Struct InferenceSession 

pub struct InferenceSession { /* private fields */ }

Inference session for optimized model execution

Implementations

impl InferenceSession

pub fn new(config: InferenceConfig) -> Self

Create a new inference session

pub fn initialize(&mut self) -> Result<()>

Initialize the session

pub fn run(&mut self, _input: &Tensor) -> Result<Tensor>

Run inference on a single input

pub fn run_batch(&mut self, _inputs: &[Tensor]) -> Result<Vec<Tensor>>

Run batch inference

pub fn cache_tensor(&mut self, name: String, tensor: Tensor)

Cache a tensor for reuse

pub fn get_cached(&self, name: &str) -> Option<&Tensor>

Get a cached tensor

pub fn clear_cache(&mut self)

Clear the cache

pub fn config(&self) -> &InferenceConfig

Get the configuration