"""Implementation of TorusE."""
from collections.abc import Mapping
from typing import Any, ClassVar
from ..nbase import ERModel
from ...constants import DEFAULT_EMBEDDING_HPO_EMBEDDING_DIM_RANGE
from ...nn.modules import TorusEInteraction
from ...typing import FloatTensor, Hint, Initializer, Normalizer
__all__ = [
"TorusE",
]
[docs]
class TorusE(ERModel[FloatTensor, FloatTensor, FloatTensor]):
r"""An implementation of TorusE from [ebisu2018]_.
---
citation:
author: Ebisu
year: 2018
link: https://www.aaai.org/ocs/index.php/AAAI/AAAI18/paper/view/16227
arxiv: 1711.05435
github: TakumaE/TorusE
"""
#: The default strategy for optimizing the model's hyper-parameters
hpo_default: ClassVar[Mapping[str, Any]] = {
"embedding_dim": DEFAULT_EMBEDDING_HPO_EMBEDDING_DIM_RANGE,
"p": {"type": int, "low": 1, "high": 2},
}
def __init__(
self,
embedding_dim: int = 256,
p: int = 2,
power_norm: bool = False,
entity_initializer: Hint[Initializer] = None,
entity_initializer_kwargs: Mapping[str, Any] | None = None,
entity_normalizer: Hint[Normalizer] = None,
entity_normalizer_kwargs: Mapping[str, Any] | None = None,
relation_initializer: Hint[Initializer] = None,
relation_initializer_kwargs: Mapping[str, Any] | None = None,
**kwargs,
) -> None:
r"""Initialize TorusE via the :class:`pykeen.nn.modules.TorusEInteraction` interaction.
:param embedding_dim: The entity embedding dimension $d$.
:param p:
The norm used with :func:`torch.linalg.vector_norm`. Typically is 1 or 2.
:param power_norm:
Whether to use the p-th power of the $L_p$ norm. It has the advantage of being differentiable around 0,
and numerically more stable.
:param entity_initializer: Entity initializer function. Defaults to None
:param entity_initializer_kwargs: Keyword arguments to be used when calling the entity initializer
:param entity_normalizer: Entity normalizer function. Defaults to None
:param entity_normalizer_kwargs: Keyword arguments to be used when calling the entity normalizer
:param relation_initializer: Relation initializer function. Defaults to None
:param relation_initializer_kwargs: Keyword arguments to be used when calling the relation initializer
:param kwargs: Remaining keyword arguments passed through to :class:`pykeen.models.ERModel`.
"""
super().__init__(
interaction=TorusEInteraction,
interaction_kwargs={"p": p, "power_norm": power_norm},
entity_representations_kwargs={
"shape": embedding_dim,
"initializer": entity_initializer,
"initializer_kwargs": entity_initializer_kwargs,
"normalizer": entity_normalizer,
"normalizer_kwargs": entity_normalizer_kwargs,
},
relation_representations_kwargs={
"shape": embedding_dim,
"initializer": relation_initializer,
"initializer_kwargs": relation_initializer_kwargs,
},
**kwargs,
)