mirror of
https://github.com/k2-fsa/icefall.git
synced 2025-08-08 09:32:20 +00:00
15c1a4a441
* workspace setup * csj prepare done * Change compute_fbank_musan.py t soft link * add description * change lhotse prepare csj command * split train-dev here * Add header * remove debug * save manifest_statistics * generate transcript in Lhotse * update comments in config file
322 lines
7.9 KiB
INI
322 lines
7.9 KiB
INI
; # This section is ignored if this file is not supplied as the first config file to
|
||
; # lhotse prepare csj
|
||
[SEGMENTS]
|
||
; # Allowed period of nonverbal noise. If exceeded, a new segment is created.
|
||
gap = 0.5
|
||
; # Maximum length of segment (s).
|
||
maxlen = 10
|
||
; # Minimum length of segment (s). Segments shorter than `minlen` will be dropped silently.
|
||
minlen = 0.02
|
||
; # Use this symbol to represent a period of allowed nonverbal noise, i.e. `gap`.
|
||
; # Pass an empty string to avoid adding any symbol. It was "<sp>" in kaldi.
|
||
; # If you intend to use a multicharacter string for gap_sym, remember to register the
|
||
; # multicharacter string as part of userdef-string in prepare_lang_char.py.
|
||
gap_sym =
|
||
|
||
[CONSTANTS]
|
||
; # Name of this mode
|
||
MODE = number
|
||
; # Suffixes to use after the word surface (no longer used)
|
||
MORPH = pos1 cForm cType2 pos2
|
||
; # Used to differentiate between A tag and A_num tag
|
||
JPN_NUM = ゼロ ０ 零 一 二 三 四 五 六 七 八 九 十 百 千 ．
|
||
; # Dummy character to delineate multiline words
|
||
PLUS = ＋
|
||
|
||
[DECISIONS]
|
||
; # TAG+'^'とは、タグが一つの転記単位に独立していない場合
|
||
; # The PLUS (fullwidth) sign '＋' marks line boundaries for multiline entries
|
||
|
||
; # フィラー、感情表出系感動詞
|
||
; # 0 to remain, 1 to delete
|
||
; # Example: '(F ぎょっ)'
|
||
F = 1
|
||
; # Example: '(L (F ン))', '比べ(F えー)る'
|
||
F^ = 1
|
||
; # 言い直し、いいよどみなどによる語断片
|
||
; # 0 to remain, 1 to delete
|
||
; # Example: '(D だ)(D だいが) 大学の学部の会議'
|
||
D = 1
|
||
; # Example: '(L (D ドゥ)＋(D ヒ))'
|
||
D^ = 1
|
||
; # 助詞、助動詞、接辞の言い直し
|
||
; # 0 to remain, 1 to delete
|
||
; # Example: '西洋 (D2 的)(F えー)(D ふ) 風というか'
|
||
D2 = 1
|
||
; # Example: '(X (D2 ノ))'
|
||
D2^ = 1
|
||
; # 聞き取りや語彙の判断に自信がない場合
|
||
; # 0 to remain, 1 to delete
|
||
; # Example: (? 字数) の
|
||
; # If no option: empty string is returned regardless of output
|
||
; # Example: '(?) で'
|
||
? = 0
|
||
; # Example: '(D (? すー))＋そう＋です＋よ＋ね'
|
||
?^ = 0
|
||
; # タグ?で、値は複数の候補が想定される場合
|
||
; # 0 for main guess with matching morph info, 1 for second guess
|
||
; # Example: '(? 次数, 実数)', '(? これ,ここで)＋(? 説明＋し＋た＋方＋が＋いい＋か＋な)'
|
||
?, = 0
|
||
; # Example: '(W (? テユクー);(? ケッキョク,テユウコトデ))', '(W マシ;(? マシ＋タ,マス))'
|
||
?,^ = 0
|
||
; # 音や言葉に関するメタ的な引用
|
||
; # 0 to remain, 1 to delete
|
||
; # Example: '助詞の (M は) は (M は) と書くが発音は (M わ)'
|
||
M = 0
|
||
; # Example: '(L (M ヒ)＋(M ヒ))', '(L (M (? ヒ＋ヒ)))'
|
||
M^ = 0
|
||
; # 外国語や古語、方言など
|
||
; # 0 to remain, 1 to delete
|
||
; # Example: '(O ザッツファイン)'
|
||
O = 0
|
||
; # Example: '(笑 (O エクスキューズ＋ミー))', '(笑 メダッ＋テ＋(O ナンボ))'
|
||
O^ = 0
|
||
; # 講演者の名前、差別語、誹謗中傷など
|
||
; # 0 to remain, 1 to delete
|
||
; # Example: '国語研の (R ××) です'
|
||
R = 0
|
||
R^ = 0
|
||
; # 非朗読対象発話（朗読における言い間違い等）
|
||
; # 0 to remain, 1 to delete
|
||
; # Example: '(X 実際は) 実際には'
|
||
X = 0
|
||
; # Example: '(L (X (D2 ニ)))'
|
||
X^ = 0
|
||
; # アルファベットや算用数字、記号の表記
|
||
; # 0 to use Japanese form, 1 to use alphabet form
|
||
; # Example: '(A シーディーアール;ＣＤ－Ｒ)'
|
||
A = 1
|
||
; # Example: 'スモール(A エヌ;Ｎ)', 'ラージ(A キュー;Ｑ)', '(A ティーエフ;ＴＦ)＋(A アイディーエフ;ＩＤＦ)' (Strung together by pron: '(W (? ティーワイド);ティーエフ＋アイディーエフ)')
|
||
A^ = 1
|
||
; # タグAで、単語は算用数字の場合
|
||
; # 0 to use Japanese form, 1 to use Arabic numerals
|
||
; # Example: (A 二千;２０００)
|
||
A_num = 1
|
||
A_num^ = 1
|
||
; # 何らかの原因で漢字表記できなくなった場合
|
||
; # 0 to use broken form, 1 to use orthodox form
|
||
; # Example: '(K たち (F えー) ばな;橘)'
|
||
K = 1
|
||
; # Example: '合(K か(?)く;格)', '宮(K ま(?)え;前)'
|
||
K^ = 1
|
||
; # 転訛、発音の怠けなど、一時的な発音エラー
|
||
; # 0 to use wrong form, 1 to use orthodox form
|
||
; # Example: '(W ギーツ;ギジュツ)'
|
||
W = 1
|
||
; # Example: '(F (W エド;エト))', 'イベント(W リレーティッド;リレーテッド)'
|
||
W^ = 1
|
||
; # 語の読みに関する知識レベルのいい間違い
|
||
; # 0 to use wrong form, 1 to use orthodox form
|
||
; # Example: '(B シブタイ;ジュータイ)'
|
||
B = 0
|
||
; # Example: 'データー(B カズ;スー)'
|
||
B^ = 0
|
||
; # 笑いながら発話
|
||
; # 0 to remain, 1 to delete
|
||
; # Example: '(笑 ナニガ)', '(笑 (F エー)＋ソー＋イッ＋タ＋ヨー＋ナ)'
|
||
笑 = 0
|
||
; # Example: 'コク(笑 サイ＋(D オン))',
|
||
笑^ = 0
|
||
; # 泣きながら発話
|
||
; # 0 to remain, 1 to delete
|
||
; # Example: '(泣 ドンナニ)'
|
||
泣 = 0
|
||
泣^ = 0
|
||
; # 咳をしながら発話
|
||
; # 0 to remain, 1 to delete
|
||
; # Example: 'シャ(咳 リン) ノ'
|
||
咳 = 0
|
||
; # Example: 'イッ(咳 パン)', 'ワズ(咳 カ)'
|
||
咳^ = 0
|
||
; # ささやき声や独り言などの小さな声
|
||
; # 0 to remain, 1 to delete
|
||
; # Example: '(L アレコレナンダッケ)', '(L (W コデ;(? コレ,ココデ))＋(? セツメー＋シ＋タ＋ホー＋ガ＋イー＋カ＋ナ))'
|
||
L = 0
|
||
; # Example: 'デ(L ス)', 'ッ(L テ＋コ)ト'
|
||
L^ = 0
|
||
|
||
[REPLACEMENTS]
|
||
; # ボーカルフライなどで母音が同定できない場合
|
||
<FV> =
|
||
; # 「うん/うーん/ふーん」の音の特定が困難な場合
|
||
<VN> =
|
||
; # 非語彙的な母音の引き延ばし
|
||
<H> =
|
||
; # 非語彙的な子音の引き延ばし
|
||
<Q> =
|
||
; # 言語音と独立に講演者の笑いが生じている場合
|
||
<笑> =
|
||
; # 言語音と独立に講演者の咳が生じている場合
|
||
<咳> =
|
||
; # 言語音と独立に講演者の息が生じている場合
|
||
<息> =
|
||
; # 講演者の泣き声
|
||
<泣> =
|
||
; # 聴衆（司会者なども含む）の発話
|
||
<フロア発話> =
|
||
; # 聴衆の笑い
|
||
<フロア笑> =
|
||
; # 聴衆の拍手
|
||
<拍手> =
|
||
; # 講演者が発表中に用いたデモンストレーションの音声
|
||
<デモ> =
|
||
; # 学会講演に発表時間を知らせるためにならすベルの音
|
||
<ベル> =
|
||
; # 転記単位全体が再度読み直された場合
|
||
<朗読間違い> =
|
||
; # 上記以外の音で特に目立った音
|
||
<雑音> =
|
||
; # 0.2秒以上のポーズ
|
||
<P> =
|
||
; # Redacted information, for R
|
||
; # It is \x00D7 multiplication sign, not your normal 'x'
|
||
× = ×
|
||
|
||
[FIELDS]
|
||
; # Time information for segment
|
||
time = 3
|
||
; # Word surface
|
||
surface = 5
|
||
; # Word surface root form without CSJ tags
|
||
notag = 9
|
||
; # Part Of Speech
|
||
pos1 = 11
|
||
; # Conjugated Form
|
||
cForm = 12
|
||
; # Conjugation Type
|
||
cType1 = 13
|
||
; # Subcategory of POS
|
||
pos2 = 14
|
||
; # Euphonic Change / Subcategory of Conjugation Type
|
||
cType2 = 15
|
||
; # Other information
|
||
other = 16
|
||
; # Pronunciation for lexicon
|
||
pron = 10
|
||
; # Speaker ID
|
||
spk_id = 2
|
||
|
||
[KATAKANA2ROMAJI]
|
||
ア = 'a
|
||
イ = 'i
|
||
ウ = 'u
|
||
エ = 'e
|
||
オ = 'o
|
||
カ = ka
|
||
キ = ki
|
||
ク = ku
|
||
ケ = ke
|
||
コ = ko
|
||
ガ = ga
|
||
ギ = gi
|
||
グ = gu
|
||
ゲ = ge
|
||
ゴ = go
|
||
サ = sa
|
||
シ = si
|
||
ス = su
|
||
セ = se
|
||
ソ = so
|
||
ザ = za
|
||
ジ = zi
|
||
ズ = zu
|
||
ゼ = ze
|
||
ゾ = zo
|
||
タ = ta
|
||
チ = ti
|
||
ツ = tu
|
||
テ = te
|
||
ト = to
|
||
ダ = da
|
||
ヂ = di
|
||
ヅ = du
|
||
デ = de
|
||
ド = do
|
||
ナ = na
|
||
ニ = ni
|
||
ヌ = nu
|
||
ネ = ne
|
||
ノ = no
|
||
ハ = ha
|
||
ヒ = hi
|
||
フ = hu
|
||
ヘ = he
|
||
ホ = ho
|
||
バ = ba
|
||
ビ = bi
|
||
ブ = bu
|
||
ベ = be
|
||
ボ = bo
|
||
パ = pa
|
||
ピ = pi
|
||
プ = pu
|
||
ペ = pe
|
||
ポ = po
|
||
マ = ma
|
||
ミ = mi
|
||
ム = mu
|
||
メ = me
|
||
モ = mo
|
||
ヤ = ya
|
||
ユ = yu
|
||
ヨ = yo
|
||
ラ = ra
|
||
リ = ri
|
||
ル = ru
|
||
レ = re
|
||
ロ = ro
|
||
ワ = wa
|
||
ヰ = we
|
||
ヱ = wi
|
||
ヲ = wo
|
||
ン = ŋ
|
||
ッ = q
|
||
ー = -
|
||
キャ = kǐa
|
||
キュ = kǐu
|
||
キョ = kǐo
|
||
ギャ = gǐa
|
||
ギュ = gǐu
|
||
ギョ = gǐo
|
||
シャ = sǐa
|
||
シュ = sǐu
|
||
ショ = sǐo
|
||
ジャ = zǐa
|
||
ジュ = zǐu
|
||
ジョ = zǐo
|
||
チャ = tǐa
|
||
チュ = tǐu
|
||
チョ = tǐo
|
||
ヂャ = dǐa
|
||
ヂュ = dǐu
|
||
ヂョ = dǐo
|
||
ニャ = nǐa
|
||
ニュ = nǐu
|
||
ニョ = nǐo
|
||
ヒャ = hǐa
|
||
ヒュ = hǐu
|
||
ヒョ = hǐo
|
||
ビャ = bǐa
|
||
ビュ = bǐu
|
||
ビョ = bǐo
|
||
ピャ = pǐa
|
||
ピュ = pǐu
|
||
ピョ = pǐo
|
||
ミャ = mǐa
|
||
ミュ = mǐu
|
||
ミョ = mǐo
|
||
リャ = rǐa
|
||
リュ = rǐu
|
||
リョ = rǐo
|
||
ァ = a
|
||
ィ = i
|
||
ゥ = u
|
||
ェ = e
|
||
ォ = o
|
||
ヮ = ʍ
|
||
ヴ = vu
|
||
ャ = ǐa
|
||
ュ = ǐu
|
||
ョ = ǐo
|
||
|