Reduction Operations

A reduction function takes an input xtensors.TensorLike object, performs reduction over one or multiple axes, and returns a new xtensors.XTensor object without the axes over which reduction is applied.

xtensors.sum(x: TensorLike, /, dim: DimLike | DimsLike | None) → XTensor

Returns:

\(\sum_{\mathrm{dim}} x\)

xtensors.mean(x: TensorLike, /, dim: DimLike | DimsLike | None) → XTensor

Returns:

\(\braket{x}_\mathrm{dim}\)

xtensors.std(x: TensorLike, /, dim: DimLike | DimsLike | None) → XTensor

Returns:

\(\sqrt{\braket{x^2}_\mathrm{dim} - \braket{x}_\mathrm{dim}^2}\)

xtensors.nansum(x: TensorLike, /, dim: DimLike | DimsLike | None) → XTensor

Same as xtensors.sum(), but nan is ignored

xtensors.nanmean(x: TensorLike, /, dim: DimLike | DimsLike | None) → XTensor

Same as xtensors.mean(), but nan is ignored

xtensors.nanstd(x: TensorLike, /, dim: DimLike | DimsLike | None) → XTensor

Same as xtensors.std(), but nan is ignored

xtensors.max(x: TensorLike, /, dim: DimLike | DimsLike | None) → XTensor

Returns:

\(\max_\mathrm{dim} x\)

xtensors.min(x: TensorLike, /, dim: DimLike | DimsLike | None) → XTensor

Returns:

\(\min_\mathrm{dim} x\)

xtensors.nanmax(x: TensorLike, /, dim: DimLike | DimsLike | None) → XTensor

Same as xtensors.max(), but nan is ignored

xtensors.nanmin(x: TensorLike, /, dim: DimLike | DimsLike | None) → XTensor

Same as xtensors.min(), but nan is ignored

xtensors.all(x: TensorLike, /, dim: DimLike | DimsLike | None) → XTensor

Used on bool-valued TensorLike objects, similar to np.all

xtensors.any(x: TensorLike, /, dim: DimLike | DimsLike | None) → XTensor

Used on bool-valued TensorLike objects, similar to np.any