Reuse params from cached_op_args

[samskalicky]

Description

Reuse the params in cached_op_args for subsequent calls to optimize_for. Currently, if a new param is added during one call to optimize_for it is not available for subsequent calls. For example, if you run multiple graph passes then you would need to export/import to get the param. This PR checks if cached_op_args is set and reuses those, otherwise it just calls collect_params like normal.

Also added a check for backend argument in optimize_for and cleaned up logic to call hybridize now that we're using a clear argument to control clearing the cached_graph (whereas before we werent calling hybridize in order to not clear the cached_graph).

FYI @Kh4L

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

• Feature1, tests, (and when applicable, API doc)
• Feature2, tests, (and when applicable, API doc)

Comments

• If this change is a backward incompatible change, why must this change be made.
• Interesting edge cases to note here

[mxnet-bot]

Hey @samskalicky , 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: [unix-gpu, centos-cpu, windows-cpu, unix-cpu, centos-gpu, edge, miscellaneous, website, clang, windows-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.

[Kh4L]

LGTM! Thanks for the PR

[waytrue17]

LGTM, thanks!

[mseth10]

LGTM

[mseth10]

LGTM