piper-plus/data-sources.yml at dev · ayutaz/piper-plus · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
# Canonical metadata for datasets used to train the 6-language multilingual
# base model + downstream fine-tunes. Consumed by
# `scripts/generate_model_card.py` to populate MODEL_CARD.md and
# LICENSE_ATTRIBUTIONS.md on HF Hub release.
#
# A `verified: false` entry means the license was inferred from upstream
# README / dataset card but has not been confirmed by a maintainer. Such
# entries are emitted in MODEL_CARD with a visible warning; once a human
# checks them, flip `verified` to `true`.
#
# Schema is forward-compatible: extra keys on a dataset are passed through
# but new top-level keys are ignored until the script learns about them.

schema_version: 1
last_reviewed: "2026-05-18"
review_period_months: 3

datasets:
  - id: moe-speech-20speakers
    languages: [ja]
    title: "MOE-Speech (20 speakers, LJSpeech format)"
    speakers: 20
    utterances: 60148
    license:
      spdx: "CC-BY-SA-4.0"
      verified: false
      url: "https://creativecommons.org/licenses/by-sa/4.0/"
    source:
      url: "https://huggingface.co/datasets/ayousanz/moe-speech-20speakers-ljspeech"
      commit_or_version: "HF Hub revision @ 2026-03-16"
    attribution_required: true
    attribution_text: |
      MOE-Speech 20 speakers corpus, distributed under CC BY-SA 4.0 by ayousanz.

  - id: libritts-r
    languages: [en]
    title: "LibriTTS-R"
    speakers: 310
    utterances: 74912
    license:
      spdx: "CC-BY-4.0"
      verified: true
      url: "https://creativecommons.org/licenses/by/4.0/"
    source:
      url: "https://www.openslr.org/141/"
      commit_or_version: "v1.0"
    attribution_required: true
    attribution_text: |
      LibriTTS-R: A Restored Multi-Speaker Text-to-Speech Corpus,
      Koizumi et al., 2023. Distributed under CC BY 4.0.

  - id: aishell-3
    languages: [zh]
    title: "AISHELL-3"
    speakers: 142
    utterances: 63223
    license:
      spdx: "Apache-2.0"
      verified: true
      url: "https://www.apache.org/licenses/LICENSE-2.0"
    source:
      url: "https://www.aishelltech.com/aishell_3"
      commit_or_version: "v1.0"
    attribution_required: true
    attribution_text: |
      AISHELL-3 multi-speaker Mandarin TTS corpus,
      Shi et al., 2020. Licensed under Apache-2.0.

  - id: cml-tts-es
    languages: [es]
    title: "CML-TTS (es subset)"
    speakers: 63
    utterances: 168374
    license:
      spdx: "CC-BY-4.0"
      verified: true
      url: "https://creativecommons.org/licenses/by/4.0/"
    source:
      url: "https://github.com/freds0/CML-TTS-Dataset"
      commit_or_version: "v1.0"
    attribution_required: true
    attribution_text: |
      CML-TTS: A Multilingual Speech Corpus, freds0 et al.
      Distributed under CC BY 4.0.

  - id: cml-tts-fr
    languages: [fr]
    title: "CML-TTS (fr subset)"
    speakers: 28
    utterances: 107464
    license:
      spdx: "CC-BY-4.0"
      verified: true
      url: "https://creativecommons.org/licenses/by/4.0/"
    source:
      url: "https://github.com/freds0/CML-TTS-Dataset"
      commit_or_version: "v1.0"
    attribution_required: true
    attribution_text: |
      CML-TTS: A Multilingual Speech Corpus, freds0 et al.
      Distributed under CC BY 4.0.

  - id: cml-tts-pt
    languages: [pt]
    title: "CML-TTS (pt subset)"
    speakers: 8
    utterances: 34066
    license:
      spdx: "CC-BY-4.0"
      verified: true
      url: "https://creativecommons.org/licenses/by/4.0/"
    source:
      url: "https://github.com/freds0/CML-TTS-Dataset"
      commit_or_version: "v1.0"
    attribution_required: true
    attribution_text: |
      CML-TTS: A Multilingual Speech Corpus, freds0 et al.
      Distributed under CC BY 4.0.

  - id: tsukuyomi-chan-corpus
    languages: [ja]
    title: "つくよみちゃんコーパス"
    speakers: 1
    utterances: 100
    license:
      spdx: "CC-BY-4.0"
      verified: false
      url: "https://tyc.rei-yumesaki.net/material/corpus/"
    source:
      url: "https://tyc.rei-yumesaki.net/material/corpus/"
      commit_or_version: "v1.0"
    attribution_required: true
    attribution_text: |
      Tsukuyomi-chan Corpus, distributed by 夢前黎.
      See https://tyc.rei-yumesaki.net/about/terms/ for terms.
    used_only_in:
      - "tsukuyomi-6lang-v2"
      - "tsukuyomi-mb-istft"