TeneT.jl

This is a Julia package for the Variational Uniform Matrix product states(VUMPS) to contract infinite two-dimension square lattice tensor network.

In this package we implemented the VUMPS algorithms including the following:

• Complex number forward and backward propagation
• NixNj Big Unit Cell
• U1-symmmetry and Z2-symmmetry

This package a updated version of the original ADVUMPS.jl package, so for more easily to learn and use, we recommend you to use ADVUMPS.jl package.

install

> git clone https://github.com/XingyuZhang2018/TeneT.jl

move to the file and run julia REPL, press ] into Pkg REPL

(@v1.7) pkg> activate .
Activating environment at `..\TeneT\Project.toml`

(TeneT) pkg> instantiate

To get back to the Julia REPL, press backspace or ctrl+C. Then Precompile TeneT

julia> using TeneT
[ Info: Precompiling TeneT [260a78e0-cbf2-49ba-8157-48058c700f32]

Examples

2D Classical Ising Model

A simple example is presentated in ./example/ising.jl to constructing the tensor for the tensor network representation of the 2d classical Ising Model.

julia> include("./example/ising.jl")

You can modify Ni and Nj to change the Unit cell size and atype Array or CuArray to change computation type.

More Project including this package

• AD-Kitaev - Optimize iPEPS of Kitaev-like model with 1x2 unit cell VUMPS.
• ADFPEPS.jl - Optimize fermionic iPEPS of Hubbard model with 2x2 unit cell VUMPS.