transformer

eTransformerEncoderLayer(in_rep, self_attn)	Equivariant Transformer encoder layer with the same API as torch.nn.TransformerEncoderLayer.
eTransformerDecoderLayer(in_rep, self_attn, ...)	Equivariant Transformer decoder layer mirroring torch.nn.TransformerDecoderLayer.
TransformerEncoderLayer(d_model, self_attn)	Transformer encoder layer with optional positional attention.
TransformerDecoderLayer(d_model, self_attn, ...)	Transformer decoder layer with optional positional self- and cross-attention.
TransformerEncoder(encoder_layer, num_layers)	Stack encoder layers and apply an optional final normalization.
TransformerDecoder(decoder_layer, num_layers)	Stack decoder layers and apply an optional final normalization.