pub fn linear_bf16( x: &[f32], rows: usize, in_f: usize, w: &[f32], out_f: usize, ) -> Vec<f32>
F.linear on bf16: f32 accumulate, bf16 output. x: [rows, in], w: [out, in].
F.linear
x
[rows, in]
w
[out, in]