SCML: Add warm_start parameter

[maxi-marufo]

Because SCML optimization procedure is based on stochastic subgradient descent, we can save the weights after fitting the model, and use them in a following fit call (with a different set of triplets). The decision to use the warm_start parameter instead of a new partial_fit method is because partial_fit in scikit-learn will only fit 1 epoch, whereas fit will fit for multiple epochs (until the loss converges or max_iter is reached), which is the case also for SCML.

[perimosocordiae]

This change looks fine to me, though I'm not sure when this warm-start option is useful in practice. Sorry for the extreme delay in reviewing!

@grudloff want to take a look?

[grudloff]

LGTM!

[bellet]

Thanks @maxi-marufo !
Maybe it is possible to design a better test, something in the spirit https://github.com/scikit-learn/scikit-learn/blob/80598905e517759b4696c74ecc35c6e2eb508cff/sklearn/linear_model/tests/test_sgd.py#L274
where the idea is to check that calling warm_start=True is equivalent to manually setting the parameters?

[bellet]

LGTM @maxi-marufo, but I would like to rerun the CIs, not sure how to do this - @perimosocordiae @terrytangyuan do you know?

One simple way sould be for you @maxi-marufo to update your branch to the current state of master and push an empty commit

[grudloff]

@bellet For a manual run you need to have a trigger on workflow_dispatch but currently it is only set up with push and pull requests to master. So it would not be possible rn
(see https://docs.github.com/en/actions/managing-workflow-runs-and-deployments/managing-workflow-runs/manually-running-a-workflow)

[terrytangyuan]

This just needs a rebase to trigger the CI.

[terrytangyuan]

CI triggered

[bellet]

@maxi-marufo @terrytangyuan well looks like checks have been cancelled and/or failed?

[terrytangyuan]

I think some failed which triggered cancellation of the rest.