ROCm上运行Transformer

10.7. Transformer --- 动手学深度学习 2.0.0 documentation (d2l.ai)

代码

import math
import pandas as pd
import torch
from torch import nn
from d2l import torch as d2l

#@save
class PositionWiseFFN(nn.Module):
    """基于位置的前馈网络"""
    def __init__(self, ffn_num_input, ffn_num_hiddens, ffn_num_outputs,
                 **kwargs):
        super(PositionWiseFFN, self).__init__(**kwargs)
        self.dense1 = nn.Linear(ffn_num_input, ffn_num_hiddens)
        self.relu = nn.ReLU()
        self.dense2 = nn.Linear(ffn_num_hiddens, ffn_num_outputs)

    def forward(self, X):
        return self.dense2(self.relu(self.dense1(X)))

ffn = PositionWiseFFN(4, 4, 8)
ffn.eval()
ffn(torch.ones((2, 3, 4)))[0]

ln = nn.LayerNorm(2)
bn = nn.BatchNorm1d(2)
X = torch.tensor([[1, 2], [2, 3]], dtype=torch.float32)
# 在训练模式下计算X的均值和方差
print('layer norm:', ln(X), '\nbatch norm:', bn(X))

#@save
class AddNorm(nn.Module):
    """残差连接后进行层规范化"""
    def __init__(self, normalized_shape, dropout, **kwargs):
        super(AddNorm, self).__init__(**kwargs)
        self.dropout = nn.Dropout(dropout)
        self.ln = nn.LayerNorm(normalized_shape)

    def forward(self, X, Y):
        return self.ln(self.dropout(Y) + X)

add_norm = AddNorm([3, 4], 0.5)
add_norm.eval()
add_norm(torch.ones((2, 3, 4)), torch.ones((2, 3, 4))).shape

#@save
class EncoderBlock(nn.Module):
    """Transformer编码器块"""
    def __init__(self, key_size, query_size, value_size, num_hiddens,
                 norm_shape, ffn_num_input, ffn_num_hiddens, num_heads,
                 dropout, use_bias=False, **kwargs):
        super(EncoderBlock, self).__init__(**kwargs)
        self.attention = d2l.MultiHeadAttention(
            key_size, query_size, value_size, num_hiddens, num_heads, dropout,
            use_bias)
        self.addnorm1 = AddNorm(norm_shape, dropout)
        self.ffn = PositionWiseFFN(
            ffn_num_input, ffn_num_hiddens, num_hiddens)
        self.addnorm2 = AddNorm(norm_shape, dropout)

    def forward(self, X, valid_lens):
        Y = self.addnorm1(X, self.attention(X, X, X, valid_lens))
        return self.addnorm2(Y, self.ffn(Y))

X = torch.ones((2, 100, 24))
valid_lens = torch.tensor([3, 2])
encoder_blk = EncoderBlock(24, 24, 24, 24, [100, 24], 24, 48, 8, 0.5)
encoder_blk.eval()
encoder_blk(X, valid_lens).shape

#@save
class TransformerEncoder(d2l.Encoder):
    """Transformer编码器"""
    def __init__(self, vocab_size, key_size, query_size, value_size,
                 num_hiddens, norm_shape, ffn_num_input, ffn_num_hiddens,
                 num_heads, num_layers, dropout, use_bias=False, **kwargs):
        super(TransformerEncoder, self).__init__(**kwargs)
        self.num_hiddens = num_hiddens
        self.embedding = nn.Embedding(vocab_size, num_hiddens)
        self.pos_encoding = d2l.PositionalEncoding(num_hiddens, dropout)
        self.blks = nn.Sequential()
        for i in range(num_layers):
            self.blks.add_module("block"+str(i),
                EncoderBlock(key_size, query_size, value_size, num_hiddens,
                             norm_shape, ffn_num_input, ffn_num_hiddens,
                             num_heads, dropout, use_bias))

    def forward(self, X, valid_lens, *args):
        # 因为位置编码值在-1和1之间，
        # 因此嵌入值乘以嵌入维度的平方根进行缩放，
        # 然后再与位置编码相加。
        X = self.pos_encoding(self.embedding(X) * math.sqrt(self.num_hiddens))
        self.attention_weights = [None] * len(self.blks)
        for i, blk in enumerate(self.blks):
            X = blk(X, valid_lens)
            self.attention_weights[
                i] = blk.attention.attention.attention_weights
        return X

encoder = TransformerEncoder(
    200, 24, 24, 24, 24, [100, 24], 24, 48, 8, 2, 0.5)
encoder.eval()
encoder(torch.ones((2, 100), dtype=torch.long), valid_lens).shape

class DecoderBlock(nn.Module):
    """解码器中第i个块"""
    def __init__(self, key_size, query_size, value_size, num_hiddens,
                 norm_shape, ffn_num_input, ffn_num_hiddens, num_heads,
                 dropout, i, **kwargs):
        super(DecoderBlock, self).__init__(**kwargs)
        self.i = i
        self.attention1 = d2l.MultiHeadAttention(
            key_size, query_size, value_size, num_hiddens, num_heads, dropout)
        self.addnorm1 = AddNorm(norm_shape, dropout)
        self.attention2 = d2l.MultiHeadAttention(
            key_size, query_size, value_size, num_hiddens, num_heads, dropout)
        self.addnorm2 = AddNorm(norm_shape, dropout)
        self.ffn = PositionWiseFFN(ffn_num_input, ffn_num_hiddens,
                                   num_hiddens)
        self.addnorm3 = AddNorm(norm_shape, dropout)

    def forward(self, X, state):
        enc_outputs, enc_valid_lens = state[0], state[1]
        # 训练阶段，输出序列的所有词元都在同一时间处理，
        # 因此state[2][self.i]初始化为None。
        # 预测阶段，输出序列是通过词元一个接着一个解码的，
        # 因此state[2][self.i]包含着直到当前时间步第i个块解码的输出表示
        if state[2][self.i] is None:
            key_values = X
        else:
            key_values = torch.cat((state[2][self.i], X), axis=1)
        state[2][self.i] = key_values
        if self.training:
            batch_size, num_steps, _ = X.shape
            # dec_valid_lens的开头:(batch_size,num_steps),
            # 其中每一行是[1,2,...,num_steps]
            dec_valid_lens = torch.arange(
                1, num_steps + 1, device=X.device).repeat(batch_size, 1)
        else:
            dec_valid_lens = None

        # 自注意力
        X2 = self.attention1(X, key_values, key_values, dec_valid_lens)
        Y = self.addnorm1(X, X2)
        # 编码器－解码器注意力。
        # enc_outputs的开头:(batch_size,num_steps,num_hiddens)
        Y2 = self.attention2(Y, enc_outputs, enc_outputs, enc_valid_lens)
        Z = self.addnorm2(Y, Y2)
        return self.addnorm3(Z, self.ffn(Z)), state

decoder_blk = DecoderBlock(24, 24, 24, 24, [100, 24], 24, 48, 8, 0.5, 0)
decoder_blk.eval()
X = torch.ones((2, 100, 24))
state = [encoder_blk(X, valid_lens), valid_lens, [None]]
decoder_blk(X, state)[0].shape

class TransformerDecoder(d2l.AttentionDecoder):
    def __init__(self, vocab_size, key_size, query_size, value_size,
                 num_hiddens, norm_shape, ffn_num_input, ffn_num_hiddens,
                 num_heads, num_layers, dropout, **kwargs):
        super(TransformerDecoder, self).__init__(**kwargs)
        self.num_hiddens = num_hiddens
        self.num_layers = num_layers
        self.embedding = nn.Embedding(vocab_size, num_hiddens)
        self.pos_encoding = d2l.PositionalEncoding(num_hiddens, dropout)
        self.blks = nn.Sequential()
        for i in range(num_layers):
            self.blks.add_module("block"+str(i),
                DecoderBlock(key_size, query_size, value_size, num_hiddens,
                             norm_shape, ffn_num_input, ffn_num_hiddens,
                             num_heads, dropout, i))
        self.dense = nn.Linear(num_hiddens, vocab_size)

    def init_state(self, enc_outputs, enc_valid_lens, *args):
        return [enc_outputs, enc_valid_lens, [None] * self.num_layers]

    def forward(self, X, state):
        X = self.pos_encoding(self.embedding(X) * math.sqrt(self.num_hiddens))
        self._attention_weights = [[None] * len(self.blks) for _ in range (2)]
        for i, blk in enumerate(self.blks):
            X, state = blk(X, state)
            # 解码器自注意力权重
            self._attention_weights[0][
                i] = blk.attention1.attention.attention_weights
            # "编码器－解码器"自注意力权重
            self._attention_weights[1][
                i] = blk.attention2.attention.attention_weights
        return self.dense(X), state

    @property
    def attention_weights(self):
        return self._attention_weights

num_hiddens, num_layers, dropout, batch_size, num_steps = 32, 2, 0.1, 64, 10
lr, num_epochs, device = 0.005, 200, d2l.try_gpu()
ffn_num_input, ffn_num_hiddens, num_heads = 32, 64, 4
key_size, query_size, value_size = 32, 32, 32
norm_shape = [32]

train_iter, src_vocab, tgt_vocab = d2l.load_data_nmt(batch_size, num_steps)

encoder = TransformerEncoder(
    len(src_vocab), key_size, query_size, value_size, num_hiddens,
    norm_shape, ffn_num_input, ffn_num_hiddens, num_heads,
    num_layers, dropout)
decoder = TransformerDecoder(
    len(tgt_vocab), key_size, query_size, value_size, num_hiddens,
    norm_shape, ffn_num_input, ffn_num_hiddens, num_heads,
    num_layers, dropout)
net = d2l.EncoderDecoder(encoder, decoder)
d2l.train_seq2seq(net, train_iter, lr, num_epochs, tgt_vocab, device)

engs = ['go .', "i lost .", 'he\'s calm .', 'i\'m home .']
fras = ['va !', 'j\'ai perdu .', 'il est calme .', 'je suis chez moi .']
for eng, fra in zip(engs, fras):
    translation, dec_attention_weight_seq = d2l.predict_seq2seq(
        net, eng, src_vocab, tgt_vocab, num_steps, device, True)
    print(f'{eng} => {translation}, ',
          f'bleu {d2l.bleu(translation, fra, k=2):.3f}')

enc_attention_weights = torch.cat(net.encoder.attention_weights, 0).reshape((num_layers, num_heads,
    -1, num_steps))
enc_attention_weights.shape

d2l.show_heatmaps(
    enc_attention_weights.cpu(), xlabel='Key positions',
    ylabel='Query positions', titles=['Head %d' % i for i in range(1, 5)],
    figsize=(7, 3.5))

dec_attention_weights_2d = [head[0].tolist()
                            for step in dec_attention_weight_seq
                            for attn in step for blk in attn for head in blk]
dec_attention_weights_filled = torch.tensor(
    pd.DataFrame(dec_attention_weights_2d).fillna(0.0).values)
dec_attention_weights = dec_attention_weights_filled.reshape((-1, 2, num_layers, num_heads, num_steps))
dec_self_attention_weights, dec_inter_attention_weights = \
    dec_attention_weights.permute(1, 2, 3, 0, 4)
dec_self_attention_weights.shape, dec_inter_attention_weights.shape

# Plusonetoincludethebeginning-of-sequencetoken
d2l.show_heatmaps(
    dec_self_attention_weights[:, :, :, :len(translation.split()) + 1],
    xlabel='Key positions', ylabel='Query positions',
    titles=['Head %d' % i for i in range(1, 5)], figsize=(7, 3.5))

d2l.show_heatmaps(
    dec_inter_attention_weights, xlabel='Key positions',
    ylabel='Query positions', titles=['Head %d' % i for i in range(1, 5)],
    figsize=(7, 3.5))

代码解析

这段代码定义了基于 Transformer 架构的编码器-解码器网络，并应用在一个简单的机器翻译任务中。以下是对此代码的主要部分的解析：

定义基于位置的前馈网络 (PositionWiseFFN)

class PositionWiseFFN(nn.Module):
    def __init__(self, ffn_num_input, ffn_num_hiddens, ffn_num_outputs, **kwargs):
        super(PositionWiseFFN, self).__init__(**kwargs)
        self.dense1 = nn.Linear(ffn_num_input, ffn_num_hiddens)
        self.relu = nn.ReLU()
        self.dense2 = nn.Linear(ffn_num_hiddens, ffn_num_outputs)
    
    def forward(self, X):
        return self.dense2(self.relu(self.dense1(X)))

此类定义了一个两层的前馈神经网络，输入通过一个全连接层、ReLU激活函数、再通过另一个全连接层传播。

定义残差连接和层规范化 (AddNorm)

class AddNorm(nn.Module):
    def __init__(self, normalized_shape, dropout, **kwargs):
        super(AddNorm, self).__init__(**kwargs)
        self.dropout = nn.Dropout(dropout)
        self.ln = nn.LayerNorm(normalized_shape)
    
    def forward(self, X, Y):
        return self.ln(self.dropout(Y) + X)

在 Transformer 结构中，用于组合输入和子层的输出以及进行层规范化的类。首先应用 dropout，然后将结果与输入 X 相加并进行层规范化。

定义 Transformer 的编码器块 (EncoderBlock)

class EncoderBlock(nn.Module):
    def __init__(self, key_size, query_size, value_size, num_hiddens, norm_shape, ffn_num_input, ffn_num_hiddens, num_heads, dropout, use_bias=False, **kwargs):
        super(EncoderBlock, self).__init__(**kwargs)
        self.attention = d2l.MultiHeadAttention(key_size, query_size, value_size, num_hiddens, num_heads, dropout, use_bias)
        self.addnorm1 = AddNorm(norm_shape, dropout)
        self.ffn = PositionWiseFFN(ffn_num_input, ffn_num_hiddens, num_hiddens)
        self.addnorm2 = AddNorm(norm_shape, dropout)
    
    def forward(self, X, valid_lens):
        Y = self.addnorm1(X, self.attention(X, X, X, valid_lens))
        return self.addnorm2(Y, self.ffn(Y))

编码器块包括两个子层：多头注意力层和基于位置的前馈网络，每个子层周围都有残差连接和层规范化。

定义 Transformer 的编码器 (TransformerEncoder)

class TransformerEncoder(d2l.Encoder):
    # ...
    def __init__(self, vocab_size, key_size, query_size, value_size, num_hiddens, norm_shape, ffn_num_input, ffn_num_hiddens, num_heads, num_layers, dropout, use_bias=False, **kwargs):
        # ...
        self.blks = nn.Sequential()
        for i in range(num_layers):
            # 添加多个编码器块
            self.blks.add_module("block"+str(i),
                EncoderBlock(key_size, query_size, value_size, num_hiddens, norm_shape, ffn_num_input, ffn_num_hiddens, num_heads, dropout, use_bias))
    
    def forward(self, X, valid_lens, *args):
        # ...
        for i, blk in enumerate(self.blks):
            X = blk(X, valid_lens)
        return X

Transformer 编码器由多个相同的编码器块组成。输入首先通过嵌入层和位置编码，然后传递给多个编码器块。

训练编码器-解码器模型并进行翻译

encoder = TransformerEncoder( ... )
decoder = TransformerDecoder( ... )
net = d2l.EncoderDecoder(encoder, decoder)
d2l.train_seq2seq(net, train_iter, lr, num_epochs, tgt_vocab, device)

for eng, fra in zip(engs, fras):
    translation, _ = d2l.predict_seq2seq(net, eng, src_vocab, tgt_vocab, num_steps, device, True)
    print(f'{eng} => {translation}, ', f'bleu {d2l.bleu(translation, fra, k=2):.3f}')

编码器和解码器被创建并组合成一个编码器-解码器模型 (seq2seq)。训练过程中使用的是机器翻译的数据集。训练完成后，进行简单的翻译任务，显示原句子、翻译后的句子以及 BLEU 分数。 BLEU 分数是评估句子质量的指标之一。

注意力权重可视化

在代码的后面部分，实现了对注意力权重的可视化。首先是编码器的注意力权重：

enc_attention_weights = torch.cat(net.encoder.attention_weights, 0).reshape((num_layers, num_heads, -1, num_steps))

这段代码将编码器中所有关注权重连结（`torch.cat`）后重塑形状，以便能够针对每个头（head）为其绘制热图。

然后，使用 d2l.show_heatmaps 函数显示热图：

d2l.show_heatmaps(
    enc_attention_weights.cpu(), xlabel='Key positions',
    ylabel='Query positions', titles=['Head %d' % i for i in range(1, num_heads + 1)],
    figsize=(7, 3.5))

这会为每个注意力头展示一个热图，显示在给定查询位置时，编码器层中各个键位的注意力分布情况。

接下来，代码段处理解码器的两组注意力权重：解码器的自注意力权重和编码器-解码器之间的注意力权重。解码器的自注意力权重表现在给定解码器中前一个词元时，下一个词元的关注点。

d2l.show_heatmaps(
    dec_self_attention_weights[:, :, :, :len(translation.split()) + 1],
    xlabel='Key positions', ylabel='Query positions',
    titles=['Head %d' % i for i in range(1, num_heads + 1)], figsize=(7, 3.5))

显示了解码器自注意力权重的热图，其中包括了当前词元和前面所有已生成的词元之间的关注关系。

最后，编码器-解码器注意力权重表现在解码器如何根据编码器的输出来选择关注输入句子的哪些词元。

d2l.show_heatmaps(
    dec_inter_attention_weights, xlabel='Key positions',
    ylabel='Query positions', titles=['Head %d' % i for i in range(1, num_heads + 1)],
    figsize=(7, 3.5))

这将为每个头展示编码器-解码器注意力的热图，显示了当生成翻译的每个词元时，模型对输入句子中不同词元的关注分布。

总之，此段代码整体上展示了如何构建 Transformer 模型，训练它进行机器翻译任务，并对其内部的注意力机制进行可视化，帮助我们了解模型是如何工作的。注意，在实际使用此模型时，应该下载一个大型的预训织模型或使用大型数据集来训练足够的时间以获得更佳的翻译效果。此段代码意在演示 Transformer 架构的实现和基础使用方式。