API REFERENCE/omnibias-binary

omnibias.binary.torch.ops

0.1.0a2AlphaPython >=3.10Apache-2.0

Binary and multilevel quantization with explicit hard-forward and surrogate-gradient contracts.

import omnibias.binary.torch.ops

Module contract & mathematical conventions

Torch quantization-gradient operator surface.

Shared formulas do not imply universal floating-point identity. Verify dtype, shape, derivative-order support, and the stated numerical assumptions for the path you use.

Module source ↗Source revision 9a2d73df0573Running reference examples →

quantizeModule ↗surrogateModule ↗

Imports & aliases

These names share a definition with another public module. Follow a name for its full contract, parameters, and examples.

Import from this module	Canonical reference
binarize	omnibias.binary.torch.ops.binarize
binarize01	omnibias.binary.torch.ops.binarize01
binarize_curvature	omnibias.binary.torch.ops.binarize_curvature
curvature_corrected_slope	omnibias.binary.torch.ops.curvature_corrected_slope
heaviside	omnibias.binary.torch.ops.heaviside
kbit_quantize	omnibias.binary.torch.ops.kbit_quantize
riccati_sigmoid_derivative	omnibias.binary.torch.ops.riccati_sigmoid_derivative
riccati_tanh_derivative	omnibias.binary.torch.ops.riccati_tanh_derivative
surrogate_jet	omnibias.binary.torch.ops.surrogate_jet
surrogate_tower	omnibias.binary.torch.ops.surrogate_tower
ternarize	omnibias.binary.torch.ops.ternarize

← Package overviewRunnable learning examples →