Transformer简版实战教程

Transformer简版实战教程

至于Transformer的理论内容可以参考Transformer 与 Attention和Transformer 与 Attention的一些Trick
本文主要实战, 这是一个简单版本的Transformer实现,也便于大家理解。

准备

需要准备的是翻译的语料集sentences以及模型参数src_vocab-输入词表, tgt_vocab目标词表,src_lentgt_len是句子的最大长度,d_model是hidden_size维度大小, d_ff是前馈网络中间CNN的输出维度。n_layers 是transformer block的个数,n_heads是头的个数, 因此多头的后最后维度为d_model/n_heads(64)。注意ich mochte ein bier P中的’P’是Padding值。

## 语料集准备
sentences = ['ich mochte ein bier P', 'S i want a beer', 'i want a beer E']
 ## 模型参数
src_vocab = {'P': 0, 'ich': 1, 'mochte': 2, 'ein': 3, 'bier': 4}
src_vocab_size = len(src_vocab)

tgt_vocab = {'P': 0, 'i': 1, 'want': 2, 'a': 3, 'beer': 4, 'S': 5, 'E': 6} ## 目标词表大小
number_dict = {i: w for i, w in enumerate(tgt_vocab)} ## 
tgt_vocab_size = len(tgt_vocab)

src_len = 5 # length of source
tgt_len = 5 # length of target

d_model = 512  # Embedding Size
d_ff = 2048  # FeedForward dimension
d_k = d_v = 64  # dimension of K(=Q), V
n_layers = 6  # number of Encoder of Decoder Layer
n_heads = 8  # number of heads in Multi-Head Attention

Transfomer

Transformer主要是由encoder、decoder和Linear组成。

输入:

  • enc_inputs:[batch_size, src_len, hidden_size]
  • dec_inputs:[batch_size, tgt_len, hidden_size]
    矩阵变换:
  • enc_outputs:[batch_size, src_len, hidden_size]
  • dec_outputs:[batch_size, tgt_len, hidden_size]
  • enc_self_attns:
  • dec_self_attns:
  • dec_logits [batch_size, src_len, tgt_len]
    输出:
    dec_logits.view(-1, dec_logits.size(-1)): [batch_size* src_len, tgt_len]
class Transformer(nn.Module):
    def __init__(self):
        super(Transformer, self).__init__()
        self.encoder = Encoder()
        self.decoder = Decoder()
        self.projection = nn.Linear(d_model, tgt_vocab_size, bias=False)
    def forward(self, enc_inputs, dec_inputs):
        enc_outputs, enc_self_attns = self.encoder(enc_inputs)
        dec_outputs, dec_self_attns, dec_enc_attns = self.decoder(dec_inputs, enc_inputs, enc_outputs)
        dec_logits = self.projection(dec_outputs) # dec_logits : [batch_size x src_vocab_size x tgt_vocab_size]
        return dec_logits.view(-1, dec_logits.size(-1)), enc_self_attns, dec_self_attns, dec_enc_attns

Encoder

主要由src_embedding, postion_embedding和n_layer个EncoderLayer组成。
输入:

  • enc_inputs:[batch_size, src_len]
  • [[1,2,3,4,0]] 其实也是输入,因为是一个batch,固定写死了, 0是padding值
    矩阵变换:
    enc_outputs :[batch_size, src_len, hidden_size] (开始时input embedding与position embedding相加)
    enc_self_attn_mask :[batch_size, src_len, src_len] (encoder层注意力掩码)
    enc_self_attn:[batch_size, n_head, src_len, src_len]
    输出:
    enc_outputs:[batch_size, src_len, hidden_size]
    enc_self_attns:[batch_size, n_head, src_len, src_len]
class Encoder(nn.Module):
    def __init__(self):
        super(Encoder, self).__init__()
        self.src_emb = nn.Embedding(src_vocab_size, d_model)
        self.pos_emb = nn.Embedding.from_pretrained(get_sinusoid_encoding_table(src_len+1, d_model),freeze=True)
        self.layers = nn.ModuleList([EncoderLayer() for _ in range(n_layers)])

    def forward(self, enc_inputs): # enc_inputs : [batch_size x source_len]
        enc_outputs = self.src_emb(enc_inputs) + self.pos_emb(torch.LongTensor([[1,2,3,4,0]]))
        enc_self_attn_mask = get_attn_pad_mask(enc_inputs, enc_inputs)
        enc_self_attns = []
        for layer in self.layers:
            enc_outputs, enc_self_attn = layer(enc_outputs, enc_self_attn_mask)
            enc_self_attns.append(enc_self_attn)
        return enc_outputs, enc_self_attns

位置编码

位置编码很好理解,偶数位置sin函数,奇数位置cos函数。
Transformer简版实战教程_第1张图片

def get_sinusoid_encoding_table(n_position, d_model):
    def cal_angle(position, hid_idx):
        return position / np.power(10000, 2 * (hid_idx // 2) / d_model)
    def get_posi_angle_vec(position):
        return [cal_angle(position, hid_j) for hid_j in range(d_model)]

    sinusoid_table = np.array([get_posi_angle_vec(pos_i) for pos_i in range(n_position)])
    sinusoid_table[:, 0::2] = np.sin(sinusoid_table[:, 0::2])  # dim 2i
    sinusoid_table[:, 1::2] = np.cos(sinusoid_table[:, 1::2])  # dim 2i+1
    return torch.FloatTensor(sinusoid_table)

attention mask

mask主要是Q与K的注意力掩码,当然在Transformer中Q,K和V是相同的,都是input值。
pad_attn_mask :[batch_size,1, k_len]
pad_attn_mask.expand(batch_size, len_q, len_k): [batch_size , q_len , k_len]

def get_attn_pad_mask(seq_q, seq_k):
    batch_size, len_q = seq_q.size()
    batch_size, len_k = seq_k.size()
    # eq(zero) is PAD token
    pad_attn_mask = seq_k.data.eq(0).unsqueeze(1)  # batch_size x 1 x len_k(=len_q), one is masking
    return pad_attn_mask.expand(batch_size, len_q, len_k)  # batch_size x len_q x len_k

EncoderLayer

主要由多头注意力层(MultiHeadAttention)和前馈层(PoswiseFeedForwardNet)组成。
输入:

  • enc_inputs:[batch_size, src_len, hidden_size]
  • enc_self_attn_mask: [batch_size, src_len, src_len] (即 [batch_size, q_len, k_len])
    矩阵变换:
  • enc_outputs:[batch_size, src_len, hidden_size]
  • attn:[batch_size, n_head, src_len, src_len]
    输出:
    enc_outputs和attn
class EncoderLayer(nn.Module):
    def __init__(self):
        super(EncoderLayer, self).__init__()
        self.enc_self_attn = MultiHeadAttention()
        self.pos_ffn = PoswiseFeedForwardNet()

    def forward(self, enc_inputs, enc_self_attn_mask):
        enc_outputs, attn = self.enc_self_attn(enc_inputs, enc_inputs, enc_inputs, enc_self_attn_mask) # enc_inputs to same Q,K,V
        enc_outputs = self.pos_ffn(enc_outputs) # enc_outputs: [batch_size x len_q x d_model]
        return enc_outputs, attn

MultiHeadAttention

主要由Q K V线性层和LayerNorm组成
输入:

  • Q, K , V : [batch_size, src_len, hidden_size]
  • attn_mask [batch_size, src_len, src_len]
    矩阵变换:
    q_s = self.W_Q(Q).view(batch_size, -1, n_heads, d_k).transpose(1,2) [batch_size, , src_len, hidden_size] 转为 [batch, n_head, src_len, hidden_size//n_head] ,这里每个头是64。
    attn_mask.unsqueeze(1).repeat(1, n_heads, 1, 1)将维度变为 [batch, n_head, src_len, hidden_size//n_head] 便于注意力掩码计算。
    ScaledDotProductAttention()(q_s, k_s, v_s, attn_mask)注意力计算,下面会介绍。
    context.transpose(1, 2).contiguous().view(batch_size, -1, n_heads * d_v)维度变为原来的[batch_size, , src_len, hidden_size]
    self.layer_norm(output + residual) 残差+layerNorm ->[batch_size, , src_len, hidden_size]
    输出:
class MultiHeadAttention(nn.Module):
    def __init__(self):
        super(MultiHeadAttention, self).__init__()
        self.W_Q = nn.Linear(d_model, d_k * n_heads)
        self.W_K = nn.Linear(d_model, d_k * n_heads)
        self.W_V = nn.Linear(d_model, d_v * n_heads)
        self.linear = nn.Linear(n_heads * d_v, d_model)
        self.layer_norm = nn.LayerNorm(d_model)

    def forward(self, Q, K, V, attn_mask):
        # q: [batch_size x len_q x d_model], k: [batch_size x len_k x d_model], v: [batch_size x len_k x d_model]
        residual, batch_size = Q, Q.size(0)
        # (B, S, D) -proj-> (B, S, D) -split-> (B, S, H, W) -trans-> (B, H, S, W)
        q_s = self.W_Q(Q).view(batch_size, -1, n_heads, d_k).transpose(1,2)  # q_s: [batch_size x n_heads x len_q x d_k]
        k_s = self.W_K(K).view(batch_size, -1, n_heads, d_k).transpose(1,2)  # k_s: [batch_size x n_heads x len_k x d_k]
        v_s = self.W_V(V).view(batch_size, -1, n_heads, d_v).transpose(1,2)  # v_s: [batch_size x n_heads x len_k x d_v]

        attn_mask = attn_mask.unsqueeze(1).repeat(1, n_heads, 1, 1) # attn_mask : [batch_size x n_heads x len_q x len_k]

        # context: [batch_size x n_heads x len_q x d_v], attn: [batch_size x n_heads x len_q(=len_k) x len_k(=len_q)]
        context, attn = ScaledDotProductAttention()(q_s, k_s, v_s, attn_mask)
        context = context.transpose(1, 2).contiguous().view(batch_size, -1, n_heads * d_v) # context: [batch_size x len_q x n_heads * d_v]
        output = self.linear(context)
        return self.layer_norm(output + residual), attn # output: [batch_size x len_q x d_model]

ScaledDotProductAttention

Transformer简版实战教程_第2张图片
每个head的Q与响应的K相乘得到注意力得分, attn_mask的作用是padding部分填充-1e9,使其在注意力中不起作用, 权重得分变小。然后进行Softmax,得到权重分布,最后乘以V 得到加权后的得分。维度一致保持[batch_size, n_head, src_len, d_k]

class ScaledDotProductAttention(nn.Module):
    def __init__(self):
        super(ScaledDotProductAttention, self).__init__()

    def forward(self, Q, K, V, attn_mask):
        scores = torch.matmul(Q, K.transpose(-1, -2)) / np.sqrt(d_k) # scores:[batch_size x n_heads x len_q(=len_k) x len_k(=len_q)]
        scores.masked_fill_(attn_mask, -1e9) # Fills elements of self tensor with value where mask is one.
        attn = nn.Softmax(dim=-1)(scores)
        context = torch.matmul(attn, V)
        return context, attn

PoswiseFeedForwardNet

主要由两层CNN、一层激活层和一层LayerNorm组成。
输入:

  • inputs:[batch_size, src_len, hidden_size]
    矩阵变换:
  • 第一层CNN:inputs.transpose(1, 2)改变维度位置,因为torch dim=1的位置是channel,输出为[batch_size, 2048, hidden_size]
  • 激活函数RELU:
  • 第二层CNN:[batch_size, src_len, hidden_size]
  • self.layer_norm(output + residual) 残差+layerNorm ->[batch_size, , src_len, hidden_size]
class PoswiseFeedForwardNet(nn.Module):
    def __init__(self):
        super(PoswiseFeedForwardNet, self).__init__()
        self.conv1 = nn.Conv1d(in_channels=d_model, out_channels=d_ff, kernel_size=1)
        self.conv2 = nn.Conv1d(in_channels=d_ff, out_channels=d_model, kernel_size=1)
        self.layer_norm = nn.LayerNorm(d_model)

    def forward(self, inputs):
        residual = inputs # inputs : [batch_size, len_q, d_model]
        output = nn.ReLU()(self.conv1(inputs.transpose(1, 2)))
        output = self.conv2(output).transpose(1, 2)
        return self.layer_norm(output + residual)

------------------------------------------ Encoder的部分结束,下面我们讲解Decoder部分 ----------------------------------------------

Decoder(也是encdder输出与decoder输入交互层,重点)


主要由tgt_input embedding、tgt_postion embedding和n_layer个DecoderLayer组成。

输入:

  • dec_inputs: [batch_size, tgt_len, hidden_size]
  • enc_inputs:[batch_size, src_len, hidden_size]
  • enc_outputs:[batch_size, src_len, hidden_size]
    矩阵变换
    dec_self_attn_pad_mask decoder开始的padding 掩码部分,与encoder掩码一样,只是输入的是decoder的输入。
    dec_self_attn_subsequent_mask 解码掩码上三角,因为在翻译的时候,当前词是看不到后面词的信息。get_attn_subsequent_mask函数中np.triu方法返回的是上三角矩阵[batch_size, src_len, src_len],如下图
    Transformer简版实战教程_第3张图片

torch.gt((dec_self_attn_pad_mask + dec_self_attn_subsequent_mask), 0) -> torch.gt(a,b)函数比较a中元素大于(这里是严格大于)b中对应元素目的是转为bool值。
get_attn_pad_mask(dec_inputs, enc_inputs) 最终的encoder-decoder padding掩码,与单纯的encoder或者decoderpadding掩码不一样, 输入的是decoder input和encoder input, 其中decoder 部分最为Q, decoder部分最为K, 最中掩码维度为[batch_size, q_len, k_len]

输出:

  • dec_outputs:[batch_size, tgt_len, hidden_size]
  • dec_self_attns:[batch_size, tgt_len, tgt_len]
  • dec_enc_attns:[batch_size, tgt_len, src_len]
class Decoder(nn.Module):
    def __init__(self):
        super(Decoder, self).__init__()
        self.tgt_emb = nn.Embedding(tgt_vocab_size, d_model)
        self.pos_emb = nn.Embedding.from_pretrained(get_sinusoid_encoding_table(tgt_len+1, d_model),freeze=True)
        self.layers = nn.ModuleList([DecoderLayer() for _ in range(n_layers)])

    def forward(self, dec_inputs, enc_inputs, enc_outputs): # dec_inputs : [batch_size x target_len]
        dec_outputs = self.tgt_emb(dec_inputs) + self.pos_emb(torch.LongTensor([[5,1,2,3,4]]))
        dec_self_attn_pad_mask = get_attn_pad_mask(dec_inputs, dec_inputs)
        dec_self_attn_subsequent_mask = get_attn_subsequent_mask(dec_inputs)
        dec_self_attn_mask = torch.gt((dec_self_attn_pad_mask + dec_self_attn_subsequent_mask), 0)

        dec_enc_attn_mask = get_attn_pad_mask(dec_inputs, enc_inputs)

        dec_self_attns, dec_enc_attns = [], []
        for layer in self.layers:
            dec_outputs, dec_self_attn, dec_enc_attn = layer(dec_outputs, enc_outputs, dec_self_attn_mask, dec_enc_attn_mask)
            dec_self_attns.append(dec_self_attn)
            dec_enc_attns.append(dec_enc_attn)
        return dec_outputs, dec_self_attns, dec_enc_attns

def get_attn_subsequent_mask(seq):
    attn_shape = [seq.size(0), seq.size(1), seq.size(1)]
    subsequent_mask = np.triu(np.ones(attn_shape), k=1)
    subsequent_mask = torch.from_numpy(subsequent_mask).byte()
    return subsequent_mask

DecoderLayer

主要由自注意力MultiHeadAttention,交互注意力MultiHeadAttention(输入的是dec_enc_attn_mask掩码),和前馈神经PoswiseFeedForwardNet组成(与encoder一样)

输入:

  • dec_inputs: [batch_size, tgt_len, hidden_size]
  • enc_outputs:[batch_size, src_len, hidden_size]
  • dec_self_attn_mask:[batch_size, tgt_len, tgt_len]
  • dec_enc_attn_mask:[batch_size, tgt_len, src_len]

矩阵变化:
对头注意力层和前馈层与encoder的结构一样, 这里不一一介绍。这里需要注意的是交互注意力层输入不一样。

class DecoderLayer(nn.Module):
    def __init__(self):
        super(DecoderLayer, self).__init__()
        self.dec_self_attn = MultiHeadAttention()
        self.dec_enc_attn = MultiHeadAttention()
        self.pos_ffn = PoswiseFeedForwardNet()

    def forward(self, dec_inputs, enc_outputs, dec_self_attn_mask, dec_enc_attn_mask):
        dec_outputs, dec_self_attn = self.dec_self_attn(dec_inputs, dec_inputs, dec_inputs, dec_self_attn_mask)
        dec_outputs, dec_enc_attn = self.dec_enc_attn(dec_outputs, enc_outputs, enc_outputs, dec_enc_attn_mask)
        dec_outputs = self.pos_ffn(dec_outputs)
        return dec_outputs, dec_self_attn, dec_enc_attn

最后是一层线性层,输出logit,维度为[batch_size, src_len, tgt_vocab_size]

完整代码

# -*- coding:UTF-8 -*-

# author:user
# contact: [email protected]
# datetime:2021/12/22 16:32
# software: PyCharm

"""
文件说明:
    
"""

import numpy as np
import torch
import torch.nn as nn
import torch.optim as optim
import matplotlib.pyplot as plt

# S: Symbol that shows starting of decoding input
# E: Symbol that shows starting of decoding output
# P: Symbol that will fill in blank sequence if current batch data size is short than time steps

def make_batch(sentences):
    input_batch = [[src_vocab[n] for n in sentences[0].split()]]
    output_batch = [[tgt_vocab[n] for n in sentences[1].split()]]
    target_batch = [[tgt_vocab[n] for n in sentences[2].split()]]
    return torch.LongTensor(input_batch), torch.LongTensor(output_batch), torch.LongTensor(target_batch)

def get_sinusoid_encoding_table(n_position, d_model):
    def cal_angle(position, hid_idx):
        return position / np.power(10000, 2 * (hid_idx // 2) / d_model)
    def get_posi_angle_vec(position):
        return [cal_angle(position, hid_j) for hid_j in range(d_model)]

    sinusoid_table = np.array([get_posi_angle_vec(pos_i) for pos_i in range(n_position)])
    sinusoid_table[:, 0::2] = np.sin(sinusoid_table[:, 0::2])  # dim 2i
    sinusoid_table[:, 1::2] = np.cos(sinusoid_table[:, 1::2])  # dim 2i+1
    return torch.FloatTensor(sinusoid_table)

def get_attn_pad_mask(seq_q, seq_k):
    batch_size, len_q = seq_q.size()
    batch_size, len_k = seq_k.size()
    # eq(zero) is PAD token
    pad_attn_mask = seq_k.data.eq(0).unsqueeze(1)  # batch_size x 1 x len_k(=len_q), one is masking
    return pad_attn_mask.expand(batch_size, len_q, len_k)  # batch_size x len_q x len_k

def get_attn_subsequent_mask(seq):
    attn_shape = [seq.size(0), seq.size(1), seq.size(1)]
    subsequent_mask = np.triu(np.ones(attn_shape), k=1)
    subsequent_mask = torch.from_numpy(subsequent_mask).byte()
    return subsequent_mask

class ScaledDotProductAttention(nn.Module):
    def __init__(self):
        super(ScaledDotProductAttention, self).__init__()

    def forward(self, Q, K, V, attn_mask):
        scores = torch.matmul(Q, K.transpose(-1, -2)) / np.sqrt(d_k) # scores : [batch_size x n_heads x len_q(=len_k) x len_k(=len_q)]
        scores.masked_fill_(attn_mask, -1e9) # Fills elements of self tensor with value where mask is one.
        attn = nn.Softmax(dim=-1)(scores)
        context = torch.matmul(attn, V)
        return context, attn

class MultiHeadAttention(nn.Module):
    def __init__(self):
        super(MultiHeadAttention, self).__init__()
        self.W_Q = nn.Linear(d_model, d_k * n_heads)
        self.W_K = nn.Linear(d_model, d_k * n_heads)
        self.W_V = nn.Linear(d_model, d_v * n_heads)
        self.linear = nn.Linear(n_heads * d_v, d_model)
        self.layer_norm = nn.LayerNorm(d_model)

    def forward(self, Q, K, V, attn_mask):
        # q: [batch_size x len_q x d_model], k: [batch_size x len_k x d_model], v: [batch_size x len_k x d_model]
        residual, batch_size = Q, Q.size(0)
        # (B, S, D) -proj-> (B, S, D) -split-> (B, S, H, W) -trans-> (B, H, S, W)
        q_s = self.W_Q(Q).view(batch_size, -1, n_heads, d_k).transpose(1,2)  # q_s: [batch_size x n_heads x len_q x d_k]
        k_s = self.W_K(K).view(batch_size, -1, n_heads, d_k).transpose(1,2)  # k_s: [batch_size x n_heads x len_k x d_k]
        v_s = self.W_V(V).view(batch_size, -1, n_heads, d_v).transpose(1,2)  # v_s: [batch_size x n_heads x len_k x d_v]

        attn_mask = attn_mask.unsqueeze(1).repeat(1, n_heads, 1, 1) # attn_mask : [batch_size x n_heads x len_q x len_k]

        # context: [batch_size x n_heads x len_q x d_v], attn: [batch_size x n_heads x len_q(=len_k) x len_k(=len_q)]
        context, attn = ScaledDotProductAttention()(q_s, k_s, v_s, attn_mask)
        context = context.transpose(1, 2).contiguous().view(batch_size, -1, n_heads * d_v) # context: [batch_size x len_q x n_heads * d_v]
        output = self.linear(context)
        return self.layer_norm(output + residual), attn # output: [batch_size x len_q x d_model]

class PoswiseFeedForwardNet(nn.Module):
    def __init__(self):
        super(PoswiseFeedForwardNet, self).__init__()
        self.conv1 = nn.Conv1d(in_channels=d_model, out_channels=d_ff, kernel_size=1)
        self.conv2 = nn.Conv1d(in_channels=d_ff, out_channels=d_model, kernel_size=1)
        self.layer_norm = nn.LayerNorm(d_model)

    def forward(self, inputs):
        residual = inputs # inputs : [batch_size, len_q, d_model]
        output = nn.ReLU()(self.conv1(inputs.transpose(1, 2)))
        output = self.conv2(output).transpose(1, 2)
        return self.layer_norm(output + residual)

class EncoderLayer(nn.Module):
    def __init__(self):
        super(EncoderLayer, self).__init__()
        self.enc_self_attn = MultiHeadAttention()
        self.pos_ffn = PoswiseFeedForwardNet()

    def forward(self, enc_inputs, enc_self_attn_mask):
        enc_outputs, attn = self.enc_self_attn(enc_inputs, enc_inputs, enc_inputs, enc_self_attn_mask) # enc_inputs to same Q,K,V
        enc_outputs = self.pos_ffn(enc_outputs) # enc_outputs: [batch_size x len_q x d_model]
        return enc_outputs, attn

class DecoderLayer(nn.Module):
    def __init__(self):
        super(DecoderLayer, self).__init__()
        self.dec_self_attn = MultiHeadAttention()
        self.dec_enc_attn = MultiHeadAttention()
        self.pos_ffn = PoswiseFeedForwardNet()

    def forward(self, dec_inputs, enc_outputs, dec_self_attn_mask, dec_enc_attn_mask):
        dec_outputs, dec_self_attn = self.dec_self_attn(dec_inputs, dec_inputs, dec_inputs, dec_self_attn_mask)
        dec_outputs, dec_enc_attn = self.dec_enc_attn(dec_outputs, enc_outputs, enc_outputs, dec_enc_attn_mask)
        dec_outputs = self.pos_ffn(dec_outputs)
        return dec_outputs, dec_self_attn, dec_enc_attn

class Encoder(nn.Module):
    def __init__(self):
        super(Encoder, self).__init__()
        self.src_emb = nn.Embedding(src_vocab_size, d_model)
        self.pos_emb = nn.Embedding.from_pretrained(get_sinusoid_encoding_table(src_len+1, d_model),freeze=True)
        self.layers = nn.ModuleList([EncoderLayer() for _ in range(n_layers)])

    def forward(self, enc_inputs): # enc_inputs : [batch_size x source_len]
        enc_outputs = self.src_emb(enc_inputs) + self.pos_emb(torch.LongTensor([[1,2,3,4,0]]))
        enc_self_attn_mask = get_attn_pad_mask(enc_inputs, enc_inputs)
        enc_self_attns = []
        for layer in self.layers:
            enc_outputs, enc_self_attn = layer(enc_outputs, enc_self_attn_mask)
            enc_self_attns.append(enc_self_attn)
        return enc_outputs, enc_self_attns

class Decoder(nn.Module):
    def __init__(self):
        super(Decoder, self).__init__()
        self.tgt_emb = nn.Embedding(tgt_vocab_size, d_model)
        self.pos_emb = nn.Embedding.from_pretrained(get_sinusoid_encoding_table(tgt_len+1, d_model),freeze=True)
        self.layers = nn.ModuleList([DecoderLayer() for _ in range(n_layers)])

    def forward(self, dec_inputs, enc_inputs, enc_outputs): # dec_inputs : [batch_size x target_len]
        dec_outputs = self.tgt_emb(dec_inputs) + self.pos_emb(torch.LongTensor([[5,1,2,3,4]]))
        dec_self_attn_pad_mask = get_attn_pad_mask(dec_inputs, dec_inputs)
        dec_self_attn_subsequent_mask = get_attn_subsequent_mask(dec_inputs)
        dec_self_attn_mask = torch.gt((dec_self_attn_pad_mask + dec_self_attn_subsequent_mask), 0)

        dec_enc_attn_mask = get_attn_pad_mask(dec_inputs, enc_inputs)

        dec_self_attns, dec_enc_attns = [], []
        for layer in self.layers:
            dec_outputs, dec_self_attn, dec_enc_attn = layer(dec_outputs, enc_outputs, dec_self_attn_mask, dec_enc_attn_mask)
            dec_self_attns.append(dec_self_attn)
            dec_enc_attns.append(dec_enc_attn)
        return dec_outputs, dec_self_attns, dec_enc_attns

class Transformer(nn.Module):
    def __init__(self):
        super(Transformer, self).__init__()
        self.encoder = Encoder()
        self.decoder = Decoder()
        self.projection = nn.Linear(d_model, tgt_vocab_size, bias=False)
    def forward(self, enc_inputs, dec_inputs):
        enc_outputs, enc_self_attns = self.encoder(enc_inputs)
        dec_outputs, dec_self_attns, dec_enc_attns = self.decoder(dec_inputs, enc_inputs, enc_outputs)
        dec_logits = self.projection(dec_outputs) # dec_logits : [batch_size x src_vocab_size x tgt_vocab_size]
        return dec_logits.view(-1, dec_logits.size(-1)), enc_self_attns, dec_self_attns, dec_enc_attns

def showgraph(attn):
    attn = attn[-1].squeeze(0)[0]
    attn = attn.squeeze(0).data.numpy()
    fig = plt.figure(figsize=(n_heads, n_heads)) # [n_heads, n_heads]
    ax = fig.add_subplot(1, 1, 1)
    ax.matshow(attn, cmap='viridis')
    ax.set_xticklabels(['']+sentences[0].split(), fontdict={'fontsize': 14}, rotation=90)
    ax.set_yticklabels(['']+sentences[2].split(), fontdict={'fontsize': 14})
    plt.show()

if __name__ == '__main__':
    sentences = ['ich mochte ein bier P', 'S i want a beer', 'i want a beer E']

    # Transformer Parameters
    # Padding Should be Zero
    src_vocab = {'P': 0, 'ich': 1, 'mochte': 2, 'ein': 3, 'bier': 4}
    src_vocab_size = len(src_vocab)

    tgt_vocab = {'P': 0, 'i': 1, 'want': 2, 'a': 3, 'beer': 4, 'S': 5, 'E': 6}
    number_dict = {i: w for i, w in enumerate(tgt_vocab)}
    tgt_vocab_size = len(tgt_vocab)

    src_len = 5 # length of source
    tgt_len = 5 # length of target

    d_model = 512  # Embedding Size
    d_ff = 2048  # FeedForward dimension
    d_k = d_v = 64  # dimension of K(=Q), V
    n_layers = 6  # number of Encoder of Decoder Layer
    n_heads = 8  # number of heads in Multi-Head Attention

    model = Transformer()

    criterion = nn.CrossEntropyLoss()
    optimizer = optim.Adam(model.parameters(), lr=0.001)

    enc_inputs, dec_inputs, target_batch = make_batch(sentences)

    for epoch in range(20):
        optimizer.zero_grad()
        outputs, enc_self_attns, dec_self_attns, dec_enc_attns = model(enc_inputs, dec_inputs)
        loss = criterion(outputs, target_batch.contiguous().view(-1))
        print('Epoch:', '%04d' % (epoch + 1), 'cost =', '{:.6f}'.format(loss))
        loss.backward()
        optimizer.step()

    # Test
    predict, _, _, _ = model(enc_inputs, dec_inputs)
    predict = predict.data.max(1, keepdim=True)[1]
    print(sentences[0], '->', [number_dict[n.item()] for n in predict.squeeze()])

    print('first head of last state enc_self_attns')
    showgraph(enc_self_attns)

    print('first head of last state dec_self_attns')
    showgraph(dec_self_attns)

    print('first head of last state dec_enc_attns')
    showgraph(dec_enc_attns)

你可能感兴趣的:(预训练模型,实战,transformer,深度学习,人工智能)