naive_attention

burn::tensor::module

Function naive_attention 

pub fn naive_attention<B>(
    query: Tensor<B, 4>,
    key: Tensor<B, 4>,
    value: Tensor<B, 4>,
    mask: Option<Tensor<B, 4, Bool>>,
) -> Tensor<B, 4>
where
    B: Backend,
Expand description

Exports naive attention to test backend’s attention against