This pull request adds a hide_counts parameter to the plot_confusion_matrix() function to allow for disabling of the overlaying of count text on the plot.
Sample code for testing:
import matplotlib.pyplot as plt
import scikitplot as skplt
from sklearn.datasets import load_digits
from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import cross_val_predict
X, y = load_digits(return_X_y=True)
random_forest_clf = RandomForestClassifier(n_estimators=5, max_depth=5, random_state=1)
predictions = cross_val_predict(random_forest_clf, X, y)
skplt.metrics.plot_confusion_matrix(y, predictions, hide_counts=True)
plt.show()
This pull request adds a hide_counts parameter to the plot_confusion_matrix() function to allow for disabling of the overlaying of count text on the plot.
Sample code for testing: