Struct TensorEntry 

pub struct TensorEntry {
    pub name: String,
    pub shape: Vec<u64>,
    pub quant_type_id: u32,
    pub offset: u64,
    pub size_bytes: u64,
}

A loaded tensor entry — contains metadata only; no weight bytes are held here. Use load_tensor_metadata to obtain a collection of these, then open the file and seek to offset to read the actual data.

Fields

name: String

Tensor name as stored in the GGUF file (e.g. "blk.0.attn_q.weight").

shape: Vec<u64>

Shape dimensions (e.g. [4096, 4096]).

quant_type_id: u32

Raw GGUF quantisation type ID.

offset: u64

Byte offset of this tensor’s data from the start of the tensor data section.

size_bytes: u64

Number of bytes occupied by this tensor in the data section.

Implementations

impl TensorEntry

pub fn element_count(&self) -> u64

Total number of elements across all dimensions.

pub fn quant_name(&self) -> &'static str

Human-readable name for the quantisation type, or "UNKNOWN".

pub fn is_known_quant(&self) -> bool

Returns true when the quantisation type ID is one Pictor recognises.

Trait Implementations

impl Clone for TensorEntry

fn clone(&self) -> TensorEntry

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for TensorEntry

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

Formats the value using the given formatter.