[Fix] method always using cuda

[hiramf]

Motivation

single_level_grid_priors has a arg to specify device, which defaults to 'cuda'. this PR fixes the function to allow using cpu.

Modification

pass in the value of mask_feat.device to the method instead of converting the tensor afterwards.

BC-breaking (Optional)

no breaking changes

Use cases (Optional)

no new features

Checklist

1. Pre-commit or other linting tools are used to fix the potential lint issues.
2. The modification is covered by complete unit tests. If not, please add more unit tests to ensure the correctness.
3. If the modification has a dependency on downstream projects of a newer version, this PR should be tested with all supported versions of downstream projects.
4. The documentation has been modified accordingly, like docstring or example tutorials.

[CLAassistant]

All committers have signed the CLA.

[hiramf]

@RunningLeon would appreciate your review

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 48.38%. Comparing base (bc75c9d) to head (5f283be).

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #2715   +/-   ##
=======================================
  Coverage   48.38%   48.38%           
=======================================
  Files         357      357           
  Lines       13850    13850           
  Branches     2046     2046           
=======================================
  Hits         6702     6702           
  Misses       6693     6693           
  Partials      455      455           

Flag	Coverage Δ
unittests	48.38% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.