Struct tract_core::analyser::rules::SharedTensorsProxy

pub struct SharedTensorsProxy {
    pub len: IntProxy,
    // some fields omitted
}

A proxy for a vector of tensors.

This is used for rules concerning the vector of input or output tensors:

solver.equals(inputs.len, 2);

When the indexing operator is used on a SharedTensorsProxy (e.g. inputs[0]), a new SharedTensorProxy is created dynamically and cached in tensors.

The solver should check the coherence of len with the indices of every SharedTensorProxy involved in inference rules, to forbid rules like:

solver.equals(inputs[i].rank, 2);

when i >= len.

Fields

len: IntProxy

Methods

impl SharedTensorsProxy

pub fn new(path: Path) -> SharedTensorsProxy

Creates a new SharedTensorsProxy instance.

Trait Implementations

impl Proxy for SharedTensorsProxy

fn get_path(&self) -> &Path

Returns the symbolic path to the value.

impl<'a> Proxy for &'a SharedTensorsProxy

fn get_path(&self) -> &Path

Returns the symbolic path to the value.

impl Debug for SharedTensorsProxy

fn fmt(&self, formatter: &mut Formatter) -> Result

Formats the value using the given formatter.

impl Index<usize> for SharedTensorsProxy

type Output = SharedTensorProxy

The returned type after indexing.

fn index(&self, index: usize) -> &SharedTensorProxy

Returns the SharedTensorProxy corresponding to the given index.

When an index is used for the first time, the SharedTensorProxy is created dynamically and cached inside self.tensors. This way, future calls to index will return the same SharedTensorProxy.