pub enum RunMode {
Forward,
Fit {
y: Option<Value>,
},
}Expand description
What a run does to the nodes it visits.
Fitting and forwarding differ in exactly one way — whether a trainable
node learns its state before it computes — and in nothing else. They
used to be two whole execution loops: run/forward walked the plan
through run_node, while fit had a second, filter-only walk that
flattened the plan and re-implemented input resolution, events, caching
and panic handling. Making the difference a value is what lets both
go through one site.
Variants§
Forward
Every node computes with the state it already has.
Fit
A trainable node learns its state from its resolved input and these
labels, then computes with it. Labels are shared by the whole run;
None means unsupervised.
Trait Implementations§
Auto Trait Implementations§
impl Freeze for RunMode
impl RefUnwindSafe for RunMode
impl Send for RunMode
impl Sync for RunMode
impl Unpin for RunMode
impl UnsafeUnpin for RunMode
impl UnwindSafe for RunMode
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