utils

Configuration

config.Config([layout, tensor_product, ...])	Global configuration for e3j.
config.config(**kwargs)	Singleton config class for e3j.
options.Layout	Memory layout for equivariant arrays.
options.MixingMode	Mixing mode for bilinear tensor products.
options.TensorProduct	Evaluation strategy for TensorProduct.
options.Convolution	Evaluation strategy for Convolution.
options.GraphOrdering	Edge ordering contract for the fused CUDA Convolution.
options.Aggregation	Aggregation method for the sparse reduction step.
options.TensorProductNormalization	Tensor product normalization options of Clebsch-Gordan coefficients.
options.LinearInitialization	Linear weight initialization options.
options.LinearIndexwiseInitialization	LinearIndexwise weight initialization options.

Utilities

sparse.SparseMixin()	Properties and methods depending on a coef attribute.
irreps.irrep_range(l_max[, pseudotensors])	Return direct sum of irreps with angular momentum l <= l_max.
factorials.fact(n)	Factorial n!
factorials.bifact(n)	Bifactorial n!!