[BUGFIX] fix numpy op fallback bug when ndarray in kwargs

[wkcn]

Description

Hi there,
This PR fixes the issue #20232
When a operator is not implemented in MXNet, it will be fallbacked to official numpy operator.
However, it triggers a bug when the NDArray only exist in kwargs.

Reproduce Example:

from mxnet import np, npx
npx.set_np()
data = np.array([[1,2,3.], [4.,5.,6]])
print(np.apply_along_axis(lambda x: x.mean(), axis=1, arr=data))

The reason is that the function _as_onp_array in multiarray.py#L380 could not discriminate the NDArray in kwargs.

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

Changes

• make the function _as_onp_array in multiarray.py#L380 support NDArray in kwargs.
• add a related unittest

[mxnet-bot]

Hey @wkcn , 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: [clang, unix-gpu, windows-cpu, centos-cpu, miscellaneous, unix-cpu, windows-gpu, website, edge, centos-gpu, sanity]

---

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.

[wkcn]

The two positions may be made some change.

[szha]

Good catch and LGTM. Feel free to apply the suggested changes too.