update

speech_asr

2023-03-21 8314c5f17eff473e08dca1729c8a1e62290c7866

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
import logging
from pathlib import Path
from typing import Iterable
from typing import List
from typing import Union
 
import sentencepiece as spm
from torch.utils.data import DataLoader
from typeguard import check_argument_types
 
from funasr.datasets.large_datasets.dataset import Dataset
from funasr.iterators.abs_iter_factory import AbsIterFactory
from funasr.text.abs_tokenizer import AbsTokenizer
 
 
def read_symbol_table(symbol_table_file):
    if isinstance(symbol_table_file, str):
        symbol_table = {}
        with open(symbol_table_file, "r", encoding="utf8") as fin:
            for i, line in enumerate(fin):
                char = line.strip()
                symbol_table[char] = i
    else:
        assert isinstance(symbol_table_file, list)
        symbol_table = {}
        for i, char in enumerate(symbol_table_file):
            symbol_table[char] = i
    return symbol_table
 
 
def load_seg_dict(seg_dict_file):
    seg_dict = {}
    assert isinstance(seg_dict_file, str)
    with open(seg_dict_file, "r", encoding="utf8") as f:
        lines = f.readlines()
        for line in lines:
            s = line.strip().split()
            key = s[0]
            value = s[1:]
            seg_dict[key] = " ".join(value)
    return seg_dict
 
 
class SentencepiecesTokenizer(AbsTokenizer):
    def __init__(self, model: Union[Path, str]):
        assert check_argument_types()
        self.model = str(model)
        self.sp = None
 
    def __repr__(self):
        return f'{self.__class__.__name__}(model="{self.model}")'
 
    def _build_sentence_piece_processor(self):
        if self.sp is None:
            self.sp = spm.SentencePieceProcessor()
            self.sp.load(self.model)
 
    def text2tokens(self, line: str) -> List[str]:
        self._build_sentence_piece_processor()
        return self.sp.EncodeAsPieces(line)
 
    def tokens2text(self, tokens: Iterable[str]) -> str:
        self._build_sentence_piece_processor()
        return self.sp.DecodePieces(list(tokens))
 
 
class ArkDataLoader(AbsIterFactory):
    def __init__(self, data_list, dict_file, dataset_conf, frontend_conf=None, seg_dict_file=None, punc_dict_file=None,
                 bpemodel_file=None, mode="train"):
        symbol_table = read_symbol_table(dict_file) if dict_file is not None else None
        if seg_dict_file is not None:
            seg_dict = load_seg_dict(seg_dict_file)
        else:
            seg_dict = None
        if punc_dict_file is not None:
            punc_dict = read_symbol_table(punc_dict_file)
        else:
            punc_dict = None
        self.dataset_conf = dataset_conf
        self.frontend_conf = frontend_conf
        logging.info("dataloader config: {}".format(self.dataset_conf))
        batch_mode = self.dataset_conf.get("batch_mode", "padding")
        if bpemodel_file is not None:
            bpe_tokenizer = SentencepiecesTokenizer(bpemodel_file)
        else:
            bpe_tokenizer = None
        self.dataset = Dataset(data_list, symbol_table, seg_dict, punc_dict, bpe_tokenizer,
                               self.dataset_conf, self.frontend_conf, mode=mode, batch_mode=batch_mode)
 
    def build_iter(self, epoch, shuffle=True):
        self.dataset.set_epoch(epoch)
        data_loader = DataLoader(self.dataset,
                                 batch_size=None,
                                 pin_memory=True,
                                 num_workers=self.dataset_conf.get("num_workers", 8))
        return data_loader