add speaker-attributed ASR task for alimeeting

smohan-speech

2023-05-06 a73123bcfc14370b74b17084bc124f00c48613e4

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
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
import torch
from torch import nn
 
from funasr.modules.layer_norm  import LayerNorm
 
 
class SpeakerAttributeSpkDecoderFirstLayer(nn.Module):
 
    def __init__(
        self,
        size,
        self_attn,
        src_attn,
        feed_forward,
        dropout_rate,
        normalize_before=True,
        concat_after=False,
    ):
        """Construct an DecoderLayer object."""
        super(SpeakerAttributeSpkDecoderFirstLayer, self).__init__()
        self.size = size
        self.self_attn = self_attn
        self.src_attn = src_attn
        self.feed_forward = feed_forward
        self.norm1 = LayerNorm(size)
        self.norm2 = LayerNorm(size)
        self.dropout = nn.Dropout(dropout_rate)
        self.normalize_before = normalize_before
        self.concat_after = concat_after
        if self.concat_after:
            self.concat_linear1 = nn.Linear(size + size, size)
            self.concat_linear2 = nn.Linear(size + size, size)
 
    def forward(self, tgt, tgt_mask, asr_memory, spk_memory, memory_mask, cache=None):
        
        residual = tgt
        if self.normalize_before:
            tgt = self.norm1(tgt)
 
        if cache is None:
            tgt_q = tgt
            tgt_q_mask = tgt_mask
        else:
            # compute only the last frame query keeping dim: max_time_out -> 1
            assert cache.shape == (
                tgt.shape[0],
                tgt.shape[1] - 1,
                self.size,
            ), f"{cache.shape} == {(tgt.shape[0], tgt.shape[1] - 1, self.size)}"
            tgt_q = tgt[:, -1:, :]
            residual = residual[:, -1:, :]
            tgt_q_mask = None
            if tgt_mask is not None:
                tgt_q_mask = tgt_mask[:, -1:, :]
 
        if self.concat_after:
            tgt_concat = torch.cat(
                (tgt_q, self.self_attn(tgt_q, tgt, tgt, tgt_q_mask)), dim=-1
            )
            x = residual + self.concat_linear1(tgt_concat)
        else:
            x = residual + self.dropout(self.self_attn(tgt_q, tgt, tgt, tgt_q_mask))
        if not self.normalize_before:
            x = self.norm1(x)
        z = x
        
        residual = x
        if self.normalize_before:
            x = self.norm1(x)
 
        skip = self.src_attn(x, asr_memory, spk_memory, memory_mask)
 
        if self.concat_after:
            x_concat = torch.cat(
                (x, skip), dim=-1
            )
            x = residual + self.concat_linear2(x_concat)
        else:
            x = residual + self.dropout(skip)
        if not self.normalize_before:
            x = self.norm1(x)
        
        residual = x
        if self.normalize_before:
            x = self.norm2(x)
        x = residual + self.dropout(self.feed_forward(x))
        if not self.normalize_before:
            x = self.norm2(x)
 
        if cache is not None:
            x = torch.cat([cache, x], dim=1)
            
        return x, tgt_mask, asr_memory, spk_memory, memory_mask, z
 
class SpeakerAttributeAsrDecoderFirstLayer(nn.Module):
    
    def __init__(
        self,
        size,
        d_size,
        src_attn,
        feed_forward,
        dropout_rate,
        normalize_before=True,
        concat_after=False,
    ):
        """Construct an DecoderLayer object."""
        super(SpeakerAttributeAsrDecoderFirstLayer, self).__init__()
        self.size = size
        self.src_attn = src_attn
        self.feed_forward = feed_forward
        self.norm1 = LayerNorm(size)
        self.norm2 = LayerNorm(size)
        self.norm3 = LayerNorm(size)
        self.dropout = nn.Dropout(dropout_rate)
        self.normalize_before = normalize_before
        self.concat_after = concat_after
        self.spk_linear = nn.Linear(d_size, size, bias=False)
        if self.concat_after:
            self.concat_linear1 = nn.Linear(size + size, size)
            self.concat_linear2 = nn.Linear(size + size, size)
 
    def forward(self, tgt, tgt_mask, memory, memory_mask, dn, cache=None):
        
        residual = tgt
        if self.normalize_before:
            tgt = self.norm1(tgt)
 
        if cache is None:
            tgt_q = tgt
            tgt_q_mask = tgt_mask
        else:
            
            tgt_q = tgt[:, -1:, :]
            residual = residual[:, -1:, :]
            tgt_q_mask = None
            if tgt_mask is not None:
                tgt_q_mask = tgt_mask[:, -1:, :]
 
        x = tgt_q
        if self.normalize_before:
            x = self.norm2(x)
        if self.concat_after:
            x_concat = torch.cat(
                (x, self.src_attn(x, memory, memory, memory_mask)), dim=-1
            )
            x = residual + self.concat_linear2(x_concat)
        else:
            x = residual + self.dropout(self.src_attn(x, memory, memory, memory_mask))
        if not self.normalize_before:
            x = self.norm2(x)
        residual = x
 
        if dn!=None:
            x = x + self.spk_linear(dn)
        if self.normalize_before:
            x = self.norm3(x)
        
        x = residual + self.dropout(self.feed_forward(x))
        if not self.normalize_before:
            x = self.norm3(x)
 
        if cache is not None:
            x = torch.cat([cache, x], dim=1)
 
        return x, tgt_mask, memory, memory_mask