hamori/notebooks/report.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "daef9995",
   "metadata": {},
   "source": [
    "# hamori — отчёт по проекту\n",
    "\n",
    "**Автор:** Владислав Аршинов, группа БСМО-11-25\n",
    "**Дата:** июнь 2026\n",
    "\n",
    "Тема работы — генерация коротких аккордовых последовательностей (я называю их\n",
    "«гармоническими периодами», 4–16 тактов) в моём композиторском стиле с помощью\n",
    "небольшого трансформера.\n",
    "\n",
    "Название *hamori* (яп. ハモリ — «вторить, петь вторым голосом») отражает идею\n",
    "проекта: модель не сочиняет музыку с нуля, а предлагает гармонию к уже\n",
    "задуманной композитором линии — добавляет к ней второй голос.\n",
    "\n",
    "Ноутбук одновременно служит отчётом: текст, код и графики собраны в одном файле,\n",
    "чтобы результаты считались на месте. Для полного прогона нужны обученные\n",
    "чекпоинты в `checkpoints/` и данные в `data/`."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "812423b0",
   "metadata": {},
   "source": [
    "## 0. Постановка задачи\n",
    "\n",
    "Я генерирую не целую песню, а один законченный гармонический период — короткий\n",
    "замкнутый фрагмент гармонии на 4–16 тактов. В такте несколько позиций (их число\n",
    "задаёт `subdivision`); в позиции может стоять аккорд, точка `.` («держать\n",
    "предыдущий»), `NC` (без аккорда) или `?` (не разобрано).\n",
    "\n",
    "Конвейер устроен так:\n",
    "\n",
    "1. Переношу свои композиции из проектов REAPER в текстовый формат `.chord`.\n",
    "2. Паршу `.chord` и перевожу в последовательности токенов.\n",
    "3. **Предобучаю** модель на большом внешнем корпусе (McGill Billboard).\n",
    "4. **Дообучаю** на собственном корпусе.\n",
    "5. Сэмплирую новые периоды, задавая лад, размер, стиль, функцию и, при желании,\n",
    "   начальные аккорды.\n",
    "6. Перевожу результат обратно в `.chord` и в MIDI для REAPER.\n",
    "\n",
    "Основная гипотеза, которую я проверяю: дообучение на небольшом авторском корпусе\n",
    "сдвигает генерации ближе к моему стилю, не разрушая при этом общие закономерности\n",
    "гармонии, выученные на большом корпусе. Проверяю это тремя способами —\n",
    "перплексией, сравнением распределений и разбором конкретных примеров."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0f49cdea",
   "metadata": {},
   "source": [
    "## 1. Окружение\n",
    "\n",
    "Ячейка ниже находит корень репозитория (папку с `src/`), добавляет его в\n",
    "`sys.path`, фиксирует random seed для воспроизводимости и импортирует модули\n",
    "проекта. Основную логику — парсинг, токенизацию, расчёт перплексии, генерацию —\n",
    "я не дублирую в ноутбуке, а беру готовую из `src/`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "eb4e5ef7",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-04T13:53:56.475231Z",
     "iopub.status.busy": "2026-06-04T13:53:56.475231Z",
     "iopub.status.idle": "2026-06-04T13:53:59.666222Z",
     "shell.execute_reply": "2026-06-04T13:53:59.666222Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROOT      = D:\\Projects\\Devel\\hamori\n",
      "device    = cpu\n",
      "vocab     = 84 токенов\n"
     ]
    }
   ],
   "source": [
    "import sys, csv, random\n",
    "from pathlib import Path\n",
    "\n",
    "import numpy as np\n",
    "import torch\n",
    "import matplotlib.pyplot as plt\n",
    "from IPython.display import Image, display\n",
    "\n",
    "# Найти корень репозитория (каталог, содержащий src/).\n",
    "ROOT = Path.cwd()\n",
    "while not (ROOT / \"src\").exists() and ROOT != ROOT.parent:\n",
    "    ROOT = ROOT.parent\n",
    "sys.path.insert(0, str(ROOT))\n",
    "\n",
    "SEED = 42\n",
    "random.seed(SEED); np.random.seed(SEED); torch.manual_seed(SEED)\n",
    "DEVICE = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "\n",
    "from src.tokenizer import VOCAB, ID_TO_TOKEN, parse_chord_file, tokenize_period\n",
    "from src.model import ChordTransformer\n",
    "from src.dataset import ChordDataset\n",
    "from src.evaluate import compute_perplexity\n",
    "from src.generate import generate_period\n",
    "\n",
    "# src.evaluate forces the headless 'Agg' backend on import (used by the CLI to\n",
    "# save PNGs); re-enable the inline backend so figures render in this notebook.\n",
    "%matplotlib inline\n",
    "\n",
    "CKPT = ROOT / \"checkpoints\"\n",
    "DATA = ROOT / \"data\"\n",
    "OUT  = ROOT / \"output\" / \"eval\"\n",
    "\n",
    "print(f\"ROOT      = {ROOT}\")\n",
    "print(f\"device    = {DEVICE}\")\n",
    "print(f\"vocab     = {len(VOCAB)} токенов\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3221b89a",
   "metadata": {},
   "source": [
    "## 2. Данные\n",
    "\n",
    "### Формат `.chord`\n",
    "\n",
    "Один файл — один период. Сверху заголовок (строки с `#`): `title`, `key`,\n",
    "`time`, `subdivision`, `style`, `function`. Дальше тело — такты через `|`, по\n",
    "`subdivision` позиций в каждом.\n",
    "\n",
    "Аккорд записывается как `<тоника><качество?><надстройка?>(/<бас>)?`. Качеств 18,\n",
    "надстроек 7, обращения (через слеш, например `C/E`) — обязательная часть записи:\n",
    "для моей музыки линия баса важна, поэтому я её не отбрасываю.\n",
    "\n",
    "### Нормализация тональности\n",
    "\n",
    "Перед токенизацией я транспонирую весь период: мажор → в C-dur, минор → в\n",
    "a-moll. Так модель не видит абсолютную тональность — только лад (`MODE_major` /\n",
    "`MODE_minor`). Это сделано намеренно: периодов всего 68, и они в разных\n",
    "тональностях. Если свести их к одной, все они учат одну и ту же гармоническую\n",
    "«грамматику», а не размазываются по 24 тональностям. После генерации результат\n",
    "транспонируется обратно в нужную тональность.\n",
    "\n",
    "### Токенизация\n",
    "\n",
    "Каждый новый аккорд кодируется ровно **4 токенами**: `ROOT_*` (тоника), `QUAL_*`\n",
    "(качество), `EXT_*` (надстройка), `BASS_*` (бас). Удержание — один токен `HOLD`.\n",
    "Границы тактов токенами не являются: при обратной сборке они восстанавливаются\n",
    "подсчётом позиций (`time` × `subdivision`). Плюс метатокены в начале. Итоговый\n",
    "словарь — **84 токена**."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "39e3ff95",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-04T13:53:59.668223Z",
     "iopub.status.busy": "2026-06-04T13:53:59.668223Z",
     "iopub.status.idle": "2026-06-04T13:53:59.699235Z",
     "shell.execute_reply": "2026-06-04T13:53:59.699235Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Корпус автора (.chord) : 68 периодов\n",
      "  → train / val        : 61 / 7\n",
      "McGill (предобучение)  : 4694 периодов (train-сплит)\n"
     ]
    }
   ],
   "source": [
    "user_files   = sorted((DATA / \"raw_user\" / \"H1K0\").glob(\"*.chord\"))\n",
    "mcgill_train = sorted((DATA / \"processed\" / \"mcgill\" / \"train\").glob(\"*.pt\"))\n",
    "user_train   = sorted((DATA / \"processed\" / \"user\" / \"train\").glob(\"*.pt\"))\n",
    "user_val     = sorted((DATA / \"processed\" / \"user\" / \"val\").glob(\"*.pt\"))\n",
    "\n",
    "print(f\"Корпус автора (.chord) : {len(user_files)} периодов\")\n",
    "print(f\"  → train / val        : {len(user_train)} / {len(user_val)}\")\n",
    "print(f\"McGill (предобучение)  : {len(mcgill_train)} периодов (train-сплит)\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "11d7b64f",
   "metadata": {},
   "source": [
    "Ниже — один период «как есть», сырой текст файла `.chord`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "35016bad",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-04T13:53:59.701235Z",
     "iopub.status.busy": "2026-06-04T13:53:59.701235Z",
     "iopub.status.idle": "2026-06-04T13:53:59.704236Z",
     "shell.execute_reply": "2026-06-04T13:53:59.704236Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# title: Celestial Sphere\n",
      "# key: Eb_major\n",
      "# time: 4/4\n",
      "# subdivision: 4\n",
      "# style: H1K0\n",
      "# function: chorus\n",
      "\n",
      "| Eb . Bb . | Ab . . Bb | Eb . Eb/G . | Fm . Bb . | Eb . . . | Ab . Bb . | Fm . . . | Bb . . . |\n",
      "\n"
     ]
    }
   ],
   "source": [
    "example = DATA / \"raw_user\" / \"H1K0\" / \"celestial_sphere-chorus.chord\"\n",
    "print(example.read_text(encoding=\"utf-8\"))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2e56ea8a",
   "metadata": {},
   "source": [
    "Тот же период после парсинга и токенизации. Здесь видна нормализация, о\n",
    "которой я писал выше: в файле тональность `Eb_major`, но первые токены —\n",
    "`MODE_major` (период уже переведён в C-dur), а первый аккорд `Eb` стал `ROOT_C`.\n",
    "Абсолютная тональность исчезла, остался только лад."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "5812ab8f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-04T13:53:59.706236Z",
     "iopub.status.busy": "2026-06-04T13:53:59.706236Z",
     "iopub.status.idle": "2026-06-04T13:53:59.711992Z",
     "shell.execute_reply": "2026-06-04T13:53:59.710999Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Тональность в файле : Eb_major\n",
      "Тактов              : 8\n",
      "Токенов всего       : 78\n",
      "\n",
      "Первые 14 токенов (метаданные + первый аккорд):\n",
      "  <BOS> MODE_major TIME_4/4 SUB_4 STYLE_H1K0 FUNC_chorus ROOT_C QUAL_maj EXT_none BASS_root HOLD ROOT_G QUAL_maj EXT_none\n"
     ]
    }
   ],
   "source": [
    "period = parse_chord_file(example)\n",
    "ids    = tokenize_period(period)\n",
    "toks   = [ID_TO_TOKEN[i] for i in ids]\n",
    "\n",
    "print(f\"Тональность в файле : {period.key}\")\n",
    "print(f\"Тактов              : {len(period.bars)}\")\n",
    "print(f\"Токенов всего       : {len(ids)}\")\n",
    "print()\n",
    "print(\"Первые 14 токенов (метаданные + первый аккорд):\")\n",
    "print(\"  \" + \" \".join(toks[:14]))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9c7392d8",
   "metadata": {},
   "source": [
    "## 3. Модель\n",
    "\n",
    "Я использую небольшой авторегрессионный трансформер — по сути ту же архитектуру,\n",
    "что и в языковых моделях, только вместо слов здесь аккордовые токены.\n",
    "\n",
    "### Принцип работы\n",
    "\n",
    "Модель учится предсказывать следующий токен по всем предыдущим: она получает\n",
    "начало периода и оценивает, какой токен идёт дальше; ошибка считается\n",
    "кросс-энтропией.\n",
    "\n",
    "Подробнее о том, что происходит внутри:\n",
    "\n",
    "- **Эмбеддинги.** Каждый из 84 токенов превращается в вектор длиной `d_model`. К\n",
    "  нему добавляется позиционная информация — без неё модель не различала бы\n",
    "  порядок токенов.\n",
    "- **Self-attention.** Основной механизм трансформера. На каждой позиции модель\n",
    "  «смотрит» на остальные позиции последовательности и определяет, какие из них\n",
    "  сейчас значимы. Например, выдавая очередной аккорд, она может учесть тонику в\n",
    "  начале периода и предыдущий аккорд.\n",
    "- **Причинная (causal) маска.** Чтобы при обучении модель не использовала будущие\n",
    "  токены, внимание ограничено «назад»: токен на позиции *i* видит только позиции\n",
    "  до *i* включительно. Без этого задача предсказания следующего токена была бы\n",
    "  тривиальной — ответ виден заранее.\n",
    "- **FFN и слои.** После внимания идёт полносвязный блок, и так несколько раз\n",
    "  (2–4 слоя). Каждый слой позволяет уловить более сложные зависимости.\n",
    "- **Выход.** На последнем слое модель выдаёт распределение вероятностей по всем\n",
    "  84 токенам — насколько вероятен каждый следующим. При генерации я сэмплирую из\n",
    "  этого распределения (раздел 6).\n",
    "\n",
    "Входные и выходные эмбеддинги связаны (tied embeddings): это экономит параметры и\n",
    "обычно немного помогает на небольших данных.\n",
    "\n",
    "Конкретные размеры (число слоёв, `d_model`, число голов) хранятся в чекпоинте —\n",
    "выведу их и посчитаю число параметров. Модель небольшая и обучается даже на\n",
    "CPU."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "187c8063",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-04T13:53:59.713992Z",
     "iopub.status.busy": "2026-06-04T13:53:59.713992Z",
     "iopub.status.idle": "2026-06-04T13:53:59.926954Z",
     "shell.execute_reply": "2026-06-04T13:53:59.926954Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Конфигурация модели:\n",
      "  vocab_size         = 84\n",
      "  d_model            = 192\n",
      "  n_layers           = 3\n",
      "  n_heads            = 6\n",
      "  d_ff               = 768\n",
      "  max_seq_len        = 320\n",
      "  dropout            = 0.1\n",
      "\n",
      "Параметров (с привязанными эмбеддингами): 1,412,544\n"
     ]
    }
   ],
   "source": [
    "def load_model(path: Path):\n",
    "    ckpt  = torch.load(path, map_location=DEVICE, weights_only=True)\n",
    "    model = ChordTransformer(**ckpt[\"model_config\"])\n",
    "    model.load_state_dict(ckpt[\"model_state\"])\n",
    "    model.to(DEVICE).eval()\n",
    "    return model, ckpt[\"model_config\"]\n",
    "\n",
    "pretrained, cfg = load_model(CKPT / \"pretrained.pt\")\n",
    "finetuned,  _   = load_model(CKPT / \"finetuned.pt\")\n",
    "\n",
    "n_params = sum(p.numel() for p in finetuned.parameters())\n",
    "print(\"Конфигурация модели:\")\n",
    "for k, v in cfg.items():\n",
    "    print(f\"  {k:18s} = {v}\")\n",
    "print(f\"\\nПараметров (с привязанными эмбеддингами): {n_params:,}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d660bc1",
   "metadata": {},
   "source": [
    "## 4. Обучение\n",
    "\n",
    "Обучение двухэтапное:\n",
    "\n",
    "1. **Предобучение** на McGill Billboard. На большом внешнем корпусе модель\n",
    "   усваивает общие закономерности: как движется гармония, какие бывают\n",
    "   надстройки, насколько часто встречаются обращения. Мой стиль здесь ни при чём\n",
    "   — задача в том, чтобы модель в принципе «понимала» гармонию.\n",
    "2. **Дообучение** на собственном корпусе. Здесь я ставлю заметно меньший learning\n",
    "   rate (порядка 3·10⁻⁵ против 3·10⁻⁴ на предобучении) и небольшое число эпох.\n",
    "   Цель — сместить модель в сторону моего стиля, но не дать ей забыть выученное на\n",
    "   большом корпусе (катастрофическое забывание).\n",
    "\n",
    "Кривые потерь беру из CSV-логов, сохранённых тренировочным скриптом."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "a38d23fd",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-04T13:53:59.928949Z",
     "iopub.status.busy": "2026-06-04T13:53:59.927948Z",
     "iopub.status.idle": "2026-06-04T13:54:00.437460Z",
     "shell.execute_reply": "2026-06-04T13:54:00.437460Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1300x450 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def load_log(path: Path):\n",
    "    e, tr, vl, pp = [], [], [], []\n",
    "    with open(path, encoding=\"utf-8\") as fh:\n",
    "        for row in csv.DictReader(fh):\n",
    "            e.append(int(row[\"epoch\"]))\n",
    "            tr.append(float(row[\"train_loss\"]))\n",
    "            vl.append(float(row[\"val_loss\"]))\n",
    "            pp.append(float(row[\"val_ppl\"]))\n",
    "    return e, tr, vl, pp\n",
    "\n",
    "fig, axes = plt.subplots(1, 2, figsize=(13, 4.5))\n",
    "for ax, name, title in [\n",
    "    (axes[0], \"pretrained\", \"Предобучение (McGill)\"),\n",
    "    (axes[1], \"finetuned\",  \"Дообучение (корпус автора)\"),\n",
    "]:\n",
    "    e, tr, vl, pp = load_log(CKPT / f\"{name}.log.csv\")\n",
    "    ax.plot(e, tr, label=\"train loss\")\n",
    "    ax.plot(e, vl, label=\"val loss\")\n",
    "    ax.set_xlabel(\"эпоха\"); ax.set_ylabel(\"loss\")\n",
    "    ax.set_title(f\"{title} · лучшая val-PPL = {min(pp):.2f}\")\n",
    "    ax.grid(alpha=0.3); ax.legend()\n",
    "fig.tight_layout(); plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c36be84c",
   "metadata": {},
   "source": [
    "Предобучение сошлось к val-перплексии около **1.32** за 50 эпох — McGill\n",
    "большой и довольно регулярный корпус, поэтому перплексия низкая. Дообучение вышло\n",
    "на плато около **2.15** на моей валидации (лучшая эпоха — 20-я). Более высокая\n",
    "перплексия после дообучения ожидаема: мой корпус меньше и стилистически\n",
    "своеобразнее, предсказывать его сложнее."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7f3fdef3",
   "metadata": {},
   "source": [
    "## 5. Оценка\n",
    "\n",
    "Гипотезу я проверяю с трёх сторон:\n",
    "\n",
    "1. **Перплексия** — насколько хорошо модель предсказывает мои отложенные периоды.\n",
    "2. **Распределения** — гистограммы по качествам аккордов, надстройкам, обращениям\n",
    "   и интервалам между тониками; по ним видно, в какую сторону дообучение двигает\n",
    "   генерации.\n",
    "3. **Примеры** — одно и то же условие, генерация до и после дообучения, сравнение\n",
    "   вручную.\n",
    "\n",
    "> **О слабом месте эксперимента.** Папку `data/holdout/` в этой версии я не\n",
    "> заполнил, поэтому перплексия считается на **валидационном** сплите (7\n",
    "> периодов), а не на отдельном тестовом наборе. Валидация косвенно участвовала в\n",
    "> выборе эпохи, так что числа стоит читать как направление и порядок величины, а\n",
    "> не как строгую несмещённую оценку. При наличии времени я заполню holdout и\n",
    "> пересчитаю."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d328eaaf",
   "metadata": {},
   "source": [
    "### 5.1 Перплексия\n",
    "\n",
    "Считаю вживую, на своей валидации, сразу для обеих моделей."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "70f978c3",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-04T13:54:00.439454Z",
     "iopub.status.busy": "2026-06-04T13:54:00.439454Z",
     "iopub.status.idle": "2026-06-04T13:54:00.929473Z",
     "shell.execute_reply": "2026-06-04T13:54:00.928957Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Перплексия на валидации автора (7 периодов):\n",
      "  pretrained = 3.58\n",
      "  finetuned  = 2.15\n",
      "  улучшение  = +39.8%\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAasAAAF2CAYAAADQsbecAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjksIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvJkbTWQAAAAlwSFlzAAAPYQAAD2EBqD+naQAAPVBJREFUeJzt3QeUE1XbB/CHvksH6W3pVToqTYqCCIgiKkiRXnxFaQoIqLAgTakqUgUURAQVRFHKi/QiXXqHpRepS8dlvvN/vjN5k2yym2SzZEz+v3MGNpOZyc1kZp65Ze5NYhiGIURERBaWNNAJICIiig+DFRERWR6DFRERWR6DFRERWR6DFRERWR6DFRERWR6DFRERWR6DFRERWR6DFRERWR6DFVEIqVWrlk6BMGjQIEmSJIlYwaeffioFCxaUZMmSSbly5XRe/vz5pW3btoFOmqXkt9A+YbBKBDNnztSTMq7p8ccfT4yPpn+5ffv26UX9xIkTgU5K0Fq2bJn06dNHqlWrJjNmzJBhw4Y9ss/+7bff9Pcl7yX3YR3y0ODBg6VAgQKx5g8dOpT7kNwGq8jISM394K42MS7Uoe6PP/6QpEmTyldffSUpU6a0zT948KDOT+xgNWHCBAYsHzBYJaL69etLpUqVYs2fNm2a/P3334n50WQR//zzjzx8+NDhougv6IP67t27Eh4e7vE6iZGOf5uLFy/qPnPeF6lSpQpYmih+LAa0kNmzZ0vFihX1RMqcObO8/vrrcurUKYdlcMeNIsRt27ZJ1apVdVnk3iZNmuSw3KpVq+IshnRVDo1tu1oWxZr2/vzzT3n++eclQ4YMkjp1aqlZs6asX78+3vqJlStX6gXhzTffdJh/5swZ6dChg+TKlUvfx/f5z3/+I/fv33coVrUvGkMAKFOmjMv0/fDDD3qTkC5dOofvMWrUKNsyDx48kAMHDsi5c+fi/V2wr9KmTSvHjh2TevXqSZo0aTStyDnbD1qA9JmfM27cOClUqJB+H+SWAJ/36quv6m8bFhamaVy0aJFtfXyP1157Tf+uXbu2Ld34LQE5rRdeeEGWLl2q6+K3nzx5sr6H4qxnnnlGsmXLpp9ZsmRJmThxYrx1VuZxMm/ePM3x58mTR9P27LPPypEjR2Kt78lvD+vWrZMnnnhCt4X9YKYzPm+//bbu69u3b8d6r3nz5pIjRw6JiYnR11u3btXfI0uWLLbzoH379nFuH98V++rWrVuxjm/n+hnzuMP369Wrl2TNmlV/+5dfflkuXboUa9u///67PP3007oMjr2GDRvK3r17be9j28hVmekwJ/vfwfytnY+pmXbHuHk84rxp3Lix/o20vffee7Z9Y3+e4FgsVaqU/hbZs2eXLl26yNWrVx2Ww3H88ccf6++P3xXHn33arYA5K4vAheLDDz+Upk2bSseOHfVk+Pzzz6VGjRqyY8cOyZgxo21ZHGgNGjTQZXEC40KDizvuFJ1P1m7duulFwx62707x4sVlwIAB+jdyfz179oxVhIIcI4LqwIEDtdjEvFCuXbtWnnzySZfb/euvv/TEQrrNExbOnj2r61y7dk06d+6sn4+TEAEHFyx3OYFZs2bJ7t27Y83fuHGj7peyZcvKiBEj9KLq6nvgM0qUKCFt2rSJFexcwUUAF+nKlSvLJ598IkuWLNHvj5wTgpY97A/kePB9EDgQnHDio44kd+7c8v777+sFDb8b9smPP/6oF0D81vi9PvvsM+nfv7+mD8z/zaIq/Oa44HTq1EmKFSum8xGYcEF68cUXJXny5PLLL7/IW2+9pRerrl27xvv9sK/wW+KCd/36df2OLVu21ODk7W+P3+W5557TCyhuWrCPsDwulPFp1qyZHh+LFy+2BW7AsYDvhAs1GkUgd2R+BvYnzg9c2H/66ac4t4/jZsqUKbJ582Yt4QDc9MXlnXfekUyZMul3wGfg4o+g+v333ztsF8cSgufIkSM1vfhNqlevrucvAiF+Mxzvy5cv1+UTIiYmRj/rqaee0puj//73vzJ69Gi9McC1wITPxPHdrl07PbaOHz8uX3zxhaYJQThFihS63EcffaTBCucnpu3bt+v+NW8YLQHjWZF/zZgxA7fbxpYtW1y+X7NmTaNUqVK21ydOnDCSJUtmDB061GG53bt3G8mTJ3eYj3Wx7dGjR9vm3bt3zyhXrpyRLVs24/79+zpv5cqVutz8+fNjfX6aNGmMNm3axJpfrVo1o3bt2rbXx48f123g+8DDhw+NIkWKGPXq1dO/Tbdv3zYKFChg1K1b1zZv4MCBuq75/XLmzGlUr17duHPnjsNntm7d2kiaNKnLfWV+hrk/kR64e/eukS9fPqN+/foO6YN+/frpvHPnzsX6Hp9++mmsea72gzMsg2Xfeecdh7Q1bNjQSJkypXHp0iWHbaZPn964ePGiwzaeffZZo3Tp0pp2+21UrVpV96kJvxe2gd/PWUREhL63ZMmSWO/hN3CG36lgwYIO83D8YDKZx0mJEiX0ODKNHz9e5+MY9Pa3b9y4sREWFmZERUXZ5u3bt0+P8fguOdh27ty5jVdeecVh/rx583TdNWvW6OsFCxbEeY7F93viHHC1f+2PB/O4q1OnjsN37tmzp36Xa9eu6evo6GgjY8aMRqdOnRy2d/78eSNDhgwO87t27epyH5i/g/Pv7nwO2h+PgwcPNuyVL1/eqFixou312rVrdblvv/3WYTkcP/bzcaziOMbxbP89+/fv7/E58iiwGNACcDeIO2DkCJALMCcUeRQpUkSLz+zhzhl3TCbkPvAad5soHvQV7qLiKrffuXOnHD58WFq0aCGXL1+2pRNFKig2WrNmjX4Pe1gOd4AoFkGRF4oiTFh24cKF0qhRI5d1e+6aOePOG9vFna6z6OhoveO3z4m6gjtdFH14kqsy4W7aPm14jX2Gu1p7r7zyit7xm65cuaK5Evy+SJ+538x9g32KnJ4nUNSFdZzZ11shZ4Tto4gORZd4HR/cedvnYlGcBVjfm98ed/wopkSOMV++fLbtIXfoKt3OsF+Ro0JDhJs3b9rmIxeDXClyKmD+vr/++qsW6SYm5JDtj0XsG3zPqKgofY2cEkoGkOO1P3+RA0TOx/n89Zc3nYrTkS7z94L58+dryULdunUd0oWcMYoOzXTh+MVxjByk/ffs0aOHWAmLAS0AFwFcOBGYXDGz6ibUl6AYyV7RokX1fxRToKjKFzjhIiIi4kwnoLjDHVwYUWRiQh0Liq5Ql+I8KDWKOm/cuOFVM35sH02NUYfgqlipSpUqWszRvXt3bZ6Mk9W5fN4XCIB4LsfdPrfn3AIUdT/47ijmxeQKbjRwMY6Pq9algCIdBG8UgzrX92CfYT/ExT6wgPkbmvvO09/+3r17cufOHZfHMoosEYQ8KQpEURtubhAcEbSwHm7IzIspAjFuCtBycuzYsVoPhwCJ5f3dUMLTfYPiUFfSp08v/hYWFuZwQ2Smy/5YR7rwm+Dcc3fMgRl0nX8zbN/+XA40BisLwB0pTkJU0OJuzBnugh6F8+fPx3n3a+aa8ECl+SBlfGlFowJ8L+Qq3n33Xa3jSAjUByBw9O7dW+/wnaFRCsrbUd+HuolAcG6dZ+431Ae527+FCxf2adtw9OhRzd2gvm/MmDGSN29ezSXhAo8LuXNu1xVXxx2YNxie/vYIVgmFmy3kfFGnh+CDuioEQAQxE84X1Gtu2rRJ30duDvW1qLfBPH+eM57uG9RDoTTEGUpC4uOuFMG5wUR8abKHdCFQffvtty7fdw52VsdgZQGoFMWBj7tm8249LqikRfGLfe7q0KFD+r+vz+acPn1ai6jsK/NdpdO8U6xTp45H28XdMYonhg8frsVmrVq10gurebJgW3v27PFoW/je48eP122hWNFVsEIgQ4UzKvlRmfzll1/KhQsX9HMTAic+iljsfx9P97mZI0MOOb795ksPD7hYI0hgX9vnAvxZ/OTpb4/fFAHVzG3YQw7bU7i5wW+NnDeKALGPXZUYYB4mNFCaM2eONgqZO3dunI2I/M3cNwgMvv6+Zg4GpRv2zFyPr+lCER8a9sT1eINZmoLfzL70ACUf/iiV8BfWWVlAkyZN9E4JRRrORWV47XxRRusq+6bAKG/Ga1woUB7tC5zgcRVlALaNEwDBwL4+weSqOa9Z94GWaWh1haIc3CWbgQVFN7jYohmyM+d9gf2Doj/nsnpnyFWhjgh3lLh44GR15k3TdROKF+3ThtcIQGbwdQcXMRRT4Tdy9Xn2+828AXG+aMXFvMu2318o/kloLtaX3x5pQe4RdZEnT560vb9//37N/XgKuSgE4K+//lpbXiJ42cNF1Pn4MHN8/sjdeQPfF0EcxdOu6s88+X0RMLDvUPdnDzdbvmratKnmzIYMGRLrPVxDzDTgHMFxjPPGfp+iKNZKmLOyAFwE0Gy0X79+Wv+BCzhyDsgZLFiwQCt4UYRkX2eF4jAsizt93HmiAhzFXs71W/FBrgN1HWjGiyI0FCW5g+CC5dB8Gc2kUSmPehY0DsBdPE5YBB53d5RYFxcUfB6aRgNOcPSqgDoIfE/k7HBBR+UwntWxbyiB5RCA4nqwFU3EUVeFJtPOTfYT0nQddQS4aGJ5VJqjaBPNq9HE3JPiFDQKQeOA0qVLa5Nz3MFi36OOCblaNO0H7B9ctPD7IuCg/sV8fsodNDHGPkFDFdwMIJhMnTpV1/EmGMfFm98eNxXYV7hRwU0KLoy4EGK9Xbt2efR5FSpU0KJRPEaB4GNfBAgIYriQo8k/zh+UCuA7Ix1oev0o4TPRTP2NN97QdOM8wjGBYI1jBDdL5o2OeTOJZuQIcvitsTzqFNGwBPsJ5wq+ExqPmPVKvsA5heMBJRG4PuA4wfUBOSicX8i54rk/8xktLIc6Zuw/NG3HMY5n2Cwj0M0Rg5G3TddNP/74ozbvRrNaTMWLF9emrgcPHoy17tatW40qVapoE2E0uf3iiy8ctuVp0/X169cbhQsXNgYNGuTQdNlds1nYsWOH0aRJE+Oxxx4zUqVKpZ/ftGlTY8WKFS6brtuLjIzU5vjbt2+3zUMTZzRhz5o1q24Pza3xvc30mPsTzfPtm9Y6pw/NwsuUKaP78J9//vFr03Xss6NHjxrPPfeckTp1aiN79uz6HWNiYuL8HHtYH98zR44cRooUKbSJ9gsvvGD88MMPDstNnTpV94HZ1Ntszoz9jObFrixatEi/O46H/PnzGyNHjjSmT5/u0OQ/rqbrzsdJQn57WL16tTajRpNofJdJkya5PSbcGTBggC6P49MZjp/mzZvrIwxIBx7bwL7EeeHvpuvO57G7ZuZ4jab9aK6O36FQoUJG27ZtHdKE4xKPQOBYT5IkicP+wCMQaLKP4ytTpkxGly5djD179rhsup7GRfrd7d8pU6bobxEeHm6kS5dOH6Ho06ePcfbsWdsyOI5xbuIREyxXq1Yt/WznfRJISfBPoAMmeQ7FSWh+6mk9DyUcHkRFZb6r4i8iejRYZ0VERJbHYEVERJbHYEVERJbHOisiIrI85qyIiMjyGKyIiMjyQu6hYHSbg257zIH5iIgocPD0FB7qRmcHePjcnZALVghU6OiTiIisA6OiY6Rid0IuWCFHZe6YxOi6P5iYvcCbHe2io1CMYotRYV11eIuukDBqq30/f8i92ncVhG5m0P8ZupVBp6voww+9sc+ePfuRd5NDRIGHzoqRgTCvze6EXLAyi/4QqBis4ubcHxv6PZs+fbr2noH+8ZyhZ2fsX3fjcgECGfqVa9iwob5GX3nffPONbhN9pBFRaEoST7UMG1iQR9B7M3pmx9AkGODQHXRJhB6kcaf00ksvacey9tDzOoayQAeoyK2hE1QMtYFONomI3Am5nBV5B+NCITjdvXtXB7RDL/AlS5Z0uSxGgkXOq0yZMtpjOIaTQHBCwDLLolH8h97V8RqD0qFCFb1l16hRgz8NEbnFYEVxQgDC8AIIPujMFUNkrF692mXAQlCzz3UhUKFuC+M4mWPqIFhhJFfkrpADQ/1V165dtSWQpwM6ElHoCbkeLFCZh7FjcPFlnZX3EFDQ4MJ+8Me4YIwe5KC+++47HXQR+x65M7POCjCqK8Z0whhIRBRabnh4TWadFXn9nJqnI7GingvFiDlz5tTXGEUVk/OzFGhxiO0SEbnDYkByCyMXY2RYNDHHQ3tour5q1Srb8OStW7fW0WIxwigMHjxYKleurCO8YsjsTz/9VKKiojTnBLhrwuilvXv31paDKAZEkSJaA44ZM4a/BBG5xWBFbmFIbQQkDI2ObDoaTiBQ1a1bV9/HsN32uaSrV6/qkO3nz5+XTJky6RDeGzZscKjfQotCBMGWLVvKlStXNGANHTpU3nzzTf4SROQW66yIiChgWGdFRERBgw0siIjI8hisiIjI8hisiIjI8tga0AdTpkzx/y9BIQVdThGR55izIiIiywtosJo4caI+u2MO14F+5TCGkjszZ87UbuTtp7CwsEeaZiIiCrFiQPS8PWLECB3/CF0Ufv311zqsxI4dO6RUqVIu10FQO3jwoO01h6YnIgp+AQ1WjRo1cniNngyQ20Kv3O6CFYJTjhw5HlEKiYjICizTwAKdns6fP9/jwf3Q8SlGrh02bJjbwAbodNW+41U8LQ1Yn52nUqDw2CPy7lxIHmyD+zlDJ6uRkZGx5l+6dEk/0xcpU6b0aT0i+34XiUi0k+x/Rd+A9+/f1w5RzcH9pk2b5nZwP2cYbgKD+zVv3tw2uJ8nOSsMuY5OV30dzwppJEoIsyd6olB348YN7fg6vvGsAp6zQi4FQ0oAeunesmWLjB8/3qPB/VKkSCHly5eXI0eOuF0mVapUOjlDb+HO4yoRPSo89oi8OxeSBtPgfkREFJySB9PgfkREFJySB9vgfkREFHwC3sDCqgN9xYV9A1JCsW9Aov/HwReJiChoWK6BBRERkTMGKyIisjwGKyIisjwGKyIisjwGKyIisjwGKyIisjwGKyIisjwGKyIisjwGKyIisjwGKyIisjwGKyIisjwGKyIisjwGKyIisjwGKyIisjwGKyIisjwGKyIisjwGKyIisjwGKyIisjwGKyIisjwGKyIisjwGKyIisjwGKyIisrzk3q5w7949+fPPPyUqKkpu374tWbNmlfLly0uBAgUSJ4VERBTyPA5W69evl/Hjx8svv/wiDx48kAwZMkh4eLhcuXJFA1jBggWlc+fO8uabb0q6dOlCfscSEdEjLgZ88cUXpVmzZpI/f35ZtmyZREdHy+XLl+X06dOauzp8+LB88MEHsmLFCilatKgsX77cj0kkIqJQ51HOqmHDhvLjjz9KihQpXL6PXBWmNm3ayL59++TcuXP+TicREYUwj4JVly5dPN5gyZIldSIiIgpoa8Br167JtGnTpF+/flpnBdu3b5czZ874LWFEREQ+twbctWuX1KlTRxtYnDhxQjp16iSZM2eWn376SU6ePCnffPONt5skIiLyb86qV69e0rZtW21UERYWZpvfoEEDWbNmjVfbmjhxopQpU0bSp0+vU5UqVeT333+Pc5358+dL8eLF9bNLly4tv/32m7dfgYiIgj1YbdmyxWUdVu7cueX8+fNebStPnjwyYsQI2bZtm2zdulWeeeYZeemll2Tv3r0ul9+wYYM0b95cOnToIDt27JDGjRvrtGfPHm+/BhERBXOwSpUqldy4cSPW/EOHDukDwt5o1KiR5siKFCmiTd6HDh0qadOmlU2bNrlcHs95Pf/889K7d28pUaKEDBkyRCpUqCBffPGFt1+DiIiCuc4Kz1wNHjxY5s2bp6+TJEmidVV9+/aVV155xeeExMTEaBHfrVu3tDjQlY0bN2oxpL169erJwoUL3W4XDyxjMpmB9uHDhzoRBQKPPSLvzgWvg9Xo0aPl1VdflWzZssmdO3ekZs2aWvyHAIOckbd2796t6969e1dzVQsWLHDb9B2fkz17dod5eB1X8ePw4cMlMjIy1vxLly7pZ/oiZcqUPq1HZLp48SJ3BpGIdjKRKMEKrQDRQ8W6deu0ZeDNmze1KA4tBH1RrFgx2blzp1y/fl1++OEHfbB49erVfntWC83r7XNjyFnlzZtXiyzRqMMX9+/f90vaKHThZo+IxKGhnl+Dlal69eo6JRRyKYULF9a/K1asqA04UDc1efLkWMvmyJFDLly44DAPrzE/rjo2TM6SJk2qE1Eg8Ngj8u5c8DpYffbZZ3G+361bN0lo+aV9HZM9FBei/8EePXrY5iGX566Oi4iIgoPXwWrs2LEu5xuGIadOnfIqWKGIrn79+pIvXz4tt5wzZ46sWrVKli5dqu+3bt1am8Sj3gm6d++udWSoN0N/hXPnztUm71OmTPH2axARUTAHq+PHj7ucj8YKqVOn9rqSGQEJHd+iLgwPCCNQ1a1bV99HK0P7LGLVqlU1oKGH9/79+2uTd7QEfPzxx739GkRE9C/ic52VMzRhx+SNr776Ks73kcty9tprr+lEREShgy0MiIgo+HJWKJbzNgdFRET0SIMVHtp19+zR66+/nqDEEBER+SVYoaNZV9w1NyciIkoo1lkREVHw5aycO5K174iWiIjIEsEK40i5U6NGjYSmh4iIKOHBauXKld6uQkRElCCssyIiouDLWTVp0iTO93/66aeEpIeIiCjhOSv04Wc/LV68WB8UNl8TEREFPGc1Y8YMh9cYMPGTTz6RggUL+jNdRERE/quzwvhT7H6JiIgslbPCsPBw584dHa4Dz1flypUrMdJGRETkW7DKmDGjLSeFIenHjBnjcth4IiKigD9nFR4eroMfZsqUyW+JISIi8kuwyps3LxtTEBGRtRtYFC5cWGrXri2zZ8/WoeyJiIgsF6y2b98uZcqU0Q5tc+TIIV26dJHNmzcnTuqIiIh8CVblypWT8ePHy9mzZ2X69Oly7tw5qV69ujz++OPa2OLSpUvcsUREZI3nrJInT65dL82fP19GjhwpR44ckffee0/rtFq3bq1BjIiIKKDBauvWrfLWW29Jzpw5NUeFQHX06FFZvny55rrcjShMRESU6K0BEZjQ5dLBgwelQYMG8s033+j/6B8QChQoIDNnzpT8+fN7nRgiIiK/BKuJEydK+/btpW3btpqrciVbtmzy1VdfebtpIiIi/wSrw4cPx7sMerZo06aNt5smIiLyT7BatGhRnO+/+OKL3m6SiIjIv8GqcePGbt9Dn4Ho2JaIiCjgrQHRLB1DgzhPDFRERGSZYGW2/CMiIrJkMSBMnTpVe1tPkyaNjmWFXi2yZMni/9QRERH5krPKly+fTJkyRT7++GPp2rWrPPfcc9pHIB4CNgdm9NTw4cPliSeekHTp0mlzd9SH4fmtuOAZLtSN2U9hYWH8MYmIgpjXwerEiRM6nTlzRqKjo+X69euydOlSOXTokPTp08erba1evVoD3qZNm7TniwcPHmjwu3XrVpzrpU+fXuvNzCkqKsrbr0FERMFeDGgPuaJnn31WvvzyS+nQoYNX6y5ZsiRWrgk5rG3btkmNGjXcrofcFHJzREQUGhIcrEwY4+rYsWMJ2gZyaZA5c+Y4l7t586ZERERoC8QKFSrIsGHDpFSpUi6XvXfvnk4ms6jSbMFIFAg89oi8Oxf8Fqz8keAePXpItWrVdLgRd4oVK6ZDk2BMLQS3UaNGSdWqVWXv3r2SJ08el/VikZGRseZjKBNfB49EDx1ECXHx4kXuQCIRrU7yRBLDMAwr7LH//Oc/8vvvv8u6detcBh13UM9VokQJad68uQwZMsSjnBWGMbl69arWffli2rRpPq1HZOrYsSN3BpH8/zUZrcuR+YjrmmyJnNXbb78tv/76q6xZs8arQAUpUqSQ8uXL63harqRKlUonV8+K8XkxChQee0TenQsBfboXmToEqgULFsgff/yhw4t4C71m7N69220P8ERE9O+XoGA1YsQIuXbtms/ro9n67NmzZc6cOdqq8Pz58zrduXPHtgxGHe7Xr5/t9eDBg2XZsmXamGP79u3SqlUrbbrOYhUiouCVoGCFVnhXrlzxeX2MjYVyylq1amnOyJy+//572zInT57UZ6lMqGvq1KmT1lNh0EeUd27YsEFKliyZkK9CREQWlqA6q4S2zfBk/VWrVjm8Hjt2rE5ERBQ62CMtERFZHoMVERFZHoMVERFZHoMVERFZHoMVERFZHoMVERFZHoMVEYUEXwZ7RQfZr7zyiuTPn1+HJho3blysZQYNGhRrQNjixYsn4jcJTQxWRBQSfBns9fbt21KwYEHtrSeuMfQwRJH9gLDokJss9FAw7iCIiP4NfBnsFTkxTPD++++73Xby5Mk5IKyVc1YWGV2EiCjRBnv1xOHDhyVXrlyaC2vZsqV2E0cWyll5OmgWEZGVeDrYqyeeeuopzaVhYFgUAWKw16efflr27Nmj9WPkH5YYz4qI6FFC3RWCiT/qlurXr2/7GyOYI3hFRETIvHnzpEOHDgnePv0/BisiCikJGezVExkzZpSiRYu6HRCWfMPWgEQUEvwx2Ksnbt68KUePHuWAsH7GnBURhUzRHwZ6/fnnn22DvUKGDBkkPDzcNthr7ty59ZksuH//vuzbt8/295kzZ2Tnzp2SNm1aKVy4sM5/7733pFGjRlr0d/bsWRk4cKAkS5ZMmjdvHrDvGox8DlbI4uLuAU0+8UPjroVN2YnIqjDYK2CwV3szZsyQtm3b6t9oxZc06f8KnBB8ypcvb3s9atQonWrWrGkba+/06dMamC5fvixZs2aV6tWr67Nc+JsCGKzwgzRr1kyz0QhOaLKJ5pqoSMyUKZOMHj3aj8kjIvIPXwZ7Rc8V8a03d+7cBKeNEqHOqmfPnvoAHO5AUqdObZuPAOb80B0REVFAclbLli2TpUuXxmpFU6RIEYmKivJLooiIiBKUs0I/WvY5KtOVK1ckVapU3m6OiIjI/8EKT2Z/8803tteot8LT4J988onUrl3b280RERH5vxgQQenZZ5+VrVu3alPOPn36aDf6yFmtX7/e280RERH5P1ihH61Dhw7JF198oc8q4AG4Jk2a6DMMOXPm9HZzRGQB1T+Ke1wnorisG1xMLPmcFR6iGzBggP9TQ0RE5I86q127drmc//fff2vX+ERERAEPVnj6Gw8EOz8Bju7xUYdFREQU8GLAqVOnyssvvywTJkzQrvA7d+6s3S5Nnz5dXnrpJb8nkIiIyOtg9corr+jwzY0bN9bGFZ06dZJffvlFO3YkIiKyzBAhGF0Tg5ah9d8///zDQEVERNbKWaEHYrN3dTwMPHnyZFm+fLlt+Obt27f7P5VERBTSvM5ZofgPdVOY8ubNq93pN2zY0DbPGxgz5oknntBAly1bNt32wYPxP+8xf/58KV68uISFhUnp0qXlt99+8/ZrEBFRMOesMLAYTJs2TTZv3qzd46MeyxerV6/Wh4kRsFCc2L9/f3nuued0sLM0adK4XGfDhg06dgwC3QsvvKCDqSHIIUeHB5aJiCj4JDE8GeTFRcCaNGmS9ryOjm2nTJmigzAm1KVLlzSHhSDmbnsYigSf+euvv9rmVa5cWcqVK6dpis+NGzf0oebr169L+vTpfUonvi9RQqAVrZWwBwsKVA8Wnl6Tvc5ZtWvXTvsARA6nQIECMm7cOM3hNG3aVEfQzJgxo8+JRmIhc+bMbpfZuHGj9OrVy2FevXr1ZOHChS6Xv3fvnk72O8asb8NEFAhWO/aSiNf3rER+OZ49XdfrYLV//34NVFmyZNHXCBwoBnzrrbe0Hun8+fM+J7hHjx7a0jCu4jxsP3v27A7z8Nrd56K4MDIy0mUu7u7duz6lNWXKlD6tR2S6ePGipXZGwQw3A50ECtHjOTo6OnGC1cqVKyU8PNxhXkREhCxevFjmzZsnvkLd1Z49e7RJvD/169fPISeGnBUahmTNmtXnYkD21EEJheJuKzl2/Vqgk0AhejyHhYUlTrByDlT2UBToi7ffflvroNasWRNrBGJneCD5woULDvPwGvNdwYCQrgaFRCtGTESBYLVjz5D/fxyF6FEfz56u63Wwcq4vcjZmzBiPt4W2He+8844sWLBAVq1apXVg8alSpYqsWLFCiwxNeM4L84mIKDh5HazQoALPRVWsWFGDjT3zYWFviv7Q9Pznn3/WbZr1TmgZYubgWrduLblz59a6J+jevbvUrFlTRo8erc93oek8BoJkCz0iouDlU0e2H330kSRPnlwDBh7K9dXEiRNtPbk79+Letm1b/fvkyZMO2cSqVatqgPvggw/0uawiRYpoS0A+Y0VEFLy8DlYdOnTQh3JHjhypLffw3NPHH38cq4WeJzx5xAvFg85ee+01nYiIKDT4VCuWOnVqbQ6OrpFiYmKkaNGiMnjwYLlz547/U0hERCHP65zVokWLHF6jqyM0Xf/000+13uj06dMhv1OJiCjAwQrByR10g0RERBTwYGW1bmKIiCj4WevJRCIiIn/krD777LM43+/WrZu3myQiIvJvsBo7dqzt71OnTunQ9njmynwomMGKiIgCHqyOHz9u+xu9TmDsqYIFC/o7XURERDassyIiIstjsCIiouArBty1a5dDd0kHDhyQmzf/N3BbmTJl/Jc6IiIiX4JVuXLltCGF2a8fhrQ3X+N/dL9ERERkmQYWRERElgxW6AeQiIjIcg0sNm3a5PEGb9++LXv37k1ImoiIiLwPVm+88YbUq1dP5s+f77az2n379ulgiIUKFZJt27Z5slkiIiL/FQMiEGFUX4zO26JFCx2/KleuXBIWFiZXr161tQh8+eWXZdmyZQkaPZiIiMinYJUiRQrtRgnT1q1bZd26dRIVFaWDLZYtW1Z69uwptWvXlsyZM3uyOSIiosRtYFGpUiWdiIiIHhX2YEFERJbHYEVERJbHYEVERJbHYEVERJbHYEVERMHXGrBXr15xvj9mzJiEpIeIiCjhwWrcuHFSpUoVSZkypb7GM1cVK1aU8PBw7XWdiIgo4MEKFixYINmyZbMNbT9nzhwObU9ERNaps0JvFvfv37e9fvDggfz444/+ThcREZHvwapAgQIyd+5c/RtBCsFr6tSp0rx5c+1xnYiIKODBqm/fvvL+++9rJ7ZNmzbVv9FfIALVk08+6fcEEhEReR2s2rVrp+NVzZo1S/78808ZMGCApE+fXn7++Wdp1aqVV9tas2aNNGrUSHtwR+OMhQsXxrn8qlWrdDnn6fz58/wliYiCmE8NLIoVK6aTM+SyvIGxsdBre/v27aVJkyYer3fw4EENkCazsQcREQUnr4PVjRs34nzfPojEp379+jp5C8EpY8aMXq9HREQhEqwyZcrkcr5hGFokFxMTI4mtXLlycu/ePXn88cdl0KBBUq1aNbfLYjlMzsH24cOHOhEFgtWOvSRiBDoJFKLH80MP103uS2vAixcvapFfXEEiMeTMmVMmTZqk42khAE2bNk1q1aqldWcVKlRwuc7w4cMlMjIy1vxLly7J3bt3fUqH+UA0ka9wDllJwQw3A50ECtHjOTo6OnGC1f79++Xzzz+XoUOHyo4dO+STTz7RAPYoONeVVa1aVY4ePSpjx47VBh+u9OvXz6GLKOSs8ubNK1mzZvWqyNKe/XNmRL6wWj3rsevXAp0ECtHjOSwsLPEeCsbF//Dhw5I7d24pU6aMvPvuu3LtWmAOdjSXP3LkiNv3U6VKpUHJfoKkSZP6PBElVEKOv8SYDC0I5MR9kMSnfZDQ48+jc8bXky1z5szaTyByVydOnJDChQvr60dt586dWjxIRETBy+tiwPLly8fqsBaNK1CHhBxWjx49PN7WzZs3HXJFx48f1+CDQJgvXz4twjtz5ox88803+j6CIYocS5UqpfVNqLP6448/ZNmyZd5+DSIiCuZg1bhxY799OHq+qF27tu21WbfUpk0bmTlzppw7d05OnjzpUFeEgIgAljp1ai2C/O9//+uwDSIiCj5JDGSLQggaWGTIkEGuX7/ucwOLKVOm+D1dFFo6d+4sVlL9o4OBTgL9i60bHLuTCH9fk72us9qyZYs2FXeGecgpERER+ZvXwapr165y6tSpWPNRNIf3iIiIAh6s9u3b5/IBXDS8wHtEREQBD1Z4bunChQux5qMxRPLkPvWLS0RE5N9g9dxzz2mTclSGmfBAcP/+/aVu3brebo6IiCheXmeFRo0aJTVq1JCIiAgt+gM8G5U9e3a3XR4RERE90mCFLpZ27dol3377rfz1118SHh6uAzJiWHt0xURERORvPlUypUmTxnLPiRARUfDyus5q6dKlLuej9/OaNWv6I01EREQJC1avvvqq/PDDDw7zxo8fr8PTFy1a1NvNERER+b8YcN68edKsWTNtDYicFOqr0H8fAtjzzz/v7eaIiIj8H6zq168vixcvlhdffFF7Wm/ZsqW+9rWfPSIiokQZz+rpp5/WoTnSpk2rI0QyUBERkaVyVk2aNLH9nStXLhkxYoRs2LBBMmXKpPN++ukn/6aQiIhCntfBCl25m/BQsPlgMBERkWWC1YwZMxInJURERP6ss/rnn390hN7JkydLdHS0zjt79qwOU09ERBTwnFVUVJQ2UUdzdbQGROe16dKlk5EjR+rrSZMm+T2RREQU2rzOWXXv3l0qVaokV69e1X4BTS+//LKsWLHC3+kjIiLyPme1du1abf2XMmVKh/n58+fX0YKJiIgCnrN6+PChxMTExJp/+vRpLQ4kIiKyxOCL48aNs71OkiSJNqwYOHCgNGjQwN/pIyIi8r4YcPTo0VKvXj0pWbKk3L17V1q0aCGHDx+WLFmyyHfffcddSkREgQ9WefLk0UEX586dq4MwIlfVoUMH7SPQvsEFERFRQAdfTJ48ubRq1cpviSAiIvJrsPrss8/ifL9bt27ebpKIiMi/wapHjx5aFJgsWbJY76GxBYMVERFZohhw69atOjQIERGRJZuuI/eEiYiIyLI5K8Mw5MMPP9ShQtKkSaNjWmGYkIoVKyZOComIKOR5nbOqUaOGHDhwQNatWyfz58+Xvn37yhNPPKEBCz2ve2PNmjXSqFEjDXjIrS1cuDDedVatWiUVKlSQVKlSSeHChWXmzJkh/yMSEQU7r4MVggWm9evXy+7du+Xy5cv6UHDq1Knl3Xff9Wpbt27dkrJly8qECRM8Wv748ePSsGFDqV27tuzcuVMbe3Ts2FGWLl3q7dcgIqJgb2DhrFChQjJ+/Hjp1auXV+vVr19fJ09h+JECBQpoLxpQokQJzeGNHTtWe9UgIqLg5JdgBRg2BMV6iWnjxo1Sp04dh3kIUshhuYMxtjCZbty4YeuQFxNRIFjt2EsiRqCTQCF6PD/0cF2vg9WiRYvcvod6J9RBJZbz589L9uzZHebhNQLQnTt3XHb3NHz4cImMjIw1/9KlS9q3oS+ch0ch8tbFixcttdMKZuAo3xSY49kcbd7vwapx48ZxBitXw4cEUr9+/RyKJxHY8ubNK1mzZpX06dP7tM379+/7MYUUiqz2nOKx69cCnQQK0eM5LCws8YoBkcMJxMmWI0cOuXDhgsM8vEbQcdeJLloNYnKWNGlSnYgCwWrHniF8dpICczx7uq61zph4VKlSRVasWOEwb/ny5TqfiIiCV0CDFYYXQRN0TGbTdPx98uRJWxFe69atbcu/+eabcuzYMenTp48+6/Xll1/KvHnzpGfPngH7DkRElPi8LgZEvRQqxNyVM3pTD4Q+BvHMlMmsW2rTpo0+7Hvu3Dlb4AI0W1+8eLEGJzSVR4e606ZNY7N1IqIg51N3S0WLFnU539sGFrVq1dL13HHVOwXW2bFjhxcpJiKikAtWK1euTJyUEBER+StY1axZ09tViIiIEsSnputXr16Vr776Svbv36+vS5YsKe3atZPMmTMnLDVERET+aA2ILpXy58+vw9sjaGHC32j8kNjdLRERUWjyOmfVtWtXadasmUycONE2tD0aVbz11lv6HnpiJyIiCmjO6siRIzoUiBmoAH+j2TneIyIiCniwwsCHZl2VPczD2FREREQBLwbs1q2bdO/eXXNRlStX1nmbNm3SARRHjBghu3btsi1bpkwZ/6aWiIhCktfBqnnz5vo/ujxy9R4eDPblAWEiIiK/BSv030dERGTpYBUREZE4KSEiIvJnr+uzZs2SatWqSa5cuSQqKkrnjRs3Tn7++WdfNkdEROTfYIXnq9BMvUGDBnLt2jVbvVTGjBk1YBEREQU8WH3++ecydepUGTBggMOzVpUqVeIDwUREZI1ghQYW5cuXjzUfQ8ffunXLX+kiIiLyPVihD0BzZF97S5YskRIlSni7OSIiIv+3BkR9FfoAvHv3rj5PtXnzZvnuu+9k+PDhOmovERFRwINVx44dJTw8XD744AO5ffu2tGjRQlsFYpj5119/3e8JJCIi8mk8q5YtW+qEYHXz5k3Jli0b9yQREVkrWJlSp06tExERkaWCFXpdj8v27dsTkh4iIqKEByu0BMR4VmnTpvV2VSIiokdXDNi7d2/WUxERkbX7BiQiIrJ8sMJYVURERJYuBvzwww/dtgIcM2ZMQtNERESUsGBVo0YNOXjwoMv3mOMiIiJLBKtVq1YlSkKIiIjcYQMLIiKyPAYrIiKyPEsEqwkTJkj+/PklLCxMnnrqKe3J3Z2ZM2dq3Zj9hPWIiCh4BTxYff/99zrsyMCBA7WrprJly0q9evXk4sWLbtdJnz69nDt3zjZFRUU90jQTEVGIBSs0de/UqZO0a9dOSpYsKZMmTdJm8dOnT3e7DnJTOXLksE3Zs2d/pGkmIqJ/Ua/rCXX//n3Ztm2b9OvXzzYvadKkUqdOHdm4caPb9TAsSUREhDx8+FA71h02bJiUKlXK5bL37t3TyXTjxg39H+tiIgoEqx17ScQIdBIoRI/nhx6uG9Bg9ffff0tMTEysnBFeHzhwwOU6xYoV01xXmTJl5Pr16zJq1CipWrWq7N27V/LkyRNreYxgHBkZGWv+pUuXdLRjX6RMmdKn9YhMcRVzB0LBDDcDnQQK0eM5Ojra+sHKF1WqVNHJhEBVokQJmTx5sgwZMiTW8si1oU7MPmeVN29eyZo1q9Z9+ZojJEoIqw1Yeuz6tUAngUL0eA7zsIFcQINVlixZJFmyZHLhwgWH+XiNuihPpEiRQsqXLy9Hjhxx+X6qVKl0cobiRkxEgWC1Y88Q9vdJgTmePV03oGcMitMqVqwoK1ascCi/xGv73FNcUIy4e/duyZkzZyKmlIiIAingxYAoomvTpo1UqlRJnnzySRk3bpzcunVLWwdC69atJXfu3Fr3BIMHD5bKlStL4cKF5dq1a/Lpp59q0/WOHTsG+JsQEVHQBqtmzZppY4ePPvpIzp8/L+XKlZMlS5bYGl2cPHnSIZt49epVbeqOZTNlyqQ5sw0bNmizdyIiCk5JDMMIqTaraGCRIUMGbUnoawOLKVOm+D1dFFo6d+4sVlL9I9cjKRB5Yt3gYpLY12Rr1fISERG5wGBFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWx2BFRESWZ4lgNWHCBMmfP7+EhYXJU089JZs3b45z+fnz50vx4sV1+dKlS8tvv/32yNJKREQhGKy+//576dWrlwwcOFC2b98uZcuWlXr16snFixddLr9hwwZp3ry5dOjQQXbs2CGNGzfWac+ePY887UREFCLBasyYMdKpUydp166dlCxZUiZNmiSpU6eW6dOnu1x+/Pjx8vzzz0vv3r2lRIkSMmTIEKlQoYJ88cUXjzztRET0aCSXALp//75s27ZN+vXrZ5uXNGlSqVOnjmzcuNHlOpiPnJg95MQWLlzocvl79+7pZLp+/br+f+3aNXn48KFP6b5z545P6xGZcPxZScy96EAngUL0eL5x44b+bxiGdYPV33//LTExMZI9e3aH+Xh94MABl+ucP3/e5fKY78rw4cMlMjIy1vyIiIgEpZ0oIXr06MEdSEEj0ycJ30Z0dLRkyJDBmsHqUUCuzT4nhtzUlStX5LHHHpMkSZIENG3BCndKefPmlVOnTkn69OkDnRyiBOHxnLiQo0KgypUrV5zLBTRYZcmSRZIlSyYXLlxwmI/XOXLkcLkO5nuzfKpUqXSylzFjxgSnneKHQMVgRcGCx3PiiStHZYkGFilTppSKFSvKihUrHHI+eF2lShWX62C+/fKwfPlyt8sTEdG/X8CLAVFE16ZNG6lUqZI8+eSTMm7cOLl165a2DoTWrVtL7ty5te4JunfvLjVr1pTRo0dLw4YNZe7cubJ161aZMmVKgL8JERElloAHq2bNmsmlS5fko48+0kYS5cqVkyVLltgaUZw8eVJbCJqqVq0qc+bMkQ8++ED69+8vRYoU0ZaAjz/+eAC/BdlDsSuem3MufiX6N+LxbA1JjPjaCxIREYX6Q8FERETxYbAiIiLLY7AiIiLLY7CiRwoPYrvrGsuf0Is/WpZScEDVeufOnSVz5sx6DOFZyWDvBeTEiRP6XXfu3BnopFhCwFsDkrXhoo+Lgr8uDOfOnZNMmTL5ZVsUOtBCeObMmbJq1SopWLCgthAODw/362e0bdtW+7h7FDdT5D0GqxCFToTxULY/oH9H3AHaP2LgjrueRojicvToUcmZM6c+ukKhicWAQaJWrVry9ttv64SuS9CV1YcffmjryRg5JAyngoes0W0MilRg3bp18vTTT+tdKvrz69atmz6UbW4zKipKevbsqcHI7EsRd7gohlm0aJEO64LnUPA83JYtW6Ru3br62UgDHt7GGGXuigHNYo6ffvpJateurUPDYDwz5x7340ojYOyzRo0a6fsFChSQb7/9NpH3Nj1KyPG88847eozheMGxjGPTPrePecOGDZP27dtLunTpJF++fLE6CkBflU2bNtVjF8WJL730kh6DMGjQIPn666/l559/th3ryMVhwt/2vYqjWA7zzHXN82Hp0qU6bFHatGl1GCOUItibNm2avo9BYzF47JdffunwPgadLV++vL6PThIwXh/ZwXNW9O9Xs2ZNI23atEb37t2NAwcOGLNnzzZSp05tTJkyRd+PiIgw0qdPb4waNco4cuSIbUqTJo0xduxY49ChQ8b69euN8uXLG23bttV1Ll++bOTJk8cYPHiwce7cOZ1gxowZRooUKYyqVavqOvi8W7duGStWrDBmzZpl7N+/39i3b5/RoUMHI3v27MaNGzds6cQht2DBAv37+PHj+rp48eLGr7/+ahw8eNB49dVXNa0PHjzQZeJLI9SvX98oW7assXHjRmPr1q2arvDwcF2H/v2uXbumxyCORRyDFy9e1OMdx7oJx0zmzJmNCRMmGIcPHzaGDx9uJE2aVI9NuH//vlGiRAmjffv2xq5du/T4bNGihVGsWDHj3r17RnR0tNG0aVPj+eeftx3rmL9y5Uo9Rq9evWr7rB07dug8HL/250OdOnWMLVu2GNu2bdPPwvZNOB9z5sxp/Pjjj8axY8f0f6R35syZ+j4+P2vWrLrOnj17jF9++cUoWLCgfg4+jwyDwSpI4OTFCfLw4UPbvL59++o882Ru3LixwzoIJp07d3aYt3btWj3J79y5Y1vP+aKPkxMn0c6dO+NMU0xMjJEuXTo98eIKVtOmTbO9v3fvXp2HgOdJGhHgsPzmzZtt72NdzGOwCh74LXEsmlwFq1atWtle4zzIli2bMXHiRH2NmygEJvvzA8EINzVLly7V123atDFeeuklh8/1NFjhNW6sTAiauFEzFSpUyJgzZ47DtocMGWJUqVJF/548ebLx2GOP2c47QNoZrP6HdVZBpHLlyg7DnqBzX/ShiDolQNGCvb/++kt27drlUGyGeILOhI8fP65FFu6gvqtMmTKxer9HN1goOkHRHD739u3bWnwTF/vtoF4CsD6KSuJL46FDhyR58uTaIbIJ67Fn/dBjfxzhPED9KI4jwHF05MgRLSK0d/fuXa0PSygUYRcqVMjhODY/G0XW+IwOHTroqOimf/75x9bb+P79+zX9KAI0sXNuRwxWISRNmjQOr2/evCldunTROiBnKPOPC+qHnMcDQ4fEly9flvHjx+vglqjLwgmHxhxxSZEihe1vc5vmKM7xpRHBisj5ODKPJfvjCDc0ruozs2bN6nYHmo2G7Hule/DggUefba6Dz4apU6fKU0895bAchkgizzBYBZE///zT4fWmTZu0o193J0SFChVk3759Urhw4ThzUGbOLD7r16/XSuMGDRrYKrQxGnRCxJdG5KJwh7pt2zZ54okndN7BgwctN2w8BRaOo++//16yZcvmdow1V8e6GcjsH7nw9rkndMqNgQWPHTsmLVu2dLkMSjFmzZqlOT0zd4Xzl/6HrQGDCIrbMOQKLtbfffedfP755zqkijt9+/aVDRs2aAtCnICHDx/W1lB4bd/Kas2aNXLmzJl4Aw8CI044FGkgcOLETOizMPGlsVixYtryCrkvfCaCVseOHf3+DA79u+FYRCtVtABcu3atFiGjuBo59tOnT9uOdRQ54/zBsY4cFG6S0AIVrQVx7C1evFiL1r0VGRmpwxx99tlnWhqwe/dumTFjhowZM0bfb9GihebGUEyIm7PffvtNRo0a5ff98G/GYBVE0Cz9zp07Oi5Y165dNVCZTdRdQRn56tWr9eRB03A0m8VQLfbDSw8ePFib6KI8Pq7iEvjqq6/k6tWrehf7xhtv6IUAd7IJ4UkacdLjNZrKN2nSRL9zQj+XggvqlHDThaJjHCPIyaAOCTkZM6eFQIGbH9Tt4lhHSQGK93Djd+DAAT0WR44cKR9//LHXn48bKDRdx7FaunRpPVbR5B2PWgCau//yyy8axHCMDxgwQD+L/odDhAQJPHeCscDYxRARBSPmrIiIyPIYrIiIyPJYDEhERJbHnBUREVkegxUREVkegxUREVkegxUREVkegxUREVkegxUREVkegxUREVkegxUREVkegxUREYnV/R8nh4hoRh8xNAAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 460x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "loader = torch.utils.data.DataLoader(\n",
    "    ChordDataset(DATA / \"processed\" / \"user\" / \"val\",\n",
    "                 max_length=pretrained.max_seq_len),\n",
    "    batch_size=8, shuffle=False, num_workers=0,\n",
    ")\n",
    "\n",
    "ppl_pre = compute_perplexity(pretrained, loader, DEVICE)\n",
    "ppl_ft  = compute_perplexity(finetuned,  loader, DEVICE)\n",
    "improve = (ppl_pre - ppl_ft) / ppl_pre * 100\n",
    "\n",
    "print(f\"Перплексия на валидации автора ({len(user_val)} периодов):\")\n",
    "print(f\"  pretrained = {ppl_pre:.2f}\")\n",
    "print(f\"  finetuned  = {ppl_ft:.2f}\")\n",
    "print(f\"  улучшение  = {improve:+.1f}%\")\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(4.6, 4))\n",
    "bars = ax.bar([\"pretrained\", \"finetuned\"], [ppl_pre, ppl_ft],\n",
    "              color=[\"#999999\", \"#3b7dd8\"])\n",
    "for rect, v in zip(bars, [ppl_pre, ppl_ft]):\n",
    "    ax.text(rect.get_x() + rect.get_width() / 2, v + 0.03,\n",
    "            f\"{v:.2f}\", ha=\"center\", va=\"bottom\")\n",
    "ax.set_ylabel(\"перплексия (ниже — лучше)\")\n",
    "ax.set_title(\"Перплексия: pretrained vs finetuned\")\n",
    "ax.grid(axis=\"y\", alpha=0.3)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "13b4009c",
   "metadata": {},
   "source": [
    "### 5.2 Распределения\n",
    "\n",
    "Картинку строит скрипт `scripts/evaluate.py`: он сравнивает четыре группы —\n",
    "McGill (предобучение), мой корпус (цель), выход модели до дообучения и после.\n",
    "Панели:\n",
    "\n",
    "- **Chord quality** — распределение качеств аккордов;\n",
    "- **Root motion intervals** — интервал между тониками соседних аккордов (0–11\n",
    "  полутонов);\n",
    "- **Extension presence** — доля аккордов с надстройками;\n",
    "- **Inversion frequency** — доля обращений (бас не тоника).\n",
    "\n",
    "Ниже — готовая картинка `output/eval/distributions.png`. Команда для пересборки:\n",
    "\n",
    "```\n",
    "python scripts/evaluate.py --pretrained checkpoints/pretrained.pt \\\n",
    "    --finetuned checkpoints/finetuned.pt\n",
    "```\n",
    "\n",
    "Ожидаемый результат: группа **finetuned ближе к «моему корпусу»**, чем\n",
    "pretrained — прежде всего по обращениям и надстройкам, которые я использую\n",
    "часто."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "f321afff",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-04T13:54:00.931012Z",
     "iopub.status.busy": "2026-06-04T13:54:00.930500Z",
     "iopub.status.idle": "2026-06-04T13:54:00.965163Z",
     "shell.execute_reply": "2026-06-04T13:54:00.964645Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dist_png = OUT / \"distributions.png\"\n",
    "if dist_png.exists():\n",
    "    display(Image(str(dist_png)))\n",
    "else:\n",
    "    print(\"Файл не найден. Сгенерируйте его командой:\")\n",
    "    print(\"  python scripts/evaluate.py --pretrained checkpoints/pretrained.pt \"\n",
    "          \"--finetuned checkpoints/finetuned.pt\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "548e4218",
   "metadata": {},
   "source": [
    "**Как читать график.** Цвета: синий — McGill (корпус предобучения), оранжевый —\n",
    "мой корпус (цель), зелёный — выход pretrained, красный — выход finetuned. Чем\n",
    "ближе красный к оранжевому, тем лучше дообучение поймало мой стиль.\n",
    "\n",
    "Что видно по факту:\n",
    "\n",
    "- **Движение тоник (панель 2)** — главный эффект. До дообучения pretrained\n",
    "  (зелёный) почти в половине случаев топчется на той же тонике (`P1` ≈ 43 %);\n",
    "  после дообучения finetuned (красный) спускается до ≈ 18 %, почти как мой корпус\n",
    "  (≈ 14 %), и набирает шаги на тон (`M2`). Гармония стала заметно подвижнее.\n",
    "- **Качества аккордов (панель 1)** — finetuned срезал доминантсептаккорды (`7`)\n",
    "  почти до моего уровня, но перекосил в сторону «голого» `maj`.\n",
    "- **Обращения и надстройки (панели 3–4)** — здесь моё ожидание не подтвердилось:\n",
    "  finetuned их, наоборот, недодаёт (обращений ≈ 3 % против моих ≈ 9 %). Скорее\n",
    "  всего, корпуса слишком мало, чтобы модель уверенно переняла мою привычку к\n",
    "  slash-аккордам."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eb683f6a",
   "metadata": {},
   "source": [
    "### 5.3 Качественные примеры\n",
    "\n",
    "Беру одно и то же условие (тональность, размер, стиль, функция, seed) и\n",
    "генерирую двумя моделями. Эталон — мой собственный период, на метаданные\n",
    "которого я опираюсь."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "a7259286",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-04T13:54:00.966713Z",
     "iopub.status.busy": "2026-06-04T13:54:00.966197Z",
     "iopub.status.idle": "2026-06-04T13:54:00.972016Z",
     "shell.execute_reply": "2026-06-04T13:54:00.971503Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Эталон (автор):\n",
      "   | Eb . Bb . | Ab . . Bb | Eb . Eb/G . | Fm . Bb . | Eb . . . | Ab . Bb . | Fm . . . | Bb . . . |\n",
      "\n",
      "pretrained:\n",
      "   | D#maj . D#maj . | G#maj/D# . A#maj . | D#maj . D#maj/G . | G#maj/D# . A#maj . |\n",
      "\n",
      "finetuned:\n",
      "   | D#maj . G#maj . | A#maj . D#maj/G . | G#maj . G#maj . | A#maj . D#maj . |\n"
     ]
    }
   ],
   "source": [
    "def one_line(path: Path) -> str:\n",
    "    p = parse_chord_file(path)\n",
    "    return \"| \" + \" | \".join(\" \".join(bar) for bar in p.bars) + \" |\"\n",
    "\n",
    "ref = DATA / \"raw_user\" / \"H1K0\" / \"celestial_sphere-chorus.chord\"\n",
    "pre = OUT / \"examples\" / \"pretrained_01_celestial_sphere-chorus.chord\"\n",
    "ftn = OUT / \"examples\" / \"finetuned_01_celestial_sphere-chorus.chord\"\n",
    "\n",
    "print(\"Эталон (автор):\")\n",
    "print(\"  \", one_line(ref))\n",
    "print(\"\\npretrained:\")\n",
    "print(\"  \", one_line(pre))\n",
    "print(\"\\nfinetuned:\")\n",
    "print(\"  \", one_line(ftn))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "80719cd6",
   "metadata": {},
   "source": [
    "Что видно на этом примере (`celestial_sphere`, припев, Eb-dur):\n",
    "\n",
    "- Обе модели держатся в тональности — правда, генерация заякорена на тонику, так\n",
    "  что это отчасти заслуга условия, а не модели.\n",
    "- **Энгармония.** Модель выдаёт диезы (`D#maj`, `G#maj`, `A#maj`), а я тот же\n",
    "  звук записал бемолями (`Eb`, `Ab`, `Bb`). Это не ошибка: в словаре все 12\n",
    "  ступеней хранятся только в диезной записи (`ROOT_*` — sharps only), и при\n",
    "  обратной сборке выбирается диезное имя. Звучит идентично; при желании имена\n",
    "  можно перевести в бемоли отдельной пост-обработкой.\n",
    "- **finetuned** движется заметнее и ближе к моему стилю (тоника → субдоминанта →\n",
    "  доминанта → тоника), тогда как **pretrained** чаще повторяет тонику. Это всего\n",
    "  один пример, поэтому вывод осторожный."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1095e953",
   "metadata": {},
   "source": [
    "## 6. Генерация\n",
    "\n",
    "Сэмплирую через **nucleus / top-p** (≈ 0.9) с температурой около 1.0. Beam\n",
    "search не использую — на генеративных задачах он обычно ухудшает результат, делая\n",
    "его однообразным. Дополнительные параметры:\n",
    "\n",
    "- **tonic anchor** — по умолчанию ставлю в начало тонику, чтобы генерация не\n",
    "  уходила из тональности;\n",
    "- **repetition penalty** — биграммный штраф за повтор переходов между тониками\n",
    "  (по умолчанию выключен, разумный диапазон 0.5–1.5);\n",
    "- **n_bars** — позволяет зафиксировать точное число тактов.\n",
    "\n",
    "Запустить генерацию можно тремя способами: из командной строки\n",
    "(`scripts/generate.py`), через веб-форму (`app.py`, небольшой Gradio) или\n",
    "напрямую вызовом `generate_period(...)`. Ниже — последний вариант."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "233b65b0",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-06-04T13:54:00.974066Z",
     "iopub.status.busy": "2026-06-04T13:54:00.973554Z",
     "iopub.status.idle": "2026-06-04T13:54:01.115739Z",
     "shell.execute_reply": "2026-06-04T13:54:01.115739Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Сгенерировано (8 тактов, тональность C_major):\n",
      "| Cmaj . . . | Cmaj . . . | Cmaj . . . | Em7 . . . | Dm7 . . . | Em7 . . . | Fmaj . Gmaj . | Cmaj . Gmaj . |\n"
     ]
    }
   ],
   "source": [
    "period = generate_period(\n",
    "    model=finetuned,\n",
    "    mode=\"major\", key=\"C_major\",\n",
    "    time=\"4/4\", subdivision=4,\n",
    "    style=\"H1K0\", function=\"chorus\",\n",
    "    prefix=[\"C\"],                 # якорь на тонику\n",
    "    temperature=1.0, top_p=0.9,\n",
    "    repetition_penalty=0.8,       # сбиваю залипание на I–V\n",
    "    n_bars=8, seed=4,\n",
    ")\n",
    "print(f\"Сгенерировано ({len(period.bars)} тактов, тональность {period.key}):\")\n",
    "print(\"| \" + \" | \".join(\" \".join(bar) for bar in period.bars) + \" |\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "59c76c8c",
   "metadata": {},
   "source": [
    "## 7. Выводы и ограничения\n",
    "\n",
    "**Что получилось.**\n",
    "\n",
    "- Конвейер работает целиком: `.chord` → токены → обучение → генерация → обратно\n",
    "  `.chord` и MIDI.\n",
    "- Дообучение снизило перплексию на валидации с **3.58** до **2.15** (−39.8 %):\n",
    "  модель заметно лучше предсказывает мои периоды.\n",
    "- Качественно генерации после дообучения ближе к моему стилю.\n",
    "\n",
    "**Ограничения.**\n",
    "\n",
    "- Корпус небольшой (68 периодов), поэтому оценки имеют высокую дисперсию.\n",
    "- `data/holdout/` не заполнен — перплексия измерена на валидации, а не на честном\n",
    "  тесте.\n",
    "- Модель сводит энгармонию к диезам — это касается записи, а не звучания.\n",
    "- Обучен по сути один стиль (`H1K0`); остальные теги стиля есть в словаре, но на\n",
    "  них модель не обучалась.\n",
    "\n",
    "**Дальнейшая работа** (за рамками срока).\n",
    "\n",
    "- Заполнить holdout и пересчитать перплексию на честном тесте.\n",
    "- Попробовать дообучение на J-Pop корпусе.\n",
    "- Добавить пост-обработку имён аккордов в бемоли для удобства чтения."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}