[BUGFIX] autograd.grad

[yifeim]

Description

When the input is a single variable, autograd is supposed to return a single variable. However, it returns a list. Upon further investigation, I realized that the input variable is overwritten as a list without getting converted back. This fix separates out the flag to convert the output back to variable.

Checklist

Essentials

• PR's title starts with a category (e.g. [BUGFIX], [MODEL], [TUTORIAL], [FEATURE], [DOC], etc)
• Changes are complete (i.e. I finished coding on this PR)
• [-] All changes have test coverage
• [-] Code is well-documented

Comments

• Potentially backward incompatible but people who use it should be able to notice the inconsistency and avoid this particular use case.
• Edge case: what if the input is neither a variable nor a list? Proper input validation should be done.

[mxnet-bot]

Hey @yifeim , Thanks for submitting the PR
All tests are already queued to run once. If tests fail, you can trigger one or more tests again with the following commands:

• To trigger all jobs: @mxnet-bot run ci [all]
• To trigger specific jobs: @mxnet-bot run ci [job1, job2]

---

CI supported jobs: [miscellaneous, windows-cpu, edge, windows-gpu, unix-cpu, unix-gpu, sanity, website, centos-cpu, centos-gpu, clang]

---

Note:
Only following 3 categories can trigger CI :PR Author, MXNet Committer, Jenkins Admin.
All CI tests must pass before the PR can be merged.

[xidulu]

Hi @yifeim , thanks for your contribution
Could you add some descriptions and test cases about this patch? (e.g. Why is this a bug, how does your patch fix the problem)

[yifeim]

Hi @yifeim , thanks for your contribution
Could you add some descriptions and test cases about this patch? (e.g. Why is this a bug, how does your patch fix the problem)

Done. Thanks for the comments.