tfft.Options

class Options(*, debugging: tensorflow_nufft.python.ops.nufft_options.DebuggingOptions = DebuggingOptions(check_points_range=False), fftw: tensorflow_nufft.python.ops.nufft_options.FftwOptions = FftwOptions(planning_rigor=<FftwPlanningRigor.AUTO: 0>), max_batch_size: typing.Optional[int] = None, points_range: tensorflow_nufft.python.ops.nufft_options.PointsRange = PointsRange.EXTENDED)

Bases: pydantic.main.BaseModel

Represents options for the nufft operator.

This object can be used to control the behavior of the nufft operator. These are advanced options which may be useful for performance tuning, but are not required for most use cases.

Example

>>> options = tfft.Options()
>>> options.max_batch_size = 4
>>> tfft.nufft(x, k, options=options)

debugging

Options for debugging. See tfft.DebuggingOptions for more information.

Type

tensorflow_nufft.python.ops.nufft_options.DebuggingOptions

fftw

Options for the FFTW library. See tfft.FftwOptions for more information.

Type

tensorflow_nufft.python.ops.nufft_options.FftwOptions

max_batch_size

An optional int. The maximum batch size to use during the vectorized NUFFT computation. If set, limits the internal vectorization batch size to this value. Smaller values may reduce memory usage, but may also reduce performance. If not set, the internal batch size is chosen automatically.

Type

Optional[int]

points_range

An optional tfft.PointsRange. Specifies the supported bounds for the nonuniform points. See tfft.PointsRange for more information. Defaults to tfft.PointsRange.EXTENDED.

Type

tensorflow_nufft.python.ops.nufft_options.PointsRange