Transformer模型自提出以来，已经成为深度学习领域，尤其是自然语言处理（NLP）中的一种革命性模型。与传统的循环神经网络（RNN）和长短期记忆网络（LSTM）不同，Transformer完全依赖于注意力机制来捕捉序列中的依赖关系。这使得它能够更高效地处理长序列数据。在本文中，我们将详细介绍Transformer模型的基本原理，并使用Python和TensorFlow/Keras实现一个简单的Transformer模型。
1. Transformer模型简介
Transformer模型由编码器（Encoder）和解码器（Decoder）组成，每个编码器和解码器层都由多头自注意力机制和前馈神经网络（Feed-Forward Neural Network）组成。
1.1 编码器（Encoder）
编码器的主要组件包括：
- 自注意力层（Self-Attention Layer）：计算序列中每个位置对其他位置的注意力分数。
- 前馈神经网络（Feed-Forward Neural Network）：对每个位置的表示进行独立的非线性变换。
1.2 解码器（Decoder）
解码器与编码器类似，但有额外的编码器-解码器注意力层，用于捕捉解码器输入与编码器输出之间的关系。
1.3 注意力机制
注意力机制的核心公式如下：
2. 使用Python和TensorFlow/Keras实现Transformer模型
下面我们将使用Python和TensorFlow/Keras实现一个简单的Transformer模型，用于机器翻译任务。
2.1 安装TensorFlow
首先，确保安装了TensorFlow：
代码语言：javascript复制pip install tensorflow2.2 数据准备
我们使用TensorFlow内置的英文-德文翻译数据集。
代码语言：javascript复制import tensorflow as tf
import tensorflow_datasets as tfds
# 加载数据集
examples, metadata = tfds.load('ted_hrlr_translate/pt_to_en', with_info=True, as_supervised=True)
train_examples, val_examples = examples['train'], examples['validation']
# 准备tokenizer
tokenizer_en = tfds.deprecated.text.SubwordTextEncoder.build_from_corpus(
(en.numpy() for pt, en in train_examples), target_vocab_size=2**13)
tokenizer_pt = tfds.deprecated.text.SubwordTextEncoder.build_from_corpus(
(pt.numpy() for pt, en in train_examples), target_vocab_size=2**13)
# 定义tokenizer函数
def encode(lang1, lang2):
lang1 = [tokenizer_pt.vocab_size] tokenizer_pt.encode(
lang1.numpy()) [tokenizer_pt.vocab_size 1]
lang2 = [tokenizer_en.vocab_size] tokenizer_en.encode(
lang2.numpy()) [tokenizer_en.vocab_size 1]
return lang1, lang2
def tf_encode(pt, en):
result_pt, result_en = tf.py_function(encode, [pt, en], [tf.int64, tf.int64])
result_pt.set_shape([None])
result_en.set_shape([None])
return result_pt, result_en
# 设置缓冲区大小
BUFFER_SIZE = 20000
BATCH_SIZE = 64
# 预处理数据
train_dataset = train_examples.map(tf_encode)
train_dataset = train_dataset.cache()
train_dataset = train_dataset.shuffle(BUFFER_SIZE).padded_batch(BATCH_SIZE)
train_dataset = train_dataset.prefetch(tf.data.experimental.AUTOTUNE)
val_dataset = val_examples.map(tf_encode)
val_dataset = val_dataset.padded_batch(BATCH_SIZE)2.3 实现Transformer模型组件
我们首先实现一些基础组件，如位置编码（Positional Encoding）和多头注意力（Multi-Head Attention）。
2.3.1 位置编码
位置编码用于在序列中加入位置信息。
代码语言：javascript复制import numpy as np
def get_angles(pos, i, d_model):
angle_rates = 1 / np.power(10000, (2 * (i // 2)) / np.float32(d_model))
return pos * angle_rates
def positional_encoding(position, d_model):
angle_rads = get_angles(np.arange(position)[:, np.newaxis],
np.arange(d_model)[np.newaxis, :],
d_model)
angle_rads[:, 0::2] = np.sin(angle_rads[:, 0::2])
angle_rads[:, 1::2] = np.cos(angle_rads[:, 1::2])
pos_encoding = angle_rads[np.newaxis, ...]
return tf.cast(pos_encoding, dtype=tf.float32)2.3.2 多头注意力
代码语言：javascript复制class MultiHeadAttention(tf.keras.layers.Layer):
def __init__(self, d_model, num_heads):
super(MultiHeadAttention, self).__init__()
self.num_heads = num_heads
self.d_model = d_model
assert d_model % self.num_heads == 0
self.depth = d_model // self.num_heads
self.wq = tf.keras.layers.Dense(d_model)
self.wk = tf.keras.layers.Dense(d_model)
self.wv = tf.keras.layers.Dense(d_model)
self.dense = tf.keras.layers.Dense(d_model)
def split_heads(self, x, batch_size):
x = tf.reshape(x, (batch_size, -1, self.num_heads, self.depth))
return tf.transpose(x, perm=[0, 2, 1, 3])
def call(self, v, k, q, mask):
batch_size = tf.shape(q)[0]
q = self.wq(q) # (batch_size, seq_len, d_model)
k = self.wk(k) # (batch_size, seq_len, d_model)
v = self.wv(v) # (batch_size, seq_len, d_model)
q = self.split_heads(q, batch_size) # (batch_size, num_heads, seq_len_q, depth)
k = self.split_heads(k, batch_size) # (batch_size, num_heads, seq_len_k, depth)
v = self.split_heads(v, batch_size) # (batch_size, num_heads, seq_len_v, depth)
scaled_attention, attention_weights = self.scaled_dot_product_attention(q, k, v, mask)
scaled_attention = tf.transpose(scaled_attention, perm=[0, 2, 1, 3])
concat_attention = tf.reshape(scaled_attention, (batch_size, -1, self.d_model))
output = self.dense(concat_attention)
return output, attention_weights
def scaled_dot_product_attention(self, q, k, v, mask):
matmul_qk = tf.matmul(q, k, transpose_b=True)
dk = tf.cast(tf.shape(k)[-1], tf.float32)
scaled_attention_logits = matmul_qk / tf.math.sqrt(dk)
if mask is not None:
scaled_attention_logits = (mask * -1e9)
attention_weights = tf.nn.softmax(scaled_attention_logits, axis=-1)
output = tf.matmul(attention_weights, v)
return output, attention_weights2.4 构建Transformer模型
代码语言：javascript复制def point_wise_feed_forward_network(d_model, dff):
return tf.keras.Sequential([
tf.keras.layers.Dense(dff, activation='relu'),
tf.keras.layers.Dense(d_model)
])
class EncoderLayer(tf.keras.layers.Layer):
def __init__(self, d_model, num_heads, dff, rate=0.1):
super(EncoderLayer, self).__init__()
self.mha = MultiHeadAttention(d_model, num_heads)
self.ffn = point_wise_feed_forward_network(d_model, dff)
self.layernorm1 = tf.keras.layers.LayerNormalization(epsilon=1e-6)
self.layernorm2 = tf.keras.layers.LayerNormalization(epsilon=1e-6)
self.dropout1 = tf.keras.layers.Dropout(rate)
self.dropout2 = tf.keras.layers.Dropout(rate)
def call(self, x, training, mask):
attn_output, _ = self.mha(x, x, x, mask)
attn_output = self.dropout1(attn_output, training=training)
out1 = self.layernorm1(x attn_output)
ffn_output = self.ffn(out1)
ffn_output = self.dropout2(ffn_output, training=training)
out2 = self.layernorm2(out1 ffn_output)
return out2
class DecoderLayer(tf.keras.layers.Layer):
def __init__(self, d_model, num_heads, dff, rate=0.1):
super(DecoderLayer, self).__init__()
self.mha1 = MultiHeadAttention(d_model, num_heads)
self.mha2 = MultiHeadAttention(d_model, num_heads)
self.ffn = point_wise_feed_forward_network(d_model, dff)
self.layernorm1 = tf.keras.layers.LayerNormalization(epsilon=1e-6)
self.layernorm2 = tf.keras.layers.LayerNormalization(epsilon=1e-6)
self.layernorm3 = tf.keras.layers.LayerNormalization(epsilon=1e-6)
self.dropout1 = tf.keras.layers.Dropout(rate)
self.dropout2 = tf.keras.layers.Dropout(rate)
self.dropout3 = tf.keras.layers.Dropout(rate)
def call(self, x, enc_output, training, look_ahead_mask, padding_mask):
attn1, attn_weights_block1 = self.mha1(x, x, x, look_ahead_mask)
attn1 = self.dropout1(attn1, training=training)
out1 = self.layernorm1(x attn1)
attn2, attn_weights_block2 = self.mha2(enc_output, enc_output, out1, padding_mask)
attn2 = self.dropout2(attn2, training=training)
out2 = self.layernorm2(out1 attn2)
ffn_output = self.ffn(out2)
ffn_output = self.dropout3(ffn_output, training=training)
out3 = self.layernorm3(out2 ffn_output)
return out3, attn_weights_block1, attn_weights_block2
class Encoder(tf.keras.layers.Layer):
def __init__(self, num_layers, d_model, num_heads, dff, input_vocab_size, maximum_position_encoding, rate=0.1):
super(Encoder, self).__init__()
self.d_model = d_model
self.num_layers = num_layers
self.embedding = tf.keras.layers.Embedding(input_vocab_size, d_model)
self.pos_encoding = positional_encoding(maximum_position_encoding, d_model)
self.enc_layers = [EncoderLayer(d_model, num_heads, dff, rate) for _ in range(num_layers)]
self.dropout = tf.keras.layers.Dropout(rate)
def call(self, x, training, mask):
seq_len = tf.shape(x)[1]
x = self.embedding(x)
x *= tf.math.sqrt(tf.cast(self.d_model, tf.float32))
x = self.pos_encoding[:, :seq_len, :]
x = self.dropout(x, training=training)
for i in range(self.num_layers):
x = self.enc_layers[i](x, training, mask)
return x
class Decoder(tf.keras.layers.Layer):
def __init__(self, num_layers, d_model, num_heads, dff, target_vocab_size, maximum_position_encoding, rate=0.1):
super(Decoder, self).__init__()
self.d_model = d_model
self.num_layers = num_layers
self.embedding = tf.keras.layers.Embedding(target_vocab_size, d_model)
self.pos_encoding = positional_encoding(maximum_position_encoding, d_model)
self.dec_layers = [DecoderLayer(d_model, num_heads, dff, rate) for _ in range(num_layers)]
self.dropout = tf.keras.layers.Dropout(rate)
def call(self, x, enc_output, training, look_ahead_mask, padding_mask):
seq_len = tf.shape(x)[1]
attention_weights = {}
x = self.embedding(x)
x *= tf.math.sqrt(tf.cast(self.d_model, tf.float32))
x = self.pos_encoding[:, :seq_len, :]
x = self.dropout(x, training=training)
for i in range(self.num_layers):
x, block1, block2 = self.dec_layers[i](x, enc_output, training, look_ahead_mask, padding_mask)
attention_weights[f'decoder_layer{i 1}_block1'] = block1
attention_weights[f'decoder_layer{i 1}_block2'] = block2
return x, attention_weights
class Transformer(tf.keras.Model):
def __init__(self, num_layers, d_model, num_heads, dff, input_vocab_size, target_vocab_size, pe_input, pe_target, rate=0.1):
super(Transformer, self).__init__()
self.encoder = Encoder(num_layers, d_model, num_heads, dff, input_vocab_size, pe_input, rate)
self.decoder = Decoder(num_layers, d_model, num_heads, dff, target_vocab_size, pe_target, rate)
self.final_layer = tf.keras.layers.Dense(target_vocab_size)
def call(self, inp, tar, training, enc_padding_mask, look_ahead_mask, dec_padding_mask):
enc_output = self.encoder(inp, training, enc_padding_mask)
dec_output, attention_weights = self.decoder(tar, enc_output, training, look_ahead_mask, dec_padding_mask)
final_output = self.final_layer(dec_output)
return final_output, attention_weights
# 设置Transformer参数
num_layers = 4
d_model = 128
dff = 512
num_heads = 8
input_vocab_size = tokenizer_pt.vocab_size 2
target_vocab_size = tokenizer_en.vocab_size 2
dropout_rate = 0.1
# 创建Transformer模型
transformer = Transformer(num_layers, d_model, num_heads, dff, input_vocab_size, target_vocab_size, pe_input=1000, pe_target=1000, rate=dropout_rate)2.5 定义损失函数和优化器
代码语言：javascript复制loss_object = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True, reduction='none')
def loss_function(real, pred):
mask = tf.math.logical_not(tf.math.equal(real, 0))
loss_ = loss_object(real, pred)
mask = tf.cast(mask, dtype=loss_.dtype)
loss_ *= mask
return tf.reduce_sum(loss_)/tf.reduce_sum(mask)
optimizer = tf.keras.optimizers.Adam(learning_rate=0.001)2.6 训练模型
代码语言：javascript复制# 定义train_step
@tf.function
def train_step(inp, tar):
tar_inp = tar[:, :-1]
tar_real = tar[:, 1:]
enc_padding_mask, look_ahead_mask, dec_padding_mask = create_masks(inp, tar_inp)
with tf.GradientTape() as tape:
predictions, _ = transformer(inp, tar_inp, True, enc_padding_mask, look_ahead_mask, dec_padding_mask)
loss = loss_function(tar_real, predictions)
gradients = tape.gradient(loss, transformer.trainable_variables)
optimizer.apply_gradients(zip(gradients, transformer.trainable_variables))
return loss
# 训练模型
EPOCHS = 20
for epoch in range(EPOCHS):
total_loss = 0
for (batch, (inp, tar)) in enumerate(train_dataset):
batch_loss = train_step(inp, tar)
total_loss = batch_loss
print(f'Epoch {epoch 1}, Loss: {total_loss/len(train_dataset)}')3. 总结
在本文中，我们详细介绍了Transformer模型的基本原理，并使用Python和TensorFlow/Keras实现了一个简单的Transformer模型。通过本文的教程，希望你能够理解Transformer模型的工作原理和实现方法，并能够应用于自己的任务中。随着对Transformer模型的理解加深，你可以尝试实现更复杂的变种，如BERT和GPT等。
