Struct TernModel 

pub struct TernModel {
    pub source_model: String,
    pub format_version: u32,
    pub layers: Vec<TernLayer>,
    pub architecture: String,
    pub vocab_size: usize,
    pub hidden_size: usize,
    pub num_layers: usize,
}

A complete ternary-quantized LLM.

Fields

source_model: String

Source model identifier (e.g. “meta-llama/Llama-3.2-1B”).

format_version: u32

ternlang-compress format version.

layers: Vec<TernLayer>

All compressed weight layers.

architecture: String

Model architecture metadata (passed through from source).

vocab_size: usize

Vocabulary size.

hidden_size: usize

Hidden dimension.

num_layers: usize

Number of transformer layers.

Implementations

impl TernModel

pub fn total_params(&self) -> usize

Total number of parameters across all layers.

pub fn total_memory_bytes(&self) -> usize

Total ternary weight memory in bytes.

pub fn mean_sparsity(&self) -> f64

Average sparsity across all layers.

pub fn compression_ratio_vs_f16(&self) -> f64

Compression ratio vs original f16 storage.

pub fn summary(&self) -> String

Human-readable summary.

Trait Implementations

impl Clone for TernModel

fn clone(&self) -> TernModel

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for TernModel

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for TernModel

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for TernModel

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.