Search code examples
pythonplotdarkmodecatboost

Is there a dark mode for Catboost fitting plot?

It seems there is no darkmode easily accessible for catboost fitting plot. The documentation does not seem to contain anything on the subject.

I am running my Jupyter Notebook into VS code and I am using these lines to get dark modes with seaborn and matplotlib:

import seaborn as sns
import matplotlib.pyplot as plt

sns.set(style="darkgrid", context="talk")
plt.style.use("dark_background")
plt.rcParams.update({"grid.linewidth":0.5, "grid.alpha":0.5})

I have look briefly at the source code of their widget but couldn't find the attribute that defines the background color.

from catboost import CatBoostRegressor, Pool

train_dataset = Pool(X_train,y_train)
test_dataset  = Pool(X_test,y_test)

model = CatBoostRegressor()
model.fit(
    X=train_dataset,
    eval_set=test_dataset,
    plot=True
)

What can I do to have darkplots here?

Solution

  • A workaround is to add the following code to a Jupyter cell:

    %%html
<style>
.catboost {
    color: white;
    background-color: black;
}

.catboost .main-svg {
    filter: invert(1);
}
</style>