Args: prev_output_tokens (LongTensor): previous decoder outputs of shape `(batch, tgt_len)`, for teacher forcing encoder_out (Tensor, optional): output from the encoder, used for encoder-side attention incremental_state (di
(self, prev_output_tokens, encoder_out=None, incremental_state=None, **unused)
source not stored for this graph (policy: none)
nothing calls this directly
no test coverage detected