Cross-validation ,plots, and eval metrics error julia 1.8.3 and julia 1.9.0

[MrBenzWorld]

not able to do Cross-validation, plots, and eval metrics error.

Please add example using these functions

[WilliamZimmerman83]

I was able to get eval metrics and cv working - but plots don't work.

#CROSS VALIDATION
cat_features = ["x", "x", "x", "x", "x"]
cv_dataset = Pool(; data=pytable(test_data), label = test_label, cat_features = cat_features)
params = PyDict(Dict("iterations" => 200, 
                    "depth" => 5, 
                    "loss_function" => "Logloss",
                     "verbose" => false
                )

And here is the eval set:

#Creating Evaluation Set
eval_dataset = Pool(; data=test_data, label=test_label, cat_features = cat_features)

#INITIAL FIT MODEL
fit!(model, train_data, train_label,
    cat_features = cat_features,
    eval_set = eval_dataset,
    verbose = true,
    plot = true
)

# GET PREDICTIONS & BEST SCORE
best_iter = model.get_best_iteration()
preds = CatBoost.predict(model, test_data)
best_score = model.get_best_score()

When I run these with "plot = true" it will run successfully but no plot is generated. I tested in jupyter notebooks and in VScode.

I did the following to get plots to stop giving me errors:

using Conda
Conda.add("traitlets")
Conda.add("IPython")
Conda.add("ipywidgets")

Not sure what else needs to be done.

[tylerjthomas9]

I not sure how to get the plots working inside Julia kernels. I think the best method, for now, would just be to visualize using python. Let me know if you manage to get the plots to show within a Julia kernel, but my guess is that it'll take some work because it isn't just a matplotlib plot.

import catboost

w = catboost.MetricVisualizer('/crossentropy/')
w.start()