pub enum CompileTarget {
TensorRt(TensorRtOptions),
Cuda(CudaOptions),
Cpu,
Custom {
name: String,
provider: ExecutionProviderDispatch,
},
}Expand description
One compilation target together with its provider-specific configuration. Compilation never selects a target automatically or falls back to another provider.
Variants§
TensorRt(TensorRtOptions)
Compile an embedded TensorRT engine with these builder/execution settings.
Cuda(CudaOptions)
Optimize an ONNX graph for CUDA with these provider settings.
Cpu
Optimize for CPU; threading is configured in CompileOptions.
Custom
Compile with an explicitly configured installed ORT provider.
CUDA/TensorRT must use their typed variants, as with Backend::Custom.
Fields
§
provider: ExecutionProviderDispatchProvider configured with its compilation settings.
Trait Implementations§
Source§impl Clone for CompileTarget
impl Clone for CompileTarget
Auto Trait Implementations§
impl !RefUnwindSafe for CompileTarget
impl !UnwindSafe for CompileTarget
impl Freeze for CompileTarget
impl Send for CompileTarget
impl Sync for CompileTarget
impl Unpin for CompileTarget
impl UnsafeUnpin for CompileTarget
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