The encoder transformer body
Your encoder-only transformer body is almost complete! It's time to combine the InputEmbeddings, PositionalEncoding, and EncoderLayer classes you've created previously into a TransformerEncoder class.
This exercise is part of the course
Transformer Models with PyTorch
Exercise instructions
- Define the token embedding, positional encoding, and encoder layers (use the list comprehension to create
num_layersencoder layers). - Perform the forward pass through these layers.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
class TransformerEncoder(nn.Module):
def __init__(self, vocab_size, d_model, num_layers, num_heads, d_ff, dropout, max_seq_length):
super().__init__()
# Define the embedding, positional encoding, and encoder layers
self.embedding = ____(vocab_size, d_model)
self.positional_encoding = ____(d_model, max_seq_length)
self.layers = nn.____([____(d_model, num_heads, d_ff, dropout) for _ in range(num_layers)])
def forward(self, x, src_mask):
# Perform the forward pass through the layers
x = self.____(x)
x = self.____(x)
for layer in ____:
x = layer(x, src_mask)
return x