v1.2.3

游雁

2025-01-24 c4e70144920401a4f5d9b569929b6914af5d324c

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
139
import logging
import torch
 
from funasr.register import tables
 
 
# @tables.register("dataloader_classes", "DataloaderMapStyle")
def DataloaderMapStyle(frontend=None, tokenizer=None, **kwargs):
    # dataset
    logging.info("Build dataloader")
    dataset_class = tables.dataset_classes.get(kwargs.get("dataset", "AudioDataset"))
    dataset_tr = dataset_class(
        kwargs.get("train_data_set_list"),
        frontend=frontend,
        tokenizer=tokenizer,
        is_training=True,
        **kwargs.get("dataset_conf"),
    )
    dataset_val = dataset_class(
        kwargs.get("valid_data_set_list"),
        frontend=frontend,
        tokenizer=tokenizer,
        is_training=False,
        **kwargs.get("dataset_conf"),
    )
 
    # dataloader
    batch_sampler = kwargs["dataset_conf"].get("batch_sampler", "BatchSampler")
    batch_sampler_val = None
    if batch_sampler is not None:
        batch_sampler_class = tables.batch_sampler_classes.get(batch_sampler)
        batch_sampler = batch_sampler_class(dataset_tr, **kwargs.get("dataset_conf"))
        batch_sampler_val = batch_sampler_class(
            dataset_val, is_training=False, **kwargs.get("dataset_conf")
        )
 
    dataloader_tr = torch.utils.data.DataLoader(
        dataset_tr, collate_fn=dataset_tr.collator, **batch_sampler
    )
    dataloader_val = torch.utils.data.DataLoader(
        dataset_val, collate_fn=dataset_val.collator, **batch_sampler_val
    )
 
    return dataloader_tr, dataloader_val
 
 
@tables.register("dataloader_classes", "DataloaderMapStyle")
class DataloaderMapStyle:
    def __init__(self, frontend=None, tokenizer=None, **kwargs):
        # dataset
        logging.info("Build dataloader")
 
        dataset_class = tables.dataset_classes.get(kwargs.get("dataset", "AudioDataset"))
        dataset_tr = None
        # split dataset
        self.data_split_num = kwargs["dataset_conf"].get("data_split_num", 1)
        if self.data_split_num == 1:
            dataset_tr = dataset_class(
                kwargs.get("train_data_set_list"),
                frontend=frontend,
                tokenizer=tokenizer,
                is_training=True,
                **kwargs.get("dataset_conf"),
            )
        dataset_val = dataset_class(
            kwargs.get("valid_data_set_list"),
            frontend=frontend,
            tokenizer=tokenizer,
            is_training=False,
            **kwargs.get("dataset_conf"),
        )
 
        self.dataset_tr = dataset_tr
        self.dataset_val = dataset_val
        self.kwargs = kwargs
 
        self.dataset_class = dataset_class
        self.frontend = frontend
        self.tokenizer = tokenizer
        self.kwargs = kwargs
 
    def build_iter(self, epoch=0, data_split_i=0, start_step=0, **kwargs):
 
        # reload dataset slice
        if self.data_split_num > 1:
            del self.dataset_tr
            self.dataset_tr = self.dataset_class(
                self.kwargs.get("train_data_set_list"),
                frontend=self.frontend,
                tokenizer=self.tokenizer,
                is_training=True,
                **self.kwargs.get("dataset_conf"),
                data_split_i=data_split_i,
            )
 
        # dataloader
        batch_sampler = self.kwargs["dataset_conf"].get("batch_sampler", "BatchSampler")
        batch_sampler_val = None
        if batch_sampler is not None:
            batch_sampler_class = tables.batch_sampler_classes.get(batch_sampler)
            batch_sampler = batch_sampler_class(
                self.dataset_tr, start_step=start_step, **self.kwargs.get("dataset_conf")
            )
            batch_sampler_val = batch_sampler_class(
                self.dataset_val, is_training=False, **self.kwargs.get("dataset_conf")
            )
 
        batch_sampler["batch_sampler"].set_epoch(epoch)
        batch_sampler_val["batch_sampler"].set_epoch(epoch)
        dataloader_tr = torch.utils.data.DataLoader(
            self.dataset_tr, collate_fn=self.dataset_tr.collator, **batch_sampler
        )
        dataloader_val = torch.utils.data.DataLoader(
            self.dataset_val, collate_fn=self.dataset_val.collator, **batch_sampler_val
        )
 
        return dataloader_tr, dataloader_val
 
 
@tables.register("dataloader_classes", "DataloaderIterable")
def DataloaderIterable(frontend=None, tokenizer=None, **kwargs):
    logging.info("Build dataloader")
    dataset_class = tables.dataset_classes.get(kwargs.get("dataset", "LargeDataset"))
    dataset_tr = dataset_class(
        kwargs.get("train_data_set_list"),
        frontend=frontend,
        tokenizer=tokenizer,
        is_training=True,
        **kwargs.get("dataset_conf"),
    )
    dataset_val = dataset_class(
        kwargs.get("valid_data_set_list"),
        frontend=frontend,
        tokenizer=tokenizer,
        is_training=False,
        **kwargs.get("dataset_conf"),
    )
 
    return dataset_tr, dataset_val