artlib.optimized.backends.torch.HypersphereARTMAP

Hypersphere ARTMAP (Torch-accelerated backend)

Classes

`_TorchHypersphereARTMAPConfig`
`_TorchHypersphereARTMAP`	GPU-accelerated Hypersphere ARTMAP with export hooks for artlib
`HypersphereARTMAP`	HypersphereARTMAP for Classification. optimized with torch.

Functions

_to_device(→ torch.Tensor)

Module Contents

artlib.optimized.backends.torch.HypersphereARTMAP._to_device(x: torch.Tensor | numpy.ndarray, device, dtype=torch.float32) → torch.Tensor

class artlib.optimized.backends.torch.HypersphereARTMAP._TorchHypersphereARTMAPConfig

input_dim: int

alpha: float = 0.001

rho: float = 0.75

beta: float = 1.0

r_hat: float = 1.0

device: str = 'cuda'

dtype: torch.dtype = Ellipsis

clamp_inputs: bool = False

class artlib.optimized.backends.torch.HypersphereARTMAP._TorchHypersphereARTMAP(cfg: _TorchHypersphereARTMAPConfig)

Bases: artlib.optimized.backends.torch._TorchSimpleARTMAP._TorchSimpleARTMAP

GPU-accelerated Hypersphere ARTMAP with export hooks for artlib synchronization.

cfg

device

dtype

input_dim

weight_dim

W: torch.Tensor | None = None

map_y: torch.Tensor | None = None

_prep_tol: float = 1e-06

property n_cat: int

_ensure_capacity()

_prep_input(X: torch.Tensor) → torch.Tensor

_validate_prepared(X: torch.Tensor)

_free_mem_bytes() → int

_choice_and_match(I: torch.Tensor) → Tuple[torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor]: Returns (T, m, I_radius, max_radius) for a single prepared input I.

_commit_new_category(I: torch.Tensor, y: int): Start a new hypersphere at sample I with radius 0.

partial_fit_and_export(X_prepared: torch.Tensor | numpy.ndarray, y: torch.Tensor | numpy.ndarray, epsilon: float = 1e-10, match_tracking: Literal['MT+', 'MT-', 'MT0', 'MT1', 'MT~'] = 'MT+') → Tuple[numpy.ndarray, list[numpy.ndarray], numpy.ndarray]

predict_ab_prepared(X_prepared: torch.Tensor | numpy.ndarray) → Tuple[numpy.ndarray, numpy.ndarray]

class artlib.optimized.backends.torch.HypersphereARTMAP.HypersphereARTMAP(rho: float, alpha: float, beta: float, r_hat: float, input_dim: int | None = None, device: str = 'cuda', dtype: torch.dtype = torch.float64, clamp_inputs: bool = False)

Bases: artlib.supervised.SimpleARTMAP.SimpleARTMAP

HypersphereARTMAP for Classification. optimized with torch.

This module implements HypersphereARTMAP

HypersphereARTMAP is a non-modular classification model which has been highly optimized for run-time performance. Fit and predict functions are implemented in torch for efficient execution. This class acts as a wrapper for the underlying torch functions and to provide compatibility with the artlib style and usage. Functionally, HypersphereARTMAP behaves as a special case of SimpleARTMAP instantiated with HypersphereART.

_device = 'cuda'

_dtype = Ellipsis

_clamp = False

_backend: _TorchHypersphereARTMAP | None = None

_declared_input_dim = None

_ensure_backend(X: numpy.ndarray)