Vectorize batch tensor network contractions

[mofeing]

Links

• https://mlir.llvm.org
• https://opensource.googleblog.com/2023/03/openxla-is-ready-to-accelerate-and-simplify-ml-development.html
• https://github.com/dfdx/Umlaut.jl
• https://github.com/Pangoraw/Coil.jl

[mofeing]

An alternative for batch TN contraction is to add a "batch" hyperindex

[mofeing]

Reactant.jl will be capable of this when we add proper support for it.