In the LabelTensor class, I made the following changes:
Added the possibility to define a LabelTensor object by passing a Torch tensor and a list of DOF, which are assigned to the last dimension.
Implemented the cat method, which allows concatenation of LabelTensor objects along a specified dimension.
Re-implemented the extract method: now, if an integer or list is passed, only the last dimension is considered to improve backward compatibility with the current version and avoid unintentional errors.
Added checks on the dimensions of the input list/dictionary and tensor to ensure consistency.
In the
LabelTensorclass, I made the following changes:LabelTensorobject by passing a Torch tensor and a list of DOF, which are assigned to the last dimension.catmethod, which allows concatenation ofLabelTensorobjects along a specified dimension.extractmethod: now, if an integer or list is passed, only the last dimension is considered to improve backward compatibility with the current version and avoid unintentional errors.