BERT_PR_v0.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "BERT-PR-v0.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyMim+ZK6Mb58aVbEaYMtjsO",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/finardi/tutos/blob/master/BERT_PR_v0.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "L9codGRHsqjh",
        "colab_type": "text"
      },
      "source": [
        "# BERT - Perguntas e Respostas PR\n",
        "\n",
        "### TL;DR\n",
        "- Quando o BERT é aplicado na task de PR, em geral treinamos o modelo em um dataset similar ao Stanford Question and Answering (SQuAD). Nesse dataset, a Pergunta e um trecho que contém a resposta é enviado ao modelo e o BERT tem a função de encontrar o trecho do texto que contém a resposta. \n",
        "\n",
        "- Exemplos do SQuAD [aqui](https://rajpurkar.github.io/SQuAD-explorer/explore/1.1/dev/Super_Bowl_50.html?model=r-net+%20(ensemble)%20(Microsoft%20Research%20Asia)&version=1.1).\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Jy7OBMBNuRZF",
        "colab_type": "text"
      },
      "source": [
        "![Input format for QA](http://www.mccormickml.com/assets/BERT/SQuAD/input_formatting.png)\n",
        "\n",
        "#### As duas partes do texto são separadas pelo token especial `[SEP]`.\n",
        "\n",
        "#### O BERT também usa \"`Embeddings de segmento`\" para diferenciar a pergunta do texto de referência. Estes são simplesmente dois embeddings (para os segmentos \"`A`\" e \"`B`\") que o BERT aprendeu e que adiciona aos embeddings de token antes de alimentá-los na camada de entrada"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "feaV__XIuZyo",
        "colab_type": "text"
      },
      "source": [
        "#### O BERT precisa destacar um \"intervalo\" de texto contendo a resposta - isso é representado simplesmente pela previsão de qual token marca o início da resposta e qual marca o fim.\n",
        "\n",
        "![Iniciar classificação de token](http://www.mccormickml.com/assets/BERT/SQuAD/start_token_classification.png)\n",
        "\n",
        "\n",
        "#### Para cada token no texto, alimentamos seu embedding final no classificador de token inicial. O classificador de token inicial tem apenas um único conjunto de pesos (representado pelo retângulo \"inicial\" azul na ilustração acima) que se aplica a cada palavra.\n",
        "\n",
        "#### Depois de pegar o produto escalar entre os embeddings de saída e os pesos 'iniciais', aplicamos a ativação softmax para produzir uma distribuição de probabilidade sobre todas as palavras. A palavra que tiver a maior probabilidade de ser o token inicial é aquela que escolhemos.\n",
        "\n",
        "#### Repetimos esse processo para o token final - com um outro vetor de peso.\n",
        "\n",
        "![Terminar classificação de token](http://www.mccormickml.com/assets/BERT/SQuAD/end_token_classification.png)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "skPsHaffwMxV",
        "colab_type": "text"
      },
      "source": [
        "# EXEMPLO"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DZSOCJTxwjt4",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!pip install -q transformers\n",
        "\n",
        "import torch \n",
        "\n",
        "#modelo\n",
        "from transformers import BertForQuestionAnswering\n",
        "model = BertForQuestionAnswering.from_pretrained('bert-large-uncased-whole-word-masking-finetuned-squad')\n",
        "\n",
        "#tokenizador\n",
        "from transformers import BertTokenizer\n",
        "tokenizer = BertTokenizer.from_pretrained('bert-large-uncased-whole-word-masking-finetuned-squad')"
      ],
      "execution_count": 93,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aAZFKD4Gwt8-",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "question = \"Qual o prazo mínimo para aumentar meu limite?\"\n",
        "\n",
        "answer_text = \"O BERT foi treinado no dataset em inglês SQuAD e não possui \\\n",
        "relação com textos de aumento de limite. O prazo mínimo para \\\n",
        "aumento do limite é de 6 meses. Isso é desnecessário.\""
      ],
      "execution_count": 94,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "byqFnINwxeJK",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "c0292c00-877a-426b-f83a-eb70c5351d38"
      },
      "source": [
        "# Tokeniza o texto da pergunta e resposta\n",
        "input_ids = tokenizer.encode(question, answer_text)\n",
        "\n",
        "print(f'A entrada total possui {format(len(input_ids)):} tokens')"
      ],
      "execution_count": 95,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "A entrada total possui 80 tokens\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "W02CTBWkx25P",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "d223ea91-4da2-473c-b106-9987a2063bb2"
      },
      "source": [
        "# O BERT somente precisa dos tokens IDs, porém, vamos inspecionar o comportamento do tokenizador.\n",
        "tokens = tokenizer.convert_ids_to_tokens(input_ids)\n",
        "\n",
        "# Para cada token e seu id...\n",
        "for token, id in zip(tokens, input_ids):\n",
        "    \n",
        "    # Se o token é [SEP], adiciona um espaço ao redor.\n",
        "    if id == tokenizer.sep_token_id:\n",
        "        print('')\n",
        "    \n",
        "    # Print os tokens como strings e os IDs em duas colunas.\n",
        "    print(f'{token:<12} {id:>6,}')\n",
        "\n",
        "    if id == tokenizer.sep_token_id:\n",
        "        print('')    "
      ],
      "execution_count": 96,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[CLS]           101\n",
            "qu           24,209\n",
            "##al          2,389\n",
            "o             1,051\n",
            "pr           10,975\n",
            "##az         10,936\n",
            "##o           2,080\n",
            "mini          7,163\n",
            "##mo          5,302\n",
            "para         11,498\n",
            "au            8,740\n",
            "##ment        3,672\n",
            "##ar          2,906\n",
            "me            2,033\n",
            "##u           2,226\n",
            "limit         5,787\n",
            "##e           2,063\n",
            "?             1,029\n",
            "\n",
            "[SEP]           102\n",
            "\n",
            "o             1,051\n",
            "bert         14,324\n",
            "f             1,042\n",
            "##oi         10,448\n",
            "tre          29,461\n",
            "##ina         3,981\n",
            "##do          3,527\n",
            "no            2,053\n",
            "data          2,951\n",
            "##set        13,462\n",
            "em            7,861\n",
            "ing          13,749\n",
            "##les         4,244\n",
            "squad         4,686\n",
            "e             1,041\n",
            "na            6,583\n",
            "##o           2,080\n",
            "po           13,433\n",
            "##ss          4,757\n",
            "##ui         10,179\n",
            "re            2,128\n",
            "##la          2,721\n",
            "##cao        20,808\n",
            "com           4,012\n",
            "text          3,793\n",
            "##os          2,891\n",
            "de            2,139\n",
            "au            8,740\n",
            "##mento      23,065\n",
            "de            2,139\n",
            "limit         5,787\n",
            "##e           2,063\n",
            ".             1,012\n",
            "o             1,051\n",
            "pr           10,975\n",
            "##az         10,936\n",
            "##o           2,080\n",
            "mini          7,163\n",
            "##mo          5,302\n",
            "para         11,498\n",
            "au            8,740\n",
            "##mento      23,065\n",
            "do            2,079\n",
            "limit         5,787\n",
            "##e           2,063\n",
            "e             1,041\n",
            "de            2,139\n",
            "6             1,020\n",
            "me            2,033\n",
            "##ses         8,583\n",
            ".             1,012\n",
            "iss          26,354\n",
            "##o           2,080\n",
            "e             1,041\n",
            "des           4,078\n",
            "##ne          2,638\n",
            "##ces         9,623\n",
            "##sari       22,740\n",
            "##o           2,080\n",
            ".             1,012\n",
            "\n",
            "[SEP]           102\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zm208EApN16k",
        "colab_type": "text"
      },
      "source": [
        "#### Nós concatenamos a `Pergunta` e `texto com a resposta` juntos, mas o BERT ainda precisa de uma maneira de distingui-los. O BERT tem dois embeddings especiais de \"Segmento\", um para o segmento \"`A`\" e outro para o segmento \"`B`\". Antes que a palavra embeddings vá para as camadas BERT, o segmento de embedding `A` precisa ser adicionado aos tokens da `Pergunta`, e o segmento de embedding `B` precisa ser adicionado a cada um dos tokens `texto com resposta`.\n",
        "\n",
        "#### Essas adições são tratadas pela biblioteca `transformer`, e tudo o que precisamos fazer é especificar um '`0`' ou '`1`' para cada token.\n",
        "\n",
        "#### Nota: A biblioteca `transformers` denota os embeddings de segmento por `token_type_ids`"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uJ7PRx2dKqFN",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Pesquise os input_ids para a primeira instância do token [SEP].\n",
        "sep_index = input_ids.index(tokenizer.sep_token_id)\n",
        "\n",
        "# O nro de tokens de segmento A inclui o token [SEP].\n",
        "num_seg_a = sep_index + 1\n",
        "\n",
        "# A sobra é o segmento B.\n",
        "num_seg_b = len(input_ids) - num_seg_a\n",
        "\n",
        "# Uma lista de 0s and 1s para os segmentos.\n",
        "segment_ids = [0]*num_seg_a + [1]*num_seg_b\n",
        "\n",
        "# Verifica se os tamanhos estão corretos\n",
        "assert len(segment_ids) == len(input_ids)"
      ],
      "execution_count": 97,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "a30sBTcqQv6X",
        "colab_type": "text"
      },
      "source": [
        "### Observação lateral: onde está o padding?\n",
        "\n",
        "> ####  O [código de exemplo original](https://huggingface.co/transformers/model_doc/bert.html?highlight=bertforquestionanswering#transformers.BertForQuestionAnswering) não executa nenhum padding (note que estamos alimentando apenas 1 *exemplo*). Se, em vez disso, inserirmos um batch de exemplos, precisaremos do padding e máscara de atenção para dizer ao BERT para ignorar os tokens de paddings."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CNwhEw0kQPBN",
        "colab_type": "text"
      },
      "source": [
        "####  Vamos alimentar o modelo com o nosso exemplo\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DQiKr6Aw-YTg",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "start_scores, end_scores = model(\n",
        "    torch.tensor([input_ids]), \n",
        "    token_type_ids=torch.tensor([segment_ids])) "
      ],
      "execution_count": 98,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mBdS_QkIbDzh",
        "colab_type": "text"
      },
      "source": [
        "#### A resposta é selecionada pelos scores (probabilidades) mais altos da palavra de início e fim da resposta. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LeUQ44hAJmn9",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "b36fb8dc-c119-487a-a16e-8e63420569a9"
      },
      "source": [
        "# Encontrar os tokens com o maior score de início e fim.\n",
        "answer_start = torch.argmax(start_scores)\n",
        "answer_end = torch.argmax(end_scores)\n",
        "\n",
        "# Combinar os tokens da resposta e printar.\n",
        "answer = ' '.join(tokens[answer_start:answer_end+1])\n",
        "\n",
        "print('Answer: \"' + answer + '\"')"
      ],
      "execution_count": 99,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Answer: \"o pr ##az ##o mini ##mo para au ##mento do limit ##e e de 6 me ##ses\"\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "twMUWmr2brRw",
        "colab_type": "text"
      },
      "source": [
        "#### Parece correto, vamos reconstruir o texto das subwords.\n",
        "\n",
        "> #### **Nota**: É um pouco ingênuo escolher os scores mais altos da palavra de início e fim. Um problema desse raciocínio é que podemos pensar que possa existir uma predição de palavra final que está antes da palavra inicial?! Um raciocínio mais refinado do que de fato acontece é a escolha de score total mais alta fim $\\geq$ início (monótona não decrescente)."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Khral6HZXCuI",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "e3511a9b-c43a-406d-d919-00832c310007"
      },
      "source": [
        "# Começa com o token de inicio.\n",
        "answer = tokens[answer_start]\n",
        "\n",
        "# Seleciona o resto dos tokens da resposta.\n",
        "for i in range(answer_start + 1, answer_end + 1):\n",
        "    \n",
        "    # Se o token é uma subword, entao recombinamos com o token anterior.\n",
        "    if tokens[i][0:2] == '##':\n",
        "        answer += tokens[i][2:]\n",
        "    \n",
        "    # Se não, adiciona um espaço ao token.\n",
        "    else:\n",
        "        answer += ' ' + tokens[i]\n",
        "\n",
        "print('Answer: \"' + answer + '\"')"
      ],
      "execution_count": 100,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Answer: \"o prazo minimo para aumento do limite e de 6 meses\"\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-hh6nkIdXq-O",
        "colab_type": "text"
      },
      "source": [
        "## Visualização  dos scores"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-hG2YCHYXtg-",
        "colab_type": "text"
      },
      "source": [
        "I was curious to see what the scores were for all of the words. The following cells generate bar plots showing the start and end scores for every word in the input."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gkKFa73eJkPE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "\n",
        "sns.set(style='darkgrid')\n",
        "plt.rcParams[\"figure.figsize\"] = (16,8)"
      ],
      "execution_count": 101,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "W7kazkb2iEuQ",
        "colab_type": "text"
      },
      "source": [
        "### Recuperando todas os scores iniciais e finais e todos os tokens rótulados no eixo x."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "C56AtMg2UBxN",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Pull the scores out of PyTorch Tensors and convert them to 1D numpy arrays.\n",
        "s_scores = start_scores.detach().numpy().flatten()\n",
        "e_scores = end_scores.detach().numpy().flatten()\n",
        "\n",
        "# We'll use the tokens as the x-axis labels. In order to do that, they all need\n",
        "# to be unique, so we'll add the token index to the end of each one.\n",
        "token_labels = []\n",
        "for (i, token) in enumerate(tokens):\n",
        "    token_labels.append('{:} - {:>2}'.format(token, i))\n"
      ],
      "execution_count": 102,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "y6OAV1dL3-UB",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 571
        },
        "outputId": "074dded1-3d2b-443f-c480-29e56f6f3ba8"
      },
      "source": [
        "# Cria um barplot que mostra o início do score da palavra para todos os tokens.\n",
        "ax = sns.barplot(x=token_labels, y=s_scores, ci=None)\n",
        "\n",
        "# xlabels na vertical.\n",
        "ax.set_xticklabels(ax.get_xticklabels(), rotation=90, ha=\"center\")\n",
        "\n",
        "# Adiciona grid vertical.\n",
        "ax.grid(True)\n",
        "\n",
        "plt.title('Scores da palavra de início')\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 103,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6tXEqIp-Tzou",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 571
        },
        "outputId": "62c72664-be63-4830-fe4a-ed75c6cc8d13"
      },
      "source": [
        "# Cria um barplot que mostra o início do score da palavra para todos os tokens.\n",
        "ax = sns.barplot(x=token_labels, y=e_scores, ci=None)\n",
        "\n",
        "# xlabels na vertical.\n",
        "ax.set_xticklabels(ax.get_xticklabels(), rotation=90, ha=\"center\")\n",
        "\n",
        "# Adiciona grid vertical.\n",
        "ax.grid(True)\n",
        "\n",
        "plt.title('Scores da palavra de fim')\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 104,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "awgi7Z_a9KSq",
        "colab_type": "text"
      },
      "source": [
        "### Visão Alternativa\n",
        "\n",
        "#### Pontuações inicial e final em um gráfico de barra única"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "m4VUk6R05uXS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import pandas as pd\n",
        "\n",
        "# Armazena os tokens e scores em um DataFrame. Cada token terá duas linhas, uma para \n",
        "# a pontuação inicial e outra para o final, a coluna \"marcador\" irá diferenciá-los.\n",
        "for (i, token_label) in enumerate(token_labels):\n",
        "\n",
        "    # Adiciona os tokens de inicio como uma única linha.\n",
        "    scores.append({'tokens': token_label, \n",
        "                   'score': s_scores[i],\n",
        "                   'marker': 'start'})\n",
        "    \n",
        "    # Adiciona os tokens do fim com uma outra linha.\n",
        "    scores.append({'tokens': token_label, \n",
        "                   'score': e_scores[i],\n",
        "                   'marker': 'end'})\n",
        "    \n",
        "df = pd.DataFrame(scores)"
      ],
      "execution_count": 105,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "07xyo-I97Ntt",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 489
        },
        "outputId": "f6ad94f7-350b-4e88-f1d7-5afa0a2321e5"
      },
      "source": [
        "g = sns.catplot(x=\"tokens\", y=\"score\", hue=\"marker\", data=df,\n",
        "                kind=\"bar\", height=6, aspect=4)\n",
        "\n",
        "g.set_xticklabels(g.ax.get_xticklabels(), rotation=90, ha=\"center\")\n",
        "\n",
        "g.ax.grid(True)"
      ],
      "execution_count": 106,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1793.47x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "8UyBYNmeegGf"
      },
      "source": [
        "## Mais Exemplos"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MWtcRpPef-Ce",
        "colab_type": "text"
      },
      "source": [
        "#### Criar uma função para testes mais rápidos."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rH8NbBlsfxZ_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def answer_question(question, answer_text):\n",
        "    '''\n",
        "    Toma uma string `pergunta` e uma string `texto que contém a resposta`, e \n",
        "    identifica as palavras dentro do `answer_text` que são a resposta. Imprime-os.\n",
        "    '''\n",
        "    # ======== Tokeniza ========\n",
        "    # Aplica o tokenizador no texto da entrada, tradando-o como um text-pair.\n",
        "    input_ids = tokenizer.encode(question, answer_text)\n",
        "\n",
        "    # Imprime o tamanho da sequência.\n",
        "    print(f'A query tem {format(len(input_ids)):} tokens.\\n')\n",
        "\n",
        "    # ======== Configura o Segment IDs ========\n",
        "    # Procura no input_ids o token [SEP].\n",
        "    sep_index = input_ids.index(tokenizer.sep_token_id)\n",
        "\n",
        "    # configura o número de segments A que inclui o token [SEP].\n",
        "    num_seg_a = sep_index + 1\n",
        "\n",
        "    # o resto é o segumento B.\n",
        "    num_seg_b = len(input_ids) - num_seg_a\n",
        "\n",
        "    # estrutura o segment_ids em uma lista de 0s e 1s.\n",
        "    segment_ids = [0]*num_seg_a + [1]*num_seg_b\n",
        "\n",
        "    # confere se os tamanhos estão corretos.\n",
        "    assert len(segment_ids) == len(input_ids)\n",
        "\n",
        "    # ======== Avalia ========\n",
        "    # roda o examplo no modelo.\n",
        "    start_scores, end_scores = model(\n",
        "        torch.tensor([input_ids]), \n",
        "        token_type_ids=torch.tensor([segment_ids]))\n",
        "\n",
        "    # ======== Reconstrói a resposta ========\n",
        "    # Encontra os tokens com o score mais alto de inicio e fim.\n",
        "    answer_start = torch.argmax(start_scores)\n",
        "    answer_end = torch.argmax(end_scores)\n",
        "\n",
        "    # Converte os ids para string.\n",
        "    tokens = tokenizer.convert_ids_to_tokens(input_ids)\n",
        "\n",
        "    # Começa com o token de inicio.\n",
        "    answer = tokens[answer_start]\n",
        "\n",
        "    # Seleciona o resto da resposta.\n",
        "    for i in range(answer_start + 1, answer_end + 1):\n",
        "        \n",
        "        # Se é um token de subword, então recombina com o token anterior.\n",
        "        if tokens[i][0:2] == '##':\n",
        "            answer += tokens[i][2:]\n",
        "        \n",
        "        # Se não, adiciona um espaço ao token.\n",
        "        else:\n",
        "            answer += ' ' + tokens[i]\n",
        "\n",
        "    print('Answer: \"' + answer + '\"')"
      ],
      "execution_count": 107,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "y4VPq6FdjxyX",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "20f8f6ad-7781-405f-9120-6a4f43de6907"
      },
      "source": [
        "new_text = \"Se o seu cartão foi furtado ou roubado, você precisa entrar em contato \\\n",
        "com a central de cartões, os telefones são 00-12345-7890 ou 0800-123-00000. Você \\\n",
        "também pode entrar em contato por whatsapp com os nossos operadores nos seguintes \\\n",
        "horários: 06h00 até as 22h30 de Segunda a Segunda. Aqui é somente uma frase sem \\\n",
        "conexão com o texto, alias vale enfatizar que os textos de respostas podem ter até \\\n",
        "512 tokens. Se precisa de maiores esclarecimentos, por favor aguarde. Voldemort é \\\n",
        "o vilão! Voldermort sempre lança o feitiço avada-kedavra. Porém, Harry Potter \\\n",
        "conseguiu vencer a batalha final. Por que o C3PO estava com o braço vermelho no 7o \\\n",
        "filme de Star Wars? O novo Hulk não parece o Hulk.\"\n",
        "\n",
        "print(len(new_text))"
      ],
      "execution_count": 168,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "690\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tEB654YCknYv",
        "colab_type": "text"
      },
      "source": [
        "-----------------------------\n",
        "#### Criando uma pergunta para o BERT: "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "wfntqRCBegGj",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        },
        "outputId": "d50d8cc0-f46f-4711-df2d-916cc2851541"
      },
      "source": [
        "question = \"Qual o telefone?\"\n",
        "answer_question(question, new_text)"
      ],
      "execution_count": 169,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "A query tem 255 tokens.\n",
            "\n",
            "Answer: \"sao 00 - 12345 - 7890 ou 0800 - 123 - 00000\"\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "i7o0kWsvDn1O",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        },
        "outputId": "5afddec3-bc8b-4499-cf82-d05bd171a131"
      },
      "source": [
        "question = \"Quantos tokens?\"\n",
        "answer_question(question, new_text)"
      ],
      "execution_count": 170,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "A query tem 253 tokens.\n",
            "\n",
            "Answer: \"512\"\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iMz4BZIBD_i8",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        },
        "outputId": "2a441672-1f99-431a-ceb7-459d2ca658d2"
      },
      "source": [
        "question = \"Cor do braço do C3PO?\"\n",
        "answer_question(question, new_text)"
      ],
      "execution_count": 171,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "A query tem 258 tokens.\n",
            "\n",
            "Answer: \"vermelho no 7o filme de star wars\"\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LZEOhu1kHsCc",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        },
        "outputId": "ef2acd85-0703-4e85-c2f1-24842d24c0cf"
      },
      "source": [
        "question = \"Quão poder o BERT possui?\"\n",
        "answer_question(question, new_text)"
      ],
      "execution_count": 172,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "A query tem 258 tokens.\n",
            "\n",
            "Answer: \"[CLS]\"\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8QO5Paf2Hvfy",
        "colab_type": "text"
      },
      "source": [
        "# FIM"
      ]
    }
  ]
}