Skip to content

eqiora.torch

← Python API index

PyTorch autograd projection of Eqiora differentiable programs.

Module authority: bindings/python/python/eqiora/torch.py

Shipped stub: bindings/python/python/eqiora/torch.pyi

This module publishes 2 literal __all__ spellings. Member entries below are exact signatures under documented owning types; they do not imply member-level behavioral prose.

Process-local functional PyTorch view of one Eqiora program.

Authority: bindings/python/python/eqiora/torch.py::TorchProgram

python
class TorchProgram:
def __init__(self, program: DifferentiableProgram) -> None: ...
@property
def program(self) -> DifferentiableProgram: ...
@property
def input_shape(self) -> tuple[int]: ...
@property
def output_shape(self) -> tuple[int]: ...
def __call__(self, parameters: torch.Tensor) -> torch.Tensor: ...

Bind a program to a process-local PyTorch operator.

Authority: bindings/python/python/eqiora/torch.py::bind

python
def bind(program: DifferentiableProgram) -> TorchProgram: ...