Struct NemotronConfig 

pub struct NemotronConfig {
    pub model_path: PathBuf,
    pub use_gpu: bool,
    pub batch_size: usize,
    pub query_prefix: String,
    pub passage_prefix: String,
    pub max_length: usize,
    pub normalize: bool,
}

Configuration for NVIDIA Embed Nemotron 8B embedder

Nemotron is based on Llama 3.1 8B and produces 4096-dimensional embeddings. It supports asymmetric retrieval with different prefixes for queries and passages.

Fields

model_path: PathBuf

Path to the GGUF model file

use_gpu: bool

Whether to use GPU acceleration (if available)

batch_size: usize

Batch size for parallel embedding

query_prefix: String

Query instruction prefix for asymmetric retrieval

passage_prefix: String

Passage/document prefix (usually empty for Nemotron)

max_length: usize

Maximum sequence length in tokens

normalize: bool

Whether to L2-normalize output embeddings

Implementations

impl NemotronConfig

pub fn new(model_path: impl AsRef<Path>) -> Self

Create a new config with a model path

pub fn with_model_path(self, path: impl AsRef<Path>) -> Self

Set the model path

pub fn with_gpu(self, use_gpu: bool) -> Self

Enable or disable GPU acceleration

pub fn with_batch_size(self, batch_size: usize) -> Self

Set the batch size for parallel embedding

pub fn with_query_prefix(self, prefix: impl Into<String>) -> Self

Set custom query prefix

pub fn with_passage_prefix(self, prefix: impl Into<String>) -> Self

Set custom passage prefix

pub fn with_max_length(self, max_length: usize) -> Self

Set maximum sequence length

pub fn with_normalize(self, normalize: bool) -> Self

Enable or disable L2 normalization

Trait Implementations

impl Clone for NemotronConfig

fn clone(&self) -> NemotronConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for NemotronConfig

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

Formats the value using the given formatter.

impl Default for NemotronConfig

Available on crate feature nemotron only.

fn default() -> Self

Returns the “default value” for a type.