MiniMaxM3SparseAttentionConfig#

class tensorrt_llm.llmapi.MiniMaxM3SparseAttentionConfig( *, algorithm: Literal['minimax_m3'] = 'minimax_m3', sparse_num_index_heads: Annotated[int, Gt(gt=0)] = 4, sparse_index_dim: Annotated[int, Gt(gt=0)] = 128, sparse_block_size: int = 128, sparse_topk_blocks: int = 16, sparse_init_blocks: int = 0, sparse_local_blocks: int = 1, sparse_score_type: Literal['max'] = 'max', sparse_disable_index_value: bool = True, indexer_kv_dtype: Literal['bf16', 'fp8'] = 'bf16', num_attention_heads: int | None = None, num_key_value_heads: int | None = None, implementation: Literal['triton', 'msa'] = 'triton', )[source]#

Bases: BaseSparseAttentionConfig

Configuration for MiniMax-M3 block-sparse attention.

Drives the two-step sparse attention used by MiniMax-M3 layers 3..N:

An index attention branch projects a per-head Q vector and a single replicated K vector, scores main K/V cache blocks, and selects the top-k blocks per (num_kv_heads, q_token) pair, with init_blocks forced at the head and local_blocks forced at the tail.

A sparse GQA attention runs only over the selected blocks.

At runtime one of the MiniMax-M3 sparse attention backends under tensorrt_llm._torch.attention_backend.sparse.minimax_m3 is selected. The chosen backend runs on top of a MiniMaxM3KVCacheManagerV2 that allocates a paged side index-K cache of shape [num_slots, 1, sparse_index_dim] parallel to the main K/V cache. The M3 checkpoint sets disable_index_value=True on every sparse layer, so no index V cache is allocated.

field algorithm: Literal['minimax_m3'] = 'minimax_m3'#

field implementation: Literal['triton', 'msa'] = 'triton'#: Sparse attention implementation: ‘triton’ reference (default) or ‘msa’ (fmha_sm100 kernels). The ‘msa’ implementation requires an SM100 GPU, the fmha_sm100 package, and sparse_block_size == 128.

field indexer_kv_dtype: Literal['bf16', 'fp8'] = 'bf16'#: Storage and score-compute dtype for normalized index Q/K. ‘fp8’ uses unscaled E4M3 values with FP32 score accumulation and is supported only by the MSA implementation.

field num_attention_heads: int | None = None#: Global number of attention (query) heads. When unset, it falls back to pretrained_config.num_attention_heads.

field num_key_value_heads: int | None = None#: Global number of key/value heads. When unset, it falls back to pretrained_config.num_key_value_heads, then to num_attention_heads.

field sparse_block_size: int = 128#: Block size used by per-block scoring + top-k selection.

field sparse_disable_index_value: bool = True#: If True, skip the index V branch (M3 checkpoint default).

field sparse_index_dim: int = 128#

Per-head index Q/K dimension.

Constraints:

gt = 0

field sparse_init_blocks: int = 0#: Number of leading blocks forced into the top-k regardless of score.

field sparse_local_blocks: int = 1#: Number of trailing blocks forced into the top-k regardless of score.

field sparse_num_index_heads: Annotated[int, Gt(gt=0)] = 4#

Number of index-attention heads (per TP rank’s view).

Constraints:

gt = 0

field sparse_score_type: Literal['max'] = 'max'#: Per-block score reduction; the M3 checkpoint sets ‘max’.

field sparse_topk_blocks: int = 16#: Number of top-k blocks per (kv_head, q_token).

class Config#

Bases: object

extra = 'forbid'#

__init__(**data: Any) → None#

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

classmethod construct(

_fields_set: set[str] | None = None,

**values: Any,

) → Self#

Returns a copy of the model.

!!! warning “Deprecated”: This method is now deprecated; use model_copy instead.

If you need include or exclude, use:

`python {test="skip" lint="skip"} data = self.model_dump(include=include, exclude=exclude, round_trip=True) data = {**data, **(update or {})} copied = self.model_validate(data) `

Parameters:

include – Optional set or mapping specifying which fields to include in the copied model.
exclude – Optional set or mapping specifying which fields to exclude in the copied model.
update – Optional dictionary of field-value pairs to override field values in the copied model.
deep – If True, the values of fields that are Pydantic models will be deep-copied.

Returns:

A copy of the model with included, excluded and updated fields as specified.

dict( *, include: set[int] | set[str] | Mapping[int, set[int] | set[str] | Mapping[int, IncEx | bool] | Mapping[str, IncEx | bool] | bool] | Mapping[str, set[int] | set[str] | Mapping[int, IncEx | bool] | Mapping[str, IncEx | bool] | bool] | None = None, exclude: set[int] | set[str] | Mapping[int, set[int] | set[str] | Mapping[int, IncEx | bool] | Mapping[str, IncEx | bool] | bool] | Mapping[str, set[int] | set[str] | Mapping[int, IncEx | bool] | Mapping[str, IncEx | bool] | bool] | None = None, by_alias: bool = False, exclude_unset: bool = False, exclude_defaults: bool = False, exclude_none: bool = False, ) → Dict[str, Any]#

classmethod from_orm( obj: Any, ) → Self#

get_indices_block_size() → int[source]#

json(

*,

include: set[int] | set[str] | Mapping[int, set[int] | set[str] | Mapping[int, IncEx | bool] | Mapping[str, IncEx | bool] | bool] | Mapping[str, set[int] | set[str] | Mapping[int, IncEx | bool] | Mapping[str, IncEx | bool] | bool] | None = None,

exclude: set[int] | set[str] | Mapping[int, set[int] | set[str] | Mapping[int, IncEx | bool] | Mapping[str, IncEx | bool] | bool] | Mapping[str, set[int] | set[str] | Mapping[int, IncEx | bool] | Mapping[str, IncEx | bool] | bool] | None = None,

by_alias: bool = False,

exclude_unset: bool = False,

exclude_defaults: bool = False,

exclude_none: bool = False,

encoder: Callable[[Any], Any] | None = PydanticUndefined,

models_as_dict: bool = PydanticUndefined,

**dumps_kwargs: Any,

) → str#

classmethod model_construct(

_fields_set: set[str] | None = None,

**values: Any,

) → Self#

Creates a new instance of the Model class with validated data.

Creates a new model setting __dict__ and __pydantic_fields_set__ from trusted or pre-validated data. Default values are respected, but no other validation is performed.

!!! note: model_construct() generally respects the model_config.extra setting on the provided model. That is, if model_config.extra == ‘allow’, then all extra passed values are added to the model instance’s __dict__ and __pydantic_extra__ fields. If model_config.extra == ‘ignore’ (the default), then all extra passed values are ignored. Because no validation is performed with a call to model_construct(), having model_config.extra == ‘forbid’ does not result in an error if extra values are passed, but they will be ignored.

Parameters:

_fields_set – A set of field names that were originally explicitly set during instantiation. If provided, this is directly used for the [model_fields_set][pydantic.BaseModel.model_fields_set] attribute. Otherwise, the field names from the values argument will be used.
values – Trusted or pre-validated data dictionary.

Returns:

A new instance of the Model class with validated data.

model_copy( *, update: Mapping[str, Any] | None = None, deep: bool = False, ) → Self#

!!! abstract “Usage Documentation”: [model_copy](../concepts/models.md#model-copy)

Returns a copy of the model.

!!! note: The underlying instance’s [__dict__][object.__dict__] attribute is copied. This might have unexpected side effects if you store anything in it, on top of the model fields (e.g. the value of [cached properties][functools.cached_property]).

Parameters:

update – Values to change/add in the new model. Note: the data is not validated before creating the new model. You should trust this data.
deep – Set to True to make a deep copy of the model.

Returns:

New model instance.

!!! abstract “Usage Documentation”: [model_dump](../concepts/serialization.md#python-mode)

Generate a dictionary representation of the model, optionally specifying which fields to include or exclude.

Parameters:

mode – The mode in which to_python should run. If mode is ‘json’, the output will only contain JSON serializable types. If mode is ‘python’, the output may contain non-JSON-serializable Python objects.
include – A set of fields to include in the output.
exclude – A set of fields to exclude from the output.
context – Additional context to pass to the serializer.
by_alias – Whether to use the field’s alias in the dictionary key if defined.
exclude_unset – Whether to exclude fields that have not been explicitly set.
exclude_defaults – Whether to exclude fields that are set to their default value.
exclude_none – Whether to exclude fields that have a value of None.
exclude_computed_fields – Whether to exclude computed fields. While this can be useful for round-tripping, it is usually recommended to use the dedicated round_trip parameter instead.
round_trip – If True, dumped values should be valid as input for non-idempotent types such as Json[T].
warnings – How to handle serialization errors. False/”none” ignores them, True/”warn” logs errors, “error” raises a [PydanticSerializationError][pydantic_core.PydanticSerializationError].
fallback – A function to call when an unknown value is encountered. If not provided, a [PydanticSerializationError][pydantic_core.PydanticSerializationError] error is raised.
serialize_as_any – Whether to serialize fields with duck-typing serialization behavior.
polymorphic_serialization – Whether to use model and dataclass polymorphic serialization for this call.

Returns:

A dictionary representation of the model.

!!! abstract “Usage Documentation”: [model_dump_json](../concepts/serialization.md#json-mode)

Generates a JSON representation of the model using Pydantic’s to_json method.

Parameters:

indent – Indentation to use in the JSON output. If None is passed, the output will be compact.
ensure_ascii – If True, the output is guaranteed to have all incoming non-ASCII characters escaped. If False (the default), these characters will be output as-is.
include – Field(s) to include in the JSON output.
exclude – Field(s) to exclude from the JSON output.
context – Additional context to pass to the serializer.
by_alias – Whether to serialize using field aliases.
exclude_unset – Whether to exclude fields that have not been explicitly set.
exclude_defaults – Whether to exclude fields that are set to their default value.
exclude_none – Whether to exclude fields that have a value of None.
exclude_computed_fields – Whether to exclude computed fields. While this can be useful for round-tripping, it is usually recommended to use the dedicated round_trip parameter instead.
round_trip – If True, dumped values should be valid as input for non-idempotent types such as Json[T].
warnings – How to handle serialization errors. False/”none” ignores them, True/”warn” logs errors, “error” raises a [PydanticSerializationError][pydantic_core.PydanticSerializationError].
fallback – A function to call when an unknown value is encountered. If not provided, a [PydanticSerializationError][pydantic_core.PydanticSerializationError] error is raised.
serialize_as_any – Whether to serialize fields with duck-typing serialization behavior.
polymorphic_serialization – Whether to use model and dataclass polymorphic serialization for this call.

Returns:

A JSON string representation of the model.

classmethod model_json_schema( by_alias: bool = True, ref_template: str = '#/$defs/{model}', schema_generator: type[~pydantic.json_schema.GenerateJsonSchema] = <class 'pydantic.json_schema.GenerateJsonSchema'>, mode: ~typing.Literal['validation', 'serialization'] = 'validation', *, union_format: ~typing.Literal['any_of', 'primitive_type_array'] = 'any_of', ) → dict[str, Any]#

Generates a JSON schema for a model class.

Parameters:

by_alias – Whether to use attribute aliases or not.
ref_template – The reference template.
union_format –
The format to use when combining schemas from unions together. Can be one of:
- ’any_of’: Use the [anyOf](https://json-schema.org/understanding-json-schema/reference/combining#anyOf)
keyword to combine schemas (the default). - ‘primitive_type_array’: Use the [type](https://json-schema.org/understanding-json-schema/reference/type) keyword as an array of strings, containing each type of the combination. If any of the schemas is not a primitive type (string, boolean, null, integer or number) or contains constraints/metadata, falls back to any_of.
schema_generator – To override the logic used to generate the JSON schema, as a subclass of GenerateJsonSchema with your desired modifications
mode – The mode in which to generate the schema.

Returns:

The JSON schema for the given model class.

classmethod model_parametrized_name( params: tuple[type[Any], ...], ) → str#

Compute the class name for parametrizations of generic classes.

This method can be overridden to achieve a custom naming scheme for generic BaseModels.

Parameters:: params – Tuple of types of the class. Given a generic class Model with 2 type variables and a concrete model Model[str, int], the value (str, int) would be passed to params.
Returns:: String representing the new class where params are passed to cls as type variables.
Raises:: TypeError – Raised when trying to generate concrete names for non-generic models.

model_post_init( context: Any, /, ) → None#: Override this method to perform additional initialization after __init__ and model_construct. This is useful if you want to do some validation that requires the entire model to be initialized.

classmethod model_rebuild( *, force: bool = False, raise_errors: bool = True, _parent_namespace_depth: int = 2, _types_namespace: MappingNamespace | None = None, ) → bool | None#

Try to rebuild the pydantic-core schema for the model.

This may be necessary when one of the annotations is a ForwardRef which could not be resolved during the initial attempt to build the schema, and automatic rebuilding fails.

Parameters:

force – Whether to force the rebuilding of the model schema, defaults to False.
raise_errors – Whether to raise errors, defaults to True.
_parent_namespace_depth – The depth level of the parent namespace, defaults to 2.
_types_namespace – The types namespace, defaults to None.

Returns:

Returns None if the schema is already “complete” and rebuilding was not required. If rebuilding _was_ required, returns True if rebuilding was successful, otherwise False.

Validate a pydantic model instance.

Parameters:

obj – The object to validate.
strict – Whether to enforce types strictly.
extra – Whether to ignore, allow, or forbid extra data during model validation. See the [extra configuration value][pydantic.ConfigDict.extra] for details.
from_attributes – Whether to extract data from object attributes.
context – Additional context to pass to the validator.
by_alias – Whether to use the field’s alias when validating against the provided input data.
by_name – Whether to use the field’s name when validating against the provided input data.

Raises:

ValidationError – If the object could not be validated.

Returns:

The validated model instance.

!!! abstract “Usage Documentation”: [JSON Parsing](../concepts/json.md#json-parsing)

Validate the given JSON data against the Pydantic model.

Parameters:

json_data – The JSON data to validate.
strict – Whether to enforce types strictly.
extra – Whether to ignore, allow, or forbid extra data during model validation. See the [extra configuration value][pydantic.ConfigDict.extra] for details.
context – Extra variables to pass to the validator.
by_alias – Whether to use the field’s alias when validating against the provided input data.
by_name – Whether to use the field’s name when validating against the provided input data.

Returns:

The validated Pydantic model.

Raises:

ValidationError – If json_data is not a JSON string or the object could not be validated.

Validate the given object with string data against the Pydantic model.

Parameters:

obj – The object containing string data to validate.
strict – Whether to enforce types strictly.
extra – Whether to ignore, allow, or forbid extra data during model validation. See the [extra configuration value][pydantic.ConfigDict.extra] for details.
context – Extra variables to pass to the validator.
by_alias – Whether to use the field’s alias when validating against the provided input data.
by_name – Whether to use the field’s name when validating against the provided input data.

Returns:

The validated Pydantic model.

classmethod parse_file( path: str | Path, *, content_type: str | None = None, encoding: str = 'utf8', proto: DeprecatedParseProtocol | None = None, allow_pickle: bool = False, ) → Self#

classmethod parse_obj( obj: Any, ) → Self#

classmethod parse_raw( b: str | bytes, *, content_type: str | None = None, encoding: str = 'utf8', proto: DeprecatedParseProtocol | None = None, allow_pickle: bool = False, ) → Self#

classmethod schema( by_alias: bool = True, ref_template: str = '#/$defs/{model}', ) → Dict[str, Any]#

classmethod schema_json(

*,

by_alias: bool = True,

ref_template: str = '#/$defs/{model}',

**dumps_kwargs: Any,

) → str#

supports_backend(backend: str) → bool[source]#: Override if the sparse attention algorithm does not support a subset of the possible backends.

to_sparse_metadata_params(**kwargs)[source]#

Lower into MiniMaxM3SparseMetadataParams for the attention metadata.

Head counts resolve as this config, then pretrained_config, then a default; num_key_value_heads falls back to num_attention_heads. Setting them on the config lets tests skip building a pretrained_config.

to_sparse_params(**kwargs)[source]#: Lower user-facing config into SparseParams.

classmethod update_forward_refs(

**localns: Any,

) → None#

classmethod validate( value: Any, ) → Self#

model_computed_fields = {}#

model_config: ClassVar[ConfigDict] = {'extra': 'forbid'}#: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

property model_extra: dict[str, Any] | None#

Get extra fields set during validation.

Returns:: A dictionary of extra fields, or None if config.extra is not set to “allow”.

model_fields = {'algorithm': FieldInfo(annotation=Literal['minimax_m3'], required=False, default='minimax_m3'), 'implementation': FieldInfo(annotation=Literal['triton', 'msa'], required=False, default='triton', description="Sparse attention implementation: 'triton' reference (default) or 'msa' (fmha_sm100 kernels). The 'msa' implementation requires an SM100 GPU, the fmha_sm100 package, and sparse_block_size == 128.", json_schema_extra={'status': 'prototype'}), 'indexer_kv_dtype': FieldInfo(annotation=Literal['bf16', 'fp8'], required=False, default='bf16', description="Storage and score-compute dtype for normalized index Q/K. 'fp8' uses unscaled E4M3 values with FP32 score accumulation and is supported only by the MSA implementation.", json_schema_extra={'status': 'prototype'}), 'num_attention_heads': FieldInfo(annotation=Union[int, NoneType], required=False, default=None, description='Global number of attention (query) heads. When unset, it falls back to pretrained_config.num_attention_heads.'), 'num_key_value_heads': FieldInfo(annotation=Union[int, NoneType], required=False, default=None, description='Global number of key/value heads. When unset, it falls back to pretrained_config.num_key_value_heads, then to num_attention_heads.'), 'sparse_block_size': FieldInfo(annotation=int, required=False, default=128, description='Block size used by per-block scoring + top-k selection.'), 'sparse_disable_index_value': FieldInfo(annotation=bool, required=False, default=True, description='If True, skip the index V branch (M3 checkpoint default).'), 'sparse_index_dim': FieldInfo(annotation=int, required=False, default=128, description='Per-head index Q/K dimension.', metadata=[Gt(gt=0)]), 'sparse_init_blocks': FieldInfo(annotation=int, required=False, default=0, description='Number of leading blocks forced into the top-k regardless of score.'), 'sparse_local_blocks': FieldInfo(annotation=int, required=False, default=1, description='Number of trailing blocks forced into the top-k regardless of score.'), 'sparse_num_index_heads': FieldInfo(annotation=int, required=False, default=4, description="Number of index-attention heads (per TP rank's view).", metadata=[Gt(gt=0)]), 'sparse_score_type': FieldInfo(annotation=Literal['max'], required=False, default='max', description="Per-block score reduction; the M3 checkpoint sets 'max'."), 'sparse_topk_blocks': FieldInfo(annotation=int, required=False, default=16, description='Number of top-k blocks per (kv_head, q_token).')}#

property model_fields_set: set[str]#

Returns the set of fields that have been explicitly set on this model instance.

Returns:

A set of strings representing the fields that have been set,: i.e. that were not filled from defaults.