Skip to main content

/

Docs

Nightly

v24.6

Log in

Python module

registry

Model registry, for tracking various model variants.

PipelineRegistry​

class max.pipelines.registry.PipelineRegistry(architectures: list[max.pipelines.registry.SupportedArchitecture])

register()​

register(architecture: SupportedArchitecture)

Add new architecture to registry.

reset()​

reset() → None

retrieve()​

retrieve(pipeline_config: PipelineConfig) → tuple[max.pipelines.interfaces.PipelineTokenizer, max.pipelines.interfaces.TokenGenerator]

retrieve_factory()​

retrieve_factory(pipeline_config: PipelineConfig) → tuple[max.pipelines.interfaces.PipelineTokenizer, Callable[[], max.pipelines.interfaces.TokenGenerator]]

validate_pipeline_config()​

validate_pipeline_config(pipeline_config: PipelineConfig) → PipelineConfig

Update pipeline config with appropriate values if not provided. If invalid config is provided, error out with detailed reason.

SupportedArchitecture​

class max.pipelines.registry.SupportedArchitecture(name: str, versions: list[max.pipelines.registry.SupportedVersion], default_version: str, pipeline_model: Type[PipelineModel], tokenizer: Type[TextTokenizer | TextAndVisionTokenizer], default_weights_format: WeightsFormat, weight_converters: dict[max.pipelines.config.WeightsFormat, Type[max.graph.weights.weights.WeightsConverter]] | None = None)

SupportedVersion​

class max.pipelines.registry.SupportedVersion(name: str, encodings: dict[max.pipelines.config.SupportedEncoding, tuple[list[max.pipelines.hf_utils.HuggingFaceFile], list[max.pipelines.kv_cache.cache_params.KVCacheStrategy]]], default_encoding: SupportedEncoding)

default_cache_strategy()​

default_cache_strategy(encoding: SupportedEncoding) → KVCacheStrategy

Get the default cache strategy for an encoding.

huggingface_files()​

huggingface_files(encoding: SupportedEncoding) → list[max.pipelines.hf_utils.HuggingFaceFile]

is_supported_cache_strategy()​

is_supported_cache_strategy(encoding: SupportedEncoding, cache_strategy: KVCacheStrategy) → bool

Identify if an encoding supports a specific cache_strategy.