nnf_multi_head_attention_forward function

Multi head attention forward

Allows the model to jointly attend to information from different representation subspaces. See reference: Attention Is All You Need

nnf_multi_head_attention_forward(
  query,
  key,
  value,
  embed_dim_to_check,
  num_heads,
  in_proj_weight,
  in_proj_bias,
  bias_k,
  bias_v,
  add_zero_attn,
  dropout_p,
  out_proj_weight,
  out_proj_bias,
  training = TRUE,
  key_padding_mask = NULL,
  need_weights = TRUE,
  attn_mask = NULL,
  avg_weights = TRUE,
  use_separate_proj_weight = FALSE,
  q_proj_weight = NULL,
  k_proj_weight = NULL,
  v_proj_weight = NULL,
  static_k = NULL,
  static_v = NULL,
  batch_first = FALSE
)

Arguments

• query: $(L, N, E)$ where L is the target sequence length, N is the batch size, E is the embedding dimension. If batch_first is TRUE, the first two dimensions are transposed.

• key: $(S, N, E)$, where S is the source sequence length, N is the batch size, E is the embedding dimension. If batch_first is TRUE, the first two dimensions are transposed.

• value: $(S, N, E)$ where S is the source sequence length, N is the batch size, E is the embedding dimension. If batch_first is TRUE, the first two dimensions are transposed.

• embed_dim_to_check: total dimension of the model.

• num_heads: parallel attention heads.

• in_proj_weight: input projection weight.

• in_proj_bias: input projection bias.

• bias_k: bias of the key and value sequences to be added at dim=0.

• bias_v: currently undocumented.

• add_zero_attn: add a new batch of zeros to the key and value sequences at dim=1.

• dropout_p: probability of an element to be zeroed.

• out_proj_weight: the output projection weight.

• out_proj_bias: output projection bias.

• training: apply dropout if is TRUE.

• key_padding_mask: $(N, S)$ where N is the batch size, S is the source sequence length. If a ByteTensor is provided, the non-zero positions will be ignored while the position with the zero positions will be unchanged. If a BoolTensor is provided, the positions with the value of True will be ignored while the position with the value of False will be unchanged.

• need_weights: output attn_output_weights.

• attn_mask: 2D mask $(L, S)$ where L is the target sequence length, S is the source sequence length. 3D mask $(N*num_heads, L, S)$ where N is the batch size, L is the target sequence length, S is the source sequence length. attn_mask ensure that position i is allowed to attend the unmasked positions. If a ByteTensor is provided, the non-zero positions are not allowed to attend while the zero positions will be unchanged. If a BoolTensor is provided, positions with True

  is not allowed to attend while False values will be unchanged. If a FloatTensor is provided, it will be added to the attention weight.

• avg_weights: Logical; whether to average attn_output_weights over the attention heads before outputting them. This doesn't change the returned value of attn_output; it only affects the returned attention weight matrix.

• use_separate_proj_weight: the function accept the proj. weights for query, key, and value in different forms. If false, in_proj_weight will be used, which is a combination of q_proj_weight, k_proj_weight, v_proj_weight.

• q_proj_weight: input projection weight and bias.

• k_proj_weight: currently undocumented.

• v_proj_weight: currently undocumented.

• static_k: static key and value used for attention operators.

• static_v: currently undocumented.

• batch_first: Logical; whether to expect query, key, and value to have batch as their first parameter, and to return output with batch first.