pub struct Placement {
pub mode: MoeMode,
pub prefix_layers: usize,
pub bank_slots: usize,
pub predicted_s: f64,
pub reason: String,
}Expand description
A placement: the mode, how many MoE layers (from the first MoE layer) run on the whole-layer prefix path, and how many bank slots to allocate.
Fields§
§mode: MoeMode§prefix_layers: usize§bank_slots: usize§predicted_s: f64Predicted seconds per token of the chosen mode (model only).
reason: StringTrait Implementations§
impl StructuralPartialEq for Placement
Auto Trait Implementations§
impl Freeze for Placement
impl RefUnwindSafe for Placement
impl Send for Placement
impl Sync for Placement
impl Unpin for Placement
impl UnsafeUnpin for Placement
impl UnwindSafe for Placement
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