chapter11_part04_sequence-to-sequence-learning.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/matthewshawnkehoe/deep-learning-with-python-notebooks/blob/master/chapter11_part04_sequence-to-sequence-learning.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "M3pwvQljtO00"
      },
      "source": [
        "This is a companion notebook for the book [Deep Learning with Python, Second Edition](https://www.manning.com/books/deep-learning-with-python-second-edition?a_aid=keras&a_bid=76564dff). For readability, it only contains runnable code blocks and section titles, and omits everything else in the book: text paragraphs, figures, and pseudocode.\n",
        "\n",
        "**If you want to be able to follow what's going on, I recommend reading the notebook side by side with your copy of the book.**\n",
        "\n",
        "This notebook was generated for TensorFlow 2.6."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-L_lvqSCtO05"
      },
      "source": [
        "## Beyond text classification: Sequence-to-sequence learning"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "You now possess all of the tools you will need to tackle most natural language process-ing  tasks.  However,  you've  only  seen  these  tools  in  action  on  a  single  problem:  textclassification.  This  is  an  extremely  popular  use  case,  but  there's  a  lot  more  to  NLPthan  classification.  In  this  section,  you'll  deepen  your  expertise  by  learning  about **sequence-to-sequence** models.\n",
        "\n",
        "A  sequence-to-sequence  model  takes  a  sequence  as  input  (often  a  sentence  orparagraph) and translates it into a different sequence. This is the task at the heart ofmany of the most successful applications of NLP:\n",
        "\n",
        "- Machine translation—Convert a paragraph in a source language to its equivalentin a target language.\n",
        "- Text summarization—Convert a long document to a shorter version that retainsthe most important information.\n",
        "- Question answering—Convert an input question into its answer.\n",
        "- Chatbots—Convert a dialogue prompt into a reply to this prompt, or convert thehistory of a conversation into the next reply in the conversation.\n",
        "- Text generation—Convert a text prompt into a paragraph that completes the prompt.\n",
        "\n",
        "The general template behind sequence-to-sequence models is described in the figure below. During training,\n",
        "\n",
        "- An **encoder** model turns the source sequence into an intermediate representation.\n",
        "- A **decoder** is trained to predict the next token `i` in the target sequence by lookingat both previous tokens (`0` to `i-1`) and the encoded source sequence.\n",
        "\n",
        "During  inference,  we  don't  have  access  to  the  target  sequence—we're  trying  to  pre-dict it from scratch. We'll have to generate it one token at a time:\n",
        "\n",
        "1. We obtain the encoded source sequence from the encoder.\n",
        "2. The decoder starts by looking at the encoded source sequence as well as an initial “seed” token (such as the string \"`[start]`\"), and uses them to predict the first real token in the sequence.\n",
        "3. The predicted sequence so far is fed back into the decoder, which generates the next  token,  and  so  on,  until  it  generates  a  stop  token  (such  as  the  string \"`[end]`\").\n",
        "\n"
      ],
      "metadata": {
        "id": "J9Z40O4Gtgaj"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![sequence_to_sequence.jpg](data:image/jpeg;base64,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)"
      ],
      "metadata": {
        "id": "Ht4g_61Euy2f"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Everything you've learned so far can be repurposed to build this new kind of model. Let's dive in."
      ],
      "metadata": {
        "id": "iEeUuJ6kvDLL"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lynRMTl8tO06"
      },
      "source": [
        "### A machine translation example"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We'll  demonstrate  sequence-to-sequence  modeling  on  a  machine  translation  task. Machine translation is precisely what Transformer was developed for! We'll start with arecurrent sequence model, and we'll follow up with the full Transformer architecture.\n",
        "\n",
        "We’ll  be  working  with  an  English-to-Spanish  translation  dataset  available  at www.manythings.org/anki/. Let's download it:"
      ],
      "metadata": {
        "id": "07gBg45GvIt9"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NlIae5iTtO07",
        "outputId": "1a2b6e0b-0fed-462c-a222-ff5070d69ec9"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2023-09-13 22:41:50--  http://storage.googleapis.com/download.tensorflow.org/data/spa-eng.zip\n",
            "Resolving storage.googleapis.com (storage.googleapis.com)... 142.251.2.207, 2607:f8b0:4023:c06::cf, 2607:f8b0:4023:c0b::cf, ...\n",
            "Connecting to storage.googleapis.com (storage.googleapis.com)|142.251.2.207|:80... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 2638744 (2.5M) [application/zip]\n",
            "Saving to: ‘spa-eng.zip’\n",
            "\n",
            "spa-eng.zip         100%[===================>]   2.52M  --.-KB/s    in 0.008s  \n",
            "\n",
            "2023-09-13 22:41:51 (307 MB/s) - ‘spa-eng.zip’ saved [2638744/2638744]\n",
            "\n"
          ]
        }
      ],
      "source": [
        "!wget http://storage.googleapis.com/download.tensorflow.org/data/spa-eng.zip\n",
        "!unzip -q spa-eng.zip"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The  text  file  contains  one  example  per  line:  an  English  sentence,  followed  by  a  tab character, followed by the corresponding Spanish sentence. Let's parse this file."
      ],
      "metadata": {
        "id": "EDSzR9pQvfVk"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "id": "zjlWHHU1tO08"
      },
      "outputs": [],
      "source": [
        "text_file = \"spa-eng/spa.txt\"\n",
        "with open(text_file) as f:\n",
        "    lines = f.read().split(\"\\n\")[:-1]\n",
        "text_pairs = []\n",
        "for line in lines:                                                              # Iterate over the lines in the file.\n",
        "    english, spanish = line.split(\"\\t\")                                         # Each line contains an English phrase and its Spanish\n",
        "                                                                                # translation, tab-separated. We prepend \"[start]\"\n",
        "    spanish = \"[start] \" + spanish + \" [end]\"                                   # and append \"[end]\" to the Spanish sentence to match\n",
        "                                                                                # the template in the figure above.\n",
        "    text_pairs.append((english, spanish))"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Our `text_pairs` look like this:"
      ],
      "metadata": {
        "id": "F8IyINw3wJoc"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9UeEq1sQtO0-",
        "outputId": "fd90e6d8-e367-4ffa-a22f-a3a2620852aa"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "('Our teacher is a real idiot.', '[start] Nuestro maestro es un perfecto idiota. [end]')\n"
          ]
        }
      ],
      "source": [
        "import random\n",
        "print(random.choice(text_pairs))"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Let's shuffle them and split them into the usual training, validation, and test sets:"
      ],
      "metadata": {
        "id": "tUnFy-PDwZNL"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "nbyQefRgtO0-"
      },
      "outputs": [],
      "source": [
        "import random\n",
        "random.shuffle(text_pairs)                                                      # Reorganizes the list\n",
        "num_val_samples = int(0.15 * len(text_pairs))\n",
        "num_train_samples = len(text_pairs) - 2 * num_val_samples\n",
        "train_pairs = text_pairs[:num_train_samples]                                    # The first 70%\n",
        "val_pairs = text_pairs[num_train_samples:num_train_samples + num_val_samples]   # The next 15%\n",
        "test_pairs = text_pairs[num_train_samples + num_val_samples:]                   # The remaining 15%"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Next, let's prepare two separate `TextVectorization` layers: one for English and one for Spanish. We`re going to need to customize the way strings are preprocessed:\n",
        "\n",
        "- We need to preserve the \"`[start]`\" and \"`[end]`\" tokens that we've inserted. By default,  the  characters  `[  and  ]`  would  be  stripped,  but  we  want  to  keep  them around so we can tell apart the word “start” and the start token \"`[start]`\".\n",
        "- Punctuation  is  different  from  language  to  language!  In  the  Spanish  `TextVectorization` layer, if we're going to strip punctuation characters, we need to also strip the character ¿.\n",
        "\n",
        "Note that for a non-toy translation model, we would treat punctuation characters as separate tokens rather than stripping them, since we would want to be able to generate correctly punctuated sentences. In our case, for simplicity, we'll get rid of all punctuation."
      ],
      "metadata": {
        "id": "IgQn5y4axApa"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SlZCFtdPtO0_"
      },
      "source": [
        "**Vectorizing the English and Spanish text pairs**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "ZKDyibGGtO1A"
      },
      "outputs": [],
      "source": [
        "import tensorflow as tf\n",
        "import string\n",
        "import re\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras import layers\n",
        "\n",
        "strip_chars = string.punctuation + \"¿\"                                          # Prepare a custom string standardization function for\n",
        "strip_chars = strip_chars.replace(\"[\", \"\")                                      # the Spanish TextVectorization layer: it preserves [ and ]\n",
        "strip_chars = strip_chars.replace(\"]\", \"\")                                      # but strips ¿ (as well as all other characters from\n",
        "                                                                                # strings.punctuation).\n",
        "def custom_standardization(input_string):\n",
        "    lowercase = tf.strings.lower(input_string)\n",
        "    return tf.strings.regex_replace(\n",
        "        lowercase, f\"[{re.escape(strip_chars)}]\", \"\")\n",
        "\n",
        "vocab_size = 15000                                                              # To keep things simple, we’ll only look at the top 15,000 words\n",
        "sequence_length = 20                                                            # in each language, and we’ll restrict sentences to 20 words.\n",
        "\n",
        "source_vectorization = layers.TextVectorization(                                # The English layer\n",
        "    max_tokens=vocab_size,\n",
        "    output_mode=\"int\",\n",
        "    output_sequence_length=sequence_length,\n",
        ")\n",
        "target_vectorization = layers.TextVectorization(                                # The Spanish layer\n",
        "    max_tokens=vocab_size,\n",
        "    output_mode=\"int\",\n",
        "    output_sequence_length=sequence_length + 1,                                 # Generate Spanish sentences that have one extra token, since\n",
        "                                                                                # we’ll need to offset the sentence by one step during training\n",
        "    standardize=custom_standardization,\n",
        ")\n",
        "train_english_texts = [pair[0] for pair in train_pairs]\n",
        "train_spanish_texts = [pair[1] for pair in train_pairs]\n",
        "source_vectorization.adapt(train_english_texts)                                 # Learn the vocabulary of every language."
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Finally,  we  can  turn  our  data  into  a  `tf.data` pipeline. We want it to return a tuple `(inputs,target)` where `inputs` is a `dict` with two keys, “encoder_inputs” (the English sentence)  and “decoder_inputs”  (the  Spanish  sentence),  and  target  is  the  Spanish sentence offset by one step ahead."
      ],
      "metadata": {
        "id": "_JVDTUEayrhJ"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FNF3NE_PtO1B"
      },
      "source": [
        "**Preparing datasets for the translation task**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "zstrLf0KtO1C"
      },
      "outputs": [],
      "source": [
        "batch_size = 64\n",
        "\n",
        "def format_dataset(eng, spa):\n",
        "    eng = source_vectorization(eng)\n",
        "    spa = target_vectorization(spa)\n",
        "    return ({\n",
        "        \"english\": eng,\n",
        "        \"spanish\": spa[:, :-1],                                                 # The input Spanish sentence doesn’t include the last token\n",
        "                                                                                # to keep inputs and targets at the same length.\n",
        "    }, spa[:, 1:])                                                              # The target Spanish sentence is one step ahead.\n",
        "                                                                                # Both are still the same length (20 words).\n",
        "\n",
        "def make_dataset(pairs):\n",
        "    eng_texts, spa_texts = zip(*pairs)\n",
        "    eng_texts = list(eng_texts)\n",
        "    spa_texts = list(spa_texts)\n",
        "    dataset = tf.data.Dataset.from_tensor_slices((eng_texts, spa_texts))\n",
        "    dataset = dataset.batch(batch_size)\n",
        "    dataset = dataset.map(format_dataset, num_parallel_calls=4)\n",
        "    return dataset.shuffle(2048).prefetch(16).cache()                           # Use in-memory caching to speed up preprocessing.\n",
        "\n",
        "train_ds = make_dataset(train_pairs)\n",
        "val_ds = make_dataset(val_pairs)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Here's what our dataset outputs look like:"
      ],
      "metadata": {
        "id": "aZQxdNa00EhI"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2qyb239qtO1D",
        "outputId": "cd896004-1bc0-4106-c479-a1064972ee0f"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "inputs['english'].shape: (64, 20)\n",
            "inputs['spanish'].shape: (64, 20)\n",
            "targets.shape: (64, 20)\n"
          ]
        }
      ],
      "source": [
        "for inputs, targets in train_ds.take(1):\n",
        "    print(f\"inputs['english'].shape: {inputs['english'].shape}\")\n",
        "    print(f\"inputs['spanish'].shape: {inputs['spanish'].shape}\")\n",
        "    print(f\"targets.shape: {targets.shape}\")"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The  data  is  now  ready—time  to  build  some  models.  We'll  start  with  a  recurrent sequence-to-sequence model before moving on to a Transformer"
      ],
      "metadata": {
        "id": "AeZSTq2z0MIf"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HaD8ZQ2RtO1E"
      },
      "source": [
        "### Sequence-to-sequence learning with RNNs"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Recurrent  neural  networks  dominated  sequence-to-sequence  learning  from  2015–2017  before  being  overtaken  by  Transformer.  They  were  the  basis  for  many  real-world machine-translation systems—as mentioned in chapter 10, Google Translate circa 2017 was powered by a stack of seven large LSTM layers. It's still worth learning about  this  approach  today,  as it provides an easy entry  point to  understanding sequence-to-sequence models.\n",
        "\n",
        " The simplest, naive way to use RNNs to turn a sequence into another sequence isto keep the output of the RNN at each time step. In Keras, it would look like this:"
      ],
      "metadata": {
        "id": "fRejuvQc0SrO"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Psuedocode\n",
        "\n",
        "inputs = keras.Input(shape=(sequence_length,), dtype=\"int64\")\n",
        "x = layers.Embedding(input_dim=vocab_size, output_dim=128)(inputs)\n",
        "x = layers.LSTM(32, return_sequences=True)(x)\n",
        "outputs = layers.Dense(vocab_size, activation=\"softmax\")(x)\n",
        "model = keras.Model(inputs, outputs)"
      ],
      "metadata": {
        "id": "nCTHTcu-0hzF"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "However, there are two major issues with this approach:\n",
        "\n",
        "1. The target sequence must always be the same length as the source sequence. In practice, this is rarely the case. Technically, this isn't critical, as you could always  pad  either  the  source  sequence  or  the  target  sequence  to  make  their lengths match.\n",
        "\n",
        "2. Due  to  the  step-by-step  nature  of  RNNs,  the  model  will  only  be  looking  at tokens 0...N in the source sequence in order to predict token N in the target sequence.  This  constraint  makes  this  setup  unsuitable  for  most  tasks,  and particularly  translation.  Consider  translating  “The  weather  is  nice  today”  to French—that  would  be  “Il  fait  beau  aujourd'hui.”  You'd  need  to  be  able to predict “Il” from just “The,” “Il fait” from just “The weather,” etc., which is simply impossible.\n",
        "\n",
        "If you're a human translator, you'd start by reading the entire source sentence beforestarting  to  translate  it.  This  is  especially  important  if  you're  dealing  with  languages that have wildly different word ordering, like English and Japanese. And that's exactly what standard sequence-to-sequence models do.\n",
        "\n",
        "In a proper sequence-to-sequence setup (see figure below), you would first use an RNN (the encoder) to turn the entire source sequence into a single vector (or set of vectors). This could be the last output of the RNN, or alternatively, its final internal state vectors. Then you would use this vector (or vectors) as the initial state of another RNN (the decoder), which would look at elements 0...N in the target sequence, and try to predict step N+1 in the target sequence.\n",
        "\n"
      ],
      "metadata": {
        "id": "hhmnZp3s0x0I"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![s2s_RNN.jpg](data:image/jpeg;base64,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)"
      ],
      "metadata": {
        "id": "hWzdGwE11tw7"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        " Let's implement this in Keras with GRU-based encoders and decoders. The choiceof  GRU  rather  than  LSTM  makes  things  a  bit  simpler,  since  GRU  only  has  a  single state vector, whereas LSTM has multiple. Let's start with the encoder."
      ],
      "metadata": {
        "id": "JKZwPYND19xK"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UwcJYzQOtO1E"
      },
      "source": [
        "**GRU-based encoder**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "id": "QwsAEcu5tO1E"
      },
      "outputs": [],
      "source": [
        "from tensorflow import keras\n",
        "from tensorflow.keras import layers\n",
        "\n",
        "embed_dim = 256\n",
        "latent_dim = 1024\n",
        "\n",
        "source = keras.Input(shape=(None,), dtype=\"int64\", name=\"english\")              # The English source sentence goes here. Specifying the name of\n",
        "                                                                                # the input enables us to fit() the model with a dict of inputs.\n",
        "x = layers.Embedding(vocab_size, embed_dim, mask_zero=True)(source)             # Don’t forget masking: it’s critical in this setup.\n",
        "encoded_source = layers.Bidirectional(\n",
        "    layers.GRU(latent_dim), merge_mode=\"sum\")(x)                                # Our encoded source sentence is the last output of a\n",
        "                                                                                # bidirectional GRU."
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Next,  let's  add  the  decoder—a  simple  GRU  layer  that  takes  as  its  initial  state  theencoded source sentence. On top of it, we add a **Dense** layer that produces for each output step a probability distribution over the Spanish vocabulary."
      ],
      "metadata": {
        "id": "wEeqslYa2u17"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HwqnevJgtO1F"
      },
      "source": [
        "**GRU-based decoder and the end-to-end model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "id": "hQte7wCWtO1F"
      },
      "outputs": [],
      "source": [
        "past_target = keras.Input(shape=(None,), dtype=\"int64\", name=\"spanish\")         # The Spanish target sentence goes here.\n",
        "x = layers.Embedding(vocab_size, embed_dim, mask_zero=True)(past_target)        # Don't forget masking.\n",
        "decoder_gru = layers.GRU(latent_dim, return_sequences=True)\n",
        "x = decoder_gru(x, initial_state=encoded_source)                  # The encoded source sentence serves as the initial state of the decoder GRU.\n",
        "x = layers.Dropout(0.5)(x)\n",
        "target_next_step = layers.Dense(vocab_size, activation=\"softmax\")(x)            # Predicts the next token.\n",
        "seq2seq_rnn = keras.Model([source, past_target], target_next_step)              # End-to-end model: maps the source sentence and the target\n",
        "                                                                                # sentence to the target sentence one step in the future."
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "During training, the decoder takes as input the entire target sequence, but thanks to the step-by-step nature of RNNs, it only looks at tokens 0...N in the input to predict token N  in  the  output  (which  corresponds  to  the  next  token  in  the  sequence,  since the output is intended to be offset by one step). This means we only use information from the past to predict the future, as we should; otherwise we'd be cheating, and our model would not work at inference time.\n",
        "\n",
        " Let's start training."
      ],
      "metadata": {
        "id": "abtAfbHC3Tns"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EHOR7ixWtO1G"
      },
      "source": [
        "**Training our recurrent sequence-to-sequence model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xIV2FM1etO1H",
        "outputId": "64784cdd-b301-4084-edfc-2dade95bcba1"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/15\n",
            "1302/1302 [==============================] - 159s 108ms/step - loss: 4.6811 - accuracy: 0.3195 - val_loss: 3.8906 - val_accuracy: 0.3887\n",
            "Epoch 2/15\n",
            "1302/1302 [==============================] - 112s 86ms/step - loss: 3.7177 - accuracy: 0.4162 - val_loss: 3.2450 - val_accuracy: 0.4706\n",
            "Epoch 3/15\n",
            "1302/1302 [==============================] - 112s 86ms/step - loss: 3.2125 - accuracy: 0.4727 - val_loss: 2.8649 - val_accuracy: 0.5206\n",
            "Epoch 4/15\n",
            "1302/1302 [==============================] - 111s 85ms/step - loss: 2.8587 - accuracy: 0.5149 - val_loss: 2.6105 - val_accuracy: 0.5561\n",
            "Epoch 5/15\n",
            "1302/1302 [==============================] - 112s 86ms/step - loss: 2.5843 - accuracy: 0.5483 - val_loss: 2.4228 - val_accuracy: 0.5842\n",
            "Epoch 6/15\n",
            "1302/1302 [==============================] - 122s 93ms/step - loss: 2.3622 - accuracy: 0.5761 - val_loss: 2.2960 - val_accuracy: 0.6023\n",
            "Epoch 7/15\n",
            "1302/1302 [==============================] - 115s 89ms/step - loss: 2.1786 - accuracy: 0.5999 - val_loss: 2.2037 - val_accuracy: 0.6171\n",
            "Epoch 8/15\n",
            "1302/1302 [==============================] - 125s 96ms/step - loss: 2.0208 - accuracy: 0.6203 - val_loss: 2.1311 - val_accuracy: 0.6281\n",
            "Epoch 9/15\n",
            "1302/1302 [==============================] - 114s 88ms/step - loss: 1.8840 - accuracy: 0.6387 - val_loss: 2.0672 - val_accuracy: 0.6377\n",
            "Epoch 10/15\n",
            "1302/1302 [==============================] - 115s 89ms/step - loss: 1.7642 - accuracy: 0.6552 - val_loss: 2.0233 - val_accuracy: 0.6443\n",
            "Epoch 11/15\n",
            "1302/1302 [==============================] - 117s 90ms/step - loss: 1.6588 - accuracy: 0.6699 - val_loss: 1.9912 - val_accuracy: 0.6499\n",
            "Epoch 12/15\n",
            "1302/1302 [==============================] - 114s 87ms/step - loss: 1.5650 - accuracy: 0.6829 - val_loss: 1.9697 - val_accuracy: 0.6545\n",
            "Epoch 13/15\n",
            "1302/1302 [==============================] - 114s 87ms/step - loss: 1.4895 - accuracy: 0.6934 - val_loss: 1.9459 - val_accuracy: 0.6576\n",
            "Epoch 14/15\n",
            "1302/1302 [==============================] - 125s 96ms/step - loss: 1.4161 - accuracy: 0.7043 - val_loss: 1.9310 - val_accuracy: 0.6610\n",
            "Epoch 15/15\n",
            "1302/1302 [==============================] - 114s 88ms/step - loss: 1.3496 - accuracy: 0.7139 - val_loss: 1.9169 - val_accuracy: 0.6621\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.src.callbacks.History at 0x78afe06bf2b0>"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ],
      "source": [
        "seq2seq_rnn.compile(\n",
        "    optimizer=\"rmsprop\",\n",
        "    loss=\"sparse_categorical_crossentropy\",\n",
        "    metrics=[\"accuracy\"])\n",
        "seq2seq_rnn.fit(train_ds, epochs=15, validation_data=val_ds)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We  picked  accuracy  as  a  crude  way  to  monitor  validation-set  performance  during training. We get to 66% accuracy: on average, the model predicts the next word in the Spanish sentence correctly 66% of the time. However, in practice, next-token accuracy isn't a great metric for machine translation models, in particular because it makes the assumption  that  the  correct  target  tokens  from  0  to  N  are  already  known  when  predicting token N+1. In reality, during inference, you're generating the target sentence from scratch, and you can't rely on previously generated tokens being 100% correct. If you  work  on  a  real-world  machine  translation  system,  you  will  likely  use  “BLEUscores”  to  evaluate  your  models—a  metric  that  looks  at  entire  generated  sequences and that seems to correlate well with human perception of translation quality.\n",
        "\n",
        "At last, let's use our model for inference. We'll pick a few sentences in the test set and check how our model translates them. We'll start from the seed token, \"`[start]`\",and  feed  it  into  the  decoder  model,  together  with  the  encoded  English  source  sentence.  We'll  retrieve  a  next-token  prediction,  and  we'll  re-inject  it  into  the  decoder repeatedly, sampling one new target token at each iteration, until we get to \"`[end]`\" or reach the maximum sentence length"
      ],
      "metadata": {
        "id": "YMHGbpDd3kx0"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eBX6Th3HtO1M"
      },
      "source": [
        "**Translating new sentences with our RNN encoder and decoder**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5H1kHGcUtO1N",
        "outputId": "c54677f4-c862-49f9-8595-02876b2610e7"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "-\n",
            "We will get married in June.\n",
            "1/1 [==============================] - 3s 3s/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "[start] nos [UNK] en octubre [end]\n",
            "-\n",
            "This is Tom's bedroom.\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "[start] esta es la habitación de tom [end]\n",
            "-\n",
            "It's a miracle that you were able to survive.\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "[start] es un riesgo que no podemos venir [end]\n",
            "-\n",
            "Do you play basketball well?\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "[start] vas a jugar bien el violín [end]\n",
            "-\n",
            "Nice to see you, Tom.\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "[start] qué bueno verte tom [end]\n",
            "-\n",
            "That was my big mistake.\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "[start] eso fue mi gran error [end]\n",
            "-\n",
            "The lessons are getting more and more difficult.\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "1/1 [==============================] - 0s 43ms/step\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "1/1 [==============================] - 0s 40ms/step\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "1/1 [==============================] - 0s 45ms/step\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "[start] las chicas se están más y más y más [end]\n",
            "-\n",
            "I ran toward the door.\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "1/1 [==============================] - 0s 91ms/step\n",
            "1/1 [==============================] - 0s 96ms/step\n",
            "[start] me quedé corriendo la puerta [end]\n",
            "-\n",
            "He stood close to her and tried to protect her from the typhoon.\n",
            "1/1 [==============================] - 0s 87ms/step\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "1/1 [==============================] - 0s 53ms/step\n",
            "1/1 [==============================] - 0s 44ms/step\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "1/1 [==============================] - 0s 49ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "[start] Él se paró y se [UNK] de su ausencia para que [UNK] [end]\n",
            "-\n",
            "You are the only one who can help me.\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 21ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "[start] eres el único que puede ayudarme [end]\n",
            "-\n",
            "This shouldn't happen.\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 21ms/step\n",
            "1/1 [==============================] - 0s 21ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "[start] esto no va a ocurrir [end]\n",
            "-\n",
            "Tom got a bottle of beer out of the fridge.\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 51ms/step\n",
            "1/1 [==============================] - 0s 97ms/step\n",
            "1/1 [==============================] - 0s 82ms/step\n",
            "1/1 [==============================] - 0s 35ms/step\n",
            "1/1 [==============================] - 0s 34ms/step\n",
            "1/1 [==============================] - 0s 51ms/step\n",
            "1/1 [==============================] - 0s 41ms/step\n",
            "1/1 [==============================] - 0s 52ms/step\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 51ms/step\n",
            "[start] tom se toma una botella de vino a la leche [end]\n",
            "-\n",
            "Tom is in love with Mary.\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 38ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "[start] tom está enamorado de mary [end]\n",
            "-\n",
            "Tom is faster than you think.\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 36ms/step\n",
            "1/1 [==============================] - 0s 32ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 30ms/step\n",
            "[start] tom es más rápido de lo que crees [end]\n",
            "-\n",
            "Do people drink tea in your country?\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "[start] la gente de té bebe tu café [end]\n",
            "-\n",
            "These shoes fit my feet.\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 29ms/step\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "1/1 [==============================] - 0s 58ms/step\n",
            "1/1 [==============================] - 0s 39ms/step\n",
            "1/1 [==============================] - 0s 50ms/step\n",
            "1/1 [==============================] - 0s 47ms/step\n",
            "[start] estos zapatos me [UNK] los zapatos [end]\n",
            "-\n",
            "Tom counts on Mary's help.\n",
            "1/1 [==============================] - 0s 45ms/step\n",
            "1/1 [==============================] - 0s 46ms/step\n",
            "1/1 [==============================] - 0s 58ms/step\n",
            "1/1 [==============================] - 0s 42ms/step\n",
            "1/1 [==============================] - 0s 47ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 33ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "[start] tom se [UNK] con la ayuda de mary [end]\n",
            "-\n",
            "If I knew the truth, I would tell you.\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 31ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 28ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "[start] si lo sabía la verdad lo que te diga [end]\n",
            "-\n",
            "Tom is in the kitchen, standing in front of the refrigerator.\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 23ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 31ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "[start] tom está en la cocina del cajón del escritorio [end]\n",
            "-\n",
            "Tom held his breath.\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 27ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "1/1 [==============================] - 0s 24ms/step\n",
            "1/1 [==============================] - 0s 26ms/step\n",
            "[start] tom se rompió la pierna [end]\n"
          ]
        }
      ],
      "source": [
        "import numpy as np\n",
        "spa_vocab = target_vectorization.get_vocabulary()                               # Prepare a dict to convert token index predictions to\n",
        "spa_index_lookup = dict(zip(range(len(spa_vocab)), spa_vocab))                  # string tokens.\n",
        "max_decoded_sentence_length = 20\n",
        "\n",
        "def decode_sequence(input_sentence):\n",
        "    tokenized_input_sentence = source_vectorization([input_sentence])\n",
        "    decoded_sentence = \"[start]\"                                                # Seed token.\n",
        "    for i in range(max_decoded_sentence_length):\n",
        "        tokenized_target_sentence = target_vectorization([decoded_sentence])\n",
        "        next_token_predictions = seq2seq_rnn.predict(                           # Sample the next token.\n",
        "            [tokenized_input_sentence, tokenized_target_sentence])\n",
        "        sampled_token_index = np.argmax(next_token_predictions[0, i, :])\n",
        "        sampled_token = spa_index_lookup[sampled_token_index]                   # Convert the next token prediction to a string and\n",
        "        decoded_sentence += \" \" + sampled_token                                 # append it to the generated sentence.\n",
        "        if sampled_token == \"[end]\":                                            # Exit condition: either hit max length or sample\n",
        "            break                                                               # a stop character.\n",
        "    return decoded_sentence\n",
        "\n",
        "test_eng_texts = [pair[0] for pair in test_pairs]\n",
        "for _ in range(20):\n",
        "    input_sentence = random.choice(test_eng_texts)\n",
        "    print(\"-\")\n",
        "    print(input_sentence)\n",
        "    print(decode_sequence(input_sentence))"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Note that this inference setup, while very simple, is rather inefficient, since we reprocess  the  entire  source  sentence  and  the  entire  generated  target  sentence  every  time we  sample  a  new  word.  In  a  practical  application,  you'd  factor  the  encoder  and  the decoder  as  two  separate  models,  and  your  decoder  would  only  run  a  single  step  at each token-sampling iteration, reusing its previous internal state.\n",
        "\n",
        "Here are some more translation results. Our model works decently well for a toy model, though it still makes many basic mistakes."
      ],
      "metadata": {
        "id": "l2lTOUTP47qR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "Who is in this room?\n",
        "[start] quién está en esta habitación [end]\n",
        "-\n",
        "That doesn't sound too dangerous.\n",
        "[start] eso no es muy difícil [end]\n",
        "-\n",
        "No one will stop me.\n",
        " [start] nadie me va a hacer [end]\n",
        " -Tom is friendly.\n",
        "  [start] tom es un buen [UNK] [end]"
      ],
      "metadata": {
        "collapsed": true,
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 145
        },
        "id": "tcWYxsHL5LDC",
        "outputId": "825602ab-a618-478e-cb35-08e9e1ef17e2"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "error",
          "ename": "SyntaxError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;36m  File \u001b[0;32m\"<ipython-input-16-80bf0bb4ec50>\"\u001b[0;36m, line \u001b[0;32m1\u001b[0m\n\u001b[0;31m    Listing Some sample results from the recurrent translation model\u001b[0m\n\u001b[0m            ^\u001b[0m\n\u001b[0;31mSyntaxError\u001b[0m\u001b[0;31m:\u001b[0m invalid syntax\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "There  are  many  ways  this  toy  model  could  be  improved:  We  could  use  a  deep  stack  of recurrent layers for both the encoder and the decoder (note that for the decoder, this makes state management a bit more involved). We could use an LSTM instead of a GRU. And so on.  Beyond  such  tweaks,  however,  the  RNN  approach  to  sequence-to-sequence learning has a few fundamental limitations:\n",
        "\n",
        "*   The source sequence representation has to be held entirely in the encoder state vector(s),  which  puts  significant  limitations  on  the  size  and  complexity  of  thesentences you can translate. It's a bit as if a human were translating a sentenc eentirely from memory, without looking twice at the source sentence while producing the translation.\n",
        "*   RNNs  have  trouble  dealing  with  very  long  sequences,  since  they  tend  to  progressively forget about the past—by the time you've reached the 100th token ineither  sequence,  little  information  remains  about  the  start  of  the  sequence. That  means  RNN-based  models  can't  hold  onto  long-term  context,  which  can be essential for translating long documents.\n",
        "\n",
        "These limitations are what has led the machine learning community to embrace theTransformer architecture for sequence-to-sequence problems. Let's take a look.\n"
      ],
      "metadata": {
        "id": "CkLlNJJz5ikp"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zQZ-t__QtO1O"
      },
      "source": [
        "### Sequence-to-sequence learning with Transformer"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Sequence-to-sequence  learning  is  the  task  where  Transformer  really  shines.  Neural attention enables Transformer models to successfully process sequences that are considerably longer and more complex than those RNNs can handle.\n",
        "\n",
        "As a human translating English to Spanish, you're not going to read the English sentence one word at a time, keep its meaning in memory, and then generate the Spanish sentence one word at a time. That may work for a five-word sentence, but it's  unlikely  to  work  for  an  entire  paragraph.  Instead,  you'll  probably  want  to  go back  and  forth  between  the  source  sentence and your translation in progress, and pay attention to different words in the source as you're writing down different parts of your translation.\n",
        "\n",
        "That's  exactly  what  you  can  achieve  with  neural  attention  and  Transformers. You're already familiar with the Transformer encoder, which uses self-attention to produce  context-aware  representations  of  each  token  in  an  input  sequence.  In  a sequence-to-sequence Transformer, the **Transformer encoder** would naturally play the role of the encoder, which reads the source sequence and produces an encoded representation  of  it.  Unlike  our  previous  RNN  encoder,  though,  the  Transformerencoder  keeps  the  encoded  representation  in  a  sequence  format:  it's  a  sequence  of context-aware embedding vectors.\n",
        "\n",
        "The second half of the model is the **Transformer decoder**. Just like the RNN decoder, it  reads  tokens  `0...N`  in  the  target  sequence  and  tries  to  predict  token  `N+1`.  Crucially, while doing this, it uses **neural attention** to identify which tokens in the encoded source sentence are most closely related to the target token it's currently trying to predict—perhaps  not  unlike  what  a  human  translator  would  do.  Recall  the  query-key-value model: in a Transformer decoder, the target sequence serves as an attention “query” that  is  used  to  to  pay  closer  attention  to  different  parts  of  the  source  sequence  (the source sequence plays the roles of both keys and values)."
      ],
      "metadata": {
        "id": "IKlPEivi59hz"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kXJa9o50tO1O"
      },
      "source": [
        "#### The Transformer decoder"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The Figure below shows the full sequence-to-sequence Transformer. Look at the decoder internals: you'll recognize that it looks very similar to the Transformer encoder, except that  an  extra  attention  block  is  inserted  between  the  self-attention  block  applied to the target sequence and the dense layers of the exit block."
      ],
      "metadata": {
        "id": "ilbbrmPe6rGK"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![Transformer.jpg](data:image/jpeg;base64,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)"
      ],
      "metadata": {
        "id": "XWl6kKyX6xQB"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "  TODO - Fix code order below"
      ],
      "metadata": {
        "id": "2SVG21O-9zkf"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xwVXZJBRtO1O"
      },
      "source": [
        "**The `TransformerDecoder`**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "id": "JtShQ-6TtO1P"
      },
      "outputs": [],
      "source": [
        "class TransformerDecoder(layers.Layer):\n",
        "    def __init__(self, embed_dim, dense_dim, num_heads, **kwargs):\n",
        "        super().__init__(**kwargs)\n",
        "        self.embed_dim = embed_dim\n",
        "        self.dense_dim = dense_dim\n",
        "        self.num_heads = num_heads\n",
        "        self.attention_1 = layers.MultiHeadAttention(\n",
        "            num_heads=num_heads, key_dim=embed_dim)\n",
        "        self.attention_2 = layers.MultiHeadAttention(\n",
        "            num_heads=num_heads, key_dim=embed_dim)\n",
        "        self.dense_proj = keras.Sequential(\n",
        "            [layers.Dense(dense_dim, activation=\"relu\"),\n",
        "             layers.Dense(embed_dim),]\n",
        "        )\n",
        "        self.layernorm_1 = layers.LayerNormalization()\n",
        "        self.layernorm_2 = layers.LayerNormalization()\n",
        "        self.layernorm_3 = layers.LayerNormalization()\n",
        "        self.supports_masking = True                                            # This attribute ensures that the layer will propagate\n",
        "                                                                                # its input mask to its outputs; masking in Keras is explicitly\n",
        "                                                                                # opt-in. If you pass a mask to a layer that doesn’t implement\n",
        "                                                                                # compute_mask() and that doesn’t expose this supports_masking\n",
        "                                                                                # attribute, that’s an error.\n",
        "   # TODO - Add Get_config()!"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The `call()` method is almost a straightforward rendering of the connectivity diagram from Transformers architecture. But there's an additional detail we need to take into account: causal padding. Causal padding is absolutely critical to successfully training a sequence-to-sequence Transformer. Unlike an RNN, which looks at its input one step at a time, and thus will only have access to steps `0...N` to generate output step `N`(which is token `N+1` in the target sequence), the TransformerDecoder is order-agnostic: it looks at the entire target sequence at once. If it were allowed to use its entire input, it would simply learn to copy input step `N+1` to location `N` in the output. The model would thus achieve perfect training accuracy, but of course, when running inference, it would be completely useless, since input steps beyond `N` aren't available.\n",
        "\n",
        "The fix is simple: we'll mask the upper half of the pairwise attention matrix to prevent the model from paying any attention to information from the future—only information  from  tokens `0...N` in the target sequence should be used when generating target token `N+1`. To do this, we'll add a `get_causal_attention_mask(self,inputs)` method to our TransformerDecoder to retrieve an attention mask that we can pass to our MultiHeadAttention layers.\n",
        "\n"
      ],
      "metadata": {
        "id": "NRBNu6zo8ucs"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def get_causal_attention_mask(self, inputs):\n",
        "        input_shape = tf.shape(inputs)\n",
        "        batch_size, sequence_length = input_shape[0], input_shape[1]\n",
        "        i = tf.range(sequence_length)[:, tf.newaxis]\n",
        "        j = tf.range(sequence_length)\n",
        "        mask = tf.cast(i >= j, dtype=\"int32\")\n",
        "        mask = tf.reshape(mask, (1, input_shape[1], input_shape[1]))  # Generate matrix of shape (sequence_length, sequence_length) with 1s\n",
        "                                                                      # in one half and 0s in the other.\n",
        "        mult = tf.concat(                                             # Replicate it along thebatch axis to get a matrix of .\n",
        "            [tf.expand_dims(batch_size, -1),                          # shape (batch_size,sequence_length,sequence_length)\n",
        "             tf.constant([1, 1], dtype=tf.int32)], axis=0)\n",
        "        return tf.tile(mask, mult)\n"
      ],
      "metadata": {
        "id": "TryBF6lu8svl"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now we can write down the full `call()` method implementing the forward pass of the decoder."
      ],
      "metadata": {
        "id": "tczG4o5r7-rq"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XhMa7Xf8tO1P"
      },
      "source": [
        "#### Putting it all together: A Transformer for machine translation"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xZ4dY5mDtO1P"
      },
      "source": [
        "**PositionalEmbedding layer**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "id": "bVh-OiF-tO1P"
      },
      "outputs": [],
      "source": [
        "class PositionalEmbedding(layers.Layer):\n",
        "    def __init__(self, sequence_length, input_dim, output_dim, **kwargs):\n",
        "        super().__init__(**kwargs)\n",
        "        self.token_embeddings = layers.Embedding(\n",
        "            input_dim=input_dim, output_dim=output_dim)\n",
        "        self.position_embeddings = layers.Embedding(\n",
        "            input_dim=sequence_length, output_dim=output_dim)\n",
        "        self.sequence_length = sequence_length\n",
        "        self.input_dim = input_dim\n",
        "        self.output_dim = output_dim\n",
        "\n",
        "    def call(self, inputs):\n",
        "        length = tf.shape(inputs)[-1]\n",
        "        positions = tf.range(start=0, limit=length, delta=1)\n",
        "        embedded_tokens = self.token_embeddings(inputs)\n",
        "        embedded_positions = self.position_embeddings(positions)\n",
        "        return embedded_tokens + embedded_positions\n",
        "\n",
        "    def compute_mask(self, inputs, mask=None):\n",
        "        return tf.math.not_equal(inputs, 0)\n",
        "\n",
        "    def get_config(self):\n",
        "        config = super(PositionalEmbedding, self).get_config()\n",
        "        config.update({\n",
        "            \"output_dim\": self.output_dim,\n",
        "            \"sequence_length\": self.sequence_length,\n",
        "            \"input_dim\": self.input_dim,\n",
        "        })\n",
        "        return config"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_wJ--XEltO1R"
      },
      "source": [
        "**End-to-end Transformer**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "dq4dFeNLtO1R",
        "outputId": "1c5ddc12-d02f-4cc0-ba22-7774a30995c3"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting flair==0.4.3\n",
            "  Downloading flair-0.4.3-py3-none-any.whl (180 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m180.0/180.0 kB\u001b[0m \u001b[31m2.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: torch>=1.1.0 in /usr/local/lib/python3.10/dist-packages (from flair==0.4.3) (2.0.1+cu118)\n",
            "Requirement already satisfied: gensim>=3.4.0 in /usr/local/lib/python3.10/dist-packages (from flair==0.4.3) (4.3.2)\n",
            "Requirement already satisfied: pytest>=3.6.4 in /usr/local/lib/python3.10/dist-packages (from flair==0.4.3) (7.4.1)\n",
            "Requirement already satisfied: tqdm>=4.26.0 in /usr/local/lib/python3.10/dist-packages (from flair==0.4.3) (4.66.1)\n",
            "Collecting segtok>=1.5.7 (from flair==0.4.3)\n",
            "  Downloading segtok-1.5.11-py3-none-any.whl (24 kB)\n",
            "Requirement already satisfied: matplotlib>=2.2.3 in /usr/local/lib/python3.10/dist-packages (from flair==0.4.3) (3.7.1)\n",
            "Collecting mpld3==0.3 (from flair==0.4.3)\n",
            "  Downloading mpld3-0.3.tar.gz (788 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m788.5/788.5 kB\u001b[0m \u001b[31m9.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "Collecting sklearn (from flair==0.4.3)\n",
            "  Downloading sklearn-0.0.post9.tar.gz (3.6 kB)\n",
            "  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "Collecting sqlitedict>=1.6.0 (from flair==0.4.3)\n",
            "  Downloading sqlitedict-2.1.0.tar.gz (21 kB)\n",
            "  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "Collecting deprecated>=1.2.4 (from flair==0.4.3)\n",
            "  Downloading Deprecated-1.2.14-py2.py3-none-any.whl (9.6 kB)\n",
            "Requirement already satisfied: hyperopt>=0.1.1 in /usr/local/lib/python3.10/dist-packages (from flair==0.4.3) (0.2.7)\n",
            "Collecting pytorch-transformers>=1.1.0 (from flair==0.4.3)\n",
            "  Downloading pytorch_transformers-1.2.0-py3-none-any.whl (176 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m176.4/176.4 kB\u001b[0m \u001b[31m9.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting bpemb>=0.2.9 (from flair==0.4.3)\n",
            "  Downloading bpemb-0.3.4-py3-none-any.whl (19 kB)\n",
            "Requirement already satisfied: regex in /usr/local/lib/python3.10/dist-packages (from flair==0.4.3) (2023.6.3)\n",
            "Requirement already satisfied: tabulate in /usr/local/lib/python3.10/dist-packages (from flair==0.4.3) (0.9.0)\n",
            "Collecting urllib3<1.25,>=1.20 (from flair==0.4.3)\n",
            "  Downloading urllib3-1.24.3-py2.py3-none-any.whl (118 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m118.8/118.8 kB\u001b[0m \u001b[31m13.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting langdetect (from flair==0.4.3)\n",
            "  Downloading langdetect-1.0.9.tar.gz (981 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m981.5/981.5 kB\u001b[0m \u001b[31m14.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "Collecting ipython==7.6.1 (from flair==0.4.3)\n",
            "  Downloading ipython-7.6.1-py3-none-any.whl (774 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m774.3/774.3 kB\u001b[0m \u001b[31m18.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: ipython-genutils==0.2.0 in /usr/local/lib/python3.10/dist-packages (from flair==0.4.3) (0.2.0)\n",
            "Requirement already satisfied: setuptools>=18.5 in /usr/local/lib/python3.10/dist-packages (from ipython==7.6.1->flair==0.4.3) (67.7.2)\n",
            "Collecting jedi>=0.10 (from ipython==7.6.1->flair==0.4.3)\n",
            "  Downloading jedi-0.19.0-py2.py3-none-any.whl (1.6 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.6/1.6 MB\u001b[0m \u001b[31m24.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: decorator in /usr/local/lib/python3.10/dist-packages (from ipython==7.6.1->flair==0.4.3) (4.4.2)\n",
            "Requirement already satisfied: pickleshare in /usr/local/lib/python3.10/dist-packages (from ipython==7.6.1->flair==0.4.3) (0.7.5)\n",
            "Requirement already satisfied: traitlets>=4.2 in /usr/local/lib/python3.10/dist-packages (from ipython==7.6.1->flair==0.4.3) (5.7.1)\n",
            "Collecting prompt-toolkit<2.1.0,>=2.0.0 (from ipython==7.6.1->flair==0.4.3)\n",
            "  Downloading prompt_toolkit-2.0.10-py3-none-any.whl (340 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m340.0/340.0 kB\u001b[0m \u001b[31m20.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: pygments in /usr/local/lib/python3.10/dist-packages (from ipython==7.6.1->flair==0.4.3) (2.16.1)\n",
            "Requirement already satisfied: backcall in /usr/local/lib/python3.10/dist-packages (from ipython==7.6.1->flair==0.4.3) (0.2.0)\n",
            "Requirement already satisfied: pexpect in /usr/local/lib/python3.10/dist-packages (from ipython==7.6.1->flair==0.4.3) (4.8.0)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.10/dist-packages (from bpemb>=0.2.9->flair==0.4.3) (1.23.5)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.10/dist-packages (from bpemb>=0.2.9->flair==0.4.3) (2.31.0)\n",
            "Collecting sentencepiece (from bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading sentencepiece-0.1.99-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.3 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.3/1.3 MB\u001b[0m \u001b[31m30.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: wrapt<2,>=1.10 in /usr/local/lib/python3.10/dist-packages (from deprecated>=1.2.4->flair==0.4.3) (1.15.0)\n",
            "Requirement already satisfied: scipy>=1.7.0 in /usr/local/lib/python3.10/dist-packages (from gensim>=3.4.0->flair==0.4.3) (1.11.2)\n",
            "Requirement already satisfied: smart-open>=1.8.1 in /usr/local/lib/python3.10/dist-packages (from gensim>=3.4.0->flair==0.4.3) (6.4.0)\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.10/dist-packages (from hyperopt>=0.1.1->flair==0.4.3) (1.16.0)\n",
            "Requirement already satisfied: networkx>=2.2 in /usr/local/lib/python3.10/dist-packages (from hyperopt>=0.1.1->flair==0.4.3) (3.1)\n",
            "Requirement already satisfied: future in /usr/local/lib/python3.10/dist-packages (from hyperopt>=0.1.1->flair==0.4.3) (0.18.3)\n",
            "Requirement already satisfied: cloudpickle in /usr/local/lib/python3.10/dist-packages (from hyperopt>=0.1.1->flair==0.4.3) (2.2.1)\n",
            "Requirement already satisfied: py4j in /usr/local/lib/python3.10/dist-packages (from hyperopt>=0.1.1->flair==0.4.3) (0.10.9.7)\n",
            "Requirement already satisfied: contourpy>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.2.3->flair==0.4.3) (1.1.0)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.2.3->flair==0.4.3) (0.11.0)\n",
            "Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.2.3->flair==0.4.3) (4.42.1)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.2.3->flair==0.4.3) (1.4.5)\n",
            "Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.2.3->flair==0.4.3) (23.1)\n",
            "Requirement already satisfied: pillow>=6.2.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.2.3->flair==0.4.3) (9.4.0)\n",
            "Requirement already satisfied: pyparsing>=2.3.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.2.3->flair==0.4.3) (3.1.1)\n",
            "Requirement already satisfied: python-dateutil>=2.7 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=2.2.3->flair==0.4.3) (2.8.2)\n",
            "Requirement already satisfied: iniconfig in /usr/local/lib/python3.10/dist-packages (from pytest>=3.6.4->flair==0.4.3) (2.0.0)\n",
            "Requirement already satisfied: pluggy<2.0,>=0.12 in /usr/local/lib/python3.10/dist-packages (from pytest>=3.6.4->flair==0.4.3) (1.3.0)\n",
            "Requirement already satisfied: exceptiongroup>=1.0.0rc8 in /usr/local/lib/python3.10/dist-packages (from pytest>=3.6.4->flair==0.4.3) (1.1.3)\n",
            "Requirement already satisfied: tomli>=1.0.0 in /usr/local/lib/python3.10/dist-packages (from pytest>=3.6.4->flair==0.4.3) (2.0.1)\n",
            "Collecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.47-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m18.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting sacremoses (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading sacremoses-0.0.53.tar.gz (880 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m880.6/880.6 kB\u001b[0m \u001b[31m35.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from torch>=1.1.0->flair==0.4.3) (3.12.2)\n",
            "Requirement already satisfied: typing-extensions in /usr/local/lib/python3.10/dist-packages (from torch>=1.1.0->flair==0.4.3) (4.5.0)\n",
            "Requirement already satisfied: sympy in /usr/local/lib/python3.10/dist-packages (from torch>=1.1.0->flair==0.4.3) (1.12)\n",
            "Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from torch>=1.1.0->flair==0.4.3) (3.1.2)\n",
            "Requirement already satisfied: triton==2.0.0 in /usr/local/lib/python3.10/dist-packages (from torch>=1.1.0->flair==0.4.3) (2.0.0)\n",
            "Requirement already satisfied: cmake in /usr/local/lib/python3.10/dist-packages (from triton==2.0.0->torch>=1.1.0->flair==0.4.3) (3.27.4.1)\n",
            "Requirement already satisfied: lit in /usr/local/lib/python3.10/dist-packages (from triton==2.0.0->torch>=1.1.0->flair==0.4.3) (16.0.6)\n",
            "Requirement already satisfied: parso<0.9.0,>=0.8.3 in /usr/local/lib/python3.10/dist-packages (from jedi>=0.10->ipython==7.6.1->flair==0.4.3) (0.8.3)\n",
            "Requirement already satisfied: wcwidth in /usr/local/lib/python3.10/dist-packages (from prompt-toolkit<2.1.0,>=2.0.0->ipython==7.6.1->flair==0.4.3) (0.2.6)\n",
            "Collecting botocore<1.32.0,>=1.31.47 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.47-py3-none-any.whl (11.2 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.2/11.2 MB\u001b[0m \u001b[31m59.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting jmespath<2.0.0,>=0.7.1 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading jmespath-1.0.1-py3-none-any.whl (20 kB)\n",
            "Collecting s3transfer<0.7.0,>=0.6.0 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading s3transfer-0.6.2-py3-none-any.whl (79 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m79.8/79.8 kB\u001b[0m \u001b[31m10.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->torch>=1.1.0->flair==0.4.3) (2.1.3)\n",
            "Requirement already satisfied: ptyprocess>=0.5 in /usr/local/lib/python3.10/dist-packages (from pexpect->ipython==7.6.1->flair==0.4.3) (0.7.0)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests->bpemb>=0.2.9->flair==0.4.3) (3.2.0)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests->bpemb>=0.2.9->flair==0.4.3) (3.4)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests->bpemb>=0.2.9->flair==0.4.3) (2023.7.22)\n",
            "Requirement already satisfied: click in /usr/local/lib/python3.10/dist-packages (from sacremoses->pytorch-transformers>=1.1.0->flair==0.4.3) (8.1.7)\n",
            "Requirement already satisfied: joblib in /usr/local/lib/python3.10/dist-packages (from sacremoses->pytorch-transformers>=1.1.0->flair==0.4.3) (1.3.2)\n",
            "Requirement already satisfied: mpmath>=0.19 in /usr/local/lib/python3.10/dist-packages (from sympy->torch>=1.1.0->flair==0.4.3) (1.3.0)\n",
            "INFO: pip is looking at multiple versions of botocore to determine which version is compatible with other requirements. This could take a while.\n",
            "Collecting requests (from bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading requests-2.31.0-py3-none-any.whl (62 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m62.6/62.6 kB\u001b[0m \u001b[31m8.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.30.0-py3-none-any.whl (62 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m62.5/62.5 kB\u001b[0m \u001b[31m8.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.29.0-py3-none-any.whl (62 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m62.5/62.5 kB\u001b[0m \u001b[31m8.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.28.2-py3-none-any.whl (62 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m62.8/62.8 kB\u001b[0m \u001b[31m9.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.28.1-py3-none-any.whl (62 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m62.8/62.8 kB\u001b[0m \u001b[31m9.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting charset-normalizer<3,>=2 (from requests->bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading charset_normalizer-2.1.1-py3-none-any.whl (39 kB)\n",
            "Collecting requests (from bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading requests-2.28.0-py3-none-any.whl (62 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m62.8/62.8 kB\u001b[0m \u001b[31m9.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting charset-normalizer~=2.0.0 (from requests->bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading charset_normalizer-2.0.12-py3-none-any.whl (39 kB)\n",
            "Collecting requests (from bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading requests-2.27.1-py2.py3-none-any.whl (63 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m63.1/63.1 kB\u001b[0m \u001b[31m8.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hINFO: pip is looking at multiple versions of botocore to determine which version is compatible with other requirements. This could take a while.\n",
            "  Downloading requests-2.27.0-py2.py3-none-any.whl (63 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m63.1/63.1 kB\u001b[0m \u001b[31m9.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.26.0-py2.py3-none-any.whl (62 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m62.3/62.3 kB\u001b[0m \u001b[31m7.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.25.1-py2.py3-none-any.whl (61 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m61.2/61.2 kB\u001b[0m \u001b[31m8.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting chardet<5,>=3.0.2 (from requests->bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading chardet-4.0.0-py2.py3-none-any.whl (178 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m178.7/178.7 kB\u001b[0m \u001b[31m22.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting idna<3,>=2.5 (from requests->bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading idna-2.10-py2.py3-none-any.whl (58 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m58.8/58.8 kB\u001b[0m \u001b[31m8.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting requests (from bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading requests-2.25.0-py2.py3-none-any.whl (61 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m61.1/61.1 kB\u001b[0m \u001b[31m9.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting chardet<4,>=3.0.2 (from requests->bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading chardet-3.0.4-py2.py3-none-any.whl (133 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m133.4/133.4 kB\u001b[0m \u001b[31m16.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting requests (from bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading requests-2.24.0-py2.py3-none-any.whl (61 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m61.8/61.8 kB\u001b[0m \u001b[31m8.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hINFO: This is taking longer than usual. You might need to provide the dependency resolver with stricter constraints to reduce runtime. See https://pip.pypa.io/warnings/backtracking for guidance. If you want to abort this run, press Ctrl + C.\n",
            "  Downloading requests-2.23.0-py2.py3-none-any.whl (58 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m58.4/58.4 kB\u001b[0m \u001b[31m7.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.22.0-py2.py3-none-any.whl (57 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m58.0/58.0 kB\u001b[0m \u001b[31m8.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting idna<2.9,>=2.5 (from requests->bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading idna-2.8-py2.py3-none-any.whl (58 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m58.6/58.6 kB\u001b[0m \u001b[31m8.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting requests (from bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading requests-2.21.0-py2.py3-none-any.whl (57 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m58.0/58.0 kB\u001b[0m \u001b[31m7.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.20.1-py2.py3-none-any.whl (57 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m58.0/58.0 kB\u001b[0m \u001b[31m7.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting idna<2.8,>=2.5 (from requests->bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading idna-2.7-py2.py3-none-any.whl (58 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m58.2/58.2 kB\u001b[0m \u001b[31m8.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting requests (from bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading requests-2.20.0-py2.py3-none-any.whl (60 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m61.0/61.0 kB\u001b[0m \u001b[31m8.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.19.1-py2.py3-none-any.whl (91 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m92.0/92.0 kB\u001b[0m \u001b[31m13.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting urllib3<1.25,>=1.20 (from flair==0.4.3)\n",
            "  Downloading urllib3-1.23-py2.py3-none-any.whl (133 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m133.3/133.3 kB\u001b[0m \u001b[31m18.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting requests (from bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading requests-2.19.0-py2.py3-none-any.whl (91 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m91.9/91.9 kB\u001b[0m \u001b[31m13.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.18.4-py2.py3-none-any.whl (88 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m88.7/88.7 kB\u001b[0m \u001b[31m13.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting idna<2.7,>=2.5 (from requests->bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading idna-2.6-py2.py3-none-any.whl (56 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m56.5/56.5 kB\u001b[0m \u001b[31m7.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting urllib3<1.25,>=1.20 (from flair==0.4.3)\n",
            "  Downloading urllib3-1.22-py2.py3-none-any.whl (132 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m132.3/132.3 kB\u001b[0m \u001b[31m15.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting requests (from bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading requests-2.18.3-py2.py3-none-any.whl (88 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m88.6/88.6 kB\u001b[0m \u001b[31m12.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting idna<2.6,>=2.5 (from requests->bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading idna-2.5-py2.py3-none-any.whl (55 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m55.1/55.1 kB\u001b[0m \u001b[31m7.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting requests (from bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading requests-2.18.2-py2.py3-none-any.whl (88 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m88.3/88.3 kB\u001b[0m \u001b[31m10.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.18.1-py2.py3-none-any.whl (88 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m88.1/88.1 kB\u001b[0m \u001b[31m11.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting urllib3<1.25,>=1.20 (from flair==0.4.3)\n",
            "  Downloading urllib3-1.21.1-py2.py3-none-any.whl (131 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m131.7/131.7 kB\u001b[0m \u001b[31m17.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting requests (from bpemb>=0.2.9->flair==0.4.3)\n",
            "  Downloading requests-2.18.0-py2.py3-none-any.whl (563 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m563.6/563.6 kB\u001b[0m \u001b[31m35.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.17.3-py2.py3-none-any.whl (87 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m87.5/87.5 kB\u001b[0m \u001b[31m10.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.17.2-py2.py3-none-any.whl (87 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m87.5/87.5 kB\u001b[0m \u001b[31m11.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.17.1-py2.py3-none-any.whl (87 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m87.5/87.5 kB\u001b[0m \u001b[31m9.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.17.0-py2.py3-none-any.whl (87 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m87.4/87.4 kB\u001b[0m \u001b[31m11.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.16.5-py2.py3-none-any.whl (87 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m87.6/87.6 kB\u001b[0m \u001b[31m11.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.16.4-py2.py3-none-any.whl (87 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m87.6/87.6 kB\u001b[0m \u001b[31m11.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.16.3-py2.py3-none-any.whl (86 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m86.3/86.3 kB\u001b[0m \u001b[31m12.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.16.2-py2.py3-none-any.whl (86 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m86.3/86.3 kB\u001b[0m \u001b[31m11.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.16.1-py2.py3-none-any.whl (85 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m85.7/85.7 kB\u001b[0m \u001b[31m8.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.16.0-py2.py3-none-any.whl (85 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m85.4/85.4 kB\u001b[0m \u001b[31m11.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Downloading requests-2.15.1-py2.py3-none-any.whl (558 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m558.7/558.7 kB\u001b[0m \u001b[31m48.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.46-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m18.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.46 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.46-py3-none-any.whl (11.2 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.2/11.2 MB\u001b[0m \u001b[31m69.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.45-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m19.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.45 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.45-py3-none-any.whl (11.2 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.2/11.2 MB\u001b[0m \u001b[31m51.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.44-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m19.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.44 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.44-py3-none-any.whl (11.2 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.2/11.2 MB\u001b[0m \u001b[31m79.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.43-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m2.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.43 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.43-py3-none-any.whl (11.2 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.2/11.2 MB\u001b[0m \u001b[31m17.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.42-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m19.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.42 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.42-py3-none-any.whl (11.2 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.2/11.2 MB\u001b[0m \u001b[31m77.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.41-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m2.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.41 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.41-py3-none-any.whl (11.2 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.2/11.2 MB\u001b[0m \u001b[31m76.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.40-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m16.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.40 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.40-py3-none-any.whl (11.2 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.2/11.2 MB\u001b[0m \u001b[31m57.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.39-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m15.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.39 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.39-py3-none-any.whl (11.2 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.2/11.2 MB\u001b[0m \u001b[31m46.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.38-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m19.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.38 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.38-py3-none-any.whl (11.2 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.2/11.2 MB\u001b[0m \u001b[31m13.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.37-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m2.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.37 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.37-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m88.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.36-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m18.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.36 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.36-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m78.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.35-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m16.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.35 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.35-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m81.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.34-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m19.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.34 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.34-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m44.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.33-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m18.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.33 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.33-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m22.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.32-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m10.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.32 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.32-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m46.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.31-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m16.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.31 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.31-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m13.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.30-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m16.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.30 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.30-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m82.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.29-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m17.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.29 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.29-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m62.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.28-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m15.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.28 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.28-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m31.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.27-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m16.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.27 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.27-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m16.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.26-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m11.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.26 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.26-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m70.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.25-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m16.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.25 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.25-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m23.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.24-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m11.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.24 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.24-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m32.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.23-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m15.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.23 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.23-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m22.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.22-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m18.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.22 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.22-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m22.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.21-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m17.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.21 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.21-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m9.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.20-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m15.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.20 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.20-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m16.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.19-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m18.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.19 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.19-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m13.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.18-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m9.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.18 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.18-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m17.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.17-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m17.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.17 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.17-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m99.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.16-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m18.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.16 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.16-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m74.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.15-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m12.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.15 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.15-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m13.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.14-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m19.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.14 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.14-py3-none-any.whl (11.1 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.1/11.1 MB\u001b[0m \u001b[31m13.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.13-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m16.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.13 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.13-py3-none-any.whl (11.0 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.0/11.0 MB\u001b[0m \u001b[31m54.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.12-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m15.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.12 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.12-py3-none-any.whl (11.0 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.0/11.0 MB\u001b[0m \u001b[31m49.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.11-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m17.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.11 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.11-py3-none-any.whl (11.0 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.0/11.0 MB\u001b[0m \u001b[31m13.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.10-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.8/135.8 kB\u001b[0m \u001b[31m18.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.10 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.10-py3-none-any.whl (11.0 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.0/11.0 MB\u001b[0m \u001b[31m14.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.9-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.7/135.7 kB\u001b[0m \u001b[31m17.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.9 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.9-py3-none-any.whl (11.0 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.0/11.0 MB\u001b[0m \u001b[31m10.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.8-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.7/135.7 kB\u001b[0m \u001b[31m17.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.8 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.8-py3-none-any.whl (11.0 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.0/11.0 MB\u001b[0m \u001b[31m37.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.7-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.7/135.7 kB\u001b[0m \u001b[31m15.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.7 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.7-py3-none-any.whl (11.0 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m11.0/11.0 MB\u001b[0m \u001b[31m15.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting boto3 (from pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading boto3-1.28.6-py3-none-any.whl (135 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m135.7/135.7 kB\u001b[0m \u001b[31m17.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting botocore<1.32.0,>=1.31.6 (from boto3->pytorch-transformers>=1.1.0->flair==0.4.3)\n",
            "  Downloading botocore-1.31.6-py3-none-any.whl (11.0 MB)\n",
            "\u001b[2K     \u001b[91m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[90m╺\u001b[0m\u001b[90m━━━━━━━━━\u001b[0m \u001b[32m8.4/11.0 MB\u001b[0m \u001b[31m20.9 MB/s\u001b[0m eta \u001b[36m0:00:01\u001b[0m\n",
            "\u001b[?25h\u001b[31mERROR: Operation cancelled by user\u001b[0m\u001b[31m\n",
            "\u001b[0m"
          ]
        },
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-17-376755bb3978>\u001b[0m in \u001b[0;36m<cell line: 8>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      6\u001b[0m \u001b[0mencoder_inputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mInput\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"int64\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"english\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mPositionalEmbedding\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msequence_length\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvocab_size\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0membed_dim\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mencoder_inputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m \u001b[0mencoder_outputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mTransformerEncoder\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0membed_dim\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdense_dim\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnum_heads\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      9\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     10\u001b[0m \u001b[0mdecoder_inputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mInput\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"int64\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"spanish\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'TransformerEncoder' is not defined"
          ]
        }
      ],
      "source": [
        "!pip install flair==0.4.3                                                       # TODO - Move above and fix!\n",
        "embed_dim = 256\n",
        "dense_dim = 2048\n",
        "num_heads = 8\n",
        "\n",
        "encoder_inputs = keras.Input(shape=(None,), dtype=\"int64\", name=\"english\")\n",
        "x = PositionalEmbedding(sequence_length, vocab_size, embed_dim)(encoder_inputs)\n",
        "encoder_outputs = TransformerEncoder(embed_dim, dense_dim, num_heads)(x)\n",
        "\n",
        "decoder_inputs = keras.Input(shape=(None,), dtype=\"int64\", name=\"spanish\")\n",
        "x = PositionalEmbedding(sequence_length, vocab_size, embed_dim)(decoder_inputs)\n",
        "x = TransformerDecoder(embed_dim, dense_dim, num_heads)(x, encoder_outputs)\n",
        "x = layers.Dropout(0.5)(x)\n",
        "decoder_outputs = layers.Dense(vocab_size, activation=\"softmax\")(x)\n",
        "transformer = keras.Model([encoder_inputs, decoder_inputs], decoder_outputs)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4fu1UaMktO1R"
      },
      "source": [
        "**Training the sequence-to-sequence Transformer**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "7TvGfTfttO1R"
      },
      "outputs": [],
      "source": [
        "transformer.compile(\n",
        "    optimizer=\"rmsprop\",\n",
        "    loss=\"sparse_categorical_crossentropy\",\n",
        "    metrics=[\"accuracy\"])\n",
        "transformer.fit(train_ds, epochs=30, validation_data=val_ds)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "j3rctndOtO1S"
      },
      "source": [
        "**Translating new sentences with our Transformer model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "EOQR7H38tO1S"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "spa_vocab = target_vectorization.get_vocabulary()\n",
        "spa_index_lookup = dict(zip(range(len(spa_vocab)), spa_vocab))\n",
        "max_decoded_sentence_length = 20\n",
        "\n",
        "def decode_sequence(input_sentence):\n",
        "    tokenized_input_sentence = source_vectorization([input_sentence])\n",
        "    decoded_sentence = \"[start]\"\n",
        "    for i in range(max_decoded_sentence_length):\n",
        "        tokenized_target_sentence = target_vectorization(\n",
        "            [decoded_sentence])[:, :-1]\n",
        "        predictions = transformer(\n",
        "            [tokenized_input_sentence, tokenized_target_sentence])\n",
        "        sampled_token_index = np.argmax(predictions[0, i, :])\n",
        "        sampled_token = spa_index_lookup[sampled_token_index]\n",
        "        decoded_sentence += \" \" + sampled_token\n",
        "        if sampled_token == \"[end]\":\n",
        "            break\n",
        "    return decoded_sentence\n",
        "\n",
        "test_eng_texts = [pair[0] for pair in test_pairs]\n",
        "for _ in range(20):\n",
        "    input_sentence = random.choice(test_eng_texts)\n",
        "    print(\"-\")\n",
        "    print(input_sentence)\n",
        "    print(decode_sequence(input_sentence))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ugQFYeCFtO1T"
      },
      "source": [
        "## Summary"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "name": "chapter11_part04_sequence-to-sequence-learning.i",
      "provenance": [],
      "toc_visible": true,
      "gpuType": "T4",
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.0"
    },
    "accelerator": "GPU"
  },
  "nbformat": 4,
  "nbformat_minor": 0
}