import torch
import torch.nn as nn
import torch.nn.functional as F
[docs]
class DecoderLayer(nn.Module):
def __init__(self, self_attention, cross_attention, d_model, d_ff=None,
dropout=0.1, activation="relu"):
super(DecoderLayer, self).__init__()
d_ff = d_ff or 4*d_model
self.self_attention = self_attention
self.cross_attention = cross_attention
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.norm1 = nn.LayerNorm(d_model)
self.norm2 = nn.LayerNorm(d_model)
self.norm3 = nn.LayerNorm(d_model)
self.dropout = nn.Dropout(dropout)
self.activation = F.relu if activation == "relu" else F.gelu
[docs]
def forward(self, x, cross, x_mask=None, cross_mask=None):
# Self-Attention y almacenar la matriz de atención
self_attn_output, self_attn_weights = self.self_attention(
x, x, x,
attn_mask=x_mask
)
x = x + self.dropout(self_attn_output)
x = self.norm1(x)
# Cross-Attention y almacenar la matriz de atención
cross_attn_output, cross_attn_weights = self.cross_attention(
x, cross, cross,
attn_mask=cross_mask
)
x = x + self.dropout(cross_attn_output)
y = x = self.norm2(x)
# Feedforward y normalización
y = self.dropout(self.activation(self.conv1(y.transpose(-1,1))))
y = self.dropout(self.conv2(y).transpose(-1,1))
return self.norm3(x+y), self_attn_weights, cross_attn_weights
[docs]
class Decoder(nn.Module):
def __init__(self, layers, norm_layer=None):
super(Decoder, self).__init__()
self.layers = nn.ModuleList(layers)
self.norm = norm_layer
[docs]
def forward(self, x, cross, x_mask=None, cross_mask=None):
self_attentions = []
cross_attentions = []
for layer in self.layers:
x, self_attn, cross_attn = layer(x, cross, x_mask=x_mask, cross_mask=cross_mask)
self_attentions.append(self_attn)
cross_attentions.append(cross_attn)
if self.norm is not None:
x = self.norm(x)
return x, self_attentions, cross_attentions