Struct GgufFileSource 

pub struct GgufFileSource { /* private fields */ }

WeightSource backed by a GGUF file, using seek-based lazy tensor loading.

Each call to load_tensor seeks to the tensor’s data region, reads only the required bytes, and dequantizes them. This means only one tensor’s worth of raw data is in memory at a time.

Implementations

impl GgufFileSource

pub fn open(path: &Path) -> ModelResult<Self>

Open a GGUF file and parse its header + tensor index.

The file is kept open after construction for subsequent lazy reads.

Trait Implementations

impl WeightSource for GgufFileSource

fn tensor_names(&self) -> Vec<String>

Return the names of all tensors available in this source.

fn load_tensor(&mut self, name: &str) -> ModelResult<Vec<f32>>

Load and dequantize the tensor identified by name, returning a flat Vec<f32>.

fn contains(&self, name: &str) -> bool

Return true if the source contains a tensor with the given name.

fn total_bytes_estimate(&self) -> u64

Return a rough estimate of the total number of bytes occupied by all tensor data in the underlying file (used for progress reporting).