Enum OptimizationLevel 

pub enum OptimizationLevel {
    None,
    Basic,
    All,
}

Graph-optimization level for the session’s optimize pipeline stage (docs/ORT2.md §18). Selected via the generic "optimization" session option (see SessionBuilder::option).

The default is OptimizationLevel::None: with optimization off the graph reaches the executor exactly as the loader produced it, so default runtime behavior is byte-identical to a build with no optimizer wired in at all.

This is a generic, model-agnostic knob — no level ever special-cases a model name or op. Higher levels simply enable more of the device-independent pass pipeline from onnx_runtime_optimizer.

Variants

None

No passes — the optimize stage is a no-op (default).

Basic

Structure-preserving passes only: constant folding then dead-node elimination. No operator fusion, so the op set the executor sees is a subset of the loaded graph’s.

All

The full device-independent pipeline: constant folding, dead-node elimination, and operator fusion (which can introduce fused com.microsoft contrib ops such as LayerNormalization).

Trait Implementations

impl Clone for OptimizationLevel

fn clone(&self) -> OptimizationLevel

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for OptimizationLevel

impl Debug for OptimizationLevel

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

Formats the value using the given formatter.

impl Default for OptimizationLevel

fn default() -> OptimizationLevel

Returns the “default value” for a type.

impl Eq for OptimizationLevel

impl PartialEq for OptimizationLevel

fn eq(&self, other: &OptimizationLevel) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for OptimizationLevel