minor tweaks to allow for torch 1.5 compatability

[coquelin77]

Description

Changes done to allow for the use of the most recent pytorch release (1.5.0) see changes proposed for more detail

Issue/s resolved: None

Changes proposed:

• memory.sanitize_memory_layout now assumes that the data is distributed in the default way, also it will not change anything if the tensor passed to the function has no elements.
• DNDarray.redistribute uses where instead of nonzero (the usage of nonzero there was depreciated)

Type of change

Compatibility with dependency update

Due Diligence

• All split configurations tested
• Multiple dtypes tested in relevant functions
• Documentation updated (if needed)
• Updated changelog.md under the title "Pending Additions"

Does this change modify the behaviour of other functions? If so, which?

yes, anything which uses

[coquelin77]

i have no idea why this is failing in QR. i have all test successful on both HDFML and my local machine. anyone have any ideas?

[codecov]

Codecov Report

Merging #558 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #558   +/-   ##
=======================================
  Coverage   96.39%   96.39%           
=======================================
  Files          75       75           
  Lines       14849    14849           
=======================================
  Hits        14313    14313           
  Misses        536      536           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update be2b8df...be2b8df. Read the comment docs.

[ClaudiaComito]

Great, thanks, the sanitize_memory part looks good to me. Maybe, while you're at it, add a line to the docs about input and output being torch tensors, I forgot at the time and indeed it can get confusing.