Struct CudaOptions 

pub struct CudaOptions {
    pub device_id: i32,
    pub tf32: bool,
    pub cuda_graph: bool,
}

Options used only by the built-in CUDA execution provider.

Fields

device_id: i32

GPU ordinal, default 0.

tf32: bool

Allow TF32 for CUDA FP32 operations, default true.

cuda_graph: bool

Capture and replay fixed-shape models through persistent I/O bindings. Defaults true. Requires fixed positive I/O, sequential execution and strict GPU placement (or a compiled EPContext). Automatic placement disables it. Independent sessions can replay concurrently on their own streams; see TensorRtOptions::cuda_graph for capture coordination.

Trait Implementations

impl Clone for CudaOptions

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Copy for CudaOptions

impl Debug for CudaOptions

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for CudaOptions

fn default() -> Self

Returns the “default value” for a type.