PyLops-GPU Utilities

Alongside with its Linear Operators and Solvers, PyLops-GPU contains also a number of auxiliary routines.

Shared

Backends

backend.device()

Automatically identify device to be used with PyTorch

Dot-test

dottest(Op, nr, nc[, tol, dtype, …])

Dot test.

Torch2Numpy

torch2numpy.numpytype_from_torchtype(torchtype)	Convert torch type into equivalent numpy type
torch2numpy.torchtype_from_numpytype(numpytype)	Convert torch type into equivalent numpy type

Complex Tensors

complex.complextorch_fromnumpy(x)	Convert complex numpy array into torch ComplexTensor
complex.complexnumpy_fromtorch(xt)	Convert torch ComplexTensor into complex numpy array
complex.conj(x)	Apply complex conjugation to torch ComplexTensor
complex.divide(x, y)	Element-wise division of torch Tensor and torch ComplexTensor.
complex.reshape(x, shape)	Reshape torch ComplexTensor
complex.flatten(x)	Flatten torch ComplexTensor