Closed ghost closed 1 year ago
The default value is 3. It is not something that should be tuned depending on your use case. Generally, the more ensembles you use, the better the result (limited however up to sth like 30). At the same time, it will take longer to generate results (approximately linear with N_ensemble_configurations).
What is N_ensemble_configurations default value? Should this parameter be tuned?
Also, is there a list of all TabPFNClassifier parameters?