pyttb
pyttb copied to clipboard
Improve ktensor.full()
Implements the more optimized version of ktensor.full() from MATLAB Tensor Toolbox. This results in an effective peak memory of tensor_size (down from rank*tensor_size) and a speedup of 5-50x depending on tensor shape and size.
See #299 for a more detailed discussion with benchmarks.
📚 Documentation preview 📚: https://pyttb--300.org.readthedocs.build/en/300/
~~Local branch issues messed up this PR as well. I'll close and reopen an appropriate replacement in a few hours.~~ Fixed with a rebase and force push.