt.legend()\n",
    "            \n",
    "            plt.show()\n",
    "\n",
    "    return epoch_loss / len(iterator)\n",
    "\n",
    "def evaluate(model, iterator, criterion):\n",
    "    model.eval()\n",
    "    \n",
    "    epoch_loss = 0\n",
    "    history = []\n",
    "    \n",
    "    with torch.no_grad():\n",
    "        for i, batch in enumerate(iterator):\n",
    "            # YOUR CODE\n",
    "            model.eval()\n",
    "            logits = model(batch[\"inputs\"])\n",
    "            target = batch[\"outputs\"]\n",
    "            prob = torch.sigmoid(logits)\n",
    "            input = torch.cat([(1-prob).unsqueeze(-1), prob.unsqueeze(-1)], dim=2)\n",
    "            input = input.view(-1,2)\n",
    "            target = target.view(-1,1)[:,0]\n",
    "            loss = criterion(input, target)\n",
    "\n",
    "            #choice_2\n",
    "            #output_1d = model(batch[\"inputs\"].to(device)).view(-1)\n",
    "            #batch_outputs_1d = batch[\"outputs\"].view(-1).to(device)\n",
    "            #output_1d = output_1d.unsqueeze(-1)\n",
    "            #output = torch.cat((-output_1d, output_1d),-1).to(device)\n",
    "            #loss = criterion(output, batch_outputs_1d)\n",
    "\n",
    "            epoch_loss += loss.item()\n",
    "    return epoch_loss / len(iterator)\n",
    "\n",
    "def epoch_time(start_time, end_time):\n",
    "    elapsed_time = end_time - start_time\n",
    "    elapsed_mins = int(elapsed_time / 60)\n",
    "    elapsed_secs = int(elapsed_time - (elapsed_mins * 60))\n",
    "    return elapsed_mins, elapsed_secs\n",
    "\n",
    "def train_with_logs(model, train_iterator, valid_iterator, optimizer, criterion, N_EPOCHS, CLIP):\n",
    "    train_history = []\n",
    "    valid_history = []\n",
    "    best_valid_loss = float('inf')\n",
    "    \n",
    "    for epoch in range(N_EPOCHS):\n",
    "        \n",
    "        start_time = time.time()\n",
    "        \n",
    "        train_loss = train(model, train_iterator, optimizer, criterion, CLIP, train_history, valid_history)\n",
    "        valid_loss = evaluate(model, valid_iterator, criterion)\n",
    "        \n",
    "        end_time = time.time()\n",
    "        \n",
    "        epoch_mins, epoch_secs = epoch_time(start_time, end_time)\n",
    "        \n",
    "        if valid_loss < best_valid_loss:\n",
    "            best_valid_loss = valid_loss\n",
    "            torch.save(model.state_dict(), 'best-val-model.pt')\n",
    "        \n",
    "        train_history.append(train_loss)\n",
    "        valid_history.append(valid_loss)\n",
    "        print(f'Epoch: {epoch+1:02} | Time: {epoch_mins}m {epoch_secs}s')\n",
    "        print(f'\\tTrain Loss: {train_loss:.3f} | Train PPL: {math.exp(train_loss):7.3f}')\n",
    "        print(f'\\t Val. Loss: {valid_loss:.3f} |  Val. PPL: {math.exp(valid_loss):7.3f}')\n",
    "        \n",
    "def punct_detokenize(text):\n",
    "    text = text.strip()\n",
    "    punctuation = \",.!?:;%\"\n",
    "    closing_punctuation = \")]}\"\n",
    "    opening_punctuation = \"([}\"\n",
    "    for ch in punctuation + closing_punctuation:\n",
    "        text = text.replace(\" \" + ch, ch)\n",
    "    for ch in opening_punctuation:\n",
    "        text = text.replace(ch + \" \", ch)\n",
    "    res = [r'\"\\s[^\"]+\\s\"', r\"'\\s[^']+\\s'\"]\n",
    "    for r in res:\n",
    "        for f in re.findall(r, text, re.U):\n",
    "            text = text.replace(f, f[0] + f[2:-2] + f[-1])\n",
    "    text = text.replace(\"' s\", \"'s\").replace(\" 's\", \"'s\")\n",
    "    text = text.strip()\n",
    "    return text\n",
    "\n",
    "def postprocess(ref, hyp, is_multiple_ref=False, detokenize_after=False, tokenize_after=True):\n",
    "    if is_multiple_ref:\n",
    "        reference_sents = ref.split(\" s_s \")\n",
    "        decoded_sents = hyp.split(\"s_s\")\n",
    "        hyp = [w.replace(\"<\", \"&lt;\").replace(\">\", \"&gt;\").strip() for w in decoded_sents]\n",
    "        ref = [w.replace(\"<\", \"&lt;\").replace(\">\", \"&gt;\").strip() for w in reference_sents]\n",
    "        hyp = \" \".join(hyp)\n",
    "        ref = \" \".join(ref)\n",
    "    ref = ref.strip()\n",
    "    hyp = hyp.strip()\n",
    "    if detokenize_after:\n",
    "        hyp = punct_detokenize(hyp)\n",
    "        ref = punct_detokenize(ref)\n",
    "    if tokenize_after:\n",
    "        hyp = hyp.replace(\"@@UNKNOWN@@\", \"<unk>\")\n",
    "        hyp = \" \".join([token.text for token in razdel.tokenize(hyp)])\n",
    "        ref = \" \".join([token.text for token in razdel.tokenize(ref)])\n",
    "    return ref, hyp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "id": "lNZN5KyEeaXR"
   },
   "outputs": [],
   "source": [
    "from sklearn.utils.class_weight import compute_class_weight\n",
    "target = np.fromiter((v for batch in train_iterator for v in batch[\"outputs\"].detach().cpu().numpy().flatten()), dtype=int)\n",
    "target = target[target != PAD_IDX]\n",
    "class_weight = torch.tensor(compute_class_weight(class_weight='balanced',\n",
    "                                    classes=np.unique(target),\n",
    "                                    y=target), dtype=torch.float).to(device)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "vuv4Sh2Vj5Yr"
   },
   "source": [
    "## Inference"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "id": "wZ96X37bb_PV"
   },
   "outputs": [],
   "source": [
    "def inference_summarunner(model, iterator, top_k=3):\n",
    "    references = []\n",
    "    predictions = []\n",
    "\n",
    "    model.eval()\n",
    "    for batch in iterator:\n",
    "        logits = model(batch['inputs'])\n",
    "        sum_in = torch.argsort(logits, dim=1)[:, -top_k:]\n",
    "        \n",
    "        for i in range(len(batch['outputs'])):\n",
    "            summary = batch['records'][i]['summary'].lower()\n",
    "\n",
    "            try:\n",
    "              pred_summary = ' '.join([batch['records'][i]['sentences'][ind] for ind in sum_in.sort(dim=1)[0][i]])\n",
    "            except (IndexError):\n",
    "              continue\n",
    "              \n",
    "            summary, pred_summary = postprocess(summary, pred_summary)\n",
    "            references.append(summary)\n",
    "            predictions.append(pred_summary)\n",
    "    calc_scores(references, predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "id": "Mhz_OTvxIHue"
   },
   "outputs": [],
   "source": [
    "def inference_summarunner_th(model, iterator, th):\n",
    "    references = []\n",
    "    predictions = []\n",
    "\n",
    "    model.eval()\n",
    "    for batch in iterator:\n",
    "        logits = model(batch['inputs'])\n",
    "        logits = minmax_scale(logits.detach().cpu(), feature_range=(-1,1), axis=1, copy=True)\n",
    "        logits = torch.FloatTensor(logits)\n",
    "\n",
    "        for i in range(len(batch['outputs'])):\n",
    "\n",
    "            sum_in = torch.nonzero(logits[i]>th, as_tuple=True)[0]\n",
    "            summary = batch['records'][i]['summary'].lower()\n",
    "\n",
    "            try:\n",
    "              pred_summary = ' '.join([batch['records'][i]['sentences'][ind] for ind in sum_in])\n",
    "            except (IndexError):\n",
    "              continue\n",
    "              \n",
    "            summary, pred_summary = postprocess(summary, pred_summary)\n",
    "            references.append(summary)\n",
    "            predictions.append(pred_summary)\n",
    "    calc_scores(references, predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "id": "n2_pP5Vq-OgR"
   },
   "outputs": [],
   "source": [
    "def train_final(use_class_weights, N_EPOCHS, CLIP, lr=1e-5):\n",
    "    optimizer = optim.Adam(model.parameters(), lr)\n",
    "    if use_class_weights:\n",
    "        # weights depend on the number of objects of class 0 and 1\n",
    "        # YOUR CODE\n",
    "        criterion = nn.CrossEntropyLoss(ignore_index=PAD_IDX, weight=class_weight)\n",
    "    else:\n",
    "        criterion = nn.CrossEntropyLoss(ignore_index=PAD_IDX)\n",
    "    train_with_logs(model, train_iterator, val_iterator, optimizer, criterion, N_EPOCHS, CLIP)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "fJthF57V9Nil"
   },
   "source": [
    "### Эксперимент со случайными весами"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "mQy9sg5fz_Lh",
    "outputId": "0fe1f09a-e3bc-40c2-adb0-b723d418d316"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Count: 218\n",
      "Ref: морские порты россии – это не только одно из ключевых звеньев транспортной системы , но и залог для увеличения грузооборота и экспорта , а соответственно , и вклада в рост российской экономики . именно поэтому портовая деятельность вошла в числе национальных приоритетов . интересно , что морские порты в россии будут развиваться в основном на частные деньги . важно не только строительство самих портов , но и развитие все транспортной инфраструктуры , включая автодорожные и железнодорожные подходы к портам .\n",
      "Hyp: морские порты россии получат новое развитие в рамках нацпроекта « комплексный план модернизации и расширения магистральной инфраструктуры » . предполагается , порты дальневосточного , северо-западного , волго-каспийского и азово-черноморского бассейнов и другие к 2024 году увеличат мощность до 1,3 млрд тонн ( прирост — не менее 336 млн т ) . развитие портовой инфраструктуры и автодорожных подходов к портам , строительство 9 ледоколов и одного аварийно-спасательного судна включены в федеральный проект « морские порты » .\n",
      "BLEU:  0.48865667313605904\n",
      "ROUGE:  {'rouge-1': {'f': 0.3137427218497682, 'p': 0.32351224866090705, 'r': 0.32133252595517364}, 'rouge-2': {'f': 0.13829981505394773, 'p': 0.1421576493102655, 'r': 0.14349685707727494}, 'rouge-l': {'f': 0.2702361174775358, 'p': 0.2909183108473266, 'r': 0.28947278158848133}}\n"
     ]
    }
   ],
   "source": [
    "# Эксперимент со случайными весами\n",
    "model = SentenceTaggerRNN(vocab_size, use_novelty=False).to(device)\n",
    "inference_summarunner(model, test_iterator, 3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "KxQI14jt9TEA"
   },
   "source": [
    "## Обучим модель (use_novelty = False, use_class_weights = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 681
    },
    "id": "rwrhG4v71yts",
    "outputId": "83749bb4-803b-44bc-c23e-f3ba9d39c6e1"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 30 | Time: 0m 8s\n",
      "\tTrain Loss: 0.392 | Train PPL:   1.480\n",
      "\t Val. Loss: 0.394 |  Val. PPL:   1.483\n",
      "Count: 256\n",
      "Ref: для борьбы с с недобросовестными азс , недоливающими топливо , будут использоваться специальные приборы , а также проверки программного обеспечения . в этом вопросе росстандарт намерен использовать опыт , полученный в сфере регулирования работы игровых автоматов .\n",
      "Hyp: росстандарт готовит арсенал для проверок автозаправочных станций на недолив топлива : проверяться будет не конкретная колонка , а ее работа как системы , в том числе ее связь с оператором за кассой . контролеров при работе методом тайного покупателя вооружат скрытым прибором , который сможет « ловить » недолив на уровне 1 % . кроме того , новый гост защитит топливораздаточные колонки от приспособлений и вредоносного по , обманывающих расходомер .\n",
      "BLEU:  0.44225303105459424\n",
      "ROUGE:  {'rouge-1': {'f': 0.31182369116706404, 'p': 0.2891008385456055, 'r': 0.3551590796406379}, 'rouge-2': {'f': 0.13644316403105217, 'p': 0.12427555106592526, 'r': 0.160949996118185}, 'rouge-l': {'f': 0.2642992200064455, 'p': 0.25821069105441935, 'r': 0.31654668903240535}}\n"
     ]
    }
   ],
   "source": [
    "N_EPOCHS = 30\n",
    "CLIP = 1\n",
    "\n",
    "model = SentenceTaggerRNN(vocab_size, use_novelty = False).to(device)\n",
    "train_final(False, N_EPOCHS, CLIP)\n",
    "model.load_state_dict(torch.load('best-val-model.pt'))\n",
    "inference_summarunner(model, test_iterator, 3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "1oJqSvqSWPsF",
    "outputId": "f505f352-751c-4d56-c3dd-3fdd38c8567e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Count: 256\n",
      "Ref: экс-депутат рады александр ковтуненко устроил скандал в эфире телеканала « россия 1 » . он поддержал идею другого бывшего нардепа рефата чубарова « отметить 450 лет со дня сожжения москвы » крымским ханом девлет-гиреем . кроме того , ковтуненко назвал россию « бывшей украинской колонией » .\n",
      "Hyp: скандал в эфире программы « 60 минут » устроил бывший депутат верховной рады александр ковтуненко .\n",
      "BLEU:  0.45896064721741514\n",
      "ROUGE:  {'rouge-1': {'f': 0.30677258590915035, 'p': 0.3475847744841105, 'r': 0.30719857890650965}, 'rouge-2': {'f': 0.1391687270211789, 'p': 0.16045343040220947, 'r': 0.142284676020296}, 'rouge-l': {'f': 0.25079537484995995, 'p': 0.3078593333329834, 'r': 0.2722590707874975}}\n"
     ]
    }
   ],
   "source": [
    "# Инференс с отсечкой по threshold\n",
    "inference_summarunner_th(model, test_iterator, .5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "9r_zwQ3RAK5W"
   },
   "source": [
    "## Обучим модель (use_novelty = False, use_class_weights = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 681
    },
    "id": "gxAtAcgc-TGk",
    "outputId": "ee7bbb12-ef0a-42a1-a183-5cd0d8604cb6"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 30 | Time: 0m 9s\n",
      "\tTrain Loss: 0.626 | Train PPL:   1.870\n",
      "\t Val. Loss: 0.595 |  Val. PPL:   1.813\n",
      "Count: 256\n",
      "Ref: американские и иракские разведслужбы , а также пентагон считают , что украинский boeing 737 , потерпевший крушение в тегеране , при котором погибло 168 человек , был поражен ударом российской зенитной ракетной системы « тор м 1 » пво ирана . по мнению источника из пентагона , инцидент был случайным — противовоздушная оборона ирана , вероятно , активно работала по всем воздушным целям после ракетного обстрела со стороны сша .\n",
      "Hyp: самолет boeing 737-800 « международных авиалиний украины » , потерпевший крушение в тегеране , при котором погибло 168 человек , был поражен ударом зенитной ракетной системы – об этом newsweek рассказали представитель пентагона , высокопоставленный сотрудник американской разведки и сотрудник иракской разведки . по их словам , самолет был поражен ракетным выстрелом зрк « тор м 1 » пво ирана , известной в нато как gauntlet . по мнению пентагона , инцидент был случайным — противовоздушная оборона ирана , вероятно , активно работала по всем воздушным целям после ракетного обстрела со стороны сша .\n",
      "BLEU:  0.44674925060378357\n",
      "ROUGE:  {'rouge-1': {'f': 0.31390996957558887, 'p': 0.2943140273165956, 'r': 0.35416244710992795}, 'rouge-2': {'f': 0.13833083065518081, 'p': 0.127741709696567, 'r': 0.16134053532067033}, 'rouge-l': {'f': 0.2667387614491306, 'p': 0.26376149889633793, 'r': 0.3162068840205881}}\n"
     ]
    }
   ],
   "source": [
    "if os.path.isfile('/content/best-val-model.pt'):\n",
    "  os.remove('/content/best-val-model.pt') \n",
    "\n",
    "model = SentenceTaggerRNN(vocab_size, use_novelty = False).to(device)\n",
    "train_final(True, N_EPOCHS, CLIP)\n",
    "model.load_state_dict(torch.load('best-val-model.pt'))\n",
    "inference_summarunner(model, test_iterator, 3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "N0J6IcqDWQzA",
    "outputId": "261c3ab9-f1dd-4966-b99f-8aa99ebbbe54"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Count: 256\n",
      "Ref: россияне могут продлить себе новогодние каникулы , присоединив к ним свой отпуск , рассказали в роструде . работник может уйти в отпуск в любое время , в том числе – и с 9 января , если это было заранее оговорено с работодателем или прописано в графике отпусков . при этом за две недели до начала отпуска работник должен быть уведомлен о нем . ранее сми сообщали , что россиянам сократят новогодние каникулы , перенеся два выходных на май .\n",
      "Hyp: часть отпуска или весь отпуск полностью можно присоединить к новогодним каникулам или майским праздникам . однако для этого необходимо соблюсти ряд условий , сообщает « российская газета » со ссылкой на роструд .\n",
      "BLEU:  0.47284726323052\n",
      "ROUGE:  {'rouge-1': {'f': 0.30863746313393825, 'p': 0.33480940711001117, 'r': 0.31383286950162087}, 'rouge-2': {'f': 0.13981691268092417, 'p': 0.15178305485995597, 'r': 0.1464790104932912}, 'rouge-l': {'f': 0.2552304770642323, 'p': 0.29797353611064326, 'r': 0.2791777099292248}}\n"
     ]
    }
   ],
   "source": [
    "# Инференс с отсечкой по threshold\n",
    "inference_summarunner_th(model, test_iterator, .5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "9FfZRpFz9a0P"
   },
   "source": [
    "## Обучим модель (use_novelty = True, use_class_weights = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 681
    },
    "id": "cXCGixLWjHB5",
    "outputId": "56a21d3c-f475-4491-abf9-ba8c6c01dd6b"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 30 | Time: 0m 11s\n",
      "\tTrain Loss: 0.386 | Train PPL:   1.471\n",
      "\t Val. Loss: 0.383 |  Val. PPL:   1.466\n",
      "Count: 247\n",
      "Ref: опасные трещины найдены уже в десятках самолетов boeing 737 ng . американскому производителю , чьи новейшие лайнеры и так прикованы к земле из-за двух авиакатастроф , теперь грозят новые убытки .\n",
      "Hyp: подозрения , павшие почти две недели назад на один из самых массовых самолетов в мире boeing 737 ng , подтвердились и теперь могут привести к отзыву треснувших самолетов , а для boeing — обернуться новыми репутационными и финансовыми потерями . как сообщают американские сми , число дефектных самолетов уже измеряется десятками . подобные трещины , вовремя не замеченные , привели к нескольким крушениям советских пассажирских ан-10 .\n",
      "BLEU:  0.4037068609939744\n",
      "ROUGE:  {'rouge-1': {'f': 0.2783340459711522, 'p': 0.25098902451333444, 'r': 0.3290264292286601}, 'rouge-2': {'f': 0.11285383231203777, 'p': 0.09976572840422328, 'r': 0.1388092642181583}, 'rouge-l': {'f': 0.2328061830912851, 'p': 0.2236231520490234, 'r': 0.29288457125501555}}\n"
     ]
    }
   ],
   "source": [
    "if os.path.isfile('/content/best-val-model.pt'):\n",
    "  os.remove('/content/best-val-model.pt') \n",
    "model = SentenceTaggerRNN(vocab_size, use_novelty = True).to(device)\n",
    "train_final(False, N_EPOCHS, CLIP)\n",
    "model.load_state_dict(torch.load('best-val-model.pt'))\n",
    "inference_summarunner(model, test_iterator, 3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "YuYlFOcmWRv9",
    "outputId": "97e4d68e-1953-4ed7-f25c-3d84d91a0f1d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Count: 256\n",
      "Ref: в риме прошли переговоры глав мид и минобороны рф сергея лаврова и сергея шойгу с их итальянскими коллегами луиджи ди майо и лоренцо гуэрини в формате « 2 + 2 » . политики обсудили ситуацию в сирии , ливии , на украине , диалог россии с нато . отдельно министры обратили внимание на действия сша в сирии , а также необходимость для европы абстрагироваться от украинского кризиса в вопросах взаимодействия с москвой .\n",
      "Hyp: в риме 18 февраля министры иностранных дел и обороны рф сергей лавров и сергей шойгу встретились с итальянскими коллегами луиджи ди майо и лоренцо гуэрини .\n",
      "BLEU:  0.40778807169219483\n",
      "ROUGE:  {'rouge-1': {'f': 0.27139913061865617, 'p': 0.35052298860057696, 'r': 0.2666771127709237}, 'rouge-2': {'f': 0.11908598087033825, 'p': 0.1612904733698099, 'r': 0.11850144327873613}, 'rouge-l': {'f': 0.20727090641014093, 'p': 0.307008066781927, 'r': 0.2334457411642894}}\n"
     ]
    }
   ],
   "source": [
    "# Инференс с отсечкой по threshold\n",
    "inference_summarunner_th(model, test_iterator, 0.001)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "OquXfoB19flU"
   },
   "source": [
    "## Обучим модель (use_novelty = True, use_class_weights = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 681
    },
    "id": "uusu3bGj-Dch",
    "outputId": "354f072a-283e-4850-b600-5d7d90e73df6"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 30 | Time: 0m 11s\n",
      "\tTrain Loss: 0.624 | Train PPL:   1.866\n",
      "\t Val. Loss: 0.595 |  Val. PPL:   1.813\n",
      "Count: 256\n",
      "Ref: экс-депутат рады александр ковтуненко устроил скандал в эфире телеканала « россия 1 » . он поддержал идею другого бывшего нардепа рефата чубарова « отметить 450 лет со дня сожжения москвы » крымским ханом девлет-гиреем . кроме того , ковтуненко назвал россию « бывшей украинской колонией » .\n",
      "Hyp: скандал в эфире программы « 60 минут » устроил бывший депутат верховной рады александр ковтуненко . инцидент произошел на телеканале « россия 1 » . экс-нардеп поддержал идею другого бывшего украинского парламентария , лидера « меджлиса крымскотатарского народа » ( экстремистская организация , запрещена в рф ) рефата чубарова .\n",
      "BLEU:  0.44674925060378357\n",
      "ROUGE:  {'rouge-1': {'f': 0.31390996957558887, 'p': 0.2943140273165956, 'r': 0.35416244710992795}, 'rouge-2': {'f': 0.13833083065518084, 'p': 0.12774170969656698, 'r': 0.1613405353206704}, 'rouge-l': {'f': 0.26673876144913067, 'p': 0.263761498896338, 'r': 0.3162068840205884}}\n"
     ]
    }
   ],
   "source": [
    "if os.path.isfile('/content/best-val-model.pt'):\n",
    "  os.remove('/content/best-val-model.pt') \n",
    "model = SentenceTaggerRNN(vocab_size, use_novelty = True).to(device)\n",
    "train_final(True, N_EPOCHS, CLIP)\n",
    "model.load_state_dict(torch.load('best-val-model.pt'))\n",
    "inference_summarunner(model, test_iterator, 3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "swhZq8gJUVg0",
    "outputId": "9d7ab9a0-1c01-4aaa-beea-ac45632c1ae0"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Count: 256\n",
      "Ref: производство второго сезона сериала « лучшие в лос-анджелесе » с джессикой альбой в главной роли приостановлено из-за лесных пожаров к северу от лос-анджелеса . возгорание было вызвано опасными ветрами в сочетании с низкой влажностью и высокими температурами .\n",
      "Hyp: из-за очередной вспышки лесных пожаров к северу от лос-анджелеса приостанавливается и переносится производство ряда сериалов , среди которых второй сезон спин-офф экшн-хита « плохие парни » — « лучшие в лос-анджелесе » с джессикой альбой в главной роли , пишет deadline . в долине санта-кларита вспыхнули три пожара , вызванные опасными ветрами в сочетании с низкой влажностью и высокими температурами .\n",
      "BLEU:  0.4393667753880199\n",
      "ROUGE:  {'rouge-1': {'f': 0.31177951342430177, 'p': 0.34949156385123625, 'r': 0.29857562360274903}, 'rouge-2': {'f': 0.1415989489127609, 'p': 0.15963186612637742, 'r': 0.1373051684867328}, 'rouge-l': {'f': 0.2605694350081897, 'p': 0.30923779412157276, 'r': 0.26384215865501615}}\n"
     ]
    }
   ],
   "source": [
    "# Инференс с отсечкой по threshold\n",
    "inference_summarunner_th(model, test_iterator, .001)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "HGc4IV8dzyBP"
   },
   "source": [
    "## Вывод:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "LnyBgLtKg8A6"
   },
   "source": [
    "<table class=\"\">\n",
    "\t<tbody>\n",
    "\t\t<tr>\n",
    "\t\t\t<td colspan=\"3\" align=\"center\"><b>BLEU</b <b> (inference without treshold)</td>\n",
    "\t\t</tr>\n",
    "\t\t<tr>\n",
    "\t\t\t<td></td>\n",
    "\t\t\t<td align=\"center\"><b>Class_weight</b <b> (True)</b</td>\n",
    "\t\t\t<td align=\"center\"><b>Class_weight</b <b> (False)</td>\n",
    "\t\t</tr>\n",
    "\t\t<tr>\n",
    "\t\t\t<td><b>Novelty</b <b> (True)</td>\n",
    "\t\t\t<td align=\"center\">0.45</td>\n",
    "\t\t\t<td align=\"center\">0.4</td>\n",
    "\t\t</tr>\n",
    "\t\t<tr>\n",
    "\t\t\t<td><b>Novelty</b <b> (False)</b</td>\n",
    "\t\t\t<td align=\"center\">0.45</td>\n",
    "\t\t\t<td align=\"center\">0.45</td>\n",
    "\t\t</tr>\n",
    "\t</tbody>\n",
    "</table>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "HfdeoE1eg8H1"
   },
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "sEwr4GVSnDj9"
   },
   "source": [
    "<table class=\"\">\n",
    "\t<tbody>\n",
    "\t\t<tr>\n",
    "\t\t\t<td colspan=\"3\" align=\"center\"><b>BLEU</b <b> (inference with treshold)</td>\n",
    "\t\t</tr>\n",
    "\t\t<tr>\n",
    "\t\t\t<td></td>\n",
    "\t\t\t<td align=\"center\"><b>Class_weight</b <b> (True)</b</td>\n",
    "\t\t\t<td align=\"center\"><b>Class_weight</b <b> (False)</td>\n",
    "\t\t</tr>\n",
    "\t\t<tr>\n",
    "\t\t\t<td><b>Novelty</b <b> (True)</td>\n",
    "\t\t\t<td align=\"center\">0.44</td>\n",
    "\t\t\t<td align=\"center\">0.41</td>\n",
    "\t\t</tr>\n",
    "\t\t<tr>\n",
    "\t\t\t<td><b>Novelty</b <b> (False)</b</td>\n",
    "\t\t\t<td align=\"center\">0.47</td>\n",
    "\t\t\t<td align=\"center\">0.46</td>\n",
    "\t\t</tr>\n",
    "\t</tbody>\n",
    "</table>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "76CxOXZ7oGrq"
   },
   "source": [
    "Максимальное значение BLEU ≈ 0.47 получено с использованием class_weight при выборе предложений, у которых logit > 0.5. Использование novelty не оказывает видимого влияния на увеличение BLEU. При этом, эксперименты со случайными весами иногда показывали еще бОльшее значение ≈ 0.49, что, конечно, не подлежит серьезному рассмотрению. При этом использование эмпирически подобранных значений treshold во всех случаях позволяет добиться более высоких значений BLEU."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "s1U_ZF1cnCqa"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "[homework]summarization.ipynb",
   "provenance": []
  },
  "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": "3.8.8"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "200c9d6fa70648b2917dcd85001740d1": {
     "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": ""
     }
    },
    "32e5625802744f3f96fada6cbb429cac": {
     "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
     }
    },
    "32f724270b614855b691cb7338c3ee6a": {
     "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
     }
    },
    "3abe49bef0b44456b86ff5c8ecaec6ae": {
     "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_49d8eed23498449ca5a6e5fd491361b5",
      "placeholder": "​",
      "style": "IPY_MODEL_7d4ef9892dc144899d2a00555aa75349",
      "value": " 256/? [00:49&lt;00:00,  5.61it/s]"
     }
    },
    "3f4692e0c3814364a2e1dba7947a53cd": {
     "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": ""
     }
    },
    "4342980193114cceb3cc871eaeb6f651": {
     "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": ""
     }
    },
    "43c237a6422b4bdeb32b65838fd033de": {
     "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
     }
    },
    "49d8eed23498449ca5a6e5fd491361b5": {
     "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
     }
    },
    "4b29e97fabd74fea9c2e45a1668cf4ed": {
     "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": "danger",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_56bc5f6da9734bcc8d6df691a64ec3c8",
      "max": 1,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_3f4692e0c3814364a2e1dba7947a53cd",
      "value": 1
     }
    },
    "56bc5f6da9734bcc8d6df691a64ec3c8": {
     "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": "20px"
     }
    },
    "5a9036fe34b44f5f86232dbe71b30cd4": {
     "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": "20px"
     }
    },
    "5ab446dfdb9f41bebb504f0f44e6fe0a": {
     "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_43c237a6422b4bdeb32b65838fd033de",
      "placeholder": "​",
      "style": "IPY_MODEL_a8d4eff3041e4b2f9ca0fe608b629fc8",
      "value": ""
     }
    },
    "5c3fe8bfd6ba46d583aa840568f43e86": {
     "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
     }
    },
    "64bccb899a7348649b9f9fb1ff169507": {
     "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
     }
    },
    "65bad71fdd534716bcf2598e0cedadce": {
     "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": ""
     }
    },
    "78a3f4154ee340acb57ba40ed14f6f93": {
     "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": ""
     }
    },
    "7d4ef9892dc144899d2a00555aa75349": {
     "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": ""
     }
    },
    "8357ebe3c260419eb58bdcae946f776a": {
     "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
     }
    },
    "85c35253f55b4c608caddc371b3d1494": {
     "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_64bccb899a7348649b9f9fb1ff169507",
      "placeholder": "​",
      "style": "IPY_MODEL_200c9d6fa70648b2917dcd85001740d1",
      "value": ""
     }
    },
    "8e3b02911605486a9728833cfcde4181": {
     "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
     }
    },
    "915270cc25724dcabb18779e4b1f8fe8": {
     "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": "danger",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_963c405588f8448f984edfb693ad2c7e",
      "max": 1,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_65bad71fdd534716bcf2598e0cedadce",
      "value": 1
     }
    },
    "963c405588f8448f984edfb693ad2c7e": {
     "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": "20px"
     }
    },
    "9f9b84b6da494f21bf9355fc2f06b833": {
     "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_8357ebe3c260419eb58bdcae946f776a",
      "placeholder": "​",
      "style": "IPY_MODEL_78a3f4154ee340acb57ba40ed14f6f93",
      "value": " 2048/? [05:42&lt;00:00,  5.00it/s]"
     }
    },
    "a0eea8faa4234c95b053bbd473a7d526": {
     "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
     }
    },
    "a1c7e56bfbe04fcebf4bc4630a8dc246": {
     "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": "danger",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_5a9036fe34b44f5f86232dbe71b30cd4",
      "max": 1,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_4342980193114cceb3cc871eaeb6f651",
      "value": 1
     }
    },
    "a8d4eff3041e4b2f9ca0fe608b629fc8": {
     "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": ""
     }
    },
    "a9044743cd094564a29fa4020df1ba7f": {
     "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_32e5625802744f3f96fada6cbb429cac",
      "placeholder": "​",
      "style": "IPY_MODEL_aaa3d9e464be467096fcbf5a722fe8d6",
      "value": " 256/? [00:47&lt;00:00,  6.84it/s]"
     }
    },
    "aaa3d9e464be467096fcbf5a722fe8d6": {
     "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": ""
     }
    },
    "ce408e42ec60499ab70b08419f4ba4a8": {
     "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_8e3b02911605486a9728833cfcde4181",
      "placeholder": "​",
      "style": "IPY_MODEL_dd07dd4af55040c9af1d3b092234c60e",
      "value": ""
     }
    },
    "dd07dd4af55040c9af1d3b092234c60e": {
     "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": ""
     }
    },
    "e2d49cdfe422414a873fcc36d13cb0ec": {
     "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_85c35253f55b4c608caddc371b3d1494",
       "IPY_MODEL_4b29e97fabd74fea9c2e45a1668cf4ed",
       "IPY_MODEL_3abe49bef0b44456b86ff5c8ecaec6ae"
      ],
      "layout": "IPY_MODEL_a0eea8faa4234c95b053bbd473a7d526"
     }
    },
    "e36e0519cb484752a0616c192903c392": {
     "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_5ab446dfdb9f41bebb504f0f44e6fe0a",
       "IPY_MODEL_915270cc25724dcabb18779e4b1f8fe8",
       "IPY_MODEL_a9044743cd094564a29fa4020df1ba7f"
      ],
      "layout": "IPY_MODEL_32f724270b614855b691cb7338c3ee6a"
     }
    },
    "f685af33f30641c3a7748055889735b9": {
     "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_ce408e42ec60499ab70b08419f4ba4a8",
       "IPY_MODEL_a1c7e56bfbe04fcebf4bc4630a8dc246",
       "IPY_MODEL_9f9b84b6da494f21bf9355fc2f06b833"
      ],
      "layout": "IPY_MODEL_5c3fe8bfd6ba46d583aa840568f43e86"
     }
    }
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}