paddlespeech.t2s.frontend.vocab module
- class paddlespeech.t2s.frontend.vocab.Vocab(symbols: Iterable[str], padding_symbol='<pad>', unk_symbol='<unk>', start_symbol='<s>', end_symbol='</s>')[source]
Bases:
object
Vocabulary.
- Args:
symbols (Iterable[str]): Common symbols. padding_symbol (str, optional): Symbol for pad. Defaults to "<pad>". unk_symbol (str, optional): Symbol for unknow. Defaults to "<unk>" start_symbol (str, optional): Symbol for start. Defaults to "<s>" end_symbol (str, optional): Symbol for end. Defaults to "</s>"
- Attributes:
end_index
The index of end symbol.
num_specials
The number of special symbols.
padding_index
The index of padding symbol
start_index
The index of start symbol.
unk_index
The index of unknow symbol.
Methods
add_symbol
(symbol)Add a new symbol in vocab.
add_symbols
(symbols)Add multiple symbols in vocab.
lookup
(symbol)The index that symbol correspond.
reverse
(index)The symbol thar index cottespond.
- property end_index
The index of end symbol.
- property num_specials
The number of special symbols.
- property padding_index
The index of padding symbol
- property start_index
The index of start symbol.
- property unk_index
The index of unknow symbol.