Aggregate value vectors with vision-attention weights

Read the equation

For each query, sum the value vectors using its attention weights. The matrix product contracts the key-position axis, leaving the value-feature axis as the output. When weights form a probability row, this is a convex combination before output projection.

Terms & definitions

The recovered passage does not supply a separate symbol table. Read the source context for the notation used in this version; familiar letters may carry different meanings in other builds.

How it works

Read the definition from its right-hand expression to the quantity on the left. Its surrounding source records the intended input, state, or parameter context. Where the source leaves an operation unspecified, this page preserves that gap.

Assumptions & limits

Source recovery establishes that the expression appears in this record. Its mathematical validity, implementation behavior, and interpretation must be assessed separately. See the linked source and research notes for the checks available for this version.

Original context

attn_weight = torch.softmax(attn_weight, dim=-1,
                                dtype=torch.float32).to(q.dtype)

    attn_output = attn_weight @ v
    attn_output = attn_output.transpose(0, 1)
    attn_output = attn_output.reshape(seq_length, -1)
    return attn_output

attn_output = attn_weight @ v

Related builds