-
Notifications
You must be signed in to change notification settings - Fork 0
/
a_self_attention.py
394 lines (325 loc) · 13 KB
/
a_self_attention.py
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
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
import torch
import torch.nn as nn
from torch.nn.modules.transformer import Transformer
class SelfAttention(nn.Module):
def __init__(self, embed_size, heads):
super(SelfAttention,self).__init__()
# Embedding size
self.embed_size = embed_size
self.heads = heads
self.head_dim = embed_size // heads
assert embed_size % heads == 0 , "embeding size needs to be div by heads"
self.values = nn.Linear(self.head_dim, self.head_dim, bias=False)
self.keys = nn.Linear(self.head_dim, self.head_dim, bias=False)
self.queries = nn.Linear(self.head_dim, self.head_dim, bias=False)
self.fc_out = nn.Linear(heads*self.head_dim, embed_size)
def forward(self, queries ,keys, values, mask):
# Original Dim of embeddings are
"""[N_batch, len, embed_size]"""
# New dim for multiple heads are [N_batch, len, heads, head_dim]
N_batch = queries.shape[0]
q_len, k_len, v_len = queries.shape[1], keys.shape[1], values.shape[1]
# Split Embedding into self.heads pieces
queries = queries.reshape(N_batch, q_len, self.heads, self.head_dim)
keys = keys.reshape(N_batch, k_len, self.heads, self.head_dim)
values = values.reshape(N_batch, v_len, self.heads, self.head_dim)
# Einsum can be thought of as a dot.product aka Multiplication
# Desired output shape must match values
# [N,h, q_len, k_len]
correlation = torch.einsum("nqhd,nkhd->nhqk", [queries,keys])
if mask is not None:
# Mask fill triangulation matrix
correlation = correlation.masked_fill(mask == 0, float("-1e20"))
attention = torch.softmax(correlation/(self.embed_size**(0.5)),dim=3)
# Softmax dim=3 is k_len key dimension is the source sentence
out = torch.einsum("nhqk,nvhd->nqhd",[attention,values])
# attention shape : (N, heads, query_len, key_len)
# Values shape : (N, value_len, heads, heads_dim)
# out (N, query_len, heads, head_dim)
out = out.reshape(N_batch,q_len,self.heads*self.head_dim)
# Reshape to fc_out input (headss*self.head_dim)
out = self.fc_out(out)
# Reshape to original Embedding shape
"""[N_batch, len, embed_size]"""
return out
class TransformerBlock(nn.Module):
def __init__(self,embed_size, heads, dropout, forward_expansion):
super(TransformerBlock, self).__init__()
self.attention = SelfAttention(embed_size, heads)
self.norm1 = nn.LayerNorm(embed_size)
self.norm2 = nn.LayerNorm(embed_size)
self.feedforward = nn.Sequential(
nn.Linear(embed_size,int(forward_expansion*embed_size)),
nn.ReLU(),
nn.Linear(int(forward_expansion*embed_size), embed_size)
)
self.dropout = nn.Dropout(dropout)
def forward(self, query, key, value, mask):
residual = query
x = self.attention.forward(query,key,value,mask)
x = self.dropout(self.norm1(residual+x))
residual = x
x = self.feedforward(x)
x = self.dropout(self.norm2(residual+x))
return x
class Encoder(nn.Module):
def __init__(
self,
src_vocab_size,
embed_size,
num_layers,
heads,
device,
forward_expansion,
dropout,
max_length
):
super(Encoder, self).__init__()
self.embed_size = embed_size
self.device = device
self.word_embedding = nn.Embedding(src_vocab_size,embed_size)
self.position_embedding = nn.Embedding(max_length,embed_size)
self.trans_layers = nn.ModuleList(
[
TransformerBlock(
embed_size,
heads,
dropout,
forward_expansion
)
]
)
# Confirmed theres only 1 transformer block
#print(self.trans_layers)
self.dropout = nn.Dropout(dropout)
def forward(self,x,mask):
# WHAT is x?
N_batch, seq_length = x.shape
positions = torch.arange(0, seq_length).expand(N_batch, seq_length).to(self.device)
#print('position shape',positions.shape)
out = self.dropout(self.word_embedding(x) + self.position_embedding(positions))
#print('x', x)
#print('position_input', positions)
#print('position output', self.position_embedding(positions))
for trans_layer in self.trans_layers:
out = trans_layer(out, out, out, mask)
return out
class DecoderBlock(nn.Module):
def __init__(self, embed_size, heads, forward_expansion, dropout, device):
super(DecoderBlock, self).__init__()
self.attention = SelfAttention(embed_size,heads)
self.norm = nn.LayerNorm(embed_size)
self.transformer_block = TransformerBlock(
embed_size,heads,dropout,forward_expansion
)
self.dropout = nn.Dropout(dropout)
def forward(self, x, value, key, src_mask, trg_mask):
attention = self.attention(x,x,x,trg_mask)
query = self.dropout(self.norm(attention+x))
out = self.transformer_block.forward(query,key,value,src_mask)
return out
class Decoder(nn.Module):
def __init__(self,
trg_vocab_size,
embed_size,
num_layers,
heads,
forward_expansion,
dropout,
device,
max_length):
super(Decoder,self).__init__()
self.device = device
self.word_embedding = nn.Embedding(trg_vocab_size, embed_size)
self.position_embedding = nn.Embedding(max_length, embed_size)
self.layers = nn.ModuleList(
[DecoderBlock(embed_size,heads,forward_expansion,dropout,device)
for _ in range(num_layers)]
)
self.fc_out = nn.Linear(embed_size, trg_vocab_size)
self.dropout = nn.Dropout(dropout)
def forward(self, x,enc_out, src_mask, trg_mask):
N_batch , seq_length = x.shape
positions = torch.arange(0,seq_length).expand(N_batch,seq_length).to(self.device)
x = self.dropout(self.word_embedding(x) + self.position_embedding(positions))
for decoderlayer in self.layers:
# [Value key] = enc out, [query] = decoder_x
decoderlayer.forward(x,enc_out,enc_out,src_mask,trg_mask)
out = self.fc_out(x)
del positions
return out
class Transformer(nn.Module):
def __init__(self,
src_vocab_size,
trg_vocab_size,
src_pad_idx,
trg_pad_idx,
embed_size=256,
num_layers = 6,
forward_expansion=4,
heads=8,
dropout=0,
device = "cuda",
max_length=100
):
super(Transformer,self).__init__()
"""
# params for transformer
src_vocab_size - encoder vocab size
trg_vocab_size - decoder vocab size
src_pad_idx - encoder pad index
trg_pad_idx - decoder pad index
embed_size - Our self-determined embed size
num_layers - number of layers of encoder and decoder block
forward_expansion - Feed-forward expansion
heads - number of attention heads
dropout - dropout lol
device = gpu or cpu
max_length - maximum length of our sentence
"""
self.encoder = Encoder(
src_vocab_size,
embed_size,num_layers,
heads,
device,
forward_expansion,
dropout,
max_length
)
# Delete the ones that are a more than max_length, and keep the ones
self.decoder = Decoder(
trg_vocab_size,
embed_size,
num_layers,
heads,
forward_expansion,
dropout,
device,
max_length
)
self.src_pad_idx = src_pad_idx
self.trg_pad_idx = trg_pad_idx
self.device = device
def make_src_mask(self,src):
# Masking in on itself
src_mask = (src != self.src_pad_idx).unsqueeze(1).unsqueeze(2)
# (N,1,1,src_len)
return src_mask.to(self.device)
def make_trg_mask(self,trg):
# trg = [batch size, trg len]
"""
1. Masking "<PAD>" tokens and
2. Upper-triangular matrix"""
trg_pad_mask = (trg != self.trg_pad_idx).unsqueeze(1).unsqueeze(2)
# trg_pad_mask = 4D [batch size, 1, 1, trg len]
trg_len = trg.shape[-1]
trg_sub_mask = torch.tril(torch.ones((trg_len, trg_len))).to(self.device).bool()
# trg_sub_mask = [trg len, trg len]
trg_mask = trg_pad_mask & trg_sub_mask
# trg_mask = [batch size , 1 , trg len, trg len]
return trg_mask
def forward(self, src, trg):
# src = [batch_size, src_sentence_len]
# trg = [batch_size, trg_sentence_len]
src_mask = self.make_src_mask(src)
trg_mask = self.make_trg_mask(trg)
# src_mask = [batch size, 1, 1 , src len]
# trg_mask = [batch size, 1, trg len, trg len]
enc_src = self.encoder.forward(src,src_mask)
out = self.decoder.forward(trg,enc_src,src_mask,trg_mask)
# enc_src = [batch size, src len, hid dim]
# output = [batch size, trg len, out dim]
# out dim = probability in each trg len
return out
# def patch_src(src, pad_idx):
# #src = src.transpose(0, 1)
# return src
# def patch_trg(trg, pad_idx):
# #trg = trg.transpose(0, 1)
# trg = trg[:, :-1]
# flat = trg[:, 1:].contiguous().view(-1)
# return trg, flat
# from torch import optim,nn
# import torch.nn.functional as F
# if __name__ == "__main__":
# device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
# print(device)
# eos = 3
# sos = 1
# src = torch.tensor(
# [
# [1,5,6,4,3,9,5,2,0],
# [1,8,7,3,4,5,6,7,2]
# ],dtype=torch.int64
# ).to(device)
# trg = torch.tensor(
# [
# [sos,7,4,3,5,9,2,eos],
# [sos,5,6,2,4,7,6,eos]
# ],dtype=torch.int64
# ).to(device)
# src_pad_idx ,trg_pad_idx = 0, 0
# src_vocab_size , trg_vocab_size = 10, 10
# model= Transformer(src_vocab_size,trg_vocab_size,src_pad_idx,trg_pad_idx).to(device)
# criterion = nn.CrossEntropyLoss(ignore_index=trg_pad_idx)
# pred_trg = trg[:, :-1]
# compare_trg = trg[:, 1:]
# predicted = model(src, pred_trg)
# predicted_vocab_size = predicted.shape[-1]
# print(predicted_vocab_size)
# print(predicted.shape)
# """
# we wanna use view aka reshape, because it uses [move] or [same_memory]
# contiguous = same memory
# # View always need contiguous
# """
# predicted = predicted.contiguous().view(-1, predicted_vocab_size)
# compare_trg2 = compare_trg.contiguous().view(-1) # 1D
# loss = criterion(predicted, compare_trg2)
# print(loss)
# print(loss.item())
# print(compare_trg.data_ptr() == compare_trg2.data_ptr())
# # [batch size, trg len, out dim]
# #from a_self_attention import Transformer
# from attention_transformer import Transformer
# SRC_VOCAB_SIZE ,TRG_VOCAB_SIZE = len(SRC.vocab) , len(TRG.vocab)
# SRC_PAD_IDX, TRG_PAD_IDX = SRC.vocab['<PAD>'] , TRG.vocab['<PAD>']
# MAX_SENTENCE_LENGTH = 100
# EMBED_SIZE , NUM_LAYERS , FORWARD_EXPANSION , HEADS = 256, 3, 2 , 8
# DROPOUT = 0.1
# model = Transformer(
# src_vocab_len=SRC_VOCAB_SIZE,
# trg_vocab_len=TRG_VOCAB_SIZE,
# src_pad_idx = SRC_PAD_IDX,
# trg_pad_idx = TRG_PAD_IDX,
# src_max_sentence_len = MAX_SENTENCE_LENGTH,
# trg_max_sentence_len = MAX_SENTENCE_LENGTH,
# hid_dim = EMBED_SIZE,
# n_layers = NUM_LAYERS,
# n_heads = HEADS,
# ff_dim_multiplier = FORWARD_EXPANSION,
# dropout = DROPOUT,
# device = DEVICE
# ).to(DEVICE)
# from attention_transformer import Transformer
# import torch.nn as nn
# SRC_VOCAB_SIZE ,TRG_VOCAB_SIZE = len(vocab_en) , len(vocab_de)
# SRC_PAD_IDX, TRG_PAD_IDX = vocab_en['<PAD>'] , vocab_de['<PAD>']
# MAX_SENTENCE_LENGTH = 110
# EMBED_SIZE , NUM_LAYERS , FORWARD_EXPANSION , HEADS = 256, 3, 2 , 8
# DEVICE = torch.device("cuda" if torch.cuda.is_available() else "cpu")
# DROPOUT = 0.1
# model = Transformer(
# src_vocab_len=SRC_VOCAB_SIZE,
# trg_vocab_len=TRG_VOCAB_SIZE,
# src_pad_idx = SRC_PAD_IDX,
# trg_pad_idx = TRG_PAD_IDX,
# src_max_sentence_len = MAX_SENTENCE_LENGTH,
# trg_max_sentence_len = MAX_SENTENCE_LENGTH,
# hid_dim = EMBED_SIZE,
# n_layers = NUM_LAYERS,
# n_heads = HEADS,
# ff_dim_multiplier = FORWARD_EXPANSION,
# dropout = DROPOUT,
# device = DEVICE
# ).to(DEVICE)