artlib.optimized.backends.torch.FuzzyARTMAP

Fuzzy ARTMAP [8].

Classes

`_TorchFuzzyARTMAPConfig`
`_TorchFuzzyARTMAP`	Torch accelerated Fuzzy ARTMAP with export hooks for artlib synchronization.
`FuzzyARTMAP`	FuzzyARTMAP for Classification. optimized with torch.

Functions

`_to_device`(→ torch.Tensor)
`_complement_code`(→ torch.Tensor)

Module Contents

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

artlib.optimized.backends.torch.FuzzyARTMAP._complement_code(x: torch.Tensor) → torch.Tensor

class artlib.optimized.backends.torch.FuzzyARTMAP._TorchFuzzyARTMAPConfig

input_dim: int

alpha: float = 0.001

rho: float = 0.75

beta: float = 1.0

epsilon: float = 1e-07

match_tracking: bool = True

device: str = 'cuda'

dtype: torch.dtype = Ellipsis

clamp_inputs: bool = True

fallback_to_choice_on_fail: bool = True

class artlib.optimized.backends.torch.FuzzyARTMAP._TorchFuzzyARTMAP(cfg: _TorchFuzzyARTMAPConfig)

Torch accelerated Fuzzy ARTMAP with export hooks for artlib synchronization.

cfg

device

dtype

input_dim

code_dim

W: torch.Tensor | None = None

map_y: torch.Tensor | None = None

_lower_bounds: torch.Tensor | None = None

_upper_bounds: torch.Tensor | None = None

_prep_tol: float = 1e-06

property n_cat: int

_ensure_capacity()

_validate_prepared(X: torch.Tensor)

_free_mem_bytes() → int

set_data_bounds(lower: torch.Tensor | numpy.ndarray, upper: torch.Tensor | numpy.ndarray)

prepare_data(X: torch.Tensor | numpy.ndarray) → torch.Tensor

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

_commit_new_category(I: torch.Tensor, y: int)

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]

Incremental training on already-prepared inputs.

Returns: labels_a_out (np.ndarray): per-sample chosen A-side category indices weights_arrays (list[np.ndarray]): per-category weights (float64) cluster_labels_out (np.ndarray): map from A categories to B labels

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

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

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

FuzzyARTMAP for Classification. optimized with torch.

This module implements FuzzyARTMAP

FuzzyARTMAP 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, FuzzyARTMAP behaves as a special case of SimpleARTMAP instantiated with FuzzyART.

_device = 'cuda'

_dtype = Ellipsis

_backend: _TorchFuzzyARTMAP | None = None

_declared_input_dim = None

_ensure_backend(X: numpy.ndarray)