Enum LoadError 

#[non_exhaustive]
pub enum LoadError {
    BackendInit(String),
    ConfigureDevices(String),
    Fit(String),
    ModelLoad(String),
    MmprojInit(String),
    InvalidPath(String),
    WorkerInitDisconnected,
    WorkerNotRunning,
}

Failure modes returned when constructing or reloading a crate::Client or crate::EmbeddingClient.

Variants are intentionally coarse-grained: each one maps to a distinct stage of model bring-up, and the embedded String carries the underlying llama.cpp / FFI message verbatim for diagnostic purposes.

Variants (Non-exhaustive)

Non-exhaustive enums could have additional variants added in future. Therefore, when matching against variants of non-exhaustive enums, an extra wildcard arm must be added to account for any future variants.

BackendInit(String)

The process-wide llama_cpp_2::llama_backend::LlamaBackend could not be initialised. This failure is sticky for the lifetime of the process — typically it means there is no usable GPU device for the selected backend (e.g. no Vulkan ICD loadable, no CUDA driver, etc.).

ConfigureDevices(String)

Configuring the requested set of GPU devices failed (e.g. an invalid device list was passed to LlamaModelParams::with_devices).

Fit(String)

Automatic parameter fitting failed — llama.cpp could not find a layer-offload allocation that fits within the supplied memory margins for the given context size.

ModelLoad(String)

Loading the GGUF model file failed (file not found, invalid format, out-of-memory during allocation, etc.).

MmprojInit(String)

Initialising the multimodal projector (mmproj) for a vision model failed. Only produced when the mtmd feature is enabled and a mmproj_path was supplied.

InvalidPath(String)

The model path contained an interior NUL byte and could not be converted to a CString for the FFI call.

WorkerInitDisconnected

The inference worker thread exited before it could report the result of initialisation. This usually means the worker panicked; check the process’s stderr for a backtrace.

WorkerNotRunning

A reload was requested, but the inference worker thread is no longer accepting commands (it has either been shut down or panicked on a previous request).

Trait Implementations

impl Debug for LoadError

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

Formats the value using the given formatter.

impl Display for LoadError

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

Formats the value using the given formatter.

impl Error for LoadError

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.