caikit.runtime.model_management.loaded_model

A LoadedModel is a metadata wrapper around an instance of a core.ModuleBase class that contains the additional information needed to manage that model in the runtime.

Attributes

`log`
`error`
`CaikitModelFuture`
`CaikitModelFutureFactory`

Classes

LoadedModel

Module Contents

caikit.runtime.model_management.loaded_model.log[source]

caikit.runtime.model_management.loaded_model.error

caikit.runtime.model_management.loaded_model.CaikitModelFuture

caikit.runtime.model_management.loaded_model.CaikitModelFutureFactory

class caikit.runtime.model_management.loaded_model.LoadedModel[source]

__doc__ = Multiline-String

Show Value

"""
A LoadedModel is a metadata wrapper around an instance of a core.ModuleBase
class that contains the additional information needed to manage that model in
the runtime.
"""

class Builder[source]

The LoadedModel.Builder allows the LoadedModel instance to be constructed in pieces with chained ‘.’ getattr semantics.

_model_to_build

model_future_factory(caikit_model_future_factory: CaikitModelFutureFactory) → LoadedModel[source]

fail_callback(callback: Callable) → LoadedModel[source]

path(model_path: str) → LoadedModel[source]

type(model_type: str) → LoadedModel[source]

id(model_id: str) → LoadedModel[source]

retries(retries: int) → LoadedModel[source]

build() → LoadedModel[source]

_caikit_model_future_factory: CaikitModelFutureFactory | None = None

_caikit_model_future: CaikitModelFuture | None = None

_model: caikit.core.ModuleBase | None = None

_fail_callback: Callable | None = None

_model_id: str = ''

_model_path: str = ''

_model_type: str = ''

_retries: int = 0

_size: int | None = None

id() → str[source]

model() → caikit.core.ModuleBase[source]

loading() → bool[source]

Return if a model load job has completed. The model job might have failed

Returns:: bool: If a local model job is still running

loaded() → bool[source]

Return if a model load job has resulted in a local instance

Returns:: bool: If a local model instance is available

wait()[source]

type() → str[source]

path() → str[source]

size() → int[source]

has_size() → bool[source]

set_size(model_size: int)[source]