(
self,
train: bool,
token_type: Optional[str] = None,
token_list: Union[Path, str, Iterable[str]] = None,
bpemodel: Union[Path, str, Iterable[str]] = None,
text_cleaner: Collection[str] = None,
g2p_type: Optional[str] = None,
unk_symbol: str = "<unk>",
space_symbol: str = "<space>",
non_linguistic_symbols: Union[Path, str, Iterable[str]] = None,
delimiter: Optional[str] = None,
singing_volume_normalize: float = None,
singing_name: str = "singing",
text_name: str = "text",
label_name: str = "label",
midi_name: str = "score",
fs: np.int32 = 0,
hop_length: np.int32 = 256,
phn_seg: dict = {
1: [1],
2: [0.25, 1],
3: [0.1, 0.5, 1],
4: [0.05, 0.1, 0.5, 1],
},
discrete_token_name: str = "discrete_token",
pos_sample_name: str = "pos_idx",
neg_sample_name: str = "neg_idx",
)
| 1517 | """Preprocessor for Sing Voice Sythesis (SVS) task.""" |
| 1518 | |
| 1519 | def __init__( |
| 1520 | self, |
| 1521 | train: bool, |
| 1522 | token_type: Optional[str] = None, |
| 1523 | token_list: Union[Path, str, Iterable[str]] = None, |
| 1524 | bpemodel: Union[Path, str, Iterable[str]] = None, |
| 1525 | text_cleaner: Collection[str] = None, |
| 1526 | g2p_type: Optional[str] = None, |
| 1527 | unk_symbol: str = "<unk>", |
| 1528 | space_symbol: str = "<space>", |
| 1529 | non_linguistic_symbols: Union[Path, str, Iterable[str]] = None, |
| 1530 | delimiter: Optional[str] = None, |
| 1531 | singing_volume_normalize: float = None, |
| 1532 | singing_name: str = "singing", |
| 1533 | text_name: str = "text", |
| 1534 | label_name: str = "label", |
| 1535 | midi_name: str = "score", |
| 1536 | fs: np.int32 = 0, |
| 1537 | hop_length: np.int32 = 256, |
| 1538 | phn_seg: dict = { |
| 1539 | 1: [1], |
| 1540 | 2: [0.25, 1], |
| 1541 | 3: [0.1, 0.5, 1], |
| 1542 | 4: [0.05, 0.1, 0.5, 1], |
| 1543 | }, |
| 1544 | discrete_token_name: str = "discrete_token", |
| 1545 | pos_sample_name: str = "pos_idx", |
| 1546 | neg_sample_name: str = "neg_idx", |
| 1547 | ): |
| 1548 | super().__init__(train) |
| 1549 | self.train = train |
| 1550 | self.singing_name = singing_name |
| 1551 | self.text_name = text_name |
| 1552 | self.label_name = label_name |
| 1553 | self.midi_name = midi_name |
| 1554 | self.fs = fs |
| 1555 | self.hop_length = hop_length |
| 1556 | self.singing_volume_normalize = singing_volume_normalize |
| 1557 | self.phn_seg = phn_seg |
| 1558 | self.time_shift = hop_length / fs |
| 1559 | self.discrete_token_name = discrete_token_name |
| 1560 | self.pos_sample_name = pos_sample_name |
| 1561 | self.neg_sample_name = neg_sample_name |
| 1562 | if token_type is not None: |
| 1563 | if token_list is None: |
| 1564 | raise ValueError("token_list is required if token_type is not None") |
| 1565 | self.text_cleaner = TextCleaner(text_cleaner) |
| 1566 | |
| 1567 | self.tokenizer = build_tokenizer( |
| 1568 | token_type=token_type, |
| 1569 | bpemodel=bpemodel, |
| 1570 | delimiter=delimiter, |
| 1571 | space_symbol=space_symbol, |
| 1572 | non_linguistic_symbols=non_linguistic_symbols, |
| 1573 | g2p_type=g2p_type, |
| 1574 | ) |
| 1575 | self.token_id_converter = TokenIDConverter( |
| 1576 | token_list=token_list, |
nothing calls this directly
no test coverage detected