fix: use SII instead of explicitly accessing SciMLFunction.syms

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Comparison is base (8d53748) 0.11% compared to head (50e1e9e) 0.00%. Report is 2 commits behind head on master.

Additional details and impacted files

```diff @@ Coverage Diff @@ ## master #626 +/- ## ========================================= - Coverage 0.11% 0.00% -0.12% ========================================= Files 40 12 -28 Lines 2689 1141 -1548 ========================================= - Hits 3 0 -3 + Misses 2686 1141 -1545 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[ChrisRackauckas]

Did this work locally with the other pieces checked out?

[AayushSabharwal]

I haven't tested it yet. There are more libraries that need a bump, but it looks like for some reason it can't pull in 0.14 Flux during tests which caps CUDA at 4, whereas LinearSolve needs 5.

EDIT: it seems test/Project.toml caps Flux at 0.13. My guess is that before this change, Optimization could pull in an older version of LinearSolve compatible with CUDA 4. Now it needs the latest LinearSolve, but can't pull in latest Flux.

[AayushSabharwal]

ERROR: LoadError: Unsatisfiable requirements detected for package SparseArrays [2f01184e]:
 SparseArrays [2f01184e] log:
 ├─possible versions are: 0.0.0 or uninstalled (package in sysimage!)
 ├─restricted to versions 1.9.0-1 by OrdinaryDiffEq [1dea7af3] — no versions left
 │ └─OrdinaryDiffEq [1dea7af3] log:
 │   ├─possible versions are: 6.61.0 or uninstalled
 │   ├─restricted to versions 6.19.1-6 by SciMLSensitivity [1ed8b502], leaving only versions: 6.61.0
 │   │ └─SciMLSensitivity [1ed8b502] log:
 │   │   ├─possible versions are: 7.47.0 or uninstalled
 │   │   └─SciMLSensitivity [1ed8b502] is fixed to version 7.47.0
 │   └─OrdinaryDiffEq [1dea7af3] is fixed to version 6.61.0
 └─SparseArrays [2f01184e] is fixed to version 0.0.0

I don't even know what to make of this

[AayushSabharwal]

This passes if Flux compat is bumped and 1.10 is used

[ChrisRackauckas]

Merging to unblock others.

[Vaibhavdixit02]

I think this broke CI https://github.com/SciML/Optimization.jl/actions/runs/7209658584/job/19641063898?

[AayushSabharwal]

Yeah it won't pass until MTK is tagged. I ran it locally and tests pass.