Rework regularized layers

[BatyLeo]

• Now regularized layers are unified under the Regularized struct, and not under the IsRegularized anymore (partially adresses #68). Every regularized layer is now a particular insance of Regularized
• Specific constructors for SparseArgmax, SoftArgmax, and RegularizedFrankWolfe
• Now we can also use Regularized with a custom optimizer (adresses #62)

TODO:

• [ ] cleanup and docstrings
• [ ] test Regularized with a custom optimizer

[codecov-commenter]

Codecov Report

Patch coverage: 77.77% and project coverage change: -0.25 :warning:

Comparison is base (b9f84b9) 80.57% compared to head (3fc2bc1) 80.33%.

:exclamation: Your organization is not using the GitHub App Integration. As a result you may experience degraded service beginning May 15th. Please install the Github App Integration for your organization. Read more.

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #73 +/- ## ========================================== - Coverage 80.57% 80.33% -0.25% ========================================== Files 19 20 +1 Lines 345 356 +11 ========================================== + Hits 278 286 +8 - Misses 67 70 +3 ``` | [Impacted Files](https://app.codecov.io/gh/axelparmentier/InferOpt.jl/pull/73?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=None) | Coverage Δ | | |---|---|---| | [src/InferOpt.jl](https://app.codecov.io/gh/axelparmentier/InferOpt.jl/pull/73?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=None#diff-c3JjL0luZmVyT3B0Lmps) | `100.00% <ø> (ø)` | | | [src/regularized/frank\_wolfe\_optimizer.jl](https://app.codecov.io/gh/axelparmentier/InferOpt.jl/pull/73?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=None#diff-c3JjL3JlZ3VsYXJpemVkL2ZyYW5rX3dvbGZlX29wdGltaXplci5qbA==) | `25.00% <25.00%> (ø)` | | | [src/regularized/regularized.jl](https://app.codecov.io/gh/axelparmentier/InferOpt.jl/pull/73?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=None#diff-c3JjL3JlZ3VsYXJpemVkL3JlZ3VsYXJpemVkLmps) | `78.57% <78.57%> (ø)` | | | [ext/InferOptFrankWolfeExt.jl](https://app.codecov.io/gh/axelparmentier/InferOpt.jl/pull/73?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=None#diff-ZXh0L0luZmVyT3B0RnJhbmtXb2xmZUV4dC5qbA==) | `100.00% <100.00%> (ø)` | | | [src/fenchel\_young/fenchel\_young.jl](https://app.codecov.io/gh/axelparmentier/InferOpt.jl/pull/73?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=None#diff-c3JjL2ZlbmNoZWxfeW91bmcvZmVuY2hlbF95b3VuZy5qbA==) | `88.00% <100.00%> (ø)` | | | [src/regularized/soft\_argmax.jl](https://app.codecov.io/gh/axelparmentier/InferOpt.jl/pull/73?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=None#diff-c3JjL3JlZ3VsYXJpemVkL3NvZnRfYXJnbWF4Lmps) | `100.00% <100.00%> (ø)` | | | [src/regularized/sparse\_argmax.jl](https://app.codecov.io/gh/axelparmentier/InferOpt.jl/pull/73?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=None#diff-c3JjL3JlZ3VsYXJpemVkL3NwYXJzZV9hcmdtYXguamw=) | `100.00% <100.00%> (ø)` | |

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Do you have feedback about the report comment? Let us know in this issue.

[gdalle]

test Regularized with a custom optimizer

What do you have in mind? I think we can use a basic QP solver from JuMP or write our own with FISTA