Is it possible to mention which metric to use in model.fit while training the model?

I am doing binary classification using Random Forest, LightGBM, XGBoost andother models.

clf = RandomForestClassifier(n_estimators=100,criterion='gini')
clf.fit(train_x, train_y,metrics='F1')

I want to use f1 metric as evalution criterion while training. I tried passing metric as a parameter in fit function and the init fucntion of the model which gave me a type error/