Wolfram Language Paclet Repository
Community-contributed installable additions to the Wolfram Language
Wolfram`TensorNetworks`
TensorNetworkContract  |
|
| | ||||
|
| | ||||
| | |||||
| | |||||
| | |||||
Examples
(9)
Scope
(4)
Input forms
(1)
The default first argument is a TensorNetwork:
In[1]:=
tn=
[{A,B},{{i,j},{j,k}}];
[tn]
 | TensorNetwork |
 | TensorNetworkContract |
Out[1]=
{{-0.188101,-0.137564,-0.328064,-0.151605,-0.112973},{-0.387793,0.469894,-0.531406,0.221123,0.179044},{-1.32872,-0.841943,-0.92898,-0.347198,-0.0682217}}
Tensor-network Graph objects (produced by ToTensorNetworkGraph) are accepted directly:
In[1]:=
tn=
[{A,B},{{i,j},{j,k}}];g=
[tn]
| TensorNetwork |
 | ToTensorNetworkGraph |
In[2]:=
| TensorNetworkContract |
Out[2]=
{{-0.188101,-0.137564,-0.328064,-0.151605,-0.112973},{-0.387793,0.469894,-0.531406,0.221123,0.179044},{-1.32872,-0.841943,-0.92898,-0.347198,-0.0682217}}
The low-level Association form, as produced by TensorNetworkData, is also accepted:
In[1]:=
tn=
[{A,B},{{i,j},{j,k}}];data=
[tn]
| TensorNetwork |
| TensorNetworkData |
In[2]:=
| TensorNetworkContract |
Out[2]=
{{-0.188101,-0.137564,-0.328064,-0.151605,-0.112973},{-0.387793,0.469894,-0.531406,0.221123,0.179044},{-1.32872,-0.841943,-0.92898,-0.347198,-0.0682217}}
Path forms
(1)
Engines
(1)
Auto-binarization
(1)
Options
(3)
Applications
(1)
Properties & Relations
(1)
 |
|
 |
|
 |
|
Contract a two-tensor network into a single tensor. Without a path argument the contraction is delegated to EinsteinSummation:
In[1]:=
A=RandomReal[{-1,1},{3,4}];B=RandomReal[{-1,1},{4,5}];tn=
[{A,B},{{i,j},{j,k}}]
| TensorNetwork |
In[2]:=
| TensorNetworkContract |
Out[2]=
{{-0.188101,-0.137564,-0.328064,-0.151605,-0.112973},{-0.387793,0.469894,-0.531406,0.221123,0.179044},{-1.32872,-0.841943,-0.92898,-0.347198,-0.0682217}}
Contract using an explicit canonical path:
In[1]:=
tn=
[{A,B},{{i,j},{j,k}}];
[tn,{{1,2}}]
| TensorNetwork |
| TensorNetworkContract |
Out[1]=
{{-0.188101,-0.137564,-0.328064,-0.151605,-0.112973},{-0.387793,0.469894,-0.531406,0.221123,0.179044},{-1.32872,-0.841943,-0.92898,-0.347198,-0.0682217}}
Use a path-as-method-string shortcut to dispatch to GreedyContractionPath:
In[1]:=
tn=
[{A,B},{{i,j},{j,k}}];
[tn,"Greedy"]
| TensorNetwork |
| TensorNetworkContract |
Out[1]=
{{-0.188101,-0.137564,-0.328064,-0.151605,-0.112973},{-0.387793,0.469894,-0.531406,0.221123,0.179044},{-1.32872,-0.841943,-0.92898,-0.347198,-0.0682217}}
""