model.py

import torch
import torch.nn as nn
import math


class InputEmbeddings(nn.Module):

    def __init__(self, d_model: int, vocab_size: int):
        super(InputEmbeddings, self).__init__()
        self.d_model = d_model
        self.vocab_size = vocab_size
        self.embedding = nn.Embedding(vocab_size, d_model)

    def forward(self, x):
        return self.embedding(x) * math.sqrt(self.d_model)


class PositionalEncoding(nn.Module):

    def __init__(self, d_model: int, seq_len: int, dropout: float):
        super(PositionalEncoding, self).__init__()
        self.d_model = d_model
        self.seq_len = seq_len
        self.dropout = nn.Dropout(dropout)

        # Compute the positional encodings once in a matrix of shape (seq_len,
        # d_model)
        pe = torch.zeros(seq_len, d_model)

        # Create a vector of positions (seq_len, 1)
        position = torch.arange(0, seq_len, dtype=torch.float).unsqueeze(1)
        # Create a vector of shape (d_model)
        div_term = torch.exp(
            torch.arange(0, d_model, 2).float()
            * (-math.log(10000.0) / d_model)  # (d_model / 2,)
        )

        # Compute the positional encodings for the even indices using sin
        # function sin(position * (10000 ** (2i / d_model))
        pe[:, 0::2] = torch.sin(position * div_term)

        # Compute the positional encodings for the odd indices using cos
        # function cos(position * (10000 ** (2i / d_model))
        pe[:, 1::2] = torch.cos(position * div_term)

        # Add a batch dimension to the positional encodings
        pe = pe.unsqueeze(0)

        # Register the positional encodings as a buffer
        self.register_buffer("pe", pe)

    def forward(self, x):
        # (batch, seq_len, d_model)
        x = x + (self.pe[:, : x.shape[1], :]).requires_grad_(False)
        return self.dropout(x)


class LayerNormalization(nn.Module):

    def __init__(self, features: int, eps: float = 10**-6):
        super(LayerNormalization, self).__init__()
        self.eps = eps
        # nn.Parameter makes the multiplicative parameter trainable
        self.alpha = nn.Parameter(torch.ones(features))
        # nn.Parameter makes the additive parameter trainable
        self.bias = nn.Parameter(torch.zeros(features))

    def forward(self, x):
        # (batch, seq_len, 1)
        mean = x.mean(dim=-1, keepdim=True)
        # (batch, seq_len, 1)
        std = x.std(dim=-1, keepdim=True)
        return self.alpha * (x - mean) / (std + self.eps) + self.bias


class FeedForwardBlock(nn.Module):

    def __init__(self, d_model: int, d_ff: int, dropout: float):
        super(FeedForwardBlock, self).__init__()
        # First linear layer with W_1 and b_1 parameters
        self.fc1 = nn.Linear(d_model, d_ff)
        # Dropout layer
        self.dropout = nn.Dropout(dropout)
        # Second linear layer with W_2 and b_2 parameters
        self.fc2 = nn.Linear(d_ff, d_model)

    def forward(self, x):
        # Apply the first linear layer (batch, seq_len, d_model) -> (batch,
        # seq_len, d_ff)
        x = self.fc1(x)
        # Apply ReLU activation function
        x = torch.relu(x)
        # Apply dropout
        x = self.dropout(x)
        # Apply the second linear layer (batch, seq_len, d_ff) -> (batch,
        # seq_len, d_model)
        x = self.fc2(x)
        return x


class MultiHeadAttentionBlock(nn.Module):

    def __init__(self, d_model: int, h: int, dropout: float):
        super(MultiHeadAttentionBlock, self).__init__()
        self.d_model = d_model
        self.h = h
        assert d_model % h == 0, "d_model must be divisible by h"

        # Dimension of the vector received by each head
        self.d_k = d_model // h
        self.w_q = nn.Linear(d_model, d_model, bias=False)
        self.w_k = nn.Linear(d_model, d_model, bias=False)
        self.w_v = nn.Linear(d_model, d_model, bias=False)
        self.w_o = nn.Linear(d_model, d_model, bias=False)
        self.dropout = nn.Dropout(dropout)

    @staticmethod
    def attention(query, key, value, mask, dropout: nn.Dropout):
        d_k = query.shape[-1]

        # (batch, h, seq_len, d_k) -> (batch, h, seq_len, seq_len)
        attention_scores = (query @ key.transpose(-2, -1)) / math.sqrt(d_k)
        if mask is not None:
            attention_scores.masked_fill(mask == 0, -1e9)
        attention_scores = attention_scores.softmax(dim=-1)
        if dropout is not None:
            attention_scores = dropout(attention_scores)

        return (attention_scores @ value), attention_scores

    def forward(self, q, k, v, mask):
        # Linearly transform the queries, keys, and values
        # (batch, seq_len, d_model) -> (batch, seq_len, d_model)
        query = self.w_q(q)
        key = self.w_k(k)
        value = self.w_v(v)

        # Split the queries, keys, and values into multiple heads
        # (batch, seq_len, d_model) -> (batch, seq_len, h, d_k) -> (batch, h, seq_len, d_k)
        query = query.view(query.shape[0], query.shape[1], self.h, self.d_k).transpose(
            1, 2
        )
        key = key.view(key.shape[0], key.shape[1], self.h, self.d_k).transpose(1, 2)
        value = value.view(value.shape[0], value.shape[1], self.h, self.d_k).transpose(
            1, 2
        )

        # Calculate the attention scores
        # Implicitly calls setattr() to set the attention_scores attribute
        x, self.attention_scores = MultiHeadAttentionBlock.attention(
            query, key, value, mask, self.dropout
        )

        # Combine the heads together
        # (batch, h, seq_len, d_k) -> (batch, seq_len, h, d_k) -> (batch, seq_len, d_model)
        x = x.transpose(1, 2).contiguous().view(x.shape[0], -1, self.h * self.d_k)

        # Multiply by W_o to get the final output
        # (batch, seq_len, d_model) -> (batch, seq_len, d_model)
        return self.w_o(x)


class ResidualConnection(nn.Module):

    def __init__(self, features: int, dropout: float):
        super(ResidualConnection, self).__init__()
        self.dropout = nn.Dropout(dropout)
        self.norm = LayerNormalization(features)

    def forward(self, x, sublayer):
        # In principle, the order of normalization and sublayer can be changed
        return x + self.dropout(sublayer(self.norm(x)))


class EncoderBlock(nn.Module):

    def __init__(
        self,
        features: int,
        self_attention_block: MultiHeadAttentionBlock,
        feed_forward_block: FeedForwardBlock,
        dropout: float,
    ):
        super(EncoderBlock, self).__init__()
        self.self_attention_block = self_attention_block
        self.feed_forward_block = feed_forward_block
        self.residual_connections = nn.ModuleList(
            [ResidualConnection(features, dropout) for _ in range(2)]
        )

    def forward(self, x, src_mask):
        x = self.residual_connections[0](
            x, lambda x: self.self_attention_block(x, x, x, src_mask)
        )
        x = self.residual_connections[1](x, self.feed_forward_block)
        return x


class Encoder(nn.Module):

    def __init__(self, features: int, layers: nn.ModuleList):
        super(Encoder, self).__init__()
        self.layers = layers
        self.norm = LayerNormalization(features)

    def forward(self, x, mask):
        for layer in self.layers:
            x = layer(x, mask)
        return self.norm(x)


class DecoderBlock(nn.Module):

    def __init__(
        self,
        features: int,
        self_attention_block: MultiHeadAttentionBlock,
        cross_attention_block: MultiHeadAttentionBlock,
        feed_forward_block: FeedForwardBlock,
        dropout: float,
    ):
        super(DecoderBlock, self).__init__()
        self.self_attention_block = self_attention_block
        self.cross_attention_block = cross_attention_block
        self.feed_forward_block = feed_forward_block
        self.residual_connections = nn.ModuleList(
            [ResidualConnection(features, dropout) for _ in range(3)]
        )

    def forward(self, x, encoder_output, src_mask, tgt_mask):
        x = self.residual_connections[0](
            x, lambda x: self.self_attention_block(x, x, x, tgt_mask)
        )
        x = self.residual_connections[1](
            x,
            lambda x: self.cross_attention_block(
                x, encoder_output, encoder_output, src_mask
            ),
        )
        x = self.residual_connections[2](x, self.feed_forward_block)
        return x


class Decoder(nn.Module):

    def __init__(self, features: int, layers: nn.ModuleList):
        super(Decoder, self).__init__()
        self.layers = layers
        self.norm = LayerNormalization(features)

    def forward(self, x, encoder_output, src_mask, tgt_mask):
        for layer in self.layers:
            x = layer(x, encoder_output, src_mask, tgt_mask)
        return self.norm(x)


class ProjectionLayer(nn.Module):

    def __init__(self, d_model: int, vocab_size: int):
        super(ProjectionLayer, self).__init__()
        self.projection = nn.Linear(d_model, vocab_size)

    def forward(self, x):
        # (batch, seq_len, d_model) -> (batch, seq_len, vocab_size)
        # return torch.log_softmax(self.projection(x), dim=-1)
        return self.projection(x)


class Transformer(nn.Module):

    def __init__(
        self,
        encoder: Encoder,
        decoder: Decoder,
        src_embed: InputEmbeddings,
        tgt_embed: InputEmbeddings,
        src_pos: PositionalEncoding,
        tgt_pos: PositionalEncoding,
        projection_layer: ProjectionLayer,
    ):
        super(Transformer, self).__init__()
        self.encoder = encoder
        self.decoder = decoder
        self.src_embed = src_embed
        self.tgt_embed = tgt_embed
        self.src_pos = src_pos
        self.tgt_pos = tgt_pos
        self.projection_layer = projection_layer

    def encode(self, src, src_mask):
        # (batch, seq_len, d_model)
        src = self.src_embed(src)
        src = self.src_pos(src)
        return self.encoder(src, src_mask)

    def decode(
        self,
        encoder_output: torch.Tensor,
        src_mask: torch.Tensor,
        tgt: torch.Tensor,
        tgt_mask: torch.Tensor,
    ):
        # (batch, seq_len, d_model)
        tgt = self.tgt_embed(tgt)
        tgt = self.tgt_pos(tgt)
        return self.decoder(tgt, encoder_output, src_mask, tgt_mask)

    def project(self, x):
        # (batch, seq_len, d_model)
        return self.projection_layer(x)


def build_transformer(
    src_vocab_size: int,
    tgt_vocab_size: int,
    src_seq_len: int,
    tgt_seq_len: int,
    d_model: int = 512,
    N: int = 6,
    h: int = 8,
    dropout: float = 0.1,
    d_ff: int = 2048,
) -> Transformer:
    # Create the embedding layers
    src_embed = InputEmbeddings(d_model, src_vocab_size)
    tgt_embed = InputEmbeddings(d_model, tgt_vocab_size)

    # Create the positional encoding layers (in principle, one could be used for
    # both source and target sequences)
    src_pos = PositionalEncoding(d_model, src_seq_len, dropout)
    tgt_pos = PositionalEncoding(d_model, tgt_seq_len, dropout)

    # Create the encoder blocks
    encoder_blocks = []
    for _ in range(N):
        encoder_self_attention_block = MultiHeadAttentionBlock(d_model, h, dropout)
        feed_forward_block = FeedForwardBlock(d_model, d_ff, dropout)
        encoder_block = EncoderBlock(
            d_model, encoder_self_attention_block, feed_forward_block, dropout
        )
        encoder_blocks.append(encoder_block)

    # Create the decoder blocks
    decoder_blocks = []
    for _ in range(N):
        decoder_self_attention_block = MultiHeadAttentionBlock(d_model, h, dropout)
        decoder_cross_attention_block = MultiHeadAttentionBlock(d_model, h, dropout)
        feed_forward_block = FeedForwardBlock(d_model, d_ff, dropout)
        decoder_block = DecoderBlock(
            d_model,
            decoder_self_attention_block,
            decoder_cross_attention_block,
            feed_forward_block,
            dropout,
        )
        decoder_blocks.append(decoder_block)

    # Create the encoder and decoder
    encoder = Encoder(d_model, nn.ModuleList(encoder_blocks))
    decoder = Decoder(d_model, nn.ModuleList(decoder_blocks))

    # Create the projection layer
    projection_layer = ProjectionLayer(d_model, tgt_vocab_size)

    # Crete the transformer
    transformer = Transformer(
        encoder, decoder, src_embed, tgt_embed, src_pos, tgt_pos, projection_layer
    )

    # Initialize the parameters with Glorot initialization
    for p in transformer.parameters():
        if p.dim() > 1:
            nn.init.xavier_uniform_(p)

    return transformer