Error when using AutoTabPFNRegressor: TypeError: TabPFNRegressor.init() got an unexpected keyword argument 'random_state'

[tortueTortue]

Hi,

I am trying to use AutoTabPFNRegressor:

from tabpfn_extensions.post_hoc_ensembles.sklearn_interface import AutoTabPFNRegressor
reg = AutoTabPFNRegressor()
reg.fit(train_dataset)

I get this error when I run the .fit() function:

Exception has occurred: TypeError
TabPFNRegressor.init() got an unexpected keyword argument 'random_state'

Line 316 from pfn_phe.py:

    self._estimators, model_family_per_estimator = self._collect_base_models(
        categorical_feature_indices=categorical_feature_indices,
    )

Full trace:

TabPFNRegressor.init() got an unexpected keyword argument 'random_state'
File "utils.py", line 65, in init
super().init(*args, **kwargs)
File "pfn_phe.py", line 485, in _get_base_models_from_random_search
bm = model_base(**param)
^^^^^^^^^^^^^^^^^^^
File "pfn_phe.py", line 354, in _collect_base_models
bm_list = _get_base_models_from_random_search(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "pfn_phe.py", line 316, in fit
self._estimators, model_family_per_estimator = self.collect_base_models(
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "sklearn_interface.py", line 222, in fit
self.predictor.fit(
File "test_tabpfn.py", line 31, in
reg.fit(X_train, y_train)
TypeError: TabPFNRegressor.init() got an unexpected keyword argument 'random_state'

Thank you!

[LeoGrin]

Hello @tortueTortue! Thanks for the report.
I cannot reproduce this error. Are you using the latest versions of the tabpfn and tabpfn-extensions package? Would also be helpful if you could share some infos about the dataset you're using (some characteristics, ideally the dataset itself if public).