File size: 6,420 Bytes
cea9202
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
167
168
169
170
from transformers import BertModel,BertConfig
import torch
import torch.nn as nn
import torch.nn.functional as F
from huggingface_hub import PyTorchModelHubMixin

time_gap=10000.0

class CSIBERT(nn.Module):
    def __init__(self,bertconfig, input_dim):
        super().__init__()
        self.bertconfig=bertconfig
        self.bert=BertModel(bertconfig)
        self.hidden_dim=bertconfig.hidden_size
        self.input_dim=input_dim
        self.len=bertconfig.max_position_embeddings

        self.Norm1 = nn.LayerNorm(self.input_dim)
        self.Norm2 = nn.LayerNorm(self.hidden_dim)
        self.Norm3 = nn.LayerNorm(self.hidden_dim)

        self.csi_emb=nn.Sequential(
            nn.Linear(input_dim, input_dim),
            nn.ReLU(),
            nn.Linear(input_dim, self.hidden_dim),
            nn.ReLU(),
            nn.Linear(self.hidden_dim, self.hidden_dim)
        )

        self.time_emb=nn.Sequential(
            nn.Linear(input_dim, input_dim),
            nn.ReLU(),
            nn.Linear(input_dim, self.hidden_dim),
            nn.ReLU(),
            nn.Linear(self.hidden_dim, self.hidden_dim)
        )

        self.fusion_emb=nn.Sequential(
            nn.Linear(self.hidden_dim*2, self.hidden_dim*2),
            nn.ReLU(),
            nn.Linear(self.hidden_dim*2, self.hidden_dim),
            nn.ReLU(),
            nn.Linear(self.hidden_dim, self.hidden_dim)
        )

        self.arl = nn.Sequential(
            nn.Linear(self.len, self.len // 2),
            nn.ReLU(),
            nn.Linear(self.len // 2, self.len // 4),
            nn.ReLU(),
            nn.Linear(self.len // 4, 1)
        )

    def forward(self,x,timestamp,attention_mask=None):
        x=x.to(torch.float32)

        x=self.attention(x)
        x=self.csi_emb(x)
        x_time=self.time_embedding(timestamp)
        x = x + x_time
        y=self.bert(inputs_embeds=x, attention_mask=attention_mask, output_hidden_states=False)
        y=y.last_hidden_state
        return y

    def time_embedding(self,timestamp,t=1):
        device=timestamp.device
        # timestamp = (timestamp - timestamp[:,0:1]) / time_gap
        # timestamp = (timestamp - timestamp[:, 0:1]) / (timestamp[:,-1:] - timestamp[:, 0:1])
        timestamp = (timestamp - timestamp[:, 0:1]) / (timestamp[:,-1:] - timestamp[:, 0:1]) * self.len

        timestamp**=t
        d_model=self.input_dim
        dim=torch.tensor(list(range(d_model))).to(device)
        batch_size,length=timestamp.shape
        timestamp=timestamp.unsqueeze(2).repeat(1, 1, d_model)
        dim=dim.reshape([1,1,-1]).repeat(batch_size,length,1)
        sin_emb = torch.sin(timestamp/10000**(dim//2*2/d_model))
        cos_emb = torch.cos(timestamp/10000**(dim//2*2/d_model))
        mask=torch.zeros(d_model).to(device)
        mask[::2]=1
        emb=sin_emb*mask+cos_emb*(1-mask)
        emb=self.time_emb(emb)

        # timestamp = torch.unsqueeze(timestamp, -1)
        # emb=self.time_emb(timestamp)

        return emb

    # def attention(self,x):
    #     y = torch.transpose(x, -1, -2)
    #     batch_size = y.shape[0]
    #     queries = self.query(y).view(batch_size, -1, self.head_num, self.head_dim).transpose(1, 2)
    #     keys = self.key(y).view(batch_size, -1, self.head_num, self.head_dim).transpose(1, 2)
    #     values = self.value(y).view(batch_size, -1, self.head_num, self.head_dim).transpose(1, 2)
    #     attention_weights = self.softmax(torch.matmul(queries, keys.transpose(-1, -2))/ (self.head_dim ** 0.5))
    #
    #     # attended_values = torch.matmul(attention_weights,values).transpose(1, 2)
    #     # attended_values = attended_values.reshape(batch_size,self.input_dim,self.len)
    #     # attended_values = self.norm(attended_values)
    #     # y = attended_values.transpose(1, 2)
    #
    #     attended_values = torch.matmul(attention_weights, values).transpose(-1, -2)
    #     attended_values = attended_values.reshape(batch_size, self.len, self.input_dim)
    #     y = self.norm(attended_values)
    #
    #     return y+x

    def attention(self, x):
        y = torch.transpose(x, -1, -2)
        attn = self.arl(y)
        y = y * attn
        y = torch.transpose(y, -1, -2)
        return y

class Token_Classifier(nn.Module):
    def __init__(self,bert,class_num=52):
        super().__init__()
        self.bert=bert
        self.classifier=nn.Sequential(
            nn.Linear(bert.hidden_dim, bert.hidden_dim//2),
            nn.ReLU(),
            nn.Linear(bert.hidden_dim//2, class_num)
        )

    def forward(self,x,timestamp,attention_mask=None):
        x=self.bert(x,timestamp,attention_mask=attention_mask)
        x=self.classifier(x)
        return x

class SelfAttention(nn.Module):
    def __init__(self, input_dim, da, r):
        super().__init__()
        self.ws1 = nn.Linear(input_dim, da, bias=False)
        self.ws2 = nn.Linear(da, r, bias=False)

    def forward(self, h):
        attn_mat = F.softmax(self.ws2(torch.tanh(self.ws1(h))), dim=1)
        attn_mat = attn_mat.permute(0, 2, 1)
        return attn_mat


class Sequence_Classifier(nn.Module):
    def __init__(self, csibert, class_num, hs=128, da=128, r=4):
        super().__init__()
        self.bert = csibert
        self.attention = SelfAttention(hs, da, r)
        self.classifier = nn.Sequential(
            nn.Linear(hs * r, hs * r // 2),
            nn.ReLU(),
            nn.Linear(hs * r // 2, class_num)
        )

    def forward(self, x, timestamp,attention_mask=None):
        x = self.bert(x, timestamp,attention_mask=attention_mask)
        attn_mat = self.attention(x)
        m = torch.bmm(attn_mat, x)
        flatten = m.view(m.size()[0], -1)
        res = self.classifier(flatten)
        return res

class CSI_BERT2(nn.Module,
    PyTorchModelHubMixin
):
    def __init__(self, max_len=100, hs=128, layers=6, heads=8, intermediate_size=512, carrier_dim=52):
        super().__init__()
        self.config = BertConfig(max_position_embeddings=max_len, hidden_size=hs, num_hidden_layers=layers,num_attention_heads=heads, intermediate_size=intermediate_size)
        self.model = CSIBERT(self.config,carrier_dim)

    def forward(self, x, timestamp=None, attn_mask=None):
        return self.model(x,timestamp,attn_mask)