Struct ModelTensor 

pub struct ModelTensor {
    pub name: String,
    pub data: Tensor<f32>,
    pub requires_grad: bool,
}

Real tensor wrapper for model weights

Fields

name: String

Tensor name

data: Tensor<f32>

Actual tensor data (f32 for simplicity, can be extended)

requires_grad: bool

Whether gradients are required

Implementations

impl ModelTensor

pub fn new_random( name: String, shape: Vec<usize>, requires_grad: bool, device: DeviceType, ) -> Result<Self>

Create a new model tensor with random initialization

pub fn from_data( name: String, data: Vec<f32>, shape: Vec<usize>, requires_grad: bool, device: DeviceType, ) -> Result<Self>

Create a new model tensor from existing data

pub fn shape(&self) -> Vec<usize>

Get the shape of the tensor

pub fn numel(&self) -> usize

Get the number of elements

pub fn to_bytes(&self) -> Result<Vec<u8>>

Convert to bytes for serialization

pub fn from_bytes( name: String, bytes: &[u8], shape: Vec<usize>, requires_grad: bool, device: DeviceType, ) -> Result<Self>

Create from bytes

Trait Implementations

impl Clone for ModelTensor

fn clone(&self) -> ModelTensor

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ModelTensor

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

Formats the value using the given formatter.