`(batch, 1, tgt_len, src_len)` where padding elements are indicated by very large negative values.
            encoder_hidden_states (`torch.FloatTensor`):
                cross attention input to the layer of shape `(batch, seq_len, embed_dim)`
            encoder_attention_mask (`torch.FloatTensor`): encoder attention mask of size
                `(batch, 1, tgt_len, src_len)` where padding elements are indicated by very large negative values.
            layer_head_mask (`torch.FloatTensor`): mask for attention heads in a given layer of size
                `(encoder_attention_heads,)`.
            cross_attn_layer_head_mask (`torch.FloatTensor`): mask for cross-attention heads in a given layer of
                size `(decoder_attention_heads,)`.
            past_key_value (`Tuple(torch.FloatTensor)`): cached past key and value projection states
            output_attentions (`bool`, *optional*):
                Whether or not to return the attentions tensors of all attention layers. See `attentions` under
                returned tensors for more detail.
        NrO