Enum LlamaError 

pub enum LlamaError {
    BackendAlreadyInitialized,
    ModelLoad(String),
    ContextCreation,
    Nul(NulError),
    InvalidPath,
    Tokenize,
    TokenToPiece,
    TokenOutOfRange {
        token: i32,
        n_vocab: i32,
    },
    Decode(i32),
    BatchOverflow {
        capacity: usize,
        index: usize,
    },
    TooManySeqIds {
        got: usize,
        max: usize,
    },
    NoLogits(i32),
    MtpInit,
    MtpBegin(i32),
    MtpStep(i32),
}

Errors returned by the safe wrapper.

Variants

BackendAlreadyInitialized

The global backend was already initialized (it is a process-wide singleton).

ModelLoad(String)

llama_model_load_from_file returned null for the given path.

ContextCreation

llama_init_from_model returned null.

Nul(NulError)

A path or prompt contained an interior NUL byte.

InvalidPath

The model path was not valid UTF-8 / could not be converted to a C string.

Tokenize

Tokenization failed (negative count from llama_tokenize).

TokenToPiece

Detokenization produced invalid UTF-8 that could not be recovered.

TokenOutOfRange

A token id was outside the model’s vocabulary range [0, n_vocab). Handing it to ik’s tokenizer would abort the process (it does an out-of-range std::vector::at), so it is rejected up front.

Fields

token: i32

The offending token id.

n_vocab: i32

The model’s vocabulary size.

Decode(i32)

llama_decode returned a non-zero status.

BatchOverflow

A batch operation exceeded the batch’s allocated capacity.

Fields

capacity: usize

Allocated token capacity of the batch.

index: usize

Index that overflowed.

TooManySeqIds

More sequence ids were supplied for a token than the batch’s n_seq_max.

Fields

got: usize

Number of seq_ids supplied.

max: usize

Batch’s configured n_seq_max.

NoLogits(i32)

Requested logits for a position that was not computed / out of range.

MtpInit

MTP speculative driver initialization failed (e.g. 0 NextN layers, or an openPangu/recurrent target, or a context not created with .with_mtp(true)).

MtpBegin(i32)

MTP begin (prompt warmup) failed.

MtpStep(i32)

MTP step (draft/verify/accept/commit) failed.

Trait Implementations

impl Debug for LlamaError

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

Formats the value using the given formatter.

impl Display for LlamaError

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

Formats the value using the given formatter.

impl Error for LlamaError

fn source(&self) -> Option<&(dyn Error + 'static)>

Returns the lower-level source of this error, if any.

fn description(&self) -> &str

👎Deprecated since 1.42.0:

use the Display impl or to_string()

fn cause(&self) -> Option<&dyn Error>

👎Deprecated since 1.33.0:

replaced by Error::source, which can support downcasting

fn provide<'a>(&'a self, request: &mut Request<'a>)

🔬This is a nightly-only experimental API. (error_generic_member_access)

Provides type-based access to context intended for error reports.

impl From<NulError> for LlamaError

fn from(source: NulError) -> Self

Converts to this type from the input type.