Juraj Hasik

Dear mtfishman, I extended the ITensor in this direction. An svd function takes through the Args system a flag, which selects an svd implementation to use. At the moment, besides...

Hi, I just did the initial commit of ?gesdd. I think there is an opportunity to make ITensor quite flexible and allow implementation of plethora of svd solvers for whomever...

The use case which led me to include Cholesky & LU comes from the full update algorithm for iPEPS networks. Applying either 2-site gate or 3-site gate leads to a...

Hi, just add --GLOBALARGS_dtype complex128 In principle this should can be inferred from input file, but right now the dtype has to be explicitly passed in case of computation in...

Hi, see https://peps-torch.readthedocs.io/en/latest/envs/generic/corrf.html But only correlations along horizontal r=(x,0) or vertical r=(0,y) direction can be computed in this way. Arbitrary coordinates for correlations, i.e. where r=(x,y) cannot computed as easily/efficiently...

Thanks ! Fortunately its in the variant of the backward, which is not used (and should be just deleted from master)

Hi, currently fermionic iPEPS are not available out-of-the-box. What are the options: One way of formulating iPEPS computations with fermions is in terms of Z2-symmetric block-sparse tensors together with swap-gates...

Hi, external magnetic field, i.e. a term in the Hamiltonian like $h_z \sum_{i\in\textrm{sites}} S^z_i$, can be easily implemented. In particular, given an iPEPS, you can compute the single-site reduced density...