pub fn sigmoid_blend(
gpu: &dyn GpuBackend,
kernel: KernelHandle,
output: DevicePtr,
src: DevicePtr,
sigmoid_gate: f32,
num_elements: u32,
stream: u64,
) -> Result<()>Expand description
Sigmoid-gated blend: output = output + sigmoid_gate * src.
Kernel: bf16_sigmoid_blend(output, src, sigmoid_gate, n)
Grid: (ceil(n/256), 1, 1) Block: (256, 1, 1)