一文读懂「Attention is All You Need」
一文读懂「Attention is All You Need」
- 1. 介绍
- 2. 模型架构
-
- 2.1 模型框架
- 2.2Encoder-Decoder 框架
-
- 2.1.1 Encoder模块
- 2.1.2 Decoder模块
- 2.1.3 Attention模块
-
- 2.1.3.1 Attention的结构
- 2.1.3.2 multi head Attention结构
- 2.1.3.3 Attention的整体计算过程
- 2.1.3.4 过程详解举例
- 2.1.3.5 代码实现
- 2.3 Position-wise Feed-Forward
- 2.4 Embeddings and Softmax
- 2.5 Positional Encoding
- 2.6 Full Model
1. 介绍
核心Transformer，下面我们介绍下Transformer
2. 模型架构
Transformer的抽象结构图
Transformer内部是encoder-decoder框架。
2.1 模型框架
每一个Transformer结构，由6个encoder和decoder构成。最后一个encoder连接到各个decoder。
展开后如下图，其中N=6
transformer中decoder和encoder的内部网络结构
Encoder input X = ( x 1 , x 2 . . . x n ) Decoder output Y = ( y 1 , y 2 . . . y n ) o u t p u t {\text{Encoder input }} X=(x_{1}, x_{2}...x_{n}) \\ {\text{Decoder output }} Y=(y_{1}, y_{2}...y_{n}) \\ output Encoder input X=(x1,x2...xn)Decoder output Y=(y1,y2...yn)output
代码实现
import numpy as np
import torch
import torch.nn as nn
import torch.nn.functional as F
import math, copy, time
from torch.autograd import Variable
import matplotlib.pyplot as plt
import seaborn
seaborn.set_context(context="talk")
%matplotlib inline
class EncoderDecoder(nn.Module):
"""
A standard Encoder-Decoder architecture. Base for this and many
other models.
"""
def __init__(self, encoder, decoder, src_embed, tgt_embed, generator):
super(EncoderDecoder, self).__init__()
self.encoder = encoder
self.decoder = decoder
self.src_embed = src_embed
self.tgt_embed = tgt_embed
self.generator = generator
def forward(self, src, tgt, src_mask, tgt_mask):
"Take in and process masked src and target sequences."
return self.decode(self.encode(src, src_mask), src_mask,
tgt, tgt_mask)
def encode(self, src, src_mask):
return self.encoder(self.src_embed(src), src_mask)
def decode(self, memory, src_mask, tgt, tgt_mask):
return self.decoder(self.tgt_embed(tgt), memory, src_mask, tgt_mask)
class Generator(nn.Module):
"Define standard linear + softmax generation step."
def __init__(self, d_model, vocab):
super(Generator, self).__init__()
self.proj = nn.Linear(d_model, vocab)
def forward(self, x):
return F.log_softmax(self.proj(x), dim=-1)
2.2Encoder-Decoder 框架
2.1.1 Encoder模块
Encoder整体结构
更详细的结构
代码实现
def clones(module, N):
"Produce N identical layers."
return nn.ModuleList([copy.deepcopy(module) for _ in range(N)])
class Encoder(nn.Module):
"Core encoder is a stack of N layers"
def __init__(self, layer, N):
super(Encoder, self).__init__()
self.layers = clones(layer, N)
self.norm = LayerNorm(layer.size)
def forward(self, x, mask):
"Pass the input (and mask) through each layer in turn."
for layer in self.layers:
x = layer(x, mask)
return self.norm(x)
N=6个Encoder，一个Encoder包括EncoderLayer，LayerNorm，self-attention（后面单独讲），还有残差网络SublayerConnection, 计算LayerNorm(x + sublayer(x))
class LayerNorm(nn.Module):
"Construct a layernorm module (See citation for details)."
def __init__(self, features, eps=1e-6):
super(LayerNorm, self).__init__()
self.a_2 = nn.Parameter(torch.ones(features))
self.b_2 = nn.Parameter(torch.zeros(features))
self.eps = eps
def forward(self, x):
mean = x.mean(-1, keepdim=True)
std = x.std(-1, keepdim=True)
return self.a_2 * (x - mean) / (std + self.eps) + self.b_2
class SublayerConnection(nn.Module):
"""
A residual connection followed by a layer norm.
Note for code simplicity the norm is first as opposed to last.
"""
def __init__(self, size, dropout):
super(SublayerConnection, self).__init__()
self.norm = LayerNorm(size)
self.dropout = nn.Dropout(dropout)
def forward(self, x, sublayer):
"Apply residual connection to any sublayer with the same size."
return x + self.dropout(sublayer(self.norm(x)))
class EncoderLayer(nn.Module):
"Encoder is made up of self-attn and feed forward (defined below)"
def __init__(self, size, self_attn, feed_forward, dropout):
super(EncoderLayer, self).__init__()
self.self_attn = self_attn
self.feed_forward = feed_forward
self.sublayer = clones(SublayerConnection(size, dropout), 2)
self.size = size
def forward(self, x, mask):
"Follow Figure 1 (left) for connections."
x = self.sublayer[0](x, lambda x: self.self_attn(x, x, x, mask))
return self.sublayer[1](x, self.feed_forward)
2.1.2 Decoder模块
代码实现
class Decoder(nn.Module):
"Generic N layer decoder with masking."
def __init__(self, layer, N):
super(Decoder, self).__init__()
self.layers = clones(layer, N)
self.norm = LayerNorm(layer.size)
def forward(self, x, memory, src_mask, tgt_mask):
for layer in self.layers:
x = layer(x, memory, src_mask, tgt_mask)
return self.norm(x)
class DecoderLayer(nn.Module):
"Decoder is made of self-attn, src-attn, and feed forward (defined below)"
def __init__(self, size, self_attn, src_attn, feed_forward, dropout):
super(DecoderLayer, self).__init__()
self.size = size
self.self_attn = self_attn
self.src_attn = src_attn
self.feed_forward = feed_forward
self.sublayer = clones(SublayerConnection(size, dropout), 3)
def forward(self, x, memory, src_mask, tgt_mask):
"Follow Figure 1 (right) for connections."
m = memory
x = self.sublayer[0](x, lambda x: self.self_attn(x, x, x, tgt_mask))
x = self.sublayer[1](x, lambda x: self.src_attn(x, m, m, src_mask))
return self.sublayer[2](x, self.feed_forward)
def subsequent_mask(size):
"Mask out subsequent positions."
attn_shape = (1, size, size)
subsequent_mask = np.triu(np.ones(attn_shape), k=1).astype('uint8')
return torch.from_numpy(subsequent_mask) == 0
2.1.3 Attention模块
2.1.3.1 Attention的结构
- 先介绍下Q、K、V三个矩阵。QKV是通过X进行线性变换得到的
- 进行计算
2.1.3.2 multi head Attention结构
- 对每个Head进行计算
- 每个X的每个Header得到对应的Z
- 所有Zconcat起来，再计算得到Z
2.1.3.3 Attention的整体计算过程
2.1.3.4 过程详解举例
- 生成q、k、v三个向量，生成方法为分别乘以三个矩阵。词向量从512维变成qkv的64维
- 计算Thinking的Attention，我们需要计算所有整个句子所有单词跟Thinking的相关性。q1和所有k分别进行点积。
- 进行归一化。除以8（paper中采用维度64的平方根，这样梯度会更稳定，可以为其他的值）。然后加上softmax操作。
- 将softmax分值与Value vector按位相乘。并加和得到z向量。
2.1.3.5 代码实现
def attention(query, key, value, mask=None, dropout=None):
"Compute 'Scaled Dot Product Attention'"
d_k = query.size(-1)
scores = torch.matmul(query, key.transpose(-2, -1)) \
/ math.sqrt(d_k)
if mask is not None:
scores = scores.masked_fill(mask == 0, -1e9)
p_attn = F.softmax(scores, dim = -1)
if dropout is not None:
p_attn = dropout(p_attn)
return torch.matmul(p_attn, value), p_attn
class MultiHeadedAttention(nn.Module):
def __init__(self, h, d_model, dropout=0.1):
"Take in model size and number of heads."
super(MultiHeadedAttention, self).__init__()
assert d_model % h == 0
# We assume d_v always equals d_k
self.d_k = d_model // h
self.h = h
self.linears = clones(nn.Linear(d_model, d_model), 4)
self.attn = None
self.dropout = nn.Dropout(p=dropout)
def forward(self, query, key, value, mask=None):
"Implements Figure 2"
if mask is not None:
# Same mask applied to all h heads.
mask = mask.unsqueeze(1)
nbatches = query.size(0)
# 1) Do all the linear projections in batch from d_model => h x d_k
query, key, value = \
[l(x).view(nbatches, -1, self.h, self.d_k).transpose(1, 2)
for l, x in zip(self.linears, (query, key, value))]
# 2) Apply attention on all the projected vectors in batch.
x, self.attn = attention(query, key, value, mask=mask,
dropout=self.dropout)
# 3) "Concat" using a view and apply a final linear.
x = x.transpose(1, 2).contiguous() \
.view(nbatches, -1, self.h * self.d_k)
return self.linears[-1](x)
2.3 Position-wise Feed-Forward
class PositionwiseFeedForward(nn.Module):
"Implements FFN equation."
def __init__(self, d_model, d_ff, dropout=0.1):
super(PositionwiseFeedForward, self).__init__()
self.w_1 = nn.Linear(d_model, d_ff)
self.w_2 = nn.Linear(d_ff, d_model)
self.dropout = nn.Dropout(dropout)
def forward(self, x):
return self.w_2(self.dropout(F.relu(self.w_1(x))))
2.4 Embeddings and Softmax
输入词向量的Embedding
class Embeddings(nn.Module):
def __init__(self, d_model, vocab):
super(Embeddings, self).__init__()
self.lut = nn.Embedding(vocab, d_model)
self.d_model = d_model
def forward(self, x):
return self.lut(x) * math.sqrt(self.d_model)
2.5 Positional Encoding
位置编码，从上面的结构看Transformer还不具备时序学习能力。为了解决这个问题，加了位置编码。
常见的模式有：
- 根据数据学习；
- 自己设计编码规则 （Transformer采用）
论文中给出的编码公式。
其中，pos表示单词的位置， i 表示单词的维度
P E p o s + k 都可以被 P E p o s 的线性函数表示，三角函数特性： c o s ( α + β ) = c o s ( α ) c o s ( β ) − s i n ( α ) s i n ( β ) s i n ( α + β ) = s i n ( α ) c o s ( β ) + c o s ( α ) s i n ( β ) PE_{pos+k} {\text{都可以被}} PE_{pos}{\text{的线性函数表示，三角函数特性：}} \\ cos(\alpha+\beta)=cos(\alpha)cos(\beta)-sin(\alpha)sin(\beta) \\ sin(\alpha+\beta)=sin(\alpha)cos(\beta)+cos(\alpha)sin(\beta) PEpos+k都可以被PEpos的线性函数表示，三角函数特性：cos(α+β)=cos(α)cos(β)−sin(α)sin(β)sin(α+β)=sin(α)cos(β)+cos(α)sin(β)
也就是Transformer可以捕获单词相对位置的信息。
代码实现
class PositionalEncoding(nn.Module):
"Implement the PE function."
def __init__(self, d_model, dropout, max_len=5000):
super(PositionalEncoding, self).__init__()
self.dropout = nn.Dropout(p=dropout)
# Compute the positional encodings once in log space.
pe = torch.zeros(max_len, d_model)
position = torch.arange(0, max_len).unsqueeze(1)
div_term = torch.exp(torch.arange(0, d_model, 2) *
-(math.log(10000.0) / d_model))
pe[:, 0::2] = torch.sin(position * div_term)
pe[:, 1::2] = torch.cos(position * div_term)
pe = pe.unsqueeze(0)
self.register_buffer('pe', pe)
def forward(self, x):
x = x + Variable(self.pe[:, :x.size(1)],
requires_grad=False)
return self.dropout(x)
2.6 Full Model
def make_model(src_vocab, tgt_vocab, N=6,
d_model=512, d_ff=2048, h=8, dropout=0.1):
"Helper: Construct a model from hyperparameters."
c = copy.deepcopy
attn = MultiHeadedAttention(h, d_model)
ff = PositionwiseFeedForward(d_model, d_ff, dropout)
position = PositionalEncoding(d_model, dropout)
model = EncoderDecoder(
Encoder(EncoderLayer(d_model, c(attn), c(ff), dropout), N),
Decoder(DecoderLayer(d_model, c(attn), c(attn),
c(ff), dropout), N),
nn.Sequential(Embeddings(d_model, src_vocab), c(position)),
nn.Sequential(Embeddings(d_model, tgt_vocab), c(position)),
Generator(d_model, tgt_vocab))
# This was important from their code.
# Initialize parameters with Glorot / fan_avg.
for p in model.parameters():
if p.dim() > 1:
nn.init.xavier_uniform(p)
return model
tmp_model = make_model(10, 10, 2)
参考：
- illustrated
- transformer-pytorch