File size: 6,882 Bytes
1dd09ef
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import torch
import torch.nn as nn
import math
import torch.nn.functional as F

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")


class Embeddings(nn.Module):
    """
    Implements embeddings of the words and adds their positional encodings. 
    """
    def __init__(self, vocab_size, d_model, max_len = 50):
        super(Embeddings, self).__init__()
        self.d_model = d_model
        self.dropout = nn.Dropout(0.1)
        self.embed = nn.Embedding(vocab_size, d_model)
        self.pe = self.create_positinal_encoding(max_len, self.d_model)
        self.dropout = nn.Dropout(0.1)
        
    def create_positinal_encoding(self, max_len, d_model):
        pe = torch.zeros(max_len, d_model).to(device)
        for pos in range(max_len):   # for each position of the word
            for i in range(0, d_model, 2):   # for each dimension of the each position
                pe[pos, i] = math.sin(pos / (10000 ** ((2 * i)/d_model)))
                pe[pos, i + 1] = math.cos(pos / (10000 ** ((2 * (i + 1))/d_model)))
        pe = pe.unsqueeze(0)   # include the batch size
        return pe
        
    def forward(self, encoded_words):
        embedding = self.embed(encoded_words) * math.sqrt(self.d_model)
        embedding += self.pe[:, :embedding.size(1)]   # pe will automatically be expanded with the same batch size as encoded_words
        embedding = self.dropout(embedding)
        return embedding



class MultiHeadAttention(nn.Module):
    
    def __init__(self, heads, d_model):
        
        super(MultiHeadAttention, self).__init__()
        assert d_model % heads == 0
        self.d_k = d_model // heads
        self.heads = heads
        self.dropout = nn.Dropout(0.1)
        self.query = nn.Linear(d_model, d_model)
        self.key = nn.Linear(d_model, d_model)
        self.value = nn.Linear(d_model, d_model)
        self.concat = nn.Linear(d_model, d_model)
        
    def forward(self, query, key, value, mask):
        """
        query, key, value of shape: (batch_size, max_len, 512)
        mask of shape: (batch_size, 1, 1, max_words)
        """
        # (batch_size, max_len, 512)
        query = self.query(query)
        key = self.key(key)        
        value = self.value(value)   
        
        # (batch_size, max_len, 512) --> (batch_size, max_len, h, d_k) --> (batch_size, h, max_len, d_k)
        query = query.view(query.shape[0], -1, self.heads, self.d_k).permute(0, 2, 1, 3)   
        key = key.view(key.shape[0], -1, self.heads, self.d_k).permute(0, 2, 1, 3)  
        value = value.view(value.shape[0], -1, self.heads, self.d_k).permute(0, 2, 1, 3)  
        
        # (batch_size, h, max_len, d_k) matmul (batch_size, h, d_k, max_len) --> (batch_size, h, max_len, max_len)
        scores = torch.matmul(query, key.permute(0,1,3,2)) / math.sqrt(query.size(-1))
        scores = scores.masked_fill(mask == 0, -1e9)    # (batch_size, h, max_len, max_len)
        weights = F.softmax(scores, dim = -1)           # (batch_size, h, max_len, max_len)
        weights = self.dropout(weights)
        # (batch_size, h, max_len, max_len) matmul (batch_size, h, max_len, d_k) --> (batch_size, h, max_len, d_k)
        context = torch.matmul(weights, value)
        # (batch_size, h, max_len, d_k) --> (batch_size, max_len, h, d_k) --> (batch_size, max_len, h * d_k)
        context = context.permute(0,2,1,3).contiguous().view(context.shape[0], -1, self.heads * self.d_k)
        # (batch_size, max_len, h * d_k)
        interacted = self.concat(context)
        return interacted 



class FeedForward(nn.Module):

    def __init__(self, d_model, middle_dim = 2048):
        super(FeedForward, self).__init__()
        
        self.fc1 = nn.Linear(d_model, middle_dim)
        self.fc2 = nn.Linear(middle_dim, d_model)
        self.dropout = nn.Dropout(0.1)

    def forward(self, x):
        out = F.relu(self.fc1(x))
        out = self.fc2(self.dropout(out))
        return out


class EncoderLayer(nn.Module):

    def __init__(self, d_model, heads):
        super(EncoderLayer, self).__init__()
        self.layernorm = nn.LayerNorm(d_model)
        self.self_multihead = MultiHeadAttention(heads, d_model)
        self.feed_forward = FeedForward(d_model)
        self.dropout = nn.Dropout(0.1)

    def forward(self, embeddings, mask):
        interacted = self.dropout(self.self_multihead(embeddings, embeddings, embeddings, mask))
        interacted = self.layernorm(interacted + embeddings)
        feed_forward_out = self.dropout(self.feed_forward(interacted))
        encoded = self.layernorm(feed_forward_out + interacted)
        return encoded


class DecoderLayer(nn.Module):
    
    def __init__(self, d_model, heads):
        super(DecoderLayer, self).__init__()
        self.layernorm = nn.LayerNorm(d_model)
        self.self_multihead = MultiHeadAttention(heads, d_model)
        self.src_multihead = MultiHeadAttention(heads, d_model)
        self.feed_forward = FeedForward(d_model)
        self.dropout = nn.Dropout(0.1)
        
    def forward(self, embeddings, encoded, src_mask, target_mask):
        query = self.dropout(self.self_multihead(embeddings, embeddings, embeddings, target_mask))
        query = self.layernorm(query + embeddings)
        interacted = self.dropout(self.src_multihead(query, encoded, encoded, src_mask))
        interacted = self.layernorm(interacted + query)
        feed_forward_out = self.dropout(self.feed_forward(interacted))
        decoded = self.layernorm(feed_forward_out + interacted)
        return decoded


class Transformer(nn.Module):
    
    def __init__(self, d_model, heads, num_layers, word_map):
        super(Transformer, self).__init__()
        
        self.d_model = d_model
        self.vocab_size = len(word_map)
        self.embed = Embeddings(self.vocab_size, d_model)
        self.encoder = nn.ModuleList([EncoderLayer(d_model, heads) for _ in range(num_layers)])
        self.decoder = nn.ModuleList([DecoderLayer(d_model, heads) for _ in range(num_layers)])
        self.logit = nn.Linear(d_model, self.vocab_size)
        
    def encode(self, src_words, src_mask):
        src_embeddings = self.embed(src_words)
        for layer in self.encoder:
            src_embeddings = layer(src_embeddings, src_mask)
        return src_embeddings
    
    def decode(self, target_words, target_mask, src_embeddings, src_mask):
        tgt_embeddings = self.embed(target_words)
        for layer in self.decoder:
            tgt_embeddings = layer(tgt_embeddings, src_embeddings, src_mask, target_mask)
        return tgt_embeddings
        
    def forward(self, src_words, src_mask, target_words, target_mask):
        encoded = self.encode(src_words, src_mask)
        decoded = self.decode(target_words, target_mask, encoded, src_mask)
        out = F.log_softmax(self.logit(decoded), dim = 2)
        return out