store_model | Work with JSON-backed attributes as ActiveRecord-ish models | JSON Processing library

I have been trying to use mlr3 to do some hyperparameter tuning for xgboost. I want to compare three different models:

1. xgboost tuned over just the alpha hyperparameter
2. xgboost tuned over alpha and lambda hyperparameters
3. xgboost tuned over alpha, lambda, and maxdepth hyperparameters.

After reading the mlr3 book, I thought that using AutoTuner for the nested resampling and benchmarking would be the best way to go about doing this. Here is what I have tried:

I have just started using mlr3 and still very unfamiliar with the syntax, I have two questions:

1. How can I access the coefficient from the trained Logistic Regression in mlr3?
2. I am dealing with a extremely imbalanced dataset, 98% vs 2%, and there are over 2million rows in this dataset, I tried to use SMOTE method, but it is very slow, because it can be done very soon in python, so is there any mistake in my code? Here is my code:

Recently I am learning about the nested resampling in mlr3 package. According to the mlr3 book, the target of nested resampling is getting an unbiased performance estimates for learners. I run a test as follow:

I try to run the resampling process with parallelization in mlr3. But I find that it always slower than the sequential plan. Here is my code and result:

This is a really basic question, but I haven't found the answer on other sites, so I am kind of forced to ask about it here.

I fitted my "classif.ranger" learner usin benchmark(design,store_models) function form mlr3 library and I need to acces the fitted parameters (obv). I found nothing about it in the benchmark documentation, so I tried to do it the hard way: -> I set store_models to TRUE -> I tried to acces the model using fitted(), but it returned NULL.

I know the question is basic and that I probably am doing smth stupid (for ex. misreading the documentation or smth like that) but I just have no idea of how to acctualy access the parameters... please help.

If it is needed in such (probably) trivial situation, here comes the code:

I'm trying to migrate a ,currently broken because of breaking changes, utility that solves a .mps using IBM's APIs.
The original code, uses an empty model.tar.gz file, creates a deployment and passes the .mps file tp a new job.
The (python)code looks like this :

I am walking through great examples from the MLR3 package (mlr3gallery:imbalanced data examples), and I was hoping to see an example that combines hyper parameter tuning AND an imbalance correction.

From the link above, as description of what I am trying to achieve:

To keep runtime low, we define the search space only for the imbalacy correction method. However, one can also jointly tune the hyperparameter of the learner along with the imbalance correction method by extending the search space with the learner’s hyperparameters.

Here is an example that comes close - mlr3 PipeOps: Create branches with different data transformations and benchmark different learners within and between branches

So we can (mis)use missuse's great example from this as a walkthough:

I have a follow-up question to this one. As in the initial question, I am using the mlr3verse, have a new dataset, and would like to make predictions using parameters that performed well during autotuning. The answer to that question says to use at$train(task). This seems to initiate tuning again. Does it take advantage of the nested resampling at all by using those parameters?

Also, looking at at$tuning_result there are two sets of parameters, one called tune_x and one called params. What is the difference between these?

Thanks.

Edit: example workflow added below

I'm migrating an application that formerly ran on IBM's DoCloud to their new API based off of Watson. Since our application doesn't have data formatted in CSV nor a separation between the model and data layers it seemed simpler to upload an LP file along with a model file that reads the LP file and solves it. I can upload and it claims to solve correctly but returns empty solve status. I've also output various model info (e.g. number of variables) and everything is zeroed out. I've confirmed the LP isn't blank - it has a trivial MILP.

Here is my model code (most of which is taken directly from the example at https://dataplatform.cloud.ibm.com/exchange/public/entry/view/50fa9246181026cd7ae2a5bc7e4ac7bd):