pub struct Glm5NextMlpWorkspace { /* private fields */ }Expand description
Scratch for one MLP site, allocated once and reused every decode step.
Sized for the widest thing this site can run: a dense layer’s intermediate_size, a routed
layer’s moe_intermediate_size, and the shared expert’s width.
Implementations§
Source§impl Glm5NextMlpWorkspace
impl Glm5NextMlpWorkspace
pub fn new( gpu: &dyn GpuBackend, cfg: &Glm5NextMlpConfig, max_rows: usize, ) -> Result<Self>
Auto Trait Implementations§
impl Freeze for Glm5NextMlpWorkspace
impl RefUnwindSafe for Glm5NextMlpWorkspace
impl Send for Glm5NextMlpWorkspace
impl Sync for Glm5NextMlpWorkspace
impl Unpin for Glm5NextMlpWorkspace
impl UnwindSafe for Glm5NextMlpWorkspace
Blanket Implementations§
Source§impl<T> BorrowMut<T> for Twhere
T: ?Sized,
impl<T> BorrowMut<T> for Twhere
T: ?Sized,
Source§fn borrow_mut(&mut self) -> &mut T
fn borrow_mut(&mut self) -> &mut T
Mutably borrows from an owned value. Read more