MPSConfig#

class cuquantum.tensornet.experimental.MPSConfig(
max_extent: int | None = None,
canonical_center: int | None = None,
abs_cutoff: float | None = None,
rel_cutoff: float | None = None,
normalization: str | None = None,
discarded_weight_cutoff: float | None = None,
algorithm: str | None = None,
mpo_application: str | None = None,
gauge_option: str | None = None,
gesvdj_tol: float | None = 0,
gesvdj_max_sweeps: int | None = 0,
gesvdr_oversampling: int | None = 0,
gesvdr_niters: int | None = 0,
)[source]#

A data class for MPS based tensor network simulation configuration that can be provided to the NetworkState object.

max_extent#

The maximal extent for truncation. If not provided, no extent truncation will be performed.

Type:

int

canonical_center#

The canonical center for the final MPS. If not provided, no canonicalization will be performed.

Type:

int

abs_cutoff#

The absolute value cutoff for MPS SVD truncation. If not provided, no truncation will be performed.

Type:

float

rel_cutoff#

The relative value cutoff for MPS SVD truncation. If not provided, no truncation will be performed.

Type:

float

normalization#

The normalization option for MPS SVD operations. If not provided, no normalization will be performed.

Type:

str

discarded_weight_cutoff#

The discarded weight cutoff for MPS SVD truncation. If not provided, no truncation will be performed.

Type:

float

algorithm#

The SVD algorithm for the MPS SVD computation. It can be "gesvd" (default), "gesvdj", "gesvdp" or "gesvdr".

Type:

str

mpo_application#

The option for MPS-MPO operations. It can be "approximate" (default) or "exact".

Type:

str

gauge_option#

The option for MPS gauge. It can be "free" (default) which consider no gauging or "simple" which uses simple update algorithm to compute gauges.

Type:

str

gesvdj_tol#

The tolerance for gesvdj computation.

Type:

float

gesvdj_max_sweeps#

The maximal number of sweeps for gesvdj computation.

Type:

int

gesvdr_oversampling#

The oversampling parameter for gesvdr computation.

Type:

int

gesvdr_niters#

The number of iterations for gesvdr computation.

Type:

int