pub struct RestoredTraining<M, O, S, U> {
pub model: M,
pub optimizer: O,
pub scheduler: S,
pub accumulator: GradientsAccumulator<M>,
pub state: U,
}Available on crate feature
std only.Expand description
Components restored together, including pending gradients and caller state.
Fields§
§model: MModel with the recorded parameter IDs.
optimizer: OOptimizer with the recorded parameter state.
scheduler: SScheduler at the recorded position.
accumulator: GradientsAccumulator<M>Pending gradients, without an implicit optimizer update or reset.
state: UCaller-defined continuation state.
Auto Trait Implementations§
impl<M, O, S, U> !RefUnwindSafe for RestoredTraining<M, O, S, U>
impl<M, O, S, U> !Sync for RestoredTraining<M, O, S, U>
impl<M, O, S, U> !UnwindSafe for RestoredTraining<M, O, S, U>
impl<M, O, S, U> Freeze for RestoredTraining<M, O, S, U>
impl<M, O, S, U> Send for RestoredTraining<M, O, S, U>
impl<M, O, S, U> Unpin for RestoredTraining<M, O, S, U>
impl<M, O, S, U> UnsafeUnpin for RestoredTraining<M, O, S, U>where
M: UnsafeUnpin,
O: UnsafeUnpin,
S: UnsafeUnpin,
GradientsAccumulator<M>: UnsafeUnpin,
U: UnsafeUnpin,
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