Random: Replaced factories.array with DNDarray #960
Conversation
|
GPU cluster tests are currently disabled on this Pull Request. |
CodeSee Review Map:Review in an interactive map View more CodeSee Maps Legend |
|
@ClaudiaComito @mtar Please tell if any changes are required. |
There was a problem hiding this comment.
Please, also look at the comments @ClaudiaComito wrote in the similar PRs #949 and #951 for more tips.
heat/core/random.py
Outdated
| return factories.array(data, dtype=x.dtype, is_split=x.split, device=x.device, comm=x.comm) | ||
| return DNDarray( | ||
| data, | ||
| gshape=tuple(data.shape), |
There was a problem hiding this comment.
Take a closer look at the gshape parameter in the documentation of DNDarray and the tensor you derive the shape from.
Hint
PyTorch tensors are local
heat/core/random.py
Outdated
| return factories.array(perm, dtype=dtype, device=device, split=split, comm=comm) | ||
| return DNDarray( | ||
| perm, | ||
| gshape=tuple(perm.shape), | ||
| dtype=dtype, | ||
| device=device, | ||
| split=split, | ||
| comm=comm, | ||
| balanced=True, | ||
| ) |
There was a problem hiding this comment.
Take a look at the split parameter in the documentation of ht.array and compare it to DNDarray.
Codecov Report
@@ Coverage Diff @@
## main #960 +/- ##
=======================================
Coverage 95.36% 95.36%
=======================================
Files 64 64
Lines 9898 9898
=======================================
Hits 9439 9439
Misses 459 459
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
|
@mtar I have made some changes, could you please review. |
|
Thank you very much @shahpratham, this looks better now. |
|
run tests |
|
run tests |
Description
Replaced factories.array with DNDarray in random.py
Issue/s resolved: #800
Changes proposed:
Type of change
enhancement
Due Diligence
Does this change modify the behaviour of other functions? If so, which?
no