notebooks/11_music/01_transformer/transformer.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "id": "b076bd1a-b236-4fbc-953d-8295b25122ae",
      "metadata": {
        "id": "b076bd1a-b236-4fbc-953d-8295b25122ae"
      },
      "source": [
        "# 🎶 트랜스포머를 사용한 음악 생성"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "38034f04",
      "metadata": {
        "id": "38034f04"
      },
      "source": [
        "<table align=\"left\">\n",
        "  <td>\n",
        "    <a href=\"https://colab.research.google.com/github/rickiepark/Generative_Deep_Learning_2nd_Edition/blob/main/notebooks/11_music/01_transformer/transformer.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>\n",
        "  </td>\n",
        "</table>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "id": "5fe8cdab",
      "metadata": {
        "id": "5fe8cdab",
        "outputId": "89946759-3f08-4fbc-c376-ea133b5bd0c6",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2024-04-08 04:42:38--  https://raw.githubusercontent.com/rickiepark/Generative_Deep_Learning_2nd_Edition/main/notebooks/11_music/01_transformer/transformer_utils.py\n",
            "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.108.133, 185.199.109.133, 185.199.110.133, ...\n",
            "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.108.133|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 6876 (6.7K) [text/plain]\n",
            "Saving to: ‘transformer_utils.py’\n",
            "\n",
            "\rtransformer_utils.p   0%[                    ]       0  --.-KB/s               \rtransformer_utils.p 100%[===================>]   6.71K  --.-KB/s    in 0s      \n",
            "\n",
            "2024-04-08 04:42:38 (69.8 MB/s) - ‘transformer_utils.py’ saved [6876/6876]\n",
            "\n",
            "Extracting templates from packages: 100%\n"
          ]
        }
      ],
      "source": [
        "import sys\n",
        "\n",
        "# 코랩의 경우 깃허브 저장소로부터 transformer_utils.py를 다운로드 합니다.\n",
        "if 'google.colab' in sys.modules:\n",
        "    !wget https://raw.githubusercontent.com/rickiepark/Generative_Deep_Learning_2nd_Edition/main/notebooks/11_music/01_transformer/transformer_utils.py\n",
        "    # output 디렉토리를 만듭니다.\n",
        "    !mkdir bach-cello\n",
        "    !mkdir parsed_data\n",
        "    !mkdir models\n",
        "    !mkdir output\n",
        "    # 악보 이미지 생성을 위해\n",
        "    !apt-get install lilypond > /dev/null"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "9235cbd1-f136-411c-88d9-f69f270c0b96",
      "metadata": {
        "id": "9235cbd1-f136-411c-88d9-f69f270c0b96"
      },
      "source": [
        "이 노트북에서는 트랜스포머 모델을 훈련시켜 바흐 첼로 모음곡 스타일로 음악을 생성합니다."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "id": "84acc7be-6764-4668-b2bb-178f63deeed3",
      "metadata": {
        "id": "84acc7be-6764-4668-b2bb-178f63deeed3"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "import glob\n",
        "import numpy as np\n",
        "import time\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras import layers, models, losses, callbacks\n",
        "\n",
        "import music21\n",
        "\n",
        "from transformer_utils import (\n",
        "    parse_midi_files,\n",
        "    load_parsed_files,\n",
        "    get_midi_note,\n",
        "    SinePositionEncoding,\n",
        ")"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "339e6268-ebd7-4feb-86db-1fe7abccdbe5",
      "metadata": {
        "id": "339e6268-ebd7-4feb-86db-1fe7abccdbe5"
      },
      "source": [
        "## 0. 파라미터 <a name=\"parameters\"></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "id": "1b2ee6ce-129f-4833-b0c5-fa567381c4e0",
      "metadata": {
        "id": "1b2ee6ce-129f-4833-b0c5-fa567381c4e0"
      },
      "outputs": [],
      "source": [
        "PARSE_MIDI_FILES = True\n",
        "PARSED_DATA_PATH = \"./parsed_data/\"\n",
        "DATASET_REPETITIONS = 1\n",
        "\n",
        "SEQ_LEN = 50\n",
        "EMBEDDING_DIM = 256\n",
        "KEY_DIM = 256\n",
        "N_HEADS = 5\n",
        "DROPOUT_RATE = 0.3\n",
        "FEED_FORWARD_DIM = 256\n",
        "LOAD_MODEL = False\n",
        "\n",
        "# 최적화\n",
        "EPOCHS = 20\n",
        "BATCH_SIZE = 256\n",
        "\n",
        "GENERATE_LEN = 50"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "7d4f5e63-e36a-4dc8-9f03-cb29c1fa5290",
      "metadata": {
        "id": "7d4f5e63-e36a-4dc8-9f03-cb29c1fa5290"
      },
      "source": [
        "## 1. 데이터 준비"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "id": "89441fa5",
      "metadata": {
        "id": "89441fa5",
        "outputId": "c23b3dc3-74a8-4458-ff75-9e085632053e",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/content/bach-cello\n",
            "/content\n"
          ]
        }
      ],
      "source": [
        "# 코랩의 경우 노트북에서 첼로 데이터셋을 다운로드합니다.\n",
        "if 'google.colab' in sys.modules:\n",
        "    %cd ./bach-cello/\n",
        "    !curl -O http://www.jsbach.net/midi/cs1-1pre.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs1-2all.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs1-3cou.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs1-4sar.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs1-5men.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs1-6gig.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs2-1pre.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs2-2all.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs2-3cou.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs2-4sar.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs2-5men.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs2-6gig.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs3-1pre.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs3-2all.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs3-3cou.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs3-4sar.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs3-5bou.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs3-6gig.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs4-1pre.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs4-2all.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs4-3cou.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs4-4sar.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs4-5bou.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs4-6gig.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs5-1pre.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs5-2all.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs5-3cou.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs5-4sar.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs5-5gav.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs5-6gig.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs6-1pre.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs6-2all.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs6-3cou.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs6-4sar.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs6-5gav.mid -s\n",
        "    !curl -O http://www.jsbach.net/midi/cs6-6gig.mid -s\n",
        "    %cd .."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "id": "73de38bd-0b92-4441-9601-ed4a3b45f924",
      "metadata": {
        "id": "73de38bd-0b92-4441-9601-ed4a3b45f924",
        "outputId": "b74ec5d2-1033-4241-a3dd-6e90b0b093a9",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "36개 midi 파일 찾음\n"
          ]
        }
      ],
      "source": [
        "# 데이터 로드\n",
        "file_list = glob.glob(\"./bach-cello/*.mid\")\n",
        "print(f\"{len(file_list)}개 midi 파일 찾음\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "id": "b1ff575d-1632-43bf-844b-e5f2cea61454",
      "metadata": {
        "id": "b1ff575d-1632-43bf-844b-e5f2cea61454"
      },
      "outputs": [],
      "source": [
        "parser = music21.converter"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "id": "b563da4f-0b08-4005-aa35-ca58a60a7def",
      "metadata": {
        "tags": [],
        "id": "b563da4f-0b08-4005-aa35-ca58a60a7def"
      },
      "outputs": [],
      "source": [
        "example_score = \\\n",
        "    music21.converter.parse(file_list[1]).splitAtQuarterLength(12)[0].chordify()"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from IPython.display import Image, Audio"
      ],
      "metadata": {
        "id": "tFO7_JYkpigE"
      },
      "id": "tFO7_JYkpigE",
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "display(Image(str(example_score.write('lily.png'))))"
      ],
      "metadata": {
        "id": "nlie-J22pNBN",
        "outputId": "5452309f-57a7-41b8-9007-215bedb030d4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 99
        }
      },
      "id": "nlie-J22pNBN",
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "id": "debce40c-2d56-4140-b65b-459c6464c1a7",
      "metadata": {
        "id": "debce40c-2d56-4140-b65b-459c6464c1a7",
        "outputId": "8038bf8a-6a4e-4ac2-fe1d-2b357e06b221",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "{0.0} <music21.metadata.Metadata object at 0x7a4ac7330dc0>\n",
            "{0.0} <music21.stream.Measure 1 offset=0.0>\n",
            "    {0.0} <music21.instrument.Violoncello 'Solo Cello: Solo Cello'>\n",
            "    {0.0} <music21.instrument.Violoncello 'Violoncello'>\n",
            "    {0.0} <music21.clef.BassClef>\n",
            "    {0.0} <music21.tempo.MetronomeMark andantino Quarter=80>\n",
            "    {0.0} <music21.key.Key of G major>\n",
            "    {0.0} <music21.meter.TimeSignature 4/4>\n",
            "    {0.0} <music21.chord.Chord G2>\n",
            "    {0.25} <music21.chord.Chord D3>\n",
            "    {0.5} <music21.chord.Chord B3>\n",
            "    {0.75} <music21.chord.Chord A3>\n",
            "    {1.0} <music21.chord.Chord B3>\n",
            "    {1.25} <music21.chord.Chord D3>\n",
            "    {1.5} <music21.chord.Chord B3>\n",
            "    {1.75} <music21.chord.Chord D3>\n",
            "    {2.0} <music21.chord.Chord G2>\n",
            "    {2.25} <music21.chord.Chord D3>\n",
            "    {2.5} <music21.chord.Chord B3>\n",
            "    {2.75} <music21.chord.Chord A3>\n",
            "    {3.0} <music21.chord.Chord B3>\n",
            "    {3.25} <music21.chord.Chord D3>\n",
            "    {3.5} <music21.chord.Chord B3>\n",
            "    {3.75} <music21.chord.Chord D3>\n",
            "{4.0} <music21.stream.Measure 2 offset=4.0>\n",
            "    {0.0} <music21.chord.Chord G2>\n",
            "    {0.25} <music21.chord.Chord E3>\n",
            "    {0.5} <music21.chord.Chord C4>\n",
            "    {0.75} <music21.chord.Chord B3>\n",
            "    {1.0} <music21.chord.Chord C4>\n",
            "    {1.25} <music21.chord.Chord E3>\n",
            "    {1.5} <music21.chord.Chord C4>\n",
            "    {1.75} <music21.chord.Chord E3>\n",
            "    {2.0} <music21.chord.Chord G2>\n",
            "    {2.25} <music21.chord.Chord E3>\n",
            "    {2.5} <music21.chord.Chord C4>\n",
            "    {2.75} <music21.chord.Chord B3>\n",
            "    {3.0} <music21.chord.Chord C4>\n",
            "    {3.25} <music21.chord.Chord E3>\n",
            "    {3.5} <music21.chord.Chord C4>\n",
            "    {3.75} <music21.chord.Chord E3>\n",
            "{8.0} <music21.stream.Measure 3 offset=8.0>\n",
            "    {0.0} <music21.chord.Chord G2>\n",
            "    {0.25} <music21.chord.Chord F#3>\n",
            "    {0.5} <music21.chord.Chord C4>\n",
            "    {0.75} <music21.chord.Chord B3>\n",
            "    {1.0} <music21.chord.Chord C4>\n",
            "    {1.25} <music21.chord.Chord F#3>\n",
            "    {1.5} <music21.chord.Chord C4>\n",
            "    {1.75} <music21.chord.Chord F#3>\n",
            "    {2.0} <music21.chord.Chord G2>\n",
            "    {2.25} <music21.chord.Chord F#3>\n",
            "    {2.5} <music21.chord.Chord C4>\n",
            "    {2.75} <music21.chord.Chord B3>\n",
            "    {3.0} <music21.chord.Chord C4>\n",
            "    {3.25} <music21.chord.Chord F#3>\n",
            "    {3.5} <music21.chord.Chord C4>\n",
            "    {3.75} <music21.chord.Chord F#3>\n"
          ]
        }
      ],
      "source": [
        "example_score.show(\"text\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "id": "77459313-2417-4c18-b938-3a9859ec9bd9",
      "metadata": {
        "tags": [],
        "id": "77459313-2417-4c18-b938-3a9859ec9bd9",
        "outputId": "757cd5bc-8d4d-446b-d17c-071d5afe4b67",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1 ./bach-cello/cs4-3cou.mid 파싱\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "<ipython-input-11-04172d0f16ff>:2: Music21DeprecationWarning: .flat is deprecated.  Call .flatten() instead\n",
            "  notes, durations = parse_midi_files(\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "922개 음표 파싱\n",
            "2 ./bach-cello/cs1-1pre.mid 파싱\n",
            "1580개 음표 파싱\n",
            "3 ./bach-cello/cs4-4sar.mid 파싱\n",
            "1900개 음표 파싱\n",
            "4 ./bach-cello/cs4-2all.mid 파싱\n",
            "2955개 음표 파싱\n",
            "5 ./bach-cello/cs3-3cou.mid 파싱\n",
            "3944개 음표 파싱\n",
            "6 ./bach-cello/cs4-5bou.mid 파싱\n",
            "5268개 음표 파싱\n",
            "7 ./bach-cello/cs5-3cou.mid 파싱\n",
            "5698개 음표 파싱\n",
            "8 ./bach-cello/cs6-3cou.mid 파싱\n",
            "6984개 음표 파싱\n",
            "9 ./bach-cello/cs6-1pre.mid 파싱\n",
            "8326개 음표 파싱\n",
            "10 ./bach-cello/cs3-4sar.mid 파싱\n",
            "8667개 음표 파싱\n",
            "11 ./bach-cello/cs2-2all.mid 파싱\n",
            "9360개 음표 파싱\n",
            "12 ./bach-cello/cs2-6gig.mid 파싱\n",
            "10095개 음표 파싱\n",
            "13 ./bach-cello/cs3-6gig.mid 파싱\n",
            "11056개 음표 파싱\n",
            "14 ./bach-cello/cs1-6gig.mid 파싱\n",
            "11482개 음표 파싱\n",
            "15 ./bach-cello/cs6-2all.mid 파싱\n",
            "12167개 음표 파싱\n",
            "16 ./bach-cello/cs2-3cou.mid 파싱\n",
            "12908개 음표 파싱\n",
            "17 ./bach-cello/cs5-1pre.mid 파싱\n",
            "14217개 음표 파싱\n",
            "18 ./bach-cello/cs6-4sar.mid 파싱\n",
            "14552개 음표 파싱\n",
            "19 ./bach-cello/cs3-5bou.mid 파싱\n",
            "15433개 음표 파싱\n",
            "20 ./bach-cello/cs5-4sar.mid 파싱\n",
            "15652개 음표 파싱\n",
            "21 ./bach-cello/cs4-6gig.mid 파싱\n",
            "16623개 음표 파싱\n",
            "22 ./bach-cello/cs3-2all.mid 파싱\n",
            "17360개 음표 파싱\n",
            "23 ./bach-cello/cs2-4sar.mid 파싱\n",
            "17697개 음표 파싱\n",
            "24 ./bach-cello/cs2-1pre.mid 파싱\n",
            "18335개 음표 파싱\n",
            "25 ./bach-cello/cs5-2all.mid 파싱\n",
            "19027개 음표 파싱\n",
            "26 ./bach-cello/cs5-6gig.mid 파싱\n",
            "19476개 음표 파싱\n",
            "27 ./bach-cello/cs6-5gav.mid 파싱\n",
            "20282개 음표 파싱\n",
            "28 ./bach-cello/cs1-2all.mid 파싱\n",
            "21203개 음표 파싱\n",
            "29 ./bach-cello/cs1-5men.mid 파싱\n",
            "21850개 음표 파싱\n",
            "30 ./bach-cello/cs3-1pre.mid 파싱\n",
            "22831개 음표 파싱\n",
            "31 ./bach-cello/cs5-5gav.mid 파싱\n",
            "24033개 음표 파싱\n",
            "32 ./bach-cello/cs4-1pre.mid 파싱\n",
            "24850개 음표 파싱\n",
            "33 ./bach-cello/cs6-6gig.mid 파싱\n",
            "26022개 음표 파싱\n",
            "34 ./bach-cello/cs2-5men.mid 파싱\n",
            "26549개 음표 파싱\n",
            "35 ./bach-cello/cs1-3cou.mid 파싱\n",
            "27369개 음표 파싱\n",
            "36 ./bach-cello/cs1-4sar.mid 파싱\n",
            "27632개 음표 파싱\n",
            "길이가 51인 시퀀스 만들기\n"
          ]
        }
      ],
      "source": [
        "if PARSE_MIDI_FILES:\n",
        "    notes, durations = parse_midi_files(\n",
        "        file_list, parser, SEQ_LEN + 1, PARSED_DATA_PATH\n",
        "    )\n",
        "else:\n",
        "    notes, durations = load_parsed_files()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "id": "99f17b33-193e-4d83-8e0a-54dc8e7b249d",
      "metadata": {
        "id": "99f17b33-193e-4d83-8e0a-54dc8e7b249d",
        "outputId": "c4795935-efad-4cec-8263-028a967fd3c9",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "음표 문자열\n",
            " E-3 G#3 F3 B2 G3 E-3 C2 G3 A3 B3 C4 B3 C4 D4 E-4 D4 E-4 C4 E-4 B-3 E-4 A3 E-4 G3 F3 E-4 D4 C4 D4 B-3 D4 A3 D4 G3 D4 F3 E-3 D4 C4 B-3 C4 G#3 C4 G3 C4 F3 C4 E-3 D3 F3 G#3 ...\n",
            "\n",
            "지속 시간 문자열\n",
            " 0.5 0.5 0.5 0.5 0.5 1.0 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 ...\n"
          ]
        }
      ],
      "source": [
        "example_notes = notes[658]\n",
        "example_durations = durations[658]\n",
        "print(\"\\n음표 문자열\\n\", example_notes, \"...\")\n",
        "print(\"\\n지속 시간 문자열\\n\", example_durations, \"...\")"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "aba2f39c-882f-423a-88eb-334502413639",
      "metadata": {
        "id": "aba2f39c-882f-423a-88eb-334502413639"
      },
      "source": [
        "## 2. 데이터 토큰화 <a name=\"tokenize\"></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "id": "244e093e-b977-4b1f-8f46-6503a55ea0fe",
      "metadata": {
        "id": "244e093e-b977-4b1f-8f46-6503a55ea0fe"
      },
      "outputs": [],
      "source": [
        "def create_dataset(elements):\n",
        "    ds = (\n",
        "        tf.data.Dataset.from_tensor_slices(elements)\n",
        "        .batch(BATCH_SIZE, drop_remainder=True)\n",
        "        .shuffle(1000)\n",
        "    )\n",
        "    vectorize_layer = layers.TextVectorization(\n",
        "        standardize=None, output_mode=\"int\"\n",
        "    )\n",
        "    vectorize_layer.adapt(ds)\n",
        "    vocab = vectorize_layer.get_vocabulary()\n",
        "    return ds, vectorize_layer, vocab\n",
        "\n",
        "\n",
        "notes_seq_ds, notes_vectorize_layer, notes_vocab = create_dataset(notes)\n",
        "durations_seq_ds, durations_vectorize_layer, durations_vocab = create_dataset(\n",
        "    durations\n",
        ")\n",
        "seq_ds = tf.data.Dataset.zip((notes_seq_ds, durations_seq_ds))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "id": "9ec4d36c-a4ad-4c32-89a2-749c21786441",
      "metadata": {
        "id": "9ec4d36c-a4ad-4c32-89a2-749c21786441",
        "outputId": "2a471375-e3ec-4f6f-9a3c-e608a1216c7e",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "note token duration token\n",
            "        10         3\n",
            "        14         3\n",
            "         5         3\n",
            "        18         3\n",
            "         2         3\n",
            "        10         4\n",
            "        34         2\n",
            "         2         2\n",
            "         3         2\n",
            "         9         2\n",
            "         6         2\n"
          ]
        }
      ],
      "source": [
        "# 동일한 예시 음표와 지속 시간을 int로 변환하여 표시합니다.\n",
        "example_tokenised_notes = notes_vectorize_layer(example_notes)\n",
        "example_tokenised_durations = durations_vectorize_layer(example_durations)\n",
        "print(\"{:10} {:10}\".format(\"note token\", \"duration token\"))\n",
        "for i, (note_int, duration_int) in enumerate(\n",
        "    zip(\n",
        "        example_tokenised_notes.numpy()[:11],\n",
        "        example_tokenised_durations.numpy()[:11],\n",
        "    )\n",
        "):\n",
        "    print(f\"{note_int:10}{duration_int:10}\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 15,
      "id": "8dc29b17-1591-4b02-98e1-7f25960350be",
      "metadata": {
        "id": "8dc29b17-1591-4b02-98e1-7f25960350be",
        "outputId": "3cf9fe03-8318-4911-a695-f3662f9dd42f",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "NOTES_VOCAB: length = 59\n",
            "0: \n",
            "1: [UNK]\n",
            "2: G3\n",
            "3: A3\n",
            "4: D3\n",
            "5: F3\n",
            "6: C4\n",
            "7: D4\n",
            "8: E3\n",
            "9: B3\n",
            "\n",
            "DURATIONS_VOCAB: length = 24\n",
            "0: \n",
            "1: [UNK]\n",
            "2: 0.25\n",
            "3: 0.5\n",
            "4: 1.0\n",
            "5: 1/3\n",
            "6: 0.75\n",
            "7: 1/12\n",
            "8: 1.5\n",
            "9: 0.0\n"
          ]
        }
      ],
      "source": [
        "notes_vocab_size = len(notes_vocab)\n",
        "durations_vocab_size = len(durations_vocab)\n",
        "\n",
        "# token:note 매핑 예시를 출력합니다.\n",
        "print(f\"\\nNOTES_VOCAB: length = {len(notes_vocab)}\")\n",
        "for i, note in enumerate(notes_vocab[:10]):\n",
        "    print(f\"{i}: {note}\")\n",
        "\n",
        "print(f\"\\nDURATIONS_VOCAB: length = {len(durations_vocab)}\")\n",
        "# Display some token:duration mappings\n",
        "for i, note in enumerate(durations_vocab[:10]):\n",
        "    print(f\"{i}: {note}\")"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "823fb0c1-ebf8-453b-be94-9a33b466cae4",
      "metadata": {
        "id": "823fb0c1-ebf8-453b-be94-9a33b466cae4"
      },
      "source": [
        "## 3. 훈련 세트 생성 <a name=\"create\"></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "id": "8f0f2a52-b157-478d-8de7-11ed6c383461",
      "metadata": {
        "tags": [],
        "id": "8f0f2a52-b157-478d-8de7-11ed6c383461"
      },
      "outputs": [],
      "source": [
        "# 시퀀스와 하나의 음표를 이동한 동일한 시퀀스로 훈련 세트를 만듭니다.\n",
        "def prepare_inputs(notes, durations):\n",
        "    notes = tf.expand_dims(notes, -1)\n",
        "    durations = tf.expand_dims(durations, -1)\n",
        "    tokenized_notes = notes_vectorize_layer(notes)\n",
        "    tokenized_durations = durations_vectorize_layer(durations)\n",
        "    x = (tokenized_notes[:, :-1], tokenized_durations[:, :-1])\n",
        "    y = (tokenized_notes[:, 1:], tokenized_durations[:, 1:])\n",
        "    return x, y\n",
        "\n",
        "\n",
        "ds = seq_ds.map(prepare_inputs).repeat(DATASET_REPETITIONS)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "id": "78248965-1716-4077-8db4-191e98c7a6a5",
      "metadata": {
        "id": "78248965-1716-4077-8db4-191e98c7a6a5",
        "outputId": "e5394506-6712-46ba-85b5-d9dbe6b68741",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "((<tf.Tensor: shape=(256, 50), dtype=int64, numpy=\n",
            "array([[ 9,  2,  5, ..., 13, 16,  8],\n",
            "       [ 2,  5,  8, ..., 16,  8,  4],\n",
            "       [ 5,  8,  2, ...,  8,  4,  8],\n",
            "       ...,\n",
            "       [ 2,  5,  6, ..., 20, 13,  4],\n",
            "       [ 5,  6,  9, ..., 13,  4, 10],\n",
            "       [ 6,  9,  6, ...,  4, 10,  4]])>, <tf.Tensor: shape=(256, 50), dtype=int64, numpy=\n",
            "array([[2, 2, 2, ..., 2, 2, 2],\n",
            "       [2, 2, 2, ..., 2, 2, 2],\n",
            "       [2, 2, 2, ..., 2, 2, 2],\n",
            "       ...,\n",
            "       [2, 2, 2, ..., 2, 2, 2],\n",
            "       [2, 2, 2, ..., 2, 2, 2],\n",
            "       [2, 2, 2, ..., 2, 2, 2]])>), (<tf.Tensor: shape=(256, 50), dtype=int64, numpy=\n",
            "array([[ 2,  5,  8, ..., 16,  8,  4],\n",
            "       [ 5,  8,  2, ...,  8,  4,  8],\n",
            "       [ 8,  2,  6, ...,  4,  8,  5],\n",
            "       ...,\n",
            "       [ 5,  6,  9, ..., 13,  4, 10],\n",
            "       [ 6,  9,  6, ...,  4, 10,  4],\n",
            "       [ 9,  6, 14, ..., 10,  4, 10]])>, <tf.Tensor: shape=(256, 50), dtype=int64, numpy=\n",
            "array([[2, 2, 2, ..., 2, 2, 2],\n",
            "       [2, 2, 2, ..., 2, 2, 2],\n",
            "       [2, 2, 2, ..., 2, 2, 2],\n",
            "       ...,\n",
            "       [2, 2, 2, ..., 2, 2, 2],\n",
            "       [2, 2, 2, ..., 2, 2, 2],\n",
            "       [2, 2, 2, ..., 2, 2, 2]])>))\n"
          ]
        }
      ],
      "source": [
        "example_input_output = ds.take(1).get_single_element()\n",
        "print(example_input_output)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "5bb6376b-c4f9-4946-a736-94f1739c3149",
      "metadata": {
        "id": "5bb6376b-c4f9-4946-a736-94f1739c3149"
      },
      "source": [
        "## 5. 코잘 어텐션 마스크 함수 만들기 <a name=\"causal\"></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "id": "afb6fa02-3b59-48eb-bcef-d51c9d32ec18",
      "metadata": {
        "id": "afb6fa02-3b59-48eb-bcef-d51c9d32ec18",
        "outputId": "ba6417b9-e1e9-4f7f-bf64-354769da6260",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[1, 1, 1, 1, 1, 1, 1, 1, 1, 1],\n",
              "       [0, 1, 1, 1, 1, 1, 1, 1, 1, 1],\n",
              "       [0, 0, 1, 1, 1, 1, 1, 1, 1, 1],\n",
              "       [0, 0, 0, 1, 1, 1, 1, 1, 1, 1],\n",
              "       [0, 0, 0, 0, 1, 1, 1, 1, 1, 1],\n",
              "       [0, 0, 0, 0, 0, 1, 1, 1, 1, 1],\n",
              "       [0, 0, 0, 0, 0, 0, 1, 1, 1, 1],\n",
              "       [0, 0, 0, 0, 0, 0, 0, 1, 1, 1],\n",
              "       [0, 0, 0, 0, 0, 0, 0, 0, 1, 1],\n",
              "       [0, 0, 0, 0, 0, 0, 0, 0, 0, 1]], dtype=int32)"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ],
      "source": [
        "def causal_attention_mask(batch_size, n_dest, n_src, dtype):\n",
        "    i = tf.range(n_dest)[:, None]\n",
        "    j = tf.range(n_src)\n",
        "    m = i >= j - n_src + n_dest\n",
        "    mask = tf.cast(m, dtype)\n",
        "    mask = tf.reshape(mask, [1, n_dest, n_src])\n",
        "    mult = tf.concat(\n",
        "        [tf.expand_dims(batch_size, -1), tf.constant([1, 1], dtype=tf.int32)], 0\n",
        "    )\n",
        "    return tf.tile(mask, mult)\n",
        "\n",
        "\n",
        "np.transpose(causal_attention_mask(1, 10, 10, dtype=tf.int32)[0])"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "52210a38-c8b4-4da7-90ce-eaa89c8fcafa",
      "metadata": {
        "id": "52210a38-c8b4-4da7-90ce-eaa89c8fcafa"
      },
      "source": [
        "## 6. 트랜스포머 블록 층 만들기 <a name=\"transformer\"></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 19,
      "id": "bc3f25e8-4e3f-4849-9b92-676ea46e3ed2",
      "metadata": {
        "id": "bc3f25e8-4e3f-4849-9b92-676ea46e3ed2"
      },
      "outputs": [],
      "source": [
        "class TransformerBlock(layers.Layer):\n",
        "    def __init__(\n",
        "        self,\n",
        "        num_heads,\n",
        "        key_dim,\n",
        "        embed_dim,\n",
        "        ff_dim,\n",
        "        name,\n",
        "        dropout_rate=DROPOUT_RATE,\n",
        "    ):\n",
        "        super(TransformerBlock, self).__init__(name=name)\n",
        "        self.num_heads = num_heads\n",
        "        self.key_dim = key_dim\n",
        "        self.embed_dim = embed_dim\n",
        "        self.ff_dim = ff_dim\n",
        "        self.dropout_rate = dropout_rate\n",
        "        self.attn = layers.MultiHeadAttention(\n",
        "            num_heads, key_dim, output_shape=embed_dim\n",
        "        )\n",
        "        self.dropout_1 = layers.Dropout(self.dropout_rate)\n",
        "        self.ln_1 = layers.LayerNormalization(epsilon=1e-6)\n",
        "        self.ffn_1 = layers.Dense(self.ff_dim, activation=\"relu\")\n",
        "        self.ffn_2 = layers.Dense(self.embed_dim)\n",
        "        self.dropout_2 = layers.Dropout(self.dropout_rate)\n",
        "        self.ln_2 = layers.LayerNormalization(epsilon=1e-6)\n",
        "\n",
        "    def call(self, inputs):\n",
        "        input_shape = tf.shape(inputs)\n",
        "        batch_size = input_shape[0]\n",
        "        seq_len = input_shape[1]\n",
        "        causal_mask = causal_attention_mask(\n",
        "            batch_size, seq_len, seq_len, tf.bool\n",
        "        )\n",
        "        attention_output, attention_scores = self.attn(\n",
        "            inputs,\n",
        "            inputs,\n",
        "            attention_mask=causal_mask,\n",
        "            return_attention_scores=True,\n",
        "        )\n",
        "        attention_output = self.dropout_1(attention_output)\n",
        "        out1 = self.ln_1(inputs + attention_output)\n",
        "        ffn_1 = self.ffn_1(out1)\n",
        "        ffn_2 = self.ffn_2(ffn_1)\n",
        "        ffn_output = self.dropout_2(ffn_2)\n",
        "        return (self.ln_2(out1 + ffn_output), attention_scores)\n",
        "\n",
        "    def get_config(self):\n",
        "        config = super().get_config()\n",
        "        config.update(\n",
        "            {\n",
        "                \"key_dim\": self.key_dim,\n",
        "                \"embed_dim\": self.embed_dim,\n",
        "                \"num_heads\": self.num_heads,\n",
        "                \"ff_dim\": self.ff_dim,\n",
        "                \"dropout_rate\": self.dropout_rate,\n",
        "            }\n",
        "        )\n",
        "        return config"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "37ee1198-3bfc-415a-b247-02b8166133fe",
      "metadata": {
        "id": "37ee1198-3bfc-415a-b247-02b8166133fe"
      },
      "source": [
        "## 7. 토큰과 위치 임베딩 만들기 <a name=\"embedder\"></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "id": "d7e30225-7fea-4a3c-afbc-b0017d5da661",
      "metadata": {
        "id": "d7e30225-7fea-4a3c-afbc-b0017d5da661"
      },
      "outputs": [],
      "source": [
        "class TokenAndPositionEmbedding(layers.Layer):\n",
        "    def __init__(self, vocab_size, embed_dim):\n",
        "        super(TokenAndPositionEmbedding, self).__init__()\n",
        "        self.vocab_size = vocab_size\n",
        "        self.embed_dim = embed_dim\n",
        "        self.token_emb = layers.Embedding(\n",
        "            input_dim=vocab_size,\n",
        "            output_dim=embed_dim,\n",
        "            embeddings_initializer=\"he_uniform\",\n",
        "        )\n",
        "        self.pos_emb = SinePositionEncoding()\n",
        "\n",
        "    def call(self, x):\n",
        "        embedding = self.token_emb(x)\n",
        "        positions = self.pos_emb(embedding)\n",
        "        return embedding + positions\n",
        "\n",
        "    def get_config(self):\n",
        "        config = super().get_config()\n",
        "        config.update(\n",
        "            {\n",
        "                \"vocab_size\": self.vocab_size,\n",
        "                \"embed_dim\": self.embed_dim,\n",
        "            }\n",
        "        )\n",
        "        return config"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "id": "96ec1c62-c375-4831-99f3-d313f98f39b6",
      "metadata": {
        "id": "96ec1c62-c375-4831-99f3-d313f98f39b6",
        "outputId": "fada0e32-f911-4604-c326-2575b117633f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "image/png": {
              "width": 547,
              "height": 354
            }
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "image/png": {
              "width": 547,
              "height": 354
            }
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "image/png": {
              "width": 547,
              "height": 354
            }
          }
        }
      ],
      "source": [
        "tpe = TokenAndPositionEmbedding(notes_vocab_size, 32)\n",
        "token_embedding = tpe.token_emb(example_tokenised_notes)\n",
        "position_embedding = tpe.pos_emb(token_embedding)\n",
        "embedding = tpe(example_tokenised_notes)\n",
        "plt.imshow(\n",
        "    np.transpose(token_embedding),\n",
        "    cmap=\"coolwarm\",\n",
        "    interpolation=\"nearest\",\n",
        "    origin=\"lower\",\n",
        ")\n",
        "plt.show()\n",
        "plt.imshow(\n",
        "    np.transpose(position_embedding),\n",
        "    cmap=\"coolwarm\",\n",
        "    interpolation=\"nearest\",\n",
        "    origin=\"lower\",\n",
        ")\n",
        "plt.show()\n",
        "plt.imshow(\n",
        "    np.transpose(embedding),\n",
        "    cmap=\"coolwarm\",\n",
        "    interpolation=\"nearest\",\n",
        "    origin=\"lower\",\n",
        ")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "fee70157-5cb1-466c-bb9d-ba4720a93173",
      "metadata": {
        "id": "fee70157-5cb1-466c-bb9d-ba4720a93173"
      },
      "source": [
        "## 8. 트랜스포머 모델 구축 <a name=\"transformer_decoder\"></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 22,
      "id": "df56f070-228b-4364-94ae-5b6a12349960",
      "metadata": {
        "id": "df56f070-228b-4364-94ae-5b6a12349960"
      },
      "outputs": [],
      "source": [
        "note_inputs = layers.Input(shape=(None,), dtype=tf.int32)\n",
        "durations_inputs = layers.Input(shape=(None,), dtype=tf.int32)\n",
        "note_embeddings = TokenAndPositionEmbedding(\n",
        "    notes_vocab_size, EMBEDDING_DIM // 2)(note_inputs)\n",
        "duration_embeddings = TokenAndPositionEmbedding(\n",
        "    durations_vocab_size, EMBEDDING_DIM // 2)(durations_inputs)\n",
        "embeddings = layers.Concatenate()([note_embeddings, duration_embeddings])\n",
        "x, attention_scores = TransformerBlock(\n",
        "    N_HEADS, KEY_DIM, EMBEDDING_DIM, FEED_FORWARD_DIM, name=\"attention\")(embeddings)\n",
        "note_outputs = layers.Dense(\n",
        "    notes_vocab_size, activation=\"softmax\", name=\"note_outputs\")(x)\n",
        "duration_outputs = layers.Dense(\n",
        "    durations_vocab_size, activation=\"softmax\", name=\"duration_outputs\")(x)\n",
        "model = models.Model(\n",
        "    inputs=[note_inputs, durations_inputs],\n",
        "    outputs=[note_outputs, duration_outputs],  # attention_scores\n",
        ")\n",
        "model.compile(\n",
        "    \"adam\",\n",
        "    loss=[\n",
        "        losses.SparseCategoricalCrossentropy(),\n",
        "        losses.SparseCategoricalCrossentropy(),\n",
        "    ],\n",
        ")\n",
        "att_model = models.Model(\n",
        "    inputs=[note_inputs, durations_inputs], outputs=attention_scores\n",
        ")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 23,
      "id": "eddbcc80-d0e9-41f0-bd70-032d489369b4",
      "metadata": {
        "id": "eddbcc80-d0e9-41f0-bd70-032d489369b4",
        "outputId": "84769928-0f9b-42df-ad39-1b1972bb427e",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                Output Shape                 Param #   Connected to                  \n",
            "==================================================================================================\n",
            " input_1 (InputLayer)        [(None, None)]               0         []                            \n",
            "                                                                                                  \n",
            " input_2 (InputLayer)        [(None, None)]               0         []                            \n",
            "                                                                                                  \n",
            " token_and_position_embeddi  (None, None, 128)            7552      ['input_1[0][0]']             \n",
            " ng_1 (TokenAndPositionEmbe                                                                       \n",
            " dding)                                                                                           \n",
            "                                                                                                  \n",
            " token_and_position_embeddi  (None, None, 128)            3072      ['input_2[0][0]']             \n",
            " ng_2 (TokenAndPositionEmbe                                                                       \n",
            " dding)                                                                                           \n",
            "                                                                                                  \n",
            " concatenate (Concatenate)   (None, None, 256)            0         ['token_and_position_embedding\n",
            "                                                                    _1[0][0]',                    \n",
            "                                                                     'token_and_position_embedding\n",
            "                                                                    _2[0][0]']                    \n",
            "                                                                                                  \n",
            " attention (TransformerBloc  ((None, None, 256),          1447424   ['concatenate[0][0]']         \n",
            " k)                           (None, 5, None, None))                                              \n",
            "                                                                                                  \n",
            " note_outputs (Dense)        (None, None, 59)             15163     ['attention[0][0]']           \n",
            "                                                                                                  \n",
            " duration_outputs (Dense)    (None, None, 24)             6168      ['attention[0][0]']           \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 1479379 (5.64 MB)\n",
            "Trainable params: 1479379 (5.64 MB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "model.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 24,
      "id": "b51a1dbc-7185-43ab-b9f9-834d267c99a0",
      "metadata": {
        "id": "b51a1dbc-7185-43ab-b9f9-834d267c99a0"
      },
      "outputs": [],
      "source": [
        "if LOAD_MODEL:\n",
        "    model.load_weights(\"./checkpoint/checkpoint.ckpt\")\n",
        "    # model = models.load_model('./models/model', compile=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "afeab12c-b871-47c0-884c-752238b9f719",
      "metadata": {
        "id": "afeab12c-b871-47c0-884c-752238b9f719"
      },
      "source": [
        "## 9. 트랜스포머 훈련하기 <a name=\"train\"></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 25,
      "id": "a2003f19-610a-4acb-a225-0f6c1a3d3a1f",
      "metadata": {
        "id": "a2003f19-610a-4acb-a225-0f6c1a3d3a1f"
      },
      "outputs": [],
      "source": [
        "# MusicGenerator 체크포인트 만들기\n",
        "class MusicGenerator(callbacks.Callback):\n",
        "    def __init__(self, index_to_note, index_to_duration, top_k=10):\n",
        "        self.index_to_note = index_to_note\n",
        "        self.note_to_index = {\n",
        "            note: index for index, note in enumerate(index_to_note)\n",
        "        }\n",
        "        self.index_to_duration = index_to_duration\n",
        "        self.duration_to_index = {\n",
        "            duration: index for index, duration in enumerate(index_to_duration)\n",
        "        }\n",
        "\n",
        "    def sample_from(self, probs, temperature):\n",
        "        probs = probs ** (1 / temperature)\n",
        "        probs = probs / np.sum(probs)\n",
        "        return np.random.choice(len(probs), p=probs), probs\n",
        "\n",
        "    def get_note(self, notes, durations, temperature):\n",
        "        sample_note_idx = 1\n",
        "        while sample_note_idx == 1:\n",
        "            sample_note_idx, note_probs = self.sample_from(\n",
        "                notes[0][-1], temperature\n",
        "            )\n",
        "            sample_note = self.index_to_note[sample_note_idx]\n",
        "\n",
        "        sample_duration_idx = 1\n",
        "        while sample_duration_idx == 1:\n",
        "            sample_duration_idx, duration_probs = self.sample_from(\n",
        "                durations[0][-1], temperature\n",
        "            )\n",
        "            sample_duration = self.index_to_duration[sample_duration_idx]\n",
        "\n",
        "        new_note = get_midi_note(sample_note, sample_duration)\n",
        "\n",
        "        return (\n",
        "            new_note,\n",
        "            sample_note_idx,\n",
        "            sample_note,\n",
        "            note_probs,\n",
        "            sample_duration_idx,\n",
        "            sample_duration,\n",
        "            duration_probs,\n",
        "        )\n",
        "\n",
        "    def generate(self, start_notes, start_durations, max_tokens, temperature):\n",
        "        attention_model = models.Model(\n",
        "            inputs=self.model.input,\n",
        "            outputs=self.model.get_layer(\"attention\").output,\n",
        "        )\n",
        "\n",
        "        start_note_tokens = [self.note_to_index.get(x, 1) for x in start_notes]\n",
        "        start_duration_tokens = [\n",
        "            self.duration_to_index.get(x, 1) for x in start_durations\n",
        "        ]\n",
        "        sample_note = None\n",
        "        sample_duration = None\n",
        "        info = []\n",
        "        midi_stream = music21.stream.Stream()\n",
        "\n",
        "        midi_stream.append(music21.clef.BassClef())\n",
        "\n",
        "        for sample_note, sample_duration in zip(start_notes, start_durations):\n",
        "            new_note = get_midi_note(sample_note, sample_duration)\n",
        "            if new_note is not None:\n",
        "                midi_stream.append(new_note)\n",
        "\n",
        "        while len(start_note_tokens) < max_tokens:\n",
        "            x1 = np.array([start_note_tokens])\n",
        "            x2 = np.array([start_duration_tokens])\n",
        "            notes, durations = self.model.predict([x1, x2], verbose=0)\n",
        "\n",
        "            repeat = True\n",
        "\n",
        "            while repeat:\n",
        "                (\n",
        "                    new_note,\n",
        "                    sample_note_idx,\n",
        "                    sample_note,\n",
        "                    note_probs,\n",
        "                    sample_duration_idx,\n",
        "                    sample_duration,\n",
        "                    duration_probs,\n",
        "                ) = self.get_note(notes, durations, temperature)\n",
        "\n",
        "                if (\n",
        "                    isinstance(new_note, music21.chord.Chord)\n",
        "                    or isinstance(new_note, music21.note.Note)\n",
        "                    or isinstance(new_note, music21.note.Rest)\n",
        "                ) and sample_duration == \"0.0\":\n",
        "                    repeat = True\n",
        "                else:\n",
        "                    repeat = False\n",
        "\n",
        "            if new_note is not None:\n",
        "                midi_stream.append(new_note)\n",
        "\n",
        "            _, att = attention_model.predict([x1, x2], verbose=0)\n",
        "\n",
        "            info.append(\n",
        "                {\n",
        "                    \"prompt\": [start_notes.copy(), start_durations.copy()],\n",
        "                    \"midi\": midi_stream,\n",
        "                    \"chosen_note\": (sample_note, sample_duration),\n",
        "                    \"note_probs\": note_probs,\n",
        "                    \"duration_probs\": duration_probs,\n",
        "                    \"atts\": att[0, :, -1, :],\n",
        "                }\n",
        "            )\n",
        "            start_note_tokens.append(sample_note_idx)\n",
        "            start_duration_tokens.append(sample_duration_idx)\n",
        "            start_notes.append(sample_note)\n",
        "            start_durations.append(sample_duration)\n",
        "\n",
        "            if sample_note == \"START\":\n",
        "                break\n",
        "\n",
        "        return info\n",
        "\n",
        "    def on_epoch_end(self, epoch, logs=None):\n",
        "        info = self.generate(\n",
        "            [\"START\"], [\"0.0\"], max_tokens=GENERATE_LEN, temperature=0.5\n",
        "        )\n",
        "        midi_stream = info[-1][\"midi\"].chordify()\n",
        "        print(info[-1][\"prompt\"])\n",
        "\n",
        "        display(Image(str(midi_stream.write('lily.png'))))\n",
        "        midi_stream.write(\n",
        "            \"midi\",\n",
        "            fp=os.path.join(\n",
        "                \"./output\",\n",
        "                \"output-\" + str(epoch).zfill(4) + \".mid\",\n",
        "            ),\n",
        "        )"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 26,
      "id": "48a82971-e609-4ced-9fd1-60eb9694c08a",
      "metadata": {
        "id": "48a82971-e609-4ced-9fd1-60eb9694c08a"
      },
      "outputs": [],
      "source": [
        "# 모델 저장 체크포인트 만들기\n",
        "model_checkpoint_callback = callbacks.ModelCheckpoint(\n",
        "    filepath=\"./checkpoint/checkpoint.ckpt\",\n",
        "    save_weights_only=True,\n",
        "    save_freq=\"epoch\",\n",
        "    verbose=0,\n",
        ")\n",
        "\n",
        "tensorboard_callback = callbacks.TensorBoard(log_dir=\"./logs\")\n",
        "\n",
        "# 시작 프롬프트 토큰화\n",
        "music_generator = MusicGenerator(notes_vocab, durations_vocab)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 27,
      "id": "69f0c057-01c8-4a6a-9d82-8e83e66aa075",
      "metadata": {
        "tags": [],
        "id": "69f0c057-01c8-4a6a-9d82-8e83e66aa075",
        "outputId": "b11f2379-f21e-44b7-a6b7-ebd8d65c80ef",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 4.8973 - note_outputs_loss: 3.4831 - duration_outputs_loss: 1.4142[['START', 'D3', 'D3', 'D3', 'C4', 'G#3', 'D3', 'A3', 'G3', 'E2', 'F3', 'G3', 'D3', 'G3', 'D3', 'D3', 'B-3', 'D3', 'D3', 'E3', 'E-3', 'G3', 'B-3', 'E-3', 'G3', 'B2', 'D3', 'E-3', 'D3', 'E3', 'B2', 'E3', 'D3', 'G3', 'E3', 'G2', 'E-3', 'E3', 'E3', 'D3', 'B3', 'D3', 'C4', 'B2', 'C4', 'C4', 'D3', 'G2', 'G3'], ['0.0', '1/3', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '1.0', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '1.0']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 24s 155ms/step - loss: 4.8973 - note_outputs_loss: 3.4831 - duration_outputs_loss: 1.4142\n",
            "Epoch 2/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 4.0851 - note_outputs_loss: 3.3099 - duration_outputs_loss: 0.7752[['START', 'F3', 'D3', 'D3', 'F#3', 'C3', 'D3', 'D3', 'D3', 'C3', 'D3', 'E-3', 'D3', 'C3', 'G#3', 'G2', 'D3', 'C3', 'D3', 'E-3', 'D3', 'D3', 'E-3', 'G3', 'G2', 'D3', 'C3', 'D3', 'D3', 'D3', 'D3', 'B-3', 'D3', 'E-3', 'G3', 'B-3', 'A3', 'D4', 'C4', 'C4', 'G3', 'A3', 'D3', 'D3', 'F3', 'G3', 'D3', 'A3', 'G3'], ['0.0', '0.5', '0.5', '1.0', '0.5', '0.5', '0.5', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 14s 128ms/step - loss: 4.0851 - note_outputs_loss: 3.3099 - duration_outputs_loss: 0.7752\n",
            "Epoch 3/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 3.6306 - note_outputs_loss: 2.9517 - duration_outputs_loss: 0.6789[['START', 'C3', 'E3', 'D3', 'B2', 'G2', 'B2', 'A2', 'F3', 'G3', 'G3', 'F#3', 'G#3', 'G3', 'A3', 'G3', 'E-3', 'D3', 'B-2', 'D3', 'F3', 'G3', 'F3', 'E-3', 'D3', 'B-2', 'C3', 'D3', 'E-3', 'D3', 'E-3', 'D3', 'E-3', 'F3', 'E-3', 'G3', 'F3', 'E3', 'D3', 'E-3', 'E-3', 'E-3', 'F3', 'G3', 'F3', 'E-3', 'F3', 'E3', 'D3'], ['0.0', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 14s 131ms/step - loss: 3.6306 - note_outputs_loss: 2.9517 - duration_outputs_loss: 0.6789\n",
            "Epoch 4/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 3.4284 - note_outputs_loss: 2.7699 - duration_outputs_loss: 0.6586[['START', 'G4', 'D4', 'F#4', 'E4', 'G3', 'A3', 'B3', 'C4', 'D4', 'C#4', 'D4', 'D4', 'E4', 'E4', 'E4', 'F#4', 'D4', 'E4', 'D4', 'E4', 'E4', 'F#4', 'E4', 'E4', 'F#4', 'F#4', 'B3', 'A3', 'D4', 'C#4', 'B3', 'D4', 'C#4', 'D4', 'D4', 'C#4', 'B3', 'D4', 'C4', 'B3', 'C4', 'B3', 'C4', 'B3', 'A3', 'F3', 'G3', 'A3'], ['0.0', '1.0', '1.0', '1.0', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '1.0', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 14s 134ms/step - loss: 3.4284 - note_outputs_loss: 2.7699 - duration_outputs_loss: 0.6586\n",
            "Epoch 5/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 3.2694 - note_outputs_loss: 2.6422 - duration_outputs_loss: 0.6272[['START', 'A2', 'C3', 'B2', 'A2', 'D3', 'C3', 'B2', 'A2', 'B2', 'C3', 'D3', 'E-3', 'F3', 'D3', 'E3', 'F3', 'E3', 'F3', 'G3', 'F3', 'E3', 'F3', 'D3', 'E3', 'C3', 'B2', 'C3', 'B-2', 'C3', 'D3', 'E3', 'F3', 'G3', 'F3', 'G3', 'G#3', 'F3', 'E-3', 'D3', 'F3', 'E-3', 'D3', 'C3', 'B-2', 'C3', 'B-2', 'C3', 'C3'], ['0.0', '1.0', '1.0', '1.0', '1.0', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 14s 128ms/step - loss: 3.2694 - note_outputs_loss: 2.6422 - duration_outputs_loss: 0.6272\n",
            "Epoch 6/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 3.1865 - note_outputs_loss: 2.5740 - duration_outputs_loss: 0.6125[['START', 'G:major', 'A4', 'B4', 'A4', 'A4', 'A4', 'B4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'E4', 'F#4', 'F#4', 'E4', 'A4', 'A4', 'A4', 'A4', 'A4', 'G4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4'], ['0.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 14s 130ms/step - loss: 3.1865 - note_outputs_loss: 2.5740 - duration_outputs_loss: 0.6125\n",
            "Epoch 7/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 3.1032 - note_outputs_loss: 2.5145 - duration_outputs_loss: 0.5887[['START', 'E-:major', '3/4TS', 'E-4', 'F#4', 'F#4', 'C#4', 'B3', 'A3', 'B3', 'A3', 'B3', 'C#4', 'D4', 'B3', 'C#4', 'B3', 'A3', 'B3', 'C#4', 'B3', 'C#4', 'B3', 'A3', 'G#3', 'G3', 'A3', 'F#3', 'A3', 'A3', 'B3', 'A3', 'F#3', 'E3', 'F#3', 'A3', 'G3', 'F#3', 'A3', 'B3', 'A3', 'D3', 'F#3', 'D3', 'G3', 'F#3', 'G3', 'F#3', 'G3'], ['0.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 15s 136ms/step - loss: 3.1032 - note_outputs_loss: 2.5145 - duration_outputs_loss: 0.5887\n",
            "Epoch 8/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 3.0672 - note_outputs_loss: 2.4832 - duration_outputs_loss: 0.5840[['START', '3/4TS', 'rest', 'C4', 'G#3', 'G3', 'F3', 'E-3', 'F3', 'D3', 'E-3', 'G3', 'F3', 'E-3', 'F3', 'G3', 'F3', 'E-3', 'F3', 'G3', 'G#3', 'F3', 'E-3', 'D3', 'E-3', 'F3', 'D3', 'E-3', 'F3', 'D3', 'E-3', 'D3', 'E-3', 'D3', 'F3', 'E-3', 'F3', 'G3', 'F3', 'E-3', 'D3', 'E-3', 'D3', 'E-3', 'F3', 'G3', 'G#3', 'F3', 'E-3'], ['0.0', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 15s 137ms/step - loss: 3.0672 - note_outputs_loss: 2.4832 - duration_outputs_loss: 0.5840\n",
            "Epoch 9/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 3.0361 - note_outputs_loss: 2.4494 - duration_outputs_loss: 0.5867[['START', 'E-:major', '4/4TS', 'rest', 'D4', 'C4', 'D4', 'C4', 'B-3', 'A3', 'G3', 'G#3', 'G3', 'F3', 'G3', 'G#3', 'F3', 'E-3', 'F3', 'G3', 'F3', 'E-3', 'D3', 'C3', 'B-2', 'C3', 'F3', 'E-3', 'F3', 'E-3', 'F3', 'E-3', 'F3', 'G3', 'F3', 'E-3', 'F3', 'G3', 'E-3', 'F3', 'D3', 'E-3', 'D3', 'C3', 'B-2', 'C3', 'B-2', 'C3', 'D3'], ['0.0', '0.0', '0.0', '1/12', '1/6', '1/12', '1/6', '1/12', '1/6', '1/12', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3', '1/3']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 15s 141ms/step - loss: 3.0361 - note_outputs_loss: 2.4494 - duration_outputs_loss: 0.5867\n",
            "Epoch 10/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 2.9801 - note_outputs_loss: 2.4215 - duration_outputs_loss: 0.5586[['START', 'E-:major', '4/4TS', 'rest', 'D4', 'C4', 'A3', 'G3', 'A3', 'B-3', 'A3', 'B-3', 'A3', 'C4', 'D4', 'A3', 'A3', 'B-3', 'A3', 'G3', 'F#3', 'G3', 'A3', 'F#3', 'A3', 'A3', 'A3', 'B-3', 'A3', 'B-3', 'A3', 'G3', 'F#3', 'G3', 'A3', 'B-3', 'A3', 'B-3', 'A3', 'B-3', 'A3', 'A3', 'B-3', 'B-3', 'A3', 'B-3', 'A3', 'B-3', 'A3'], ['0.0', '0.0', '0.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 14s 134ms/step - loss: 2.9801 - note_outputs_loss: 2.4215 - duration_outputs_loss: 0.5586\n",
            "Epoch 11/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 2.9636 - note_outputs_loss: 2.3938 - duration_outputs_loss: 0.5698[['START', '4/4TS', 'rest', 'G2', 'rest', 'rest', 'G2', 'F2', 'G2', 'D2', 'F2', 'D3', 'G2', 'G2', 'D3', 'F3', 'G3', 'F3', 'E-3', 'D3', 'E-3', 'F3', 'G3', 'F3', 'E-3', 'D3', 'E-3', 'D3', 'E-3', 'G3', 'C3', 'D3', 'E-3', 'G3', 'F3', 'G3', 'G#3', 'F3', 'G3', 'F3', 'E-3', 'D3', 'E-3', 'F3', 'G3', 'G#3', 'G3', 'F3', 'E-3'], ['0.0', '0.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 15s 139ms/step - loss: 2.9636 - note_outputs_loss: 2.3938 - duration_outputs_loss: 0.5698\n",
            "Epoch 12/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 2.9669 - note_outputs_loss: 2.3901 - duration_outputs_loss: 0.5768[['START', '4/4TS', 'rest', 'E4', 'D4', 'C4', 'B3', 'C4', 'B3', 'C4', 'B3', 'C4', 'E3', 'F3', 'G3', 'E-3', 'D3', 'E-3', 'F3', 'D3', 'C3', 'D3', 'G2', 'E-3', 'G2', 'G2', 'F3', 'E-3', 'F3', 'G3', 'E-3', 'E-3', 'F3', 'D3', 'E-3', 'F3', 'E-3', 'E-3', 'F3', 'E-3', 'D3', 'E-3', 'F3', 'E-3', 'D3', 'E-3', 'F3', 'E-3', 'F3'], ['0.0', '0.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAw8AAACPCAIAAAAz5nDEAAAcgElEQVR4nO3dT2zb5vkH8DeLg9WY40ou4C5OAUeMna2H/bGUOT0MWQHJ69CzpPbQ0zrZQC/rMIDSbZcNkAwU22mA5B13aC1fF2yQBAQDOiCZKF+KDUhk2R7gZskQ0rUDpICC6nd4Mf40iXxFkQ//Sd/PKZHpVw9fvaRfke/z8EKv12MAAAAAYOJrfgcAAAAAEGiYLQEAAACIYLYEAAAAIILZEgAAAIAIZksAAAAAIpgtAQAAAIhgtgQAAAAggtkSAAAAgAhmSwAAAAAimC1NPk3T8vn8hf+VSqX8jgsAACAcLuDJJ5NN07RUKtVqtYZ/hI8eAADACsyWJpOiKLu7u4yxu3fv3r9/f3iDr3/96z//+c8XFhby+bzn0QEAAIQJZkuTKZPJ7O3tWdmy2WwmEomxGudTMUVRGGOJRKJQKESjUTtRAgAAhAFmS5NJ0zRFUf70pz/97ne/G/7p6upqqVS6fPlyNBoda6qkaVqxWNze3u5/MZlM1ut1hwEDAAAEFmZLkyyfz//xj3/8/PPP+19Mp9OVSsXG1SAsgQIAgOk043cA4K6VlZX33ntPv2uWzWbHve+mKxaLhlOlSCRir8Fw3dELZrTBjMpMuKIFAPh/PZhcsiwnk0mq1mKxmOEQyuVy4zalqqosywPtEIZKK5jRBjMqM+GKFgBgAK4tgSX5fP7w8HD49Xg8XiqVxmrK7I5eo9GwH59rghltMKMyE65oAQCGYd3SZOK3PO7evfv48eN33nnHbDMrFQSq1Wqz2bx79+4//vGPZ8+e9f/oW9/61ltvvfXSSy8xxm7evJnJZCw2JShqYLEpDwQwWr7Kno0qDMEYC8hNrlKppKrqyGid3CAGAPAAZkuTibCCQCqVsnINIBKJaJom3iYajZ6eno5sijbJzvZaGYvRWtlxqpDq9frGxoaVLYvFIlUlLdvRKopy8+ZNK1um0+lqteooSgD/XLhwoVar4QEJkw2zpcnEKwjs7Ow8fPhwIOG/n5UKAp1Op9PpjGxKkiRJkqw0Zfj3Xi9qYLEpKxzWO3AjWuclGPiWI6MiOXE7j1ZRFE3TRkabSCSCcCUMwB7MlqaCz+umwE2Eq7wJm1paWhoYhOl0WlVVksZ1qqrG43HnY54wWqqQekbfcMj7MFzRAviIMVar1fyOAtyFVd7gtZWVlddee02/qODSmhWqegerq6tU0dKWYMhms0+fPnUelZlwRQsA4CrMlsBrly5deuONN8bNpBuX2TqYcVdkz8zMvPnmmyTRUoXE5XI5V6/8hytaAABXfc3vAADoEdY7IBTAkATCFe0EyGQyQUgFhWmGQSgwFdeW+JJnwQb6HYe5ubkbN254EpQXjo6OVFUlyS8jbEpV1aOjI1cfLXd0dPTKK6/wWz+627dvf/jhh+KRMMzVaO2FxBneJiNE1YGc29FOgIODA/bfVfwQOpMxwqd5EI7M15mKnLhqtVoul4df73a7x8fHqqouLCzwV05OThhji4uL165dm52d9TRKF7Tb7bOzM7O1un411Wq15ufnV1ZWnDRyfn7+8OHD1dVVvqJoQLvd/uKLL15++eWzszPG2Pz8/OLi4vz8vF/Rco1GY3FxsdvtOgyJN7W2tqaPW3tOTk4eP35s9pkqivLVV19dvHiRBSPaoGm324yxkQPj7OzsyZMn+jhcXl6+dOmS4Zb8zy3JIRYi4kEYFn6NcPFp0IbpHISMsVartb6+/uc//1m0kd/LzH3DU35yuVx/bo6eBxSJRMrlso/hkQhmTlwymZRl2WEjtVqNmeehBC1aThCwL03xR5GY/TSYOx4cyWRSPMbGfdjLyAYnkngQhoVfI1x8GrRhOgdhr9djjI083U3vuqVMJqNpWqlU6i/0Eo1G6/V6JBI5PT3d2triC107nU4ikUilUuMWIQSA6cQf9jJcnwwPewEIqSmdLVUqlUajkclkhhdhRKPRQqHA/82rIZfL5Var1Wg07K3YcKJUKm1ubnr8pgBTQlGUfD6fSqVSqVQ+nyf8OkRbfwHc+6SmTb1er1QqfkcxKJ/PdzodqqZI2hk2pbMlfgFze3t7Y2NDkqSBaZCeFHB4eFitVre2tmKxWDKZ9L5IDP/gkYUEQEvTtHw+f/Pmze3t7Uaj0Wg0tre3CbOBbNRfkCQJBc2Huf1JTZtarba7u2v2U78G4fb2NslsqV6vCx444dBU5MQN61/zf3h4WCwW+89ukiTFYjGeQd1sNjOZDNW014ZSqcTX6uMEAUCC3yYbvvZDeJvMRv2FaDSK2dIADz4p6IdBKDAVs6XhCgIDT0vd29sbyJnUL5h/9tlnTtIpHzx48OzZM8aYk7SFWCz205/+9D//+c+41Q0muIIAP4GaZe0GLVodYZqx86aOjo6YebZwYHfcuZ2dHcN45ubmrO+vqqrMvPds1F8QfxyTSrzXJJ+UN3wZ4eLToCHxudHHQUjSgbwRe/GPfIb6lFYQaLVa/Xe+FxcXv/Od7xhusLy8bCN7nNcmODk5WVxc5AnDz58/n52dFeQPC9y/f//8/Pyll15aX18f69cnuIKAqqr7+/tmWbtBi5YjTDMmaardbh8fHyeTScOfBnPH2Tg5+WY+/fTTL7/8cvj1paWl119/3WIj4lxrG/UX2u12t9t98eLFkydPmMmZx/m+u8FJVOJBSPJJecCvCgLi06Ah8bnRr0FI1YG8Q8yGkziA9fX1e/fuiTbyJDsvcJrNpn71KBKJNJvNgQ30nxaLRRuNx2IxNpSRWKvV0un0uK3xJVbcuL+OCgLOBTORnqSp0FUQGDcn31Cz2TQ8E8bj8bEe9CvOtbbRe3zXisWifnbq7zSSfSfnPCrxICT5pDxAeGiPxUYFAfG50a9BSNWBvEPsBYAKAsYSiUSn0ymXy+VymRcI6P9pp9PRL8qlUqlOpyNJkvUKApubm3zVgp5bxxirVCqZTGZvb2+sfIRqtdq/XElcaRRgglHl5Bsexel0ul6v+75iIx6P5/P5RCLBj3r9m1Iw6xH4ElVAPqkJFq5B6KWpWLdkKBqNmiXn6yu+Y7FYIpHgDx07PDxUFGXkk0ErlQq/RB+JRPoP6Xw+z2dgu7u7FosCVCqVra0t/b+5XA7JcTC1aHPyZVnmS4gSiUQ2m/U+3dXQwCRAvzHhRj0CRVF2d3f1TigUCuNOQbypkpDNZvkKMOeflPNddkPQovJyEIbL9M6WBPRFTvzi0NbWFv82Y+VA1WfiAxuPXEE2oNPp9E+VisWie2UkAILPRk6+QJC/eGiaVq1WI5GI/rWKdt81TSsWi/0XCXgxuXHXxtJGZSaXy438jjoS1S7TCmZUnNuDMIwwWxpUKpX4fbR4PM4HynBBJgGzgS7LMj8qstmslXYkSYrH43wun8vlMFWCKWcjJz+kNjc3o9FotVrVv+UT7jtVTn69Xg/LJxLMMgTBjErn6iAMqamYLQ1XEDDz4MGDX//614yxubm5Dz/80MYcX7+GNJClubGxsby8zBiTJMlis++//z4vQPDqq6/a+7aBCgJO3oILbCI9KgiMzMkXROJ8v8QVBGz03tHR0b/+9a94PH758uWPPvpI0zT9123UIzBDlZNv2IiNqEamrDsfOd6UIRg3TpKo3Kgg4MEgNIQKAoEwXEHAULfbvX///pdffjkzM7O2tmbviev6N4NoNOo8g/3s7Ozvf//7t7/97atXr9r4dVVVnz9/bu933WsKFQQC0lS4Kgg0Go3FxcVut8ss5+Qbsp1jPEBcQcBG7/GPwzBn20Y9AjNUOfm8G5eXl/VMcntRiQchychxowzBQBb98fHxuHGSROVGBQEPBuEwVBAIE1VV+QAyLChgnb7kjSq5t1wu+54nTAsVBALSVLgqCFD1nu0c4wHkFQQODg7M6pXQfhbDbOTkU3XjyAoCJCOHZJc5wyz6ceOkKmNBXkHAm0E4zPejm6GCgEWKokiS1Gq1ksmkoihO0i5GLki0uAKpUqlcuHAhkUgcHBwwxvBsXYDJJkmSL8sTJz4nf2lpaeAV27tslkXPhlLJRrYz/GIQPgi/BmEoTMW6JbFKpcLHhyzLzhesbWxs7O3tMcYM7+b23wYW418+Wq0Wv+DfaDRu3ryJORMAOESYkx8KKysrr732Gn/wlMNdFmTR22gzmGUswMxUz5Y0Tdvc3Nzb28vlcoVCgaT24+bmZrlcbrVap6enmqYNfFEoFosWs2GvX7/OGIvFYnx2Va/Xi8UiL5LpPEh/BSE/1rpgRtsL1XLDcEU78Uhy8kPk0qVLb7zxBknq1vSUsYBh03snrlKpSJKkaVqtVuP/JmyZP/mkWCwOvK5pmsUjhF9GymQykiRJkrS5uenwFiEAADiBLPppNhXXlgYqCDx69Oi3v/3tlStXfvOb39y4cYO5cP3go48++vjjj3//+99/9tlnc3NzjLF///vf3/zmN3O5nPX3evvttwlTuCeMZxUEaPnyoHIzXlYQIOF7jnE/8goC4vcKWjUHqm70oIIAYe+9/fbbd+7c6X/F3zIW5BUEBNw+Ifg+LFFBgLGhCgL8Gcuzs7MevDU/pTLGLl++PO6DmrvdbrvdDtSjtoNDnDpLWO+AULvdvnr1qjcDzwpx8vbJycns7Kz3j1U343uO8QDyCgLi9wpUNQdG140eVBAg7L1ut3t8fOywaAKj6z3yCgICtEN6gO/DEhUEQq9cLh8cHPgdRRDZSJ2FAeLk7aCh+rgDW0FA/F6BqubQC1UFAVdT3+2h6j3yCgICqCAwFXfiwgtJcAAAAL7DbAkAAGAyKYqyu7urlyowLCDuvYGoqJp65ZVXaOIzgtkSAADApNE0rVgs9tfSbDQahk818T0qqqZchdkSAADAROFlx4cTzY6Pj32cLZlF5W9TFk3FbGmgggBMABupszBgZPJ20PieY9wPFQTY9FUQoOJBBYGdnR3D1+fm5gqFgl8VBARRsTE7RNAUKgjYN1BBACbA8+fPT05OApWQHzqqqqqq6u9leet8zzEecHJywhgzq1JBW39hgisIiAdh0CoIUKHqPcFp8NNPPzVcorS0tGSvKg1JNxJGRbuD7Xb71q1bn3zyiWgjO1l6AADeYn7nGPtogisIiJFEO8EVBAQM/9zH43FVVe01SNKNhFGR7+BI0/vkEwAAgImUy+UGXkmn0/V6feDRpR4jjMr7HZyKdUsAAADTo1QqRaNRPUU/m80G4TGjhFF5v4OYLQHAFEmlUmaX8SFo8EnZFo1GA/isX8KovN9B3IkDAAAAEMFsCQAAAEAEd+IAIARkWZYkye8o/JHNZoO275Ik8QfiwpQI4CD0GGZLABACAVyE4ZkAPl1bkqRp/kSmUAAHocdwJw4AAABAZCpqeQMAAADYhmtLAAAAACKYLQEAAACIYLYEAAAAIILZEgAAAIAIZksAAAAAIpgtAQAAAIhgtgQAAAAggtkSAAAAgAhmSwAAAAAimC0BAAAAiGC2BAAAACCC2RIAAACAyMzILfL5fP9/t7a2JElyLR4AAACAYLnQ6/XEW0iSlMlkGGOKonQ6nU6n40lgAAAAAIEw4tqSoiiZTKZUKjHG6vU6pkoAAAAwbUZcW+LTI37rLZPJVCqVaDTqUWgAAAAAATDi2pK+RKler0ejUUyVAAAAYNpYzYkrl8vZbNbVUAAAAAACaPQqb8ZYp9O5fv26qqq4tgQAAADTxtK1pXq9zhjDVAkAAACmkKXZkqZpkUjE5UgAAAAAgsjSnTjGWKfTQVFKAAAAmEJWZ0sAAAAA0wnPiQMAAAAQwWwJAAAAQASzJQAAAACRaZktZTIZ/mxgMIT+8Rg6PIzwqXkD/ewN9PNYRjz5ZGJomuZ3CIGG/vEYOjyM8Kl5A/3sDfTzWOzkxKVSKfEGqqryf8zMzMzPz9uJi1qr1WKMxeNxvwNx6uzs7MmTJ2dnZ4yx+fn55eXlS5cujdVCu91eWFhYWFjof3Fi+idoVFVVVXVlZWXgdXS4e8z63Dl8av0MzyQk0M869wYzQz/3abfbt27d+uSTTwTb2Lm2VCgUDF8/Pz//+OOP9/f319bW+Ct37txhjN2+ffvdd9+9cuWKjfeiIssyM488FHj3NhoN/RVN0xYWFra3t8dqZ2Nj49atW7lcrv/FCeifYNrZ2dnd3f3DH/4w8Do63D1mfe4cPrV+hmcSEuhnnXuDmaGf+2xsbLz11lvibezMlgyvLWmalkqlEonEvXv39Gek8Bfv3Lnzt7/9rVQqbW5u2ng7EvwL0MirYoHFe5J/Fei3v79vY6euXbs28Fth75/AqtVqzKhj0eHuMetz5/CpDRg+k5BAP+vcG8wM/fy/Rj6whGyVdyaT0TStVCr1P04uGo3W6/VIJHJ6erq1tVWtVhljnU4nkUikUincNLWuWCwOT5WYhQ8YAAAAHKJZ5V2pVBqNhizLiqIMTFSj0WihUODX+vL5fCaTKZfL/A//8MbWKYqyu7urKApjLJFIFAoF8UN/JUnyfnI2bpACfKI5jCqjwZf+cQNhn7sq7ANyMoTiNDIB0M8esHF04yw0nh6FdDqtNxiLxZrNZv9PDw4O9J/u7u4eHBzEYrFkMqmqqo33UlWV323tl0wmxb8ly7Isyzbezh57QZopl8uGn108HrfRh4yx4a7wuH/cQNvnVHhIhq+Hd0AGnFmf9wvFaSTgDM8kA9DPDrk3mHs4C/WxMphpri3V63X934eHh8Visf9aiCRJsVjs8PCQMdZsNjOZTKfTsfdGZst3+tc++448yIODg6Wlpc8//7z/xXQ6XalUQjMrd1koBoZf0DkD0CHeQD97ICydHJY4BWgqCAzvcDKZ7P9vq9XiV/yWlpZef/31cd9R1263j4+Ph1+fmZn50Y9+JP7Fbrf74sWLJ0+eMMaWl5eHczKdJ+c7DFLQ4BdffPHyyy/rsS0uLtouzdBoNIZ338v+caNB8j6nwgMbOBxYyAckeYS0bZr1+cAGw68H6jTiKpIgDc8k/dDPQR7MDGehPo1GY319/d69e4JtaCoIqKq6v7+v//f27dsD28iyzGdLP/zhD21nnD548OBXv/qV4Y9+/OMf/+IXvxD8Ls/D/NnPfra2tpbP54+Pjz/44AO9zgRVcj73zjvv2AtSEPzDhw/tBTOs0WgM5/162T/kDT569MisY233OZWdnZ3j4+PhQ8bLDs9ms4avO+kc8iFB2KZZn+tsH6EeHyZuIAzS8Eyic3JUop91Iwezk6Pbs34+Pz9///33bccpaJZwMH//+98fsRHJPb9ms6knZ0UikYF1S71eT/9psVi0/S5m82sry3dkWY7H4/zf/PDWb1KqqmpWnstenLaDFARPeHOXmaxb8qx/yBscvhfuvM+pCNYtedPhPAOZtnPIP0HaNsVLPZx0iJeHiRtog2TCpR5Ojkr0s27kuiXbndzDWagPs7BuiaaCQCKR6HQ65XK5XC7zAgH9P+10Oqenp/zfqVSq0+lIkmSvgsB3v/vdgVfS6XS9XreyfGdgG70ErQfJ+daD9JFn/eNGhy8tLQ28Evw+D++AdCNCf2tkhOI04pzHQTo5KtHPto11dOMsZB3Zc+Ki0ahZ8Ul9xXcsFkskEvl8/vDw8PDw0EYFgTfffPMnP/mJnnyYzWYHZmYjaZpWrVYjkYgerRvJ+dls9unTp7aD9JEH/eNGh6+urr733ntOBsYwbzJdvRmQvLoHI+ocNyJ0u0bGAIcd4s2npqMaih53svOj0st+DlHBlwHO/9zgLGSJrWtg44nFYvy9yuVyr9c7ODiIx+M2Kggkk0nb6Y76zax0Ot1f46DZbBp2i5MrhIyxWq1m73cNeXMnzrB/ei7cWCRvkLZ/eqSZroI7cYYdTlstovffa+A2flHAjU+QsE0rd+Jsx+nZYcIRDkXycx0bdSfO9lHp5em6F+yCL1buxNn+c4OzUH9rHt2JEyiVSrx2QDwe5/NWSZIURfH+LgnPYGSMKYqiz2oN7wYG/yaOGwz7x5DD/gn4XTPeD8PrBMkzXQ07nFeLGNgyUP1jyGGEiqJ873vfo23TJWaHiRufGu1QDNe5zrPTNfnxHq5DGGch62gqCJg5Ozvb399/8eLFzMzM+vr67OzsuO/Vr9Vqzc/P23saM09fHM6Q5Nl8y8vLJMn5XKPRWFtbI3w0d7vdPjs7o3pStFkFAcP+4dsvLi52u11G1D+Konz11VcXL16kapC2f2gzXQUVBAw7vN1un56eRiIRqgHJR7ggA9kG8iHRarW+8Y1vXLx4kWSvxUnXTjpEcJiQj2pGPRTJz3XiCgJOjkovT9cBL/gysoKAkz83grMQ4S6wMJyFvKsgYOj8/PyDDz548eLF3NxcqVS6ceOGjffqJ8vy6uqqvQIEjx49+utf/zqcOdxqtfb392mf8NxoNN59912qP97svxUEqJ4UbZj3a9Y/fPtf/vKXhLvj5HM0RNs/tPn2ZgnAZh1OWy2C/XeE0z5mPOBDQpx07aRDBIcJ+agmr4tBfq4TVxBwclR6ebomr69BewiPrCDg5M8NzkL9rXlUQWCYnt1nWFDAHifrlsy4dD81dOuWxNvT7g7550jeP8Ns3w638uCCge1p12CFYoTTDgn31i0JuDGqaYci+Y6LzyTkI7kXhuUvPeodd3XdkuBNp+0sZOXPoivrlhRFkSSp1Wolk8mRi2AAAi74N+xhIoVx7cgEQCeDIbIKArpKpZLP5xljsiyXSiXy9gHcFt4CEDBJ3KiLAcNwvIMVlLMlTdM2Nzf39vZyuVyhUJAkibBxAM/kcrlxK4EBkJuZmcEXTg/geAcryO7EVSoVSZI0TavVavzfVC0DAAAA+IiggsDz58//+c9/zs7OXr161XnSrICTCgJmXMpsDF0FAfH2tLtD/jmS9w/h/o5MAB7ennBfWEhGOO2QcK+CgEDARzVzYcfdqyBgJhSD2Y2CJi5VEBC8acDHHqPecY8qCJyfnz979uzKlSs2mhoLeY4ucy2zMXQVBMTb0+5OwCsI0O7vyATg4e0J94WFZISHpYKAQMBHNXNhx92rIGAmFIOZdsddrSAgeNOAjz1GveNWKgjYmS35dYt3YWHh2rVrbrw7eZvxeJywzVqt9vTpU8IGx+1G2t0h/xzJ+4dwf3nqrPXWyPeFC/gIpx0SVvqcvEOCP6o52gYFu+xS/Czwg5l2x60MZtr4WUjGHqPe8ZGP43X9yScAAAAAoYbZEgAAAIAIZksAAAAAIpgtAQAAAIgQVBDwDCoIkLTWarVeffXVq1evWtw+4OniLNgVBE5OTh4/fmw9tinM3WXUQ0Lc56ggQNWg+EyCCgIkrY08gaCCAElrrVbrBz/4wV/+8hfBNgQVBDwztRUE/BXwdHEW7AoC45rC3F3mzqFtBhUEqBoUQwUBqgbFUEGAqsGRJbVRQYCxwGc2+i7I6eIs2BUExjWdubvuHdpmUEHAbaggQNXgSKgg4A2sWwIAAAAQoXyqrtuy2WwoHj8ny3Io4pwYGxsb169f9zuKgEqlUrIs+x1FgKBDvIGj0huyLAf/PoYkSRNw0IVptrS5uel3CJbgseEeC/7Jwl8YkAPQIR7AUemNUAxmSZJCEacY7sQBAAAAiGC2BAAAACBip94SwAS7cOFCrVbDfQTwVz6fVxSlXq/7HciEw/EOFuHaEgAAAIAIri0BAAAAiODaEgAAAIAIZksAAAAAIpgtAQAAAIhgtgQAAAAggtkSAAAAgAhmSwAAAAAimC0BAAAAiGC2BAAAACCC2RIAAACACGZLAAAAACL/Bxkx3SZoiPFKAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 14s 133ms/step - loss: 2.9669 - note_outputs_loss: 2.3901 - duration_outputs_loss: 0.5768\n",
            "Epoch 13/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 2.9303 - note_outputs_loss: 2.3697 - duration_outputs_loss: 0.5605[['START', '3/4TS', 'rest', 'F#4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4', 'A4'], ['0.0', '0.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 14s 133ms/step - loss: 2.9303 - note_outputs_loss: 2.3697 - duration_outputs_loss: 0.5605\n",
            "Epoch 14/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 2.9078 - note_outputs_loss: 2.3617 - duration_outputs_loss: 0.5462[['START', 'G:major', '3/4TS', 'rest'], ['0.0', '0.0', '0.0', '3.75']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAIEAAAAbCAIAAADUP3siAAAAVklEQVR4nO3RQREAIAzAMMC/501GeSQKetc7M4fUqwPw4AMe9DzoedDzoOdBz4OeBz0Peh70POh50POg50HPg54HPQ96HvQ86HnQ86DnQc+Dngc9D3oLxHQDM0A4qxUAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 9s 84ms/step - loss: 2.9078 - note_outputs_loss: 2.3617 - duration_outputs_loss: 0.5462\n",
            "Epoch 15/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 2.8832 - note_outputs_loss: 2.3290 - duration_outputs_loss: 0.5542[['START', 'E-:major', 'G:major', '4/4TS', 'B3', 'C4', 'D4', 'B3', 'C4', 'B3', 'C4', 'B3', 'C4', 'D4', 'C4', 'D4', 'C4', 'B3', 'C4', 'D4', 'B3', 'A3', 'B3', 'C4', 'D4', 'E4', 'F4', 'E4', 'D4', 'E4', 'D4', 'C4', 'D4', 'E4', 'F4', 'E4', 'D4', 'D4', 'C4', 'B3', 'A3', 'G3', 'A3', 'G3', 'A3', 'G3', 'F3', 'E3', 'F3'], ['0.0', '0.0', '0.0', '0.0', '3.75', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25', '0.25']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 15s 143ms/step - loss: 2.8832 - note_outputs_loss: 2.3290 - duration_outputs_loss: 0.5542\n",
            "Epoch 16/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 2.8611 - note_outputs_loss: 2.3185 - duration_outputs_loss: 0.5426[['START', '3/4TS', 'rest', 'rest'], ['0.0', '0.0', '3.75', '0.25']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAHYAAAAbCAIAAAAmr+zCAAAAV0lEQVR4nO3QAQkAIBDAQLV/5zfFEOQuwdiemUXpvA74n8U5i3MW5yzOWZyzOGdxzuKcxTmLcxbnLM5ZnLM4Z3HO4pzFOYtzFucszlmcszhncc7inMW5C/vMAzNzmnwhAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 10s 89ms/step - loss: 2.8611 - note_outputs_loss: 2.3185 - duration_outputs_loss: 0.5426\n",
            "Epoch 17/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 2.8628 - note_outputs_loss: 2.3057 - duration_outputs_loss: 0.5572[['START', '4/4TS', 'rest', 'E4', 'D4', 'B3', 'C4', 'B3', 'A3', 'B3', 'A3', 'C4', 'D4', 'B3', 'C4', 'D4', 'E4', 'D4', 'C4', 'A3', 'D4', 'C4', 'B3', 'A3', 'G#3', 'A3', 'B3', 'A3', 'B3', 'C4', 'D4', 'B3', 'A3', 'B3', 'G3', 'E3', 'F#3', 'C4', 'A3', 'B3', 'A3', 'G3', 'F#3', 'G#3', 'A3', 'B3', 'C4', 'A3', 'B3'], ['0.0', '0.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 15s 141ms/step - loss: 2.8628 - note_outputs_loss: 2.3057 - duration_outputs_loss: 0.5572\n",
            "Epoch 18/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 2.8414 - note_outputs_loss: 2.2949 - duration_outputs_loss: 0.5465[['START', 'E-:major', '4/4TS', 'rest', 'G3', 'G#3', 'G#3', 'G3', 'F3', 'G3', 'G#3', 'B-3', 'G#3', 'G3', 'F3', 'E-3', 'D3', 'E-3', 'D3', 'E-3', 'F3', 'D3', 'E-3', 'D3', 'C3', 'B2', 'C3', 'D3', 'E-3', 'D3', 'E-3', 'D3', 'C3', 'B-2', 'G#2', 'C3', 'D3', 'E-3', 'D3', 'C3', 'D3', 'E-3', 'F3', 'D3', 'C3', 'D3', 'F3', 'E-3', 'D3'], ['0.0', '0.0', '0.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 15s 139ms/step - loss: 2.8414 - note_outputs_loss: 2.2949 - duration_outputs_loss: 0.5465\n",
            "Epoch 19/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 2.8404 - note_outputs_loss: 2.2899 - duration_outputs_loss: 0.5504[['START', '4/4TS', 'rest', 'rest', 'A3', 'rest', 'A3', 'A3', 'A3', 'B3', 'C4', 'B3', 'A3', 'G#3', 'A3', 'B3', 'A3', 'G#3', 'A3', 'C#3', 'A3', 'E4', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3', 'A3'], ['0.0', '0.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0', '1.0']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 15s 139ms/step - loss: 2.8404 - note_outputs_loss: 2.2899 - duration_outputs_loss: 0.5504\n",
            "Epoch 20/20\n",
            "107/107 [==============================] - ETA: 0s - loss: 2.8170 - note_outputs_loss: 2.2675 - duration_outputs_loss: 0.5496[['START', 'E-:major', '4/4TS', 'rest', 'G3', 'G#3', 'F3', 'G3', 'G#3', 'G#3', 'F3', 'G3', 'E-3', 'E-3', 'D3', 'E-3', 'F3', 'D3', 'E-3', 'F3', 'D3', 'E-3', 'F3', 'E-3', 'D3', 'E-3', 'F3', 'D3', 'E-3', 'D3', 'C3', 'D3', 'E-3', 'F3', 'D3', 'E-3', 'F3', 'D3', 'E-3', 'D3', 'E-3', 'F3', 'D3', 'E-3', 'D3', 'C3', 'D3', 'E-3', 'F3'], ['0.0', '0.0', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5', '0.5']]\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAs0AAABCCAIAAAAjVHzyAAAX6klEQVR4nO2dT0wbx9vHJ8EEWwFikwbIHxVsSKFSWxo7SlBblUiQ9B4Z0kPTXgpIuVRVJRup6qmHAjlE7c0mp/bQYjj00EMVG5W0jZoibELaNFCMwUogkDZeg0kwwYH3MHr3t91d787uzq6X8HxOsJ59nmef/Tc7M9+ZPdvb2wgAAAAAAEAH9hY6AAAAAAAAnlugngEAAAAAgF5APQMAAAAAAL2AegYAAAAAAHoB9QwAAAAAAPQC6hkAAAAAAOgF1DMAAAAAANALqGcAAAAAAKAXUM8AAAAAAEAvoJ6hO4lEAv+xZ88e6sYjkciePXsikYhqC7qGBwAAAOxyoJ6hL5FIpKurq9BR5MXk4QEAAAA7HahnAAAAAACgF1DPMBSGYdh+Coq4XC4qdnQKDwAAANi1mKue0d7e3t7eXugodCEYDCKEKioq/H4/deNDQ0NdXV1tbW2qjesaHgAAALBrsRQ6gP/AMEyhQ6BMIpGIRqMOhyOdTuMtehxjT0+Puh2NCQ8AAADYtcjUM6LRaCgUkigwPz+P/7BardXV1RqjicfjCKGd9Um9tLQ0NTW1vLyMEKqqqmpubrZarfinQCCwsrLCLVxVVbV//37eAUpYkCWbzXL/jcfj5MYphqflEJaWlhBC+S4eLZafP9LpdDqdrq2tLXQgOwNIlyKk70SAy+7MVb6n8cmTJ6U7IlS2Z2Sz2Zs3byaTyZqaGrxlcnISIdTY2Hj69Gm73a7O7M4CJ2FsbIzdkkwml5eXL1y4gP9l3+IHDhxoaGgYGxv74IMPFFmQBZ/mjo6O6urqpaWl33//3eDwtB/C9evXEULC8totP3/cunVrbGzM5/MVOpCdAaRLEfnuREDIbsuV1qfxtnJSqZTb7e7s7EylUryNCCG73R4IBFSY3d7ebm1tbW1tVbevwbDHK5FSt9vNPSJetkkskMCWD4VCrC9jwqNyCKInnVZynjPwK7PQUewYIF2K2EGP34Kzq3Kl/WmsZhxoe3s7wzB9fX0Oh4Pd6HA4IpGI3W5Pp9Pd3d1DQ0MIoUQi4fF42tranr9e/97e3lgsJtzObcuJRqMSIydILChifHzc4PCoH4IBlgEAAABytD+NFfebBIPBkZERn88XjUbb2tq4Pzkcjp6eHvzq8vv97e3tgUAAxycsLIrL5dopNRJckRJCrpdRZAEPlIlGowghj8fT09PDreQVKjz9DkF7/IVC6Znaoa4LeJgmD6bg7JIrkAo7LmD9kEgFhaex0iYUr9fL7ut0OsfHx7m/zs7Osr+GQqHZ2Vmn09na2srtYZHA5/P5fD6lIRkP9zC5uN1u3pGGw2HcvJZKpRBCra2t+ADJLaRSKWEHM7fJjj2JPp+P2wmiJTxCC1QOQbQFktCyqZA9U9rJ1xFggOuC+NIYzG7rN9F4arT0BZjqqiChgLkyG9KpCIfD2p/Gim9CXlOJ1+vlFXA6nfgnFTWGnVLPEE291+sV5j0cDjudzs7OTt75I7RA0jGG/x4fH+/s7JSoZygKj9AClUMgrGeIxm8ejBlQIvriNHIsi6nGzcgGs6vqGdpPjep3p6muChIKmCuzQXITHTlyhPeT0qfxnu08344Yoa61v7+fV4ZXFRocHEwmkwihpqamd955R8K4kNHR0Y2NjWw2Oz09jRA6derUmTNnFFngsrS0hGWf1dXVdMWQ8/PzoVDo1KlTrMKnsbFRqHEaHR2dnp7maUfxQZFb4A7xZSkpKfnoo4/w39wzUlNTgwcA9/f3NzQ04MNXER6hBUIv0ocwODiIBCO3SSxj9JbdEtohOVPaHWEvvDuOrmsJ70p9UUm7lmBE00XLe0HQ9QoUvRONcU3oiJadAuaqIGi8iRYXF48cOSLxqqKva62pqcHVCExDQ0O+kiUlJUqNI4QmJydbWlpOnz49ODg4NjZWW1urVP6OFTiTk5ONjY04hsnJyQMHDlA/2bJ1IO75O3DgwMrKSldXF7dBSNYCrm8JaWxslA2vqalJOnWy4claICmj+hBkLRsjuyW0o+VMKXIkhJZrEu+Evqik3bCcq/NuMAXMhpkv/sIGrMK1wVBJRVFRkZYPfoSUt2uNj4+zryK73c4bn7HN6Vjp7e2VNuX1enndLj6fz+12479xY77SbpTx8XHcccPbMRwOC7t4ZE15vV6efJe1RpI6nAd2FAt3FxILhB1jeASMz+fzer3cXo9wOKw6PHILhF4kDiFfv4m0ZcNkt4R2SM6UdkeiHQFUXJN4b21tJfFFJe1Ucq6638RsHQGy8VDpRFenMM/XXET94qdlp4C5Mhgq54477I8lFAopOrmKda0ejyeRSAQCgUAggGWr3F8TiQQ7g3VbW1sikXC5XPl0rQzDCLfzhvtWVFQoCq+rq2tubg79dyruYDDY3t4+PDyMV/EgBO8yMDDQ29uLo1UqdhgfHw+Hwy6Xi5clLXi93kgkws1SIpGIRCJ9fX1KV1PTIzwShIegAiNlt+rsqDhMdY5Ee09VZJjEe1NTk6wvKmknMULrwNV5NxLDrkB1rqmciB13t9JyrR/6nbtAIKBo2m4184E6HI6uri7Rn1gBjNPp9Hg8fr9/bm5ubm6OUNfKwjDM0NCQ3W7P50iUYDCI02q327lp8vv9uPYTCoXIDeL6CkIoEokwDNPW1qb0fezxeCKRiKJdhGAJMbbW0dFBsU5AJTwSOjo6Hj16hKgegpGyW0I7hIdJXT9WX19/7NixsrIyWdfavbe0tLzzzjvSFyT5UWgMRtGBC6F1ImhpI7XHQ/isoK4wr6+vf++99zS6NvhuLVSu1FmWQPtNRHjuuHR0dPj9fgXf7VTaZ1hYsQmeEnR2dhZPOinaxiJsemKbaLxeL27PF/atSMBqbnlm2YNVNEKYq8LAhEIh/BNhv8k2Rzi6rbbfhDxgnq5VtkdDIjxCC1TKqOs3Eb2Y9ZDdkvuSTYV2R6IdAVw1smrvJC2oJI4IW9Gp5Fw2HnUyYO58dxLeSY6CEO3xED4rlCrMA4GArGvRdnU9jvE5yBWJZXKk7ag+d7w+EdEyeB4E4aiJfNBcr7Wvrw+3Abjdbtxs4HK5cEWJHNxy4HA48Dqiiqbtyvd17vP5sCijo6OD3FowGDx58uTs7GxdXR1uzFDUJCMk35wZJsHk4fE4cuTI4uIid4vX6w0Gg9xvApIy+HoTti6OjIxIeBfakYXEEUnA6pD1Ho1Gm5qa8CpFWryT2DF5KkSfObSuHP3i0e5IyOzsLJUTYVjOzZ8rwy4b1ecuEAiEw2HptgqHw+F0OoUDJ/Iho2sdGhrKVy3isbq6OjExkcvlLBbLqVOnbDab7C44R9yBKvF4HK/NVl9fn6+MBGyKHQ4Hb5fV1VWEUHl5OYkdIcvLyw8fPnz11Vfxv6lUamJiIt+wOC6bm5upVKqqqoq3ncQCuRe2/OrqKtZojIyMnDhxQnZ0S77wCC1QKSN6imX3wlXpoqIihFB5eXllZaXw5JKUwZec0L7FYmlpaWGDqays3NzclLAjGzCJo2g0urW1JREwNsK7JGKxWHl5OXvLqPMei8X2799fVFTE3ilC7ySOSOyQpILk3MnGI5ouWe/4vqupqZE4BMKjYFldXX348CFrsKampri4mGI8JM8KkmsACR7I6XTabrdLn83V1VXpR7RhOTd5rkgsc9Fy2ag+dwsLC/F4nA0m3/nl3n0ej6evr08YzP+ikvgNIYSbFqTLIIQymcylS5dyuVxpaWlfX99LL70kuwtC6MqVK2tra9wBmw8ePPj555+5EiDcLiSxDAcXtp5RW1tLuAshn3/+eXNzMxtYLBabmJjQ4oLEghYvIyMj7777LmH9TLUFKmVETzHJXsePHxf2bSktc/HiRdHt586d+/jjj9lgPvnkE42HSeJINuCBgYFkMsnLFZXDpJVPWjmn4ks0XbLe8X139epVCdeER4EQymQy3333HfdrlWGYiooKdtobKvGQPCtIrgH03ztxYGBgZmZGOGcSF1xGo2taOTd5rkgsY7RfNqrPXSaTOX/+/Pnz5/F7PN/5vXTpUktLC34nyksQlPYJCeGu1EreYbNNNvunomnX2GG0dGdqw31Rs7Oz7BalIyeE6DE+gwsiG5+h0QKVMurGZ5CMFSApI3pTqBh7IVuGxJG6AQeEe0l7p5VPWjmn4kuYLhJNI0XNORWNJZVnBYkjiQFzEsiWMTLn5s+VrOVtA/Wo+WJ2Op3sYETRMriTnfx1r2a9Vi7RaNTlcsVisdbW1mg0arBCkofs+AlFUhyWoaEhp9OpVDUK7Aj000YWypEJvZs5GIreaUl8DctGAdOuX851gpYjkou/4HpUl8uVb0Qtxu/3Y0mprCmMpnGgwWAQB+3z+aS7Z/KBJ6UYHh7OZ0TRCq5nz57FpkQHnzIMo07GGQqFzL9MKKCO48ePa9FGmtCRKOrUa/oFU8BUIHpacVk7dDWW2jHMkX6ud3quSO7EgutRcR9Nvl+DweDw8DBvQRJpVNYzGIbp6uoaHh7u7Ozs6elR/a2PZ8Hq6elpa2vr7+8/e/ascK158ipkV1cXXok+nU4zDMPbsbe3V4VgJJFIjIyM0B3tAZgHi8Vy5swZdbVkczoSpbi4uFCuhRQXFzc3NxcwHlqupe1Eo1F2Dh4ubrebt6NhqXjuc04RnRzJ3onBYJDkslF3R7e3t3d3d0v3PCQSiVgslq8u0tfX19PT4/V6FX17q+k3CQaDuJkBq1+0dCi43W6/3+/xeHDQ+TrGFMWG5/DAk3hytzMMo+LE4OnCNCpaAQDYbeTTWBawhwgwP1iPyttI67Jh9agSZfL1ieDGhZ6eHrfbrWhmbaRU17q+vn737l2bzXb06FHVGlEWrmDm7t27i4uLdXV1vKWz4vH45uZmLpd7+PAhQogrec3H5uZmMplcWFiorKzEKqD19XWbzSa7oyg3btwoLy9nFa0YpYpTIXroWrkQ6lo1WiigrpVQY0mlDK3D1B6MOl0rifLQyHwaVkaYLlr3HXkZ7RpLKmVIjIhqNUk0q9JljM/5zs0VoqolltWjCsssLCxMTU298sor7GQHuExtbe3du3ez2azFYjlx4gQvHsq61kwms7a2dvjwYem9CGEFM5lM5v333y8tLb18+TLuteWWCYVCH3744YkTJ/x+fzKZvHTpEqFckx1Kc/z4cZ5Zcv7+++/Dhw/zdgddK8UyhdW1kpShdZg6CTVJ9pJVHhqZT8PKCNNF674jL6NdY0mlDImRfFpN6b1kyxif852bK0RPS0yiR+WVGRwcHBkZefvttz/77DOundHR0YmJCfzvl19+KZy3QrZPQ6ae4XA49OsyCIfDjx49amtra29vP3To0NDQkLCtJhwOu93ugYEBhNCtW7cGBgbwLiT2qUQuYUS7fRILqr243W6NEZJY0F4GtwEIC8juVVtbK+2aVhnZYAwLGHcs8gqQ7CV71xiZT8PKiKZLdIuQ57KMbD55ZUiuHJIyJLGZrUyhcqVrmUQiMTMz8+233+JXLVuGYRi/33/16lW32/39999zGxfC4TA762YgEFC04hgLzXnHVcCbZVy0jMYVXAEAAAAAEB17EYlE8DrndrtdYlbyzs5OdZUMVPB6RiwWI9TEqlvBVScymQziLKficrlMOLsGT4ENQ1kBAAB2LaJ61Hg8fvbsWfx3JBLJp0Ox2+1aBDiFrGd0d3dXVFQQzp3V1dXlcDiGhobMMFR7ZmYGIYRPj8Vi4Y3ewOPMa2pquBttNttrr73G/ru8vIwQ6uvrE55Xj8dD5Ri5iRU1uL6+ns1mS0pKqqurRS18+umnuMeKR3V19VtvvYVtCueTgQoNAACAqZDVowYCAQmxK37/qvZeyHqGy+UiqWSQ9K0YTGlpKft3LpcTFbAJF7mZmpribREO0rFarWxnGF64Sx0WiyWXy7H/Ssx1trGxIboeD0JobGxsbGxM9KevvvoK/8E7gyY5QQAAAABCCI+9GBgYkNCjyvaJKFrqXAi19Vp1QriCqxnAwidjfKnTtf70009bW1vUg1GHOl2r3W4/duyYzWYTVVCDrlXRXqBrxZhKP2lkGdC1kpcpuK6VYhlpPWo8Hl9eXs7lcm+88Qa7EqzQzvLy8ptvvinhy6D1WvVDuIKrGchkMjMzMw8ePPjzzz95rQ7JZPLJkyfcLY8fP378+LHNZrNY/pftXC63vr4u66iyslKdrvXcuXMIoZmZmfn5edF2kfv37+NRJjw2Nzdx44eoEnhjY+Pp06eKIrHZbLJrGAoZGRlJp9PpdHrfvn1FRUVWq9Vut7/44osvv/yy0+l0u927U9eqbq9dq2sVYir9pJFl1InkQddKDl1dK5UysnpUXObMmTNKnzk8TK1rJSTfArg7GuFKK4lE4saNG9lslrtR9VSkepy1RCKRbyK5aDT6119/8YLHkOuQuZSUlGxsbCCEcLVmfX2dYZi5ubnr168XFxeXlpaurq7iNQNramrOnTsnOqjl+dO1qttr1+pa82Eq/aSRZZTmCnSt5BRcsyosI6tHxWV8Pp/eK90UWG+yaxG9RMwgpZFAQlZDvVpjtVr37NkjWnFhW1xWV1fx2OnLly87HI6NjY1Dhw7V19cfPXr04sWLJlQAAQAAGIzs2AsDltODegZgRtLpdDQaTSQSw8PDi4uL9+7d++eff4qKikT7etD/D3RNJpN4TOvXX39dVlb2+PHj3377LRKJ1NfXe71el8ul+o4SXaZZUZlUKjU/Py+9YnAqlUqn00pjy2Qy0pbn5+dTqZSsa53Co0s0GmUHNauIh1WkOxyOAq5Yu1NIp9OyV47s1YXvC2kjJPeXyclms7w8FLwrAMnpUevq6owZkwD1DMCkeDwedoE9DO64+eabb9Lp9OTk5IMHD/bv359PSoPfKE+ePInFYrFYLBQKFRcXW63WvXv3vvDCCx6Pp7m5mbdsjQQkwijZMhMTE7KLKfMG95Bw+/btsbExWcusSj4f1MNjGAavW83bODU1df/+/fn5eXbj2tra9PQ0VyG1tLQ0Ojra399fVlbGHdiE/iueUpouVpHOmmUYpqqqymq1IoTq6+sPHjyINecmnx3HGK5fvz49PS175SCCq4vEyI7m5s2bExMT7GHyunFzuVwmkykqKrJYLNw1vCwWS0NDA1fAODk5mclkRBcoVTHlgbQe1bAWdKhnADsG/MTnfiXgN9m1a9eSyWQ0Gv3333+3tray2azo0NfNzU28fWVlZXZ2Fr9T9+7de/DgQWm/VqtVtAdHD7hPHEL27dtnnvDu3Lnzyy+/sNOu4MeraMnp6WkSj/l2xyhdt4iNn2sWVyzQf7XovBdGLpd7+vQpnmnGYrEcO3YMIXThwgVubVXF9yv+jpeWMuHql8Yy3DocOaqXhdIDklzhG1xjGXUTCtjtdu6/ot8/z549e/bsGW8eATzOjEd3dzdvC3fKg/X19XzyEB4a9ai0gHoGsIPB45R5z/dIJHLv3r0ff/wxHo8vLCysrKwghPK9iUnUv1VVVfmmGKFOXV2d0l0cDgde2tEAZMPb2tp6+vSpUlGSapS2NAiH3MvCfWGwlwF+N4RCIbYFiJ2vb319vaSkxG63kwjKhC70LqMIl8s1Pj5O16ZGTJur119//dq1a3Rtcslms9yHmPSEFCwm6RyUmT8DAJ4DcKc+XoQvHo8vLi5arVb2izYcDkt/iXLHBOiNilZ6U4V34cIF2c4XjGjD8oEDB3jfhTabzeVy8dpRsA7L4XAoTRduABsaGsIDO9bW1u7cuYN/SiaTbNOFdCOKIqQfsJFIxLAOhbKyMqX1UQmJmR5I34YmzxW+tBiG+fXXX5eWlni/3r59m1fvTKfTKysrwm7BTCZD0vhUU1PD7Xk0OVDPAHYj+AGKO1y++OKLXdsBT50rV67cvHmTt7G2tlb4hjDtSEy23sYwzA8//IA/IuPx+KNHjxBC6XR6Y2MDt2DjmfulrUk/YBOJhNfrpRa6JBUVFdKDMU3ObsiV6DdDNBoVjpO12+2FnUJTEVDPAAAAUAP7Kvrjjz/Y2lU0GsXfo/ij1rChMwBgWqCeAQAAAACAXuwtdAAAAAAAADy3QD0DAAAAAAC9gHoGAAAAAAB6AfUMAAAAAAD0AuoZAAAAAADoBdQzAAAAAADQC6hnAAAAAACgF1DPAAAAAABAL6CeAQAAAACAXkA9AwAAAAAAvfg/LyjBsj7segoAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r107/107 [==============================] - 16s 148ms/step - loss: 2.8170 - note_outputs_loss: 2.2675 - duration_outputs_loss: 0.5496\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.src.callbacks.History at 0x7a4ac5701000>"
            ]
          },
          "metadata": {},
          "execution_count": 27
        }
      ],
      "source": [
        "model.fit(\n",
        "    ds,\n",
        "    epochs=EPOCHS,\n",
        "    callbacks=[\n",
        "        model_checkpoint_callback,\n",
        "        tensorboard_callback,\n",
        "        music_generator,\n",
        "    ],\n",
        ")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 28,
      "id": "4de2131e-f13c-45af-ab8d-c361b0be8640",
      "metadata": {
        "id": "4de2131e-f13c-45af-ab8d-c361b0be8640"
      },
      "outputs": [],
      "source": [
        "# 최종 모델 저장\n",
        "model.save(\"./models/model\")"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "7f0445d3-5513-428b-995a-b337547d2a71",
      "metadata": {
        "id": "7f0445d3-5513-428b-995a-b337547d2a71"
      },
      "source": [
        "# 3. 트랜스포머를 사용하여 음악 생성"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 29,
      "id": "a0776a68-d4c7-439b-a3fd-d9397b357662",
      "metadata": {
        "id": "a0776a68-d4c7-439b-a3fd-d9397b357662",
        "outputId": "1d6400e5-c1f2-4e23-eace-48517991eec4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 90
        }
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {}
        }
      ],
      "source": [
        "info = music_generator.generate(\n",
        "    [\"START\"], [\"0.0\"], max_tokens=50, temperature=0.5\n",
        ")\n",
        "midi_stream = info[-1][\"midi\"].chordify()\n",
        "display(Image(str(midi_stream.write('lily.png'))))"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "8932e58a-5a9c-44a3-bff4-39c564fafdf4",
      "metadata": {
        "id": "8932e58a-5a9c-44a3-bff4-39c564fafdf4"
      },
      "source": [
        "## MIDI 파일로 음악 저장하기"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "id": "2402314a-ca87-468e-a9d1-b79403fe7065",
      "metadata": {
        "id": "2402314a-ca87-468e-a9d1-b79403fe7065",
        "outputId": "565adb14-5025-4191-de7e-f28db5aa29b8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'./output/output-20240408-045033.mid'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 30
        }
      ],
      "source": [
        "timestr = time.strftime(\"%Y%m%d-%H%M%S\")\n",
        "midi_stream.write(\n",
        "    \"midi\",\n",
        "    fp=os.path.join(\n",
        "        \"./output\",\n",
        "        \"output-\" + timestr + \".mid\",\n",
        "    ),\n",
        ")"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "24eb14c3-9dab-42b9-8916-b9b3a4d38590",
      "metadata": {
        "id": "24eb14c3-9dab-42b9-8916-b9b3a4d38590"
      },
      "source": [
        "## 음표 확률"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 31,
      "id": "c92ad9f2-74ed-4432-b633-f74fd17b6a49",
      "metadata": {
        "id": "c92ad9f2-74ed-4432-b633-f74fd17b6a49"
      },
      "outputs": [],
      "source": [
        "max_pitch = 70\n",
        "seq_len = len(info)\n",
        "grid = np.zeros((max_pitch, seq_len), dtype=np.float32)\n",
        "\n",
        "for j in range(seq_len):\n",
        "    for i, prob in enumerate(info[j][\"note_probs\"]):\n",
        "        try:\n",
        "            pitch = music21.note.Note(notes_vocab[i]).pitch.midi\n",
        "            grid[pitch, j] = prob\n",
        "        except:\n",
        "            pass  # 키와 박자표를 출력하지 않습니다"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 32,
      "id": "0d4ce565-a5d4-4cec-b321-1e02d814b65c",
      "metadata": {
        "id": "0d4ce565-a5d4-4cec-b321-1e02d814b65c",
        "outputId": "56d1a283-6b33-4747-ccd7-4d7ad819fbf0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 503
        }
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 800x800 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "image/png": {
              "width": 667,
              "height": 486
            }
          }
        }
      ],
      "source": [
        "fig, ax = plt.subplots(figsize=(8, 8))\n",
        "ax.set_yticks([int(j) for j in range(35, 70)])\n",
        "plt.imshow(\n",
        "    grid[35:70, :],\n",
        "    origin=\"lower\",\n",
        "    cmap=\"coolwarm\",\n",
        "    vmin=-0.5,\n",
        "    vmax=0.5,\n",
        "    extent=[0, seq_len, 35, 70],\n",
        ")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "id": "bdae49db-6e0f-4071-a289-74310d684fab",
      "metadata": {
        "id": "bdae49db-6e0f-4071-a289-74310d684fab"
      },
      "source": [
        "## 어텐션 그래프"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 33,
      "id": "cf270a8f-da3f-4d21-b4f9-c19fcf7c4332",
      "metadata": {
        "id": "cf270a8f-da3f-4d21-b4f9-c19fcf7c4332"
      },
      "outputs": [],
      "source": [
        "plot_size = 20\n",
        "\n",
        "att_matrix = np.zeros((plot_size, plot_size))\n",
        "prediction_output = []\n",
        "last_prompt = []"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 34,
      "id": "6a3fcec1-f2fc-4dbc-b997-d00a958b6388",
      "metadata": {
        "tags": [],
        "id": "6a3fcec1-f2fc-4dbc-b997-d00a958b6388"
      },
      "outputs": [],
      "source": [
        "for j in range(plot_size):\n",
        "    atts = info[j][\"atts\"].max(axis=0)\n",
        "    att_matrix[: (j + 1), j] = atts\n",
        "    prediction_output.append(info[j][\"chosen_note\"][0])\n",
        "    last_prompt.append(info[j][\"prompt\"][0][-1])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 35,
      "id": "6b2742cf-ad18-4c23-8181-5d868ca03c0f",
      "metadata": {
        "id": "6b2742cf-ad18-4c23-8181-5d868ca03c0f",
        "outputId": "fd07ed6d-9cbb-4b9c-b546-c32b1e9c954f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 702
        }
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 800x800 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "image/png": {
              "width": 688,
              "height": 685
            }
          }
        }
      ],
      "source": [
        "fig, ax = plt.subplots(figsize=(8, 8))\n",
        "im = ax.imshow(att_matrix, cmap=\"Greens\", interpolation=\"nearest\")\n",
        "\n",
        "ax.set_xticks(np.arange(-0.5, plot_size, 1), minor=True)\n",
        "ax.set_yticks(np.arange(-0.5, plot_size, 1), minor=True)\n",
        "ax.grid(which=\"minor\", color=\"black\", linestyle=\"-\", linewidth=1)\n",
        "ax.set_xticks(np.arange(plot_size))\n",
        "ax.set_yticks(np.arange(plot_size))\n",
        "ax.set_xticklabels(prediction_output[:plot_size])\n",
        "ax.set_yticklabels(last_prompt[:plot_size])\n",
        "ax.xaxis.tick_top()\n",
        "\n",
        "plt.setp(\n",
        "    ax.get_xticklabels(),\n",
        "    rotation=90,\n",
        "    ha=\"left\",\n",
        "    va=\"center\",\n",
        "    rotation_mode=\"anchor\",\n",
        ")\n",
        "plt.show()"
      ]
    }
  ],
  "metadata": {
    "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": "3.10.9"
    },
    "vscode": {
      "interpreter": {
        "hash": "aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49"
      }
    },
    "colab": {
      "provenance": [],
      "gpuType": "T4"
    },
    "accelerator": "GPU"
  },
  "nbformat": 4,
  "nbformat_minor": 5
}