Struct GraphStats 

pub struct GraphStats {
    pub total_nodes: usize,
    pub num_placeholders: usize,
    pub num_variables: usize,
    pub num_operations: usize,
    pub max_depth: usize,
    pub num_edges: usize,
    pub num_differentiable: usize,
    pub op_breakdown: Vec<(String, usize)>,
    pub max_fan_in: usize,
    pub max_fan_out: usize,
}

Statistics about a computation graph.

Fields

total_nodes: usize

Total number of nodes

num_placeholders: usize

Number of placeholder (input) nodes

num_variables: usize

Number of variable (trainable parameter) nodes

num_operations: usize

Number of operation nodes

max_depth: usize

Maximum topological rank (depth of the graph)

num_edges: usize

Number of edges (connections between nodes)

num_differentiable: usize

Number of differentiable nodes

op_breakdown: Vec<(String, usize)>

Operation type breakdown: (op_name, count)

max_fan_in: usize

Maximum fan-in (most inputs to a single node)

max_fan_out: usize

Maximum fan-out (most nodes that depend on a single node)

Implementations

impl GraphStats

pub fn from_tensor<'g, F: Float>( root: &Tensor<'g, F>, ctx: &'g Context<'g, F>, ) -> Self

Compute graph statistics from a root tensor.

pub fn display(&self) -> String

Format the statistics as a human-readable string.

Trait Implementations

impl Clone for GraphStats

fn clone(&self) -> GraphStats

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for GraphStats

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

Formats the value using the given formatter.