transformer

Attention Is All You Need

Classes

Decoder(*args, **kwargs)	Transformer Decoder that supports both one-time and distributed decoding strategies.
DecoderLayer(*args, **kwargs)
Encoder(*args, **kwargs)
Transformer([predict_sequence_length, config])	Transformer model
TransformerBlock(*args, **kwargs)	Basic Transformer block with attention and feed-forward layers.
TransformerConfig([hidden_size, num_layers, ...])	Initializes the configuration for the Transformer model with the specified parameters.