Merge pull request #10266 from nero-dv/dev

            value=value,

        )

    # TODO: maybe we should use torch.empty_like(query) to allocate storage in-advance,

    # and pass slices to be mutated, instead of torch.cat()ing the returned slices

    res = torch.cat([

        compute_query_chunk_attn(

    # slices of res tensor are mutable, modifications made

    # to the slices will affect the original tensor.

    # if output of compute_query_chunk_attn function has same number of

    # dimensions as input query tensor, we initialize tensor like this:

    num_query_chunks = int(np.ceil(q_tokens / query_chunk_size))

    query_shape = get_query_chunk(0).shape

    res_shape = (query_shape[0], query_shape[1] * num_query_chunks, *query_shape[2:])

    res_dtype = get_query_chunk(0).dtype

    res = torch.zeros(res_shape, dtype=res_dtype)

    for i in range(num_query_chunks):

        attn_scores = compute_query_chunk_attn(

            query=get_query_chunk(i * query_chunk_size),

            key=key,

            value=value,

        ) for i in range(math.ceil(q_tokens / query_chunk_size))

    ], dim=1)

        )

        res[:, i * query_chunk_size:(i + 1) * query_chunk_size, :] = attn_scores

    return res