Problem about param ml_task="regression",

[jiaqizheng2000]

If I use this param, it will raise an issue below for all of the model, if it is deleted, the model works fine.

'<' not supported between instances of 'numpy.ndarray' and 'str'
Traceback (most recent call last):
File "C:\Users\ZHENGJ\AppData\Local\Programs\Python\Python39\lib\site-packages\supervised\base_automl.py", line 1195, in _fit
trained = self.train_model(params)
File "C:\Users\ZHENGJ\AppData\Local\Programs\Python\Python39\lib\site-packages\supervised\base_automl.py", line 404, in train_model
self.keep_model(mf, model_subpath)
File "C:\Users\ZHENGJ\AppData\Local\Programs\Python\Python39\lib\site-packages\supervised\base_automl.py", line 317, in keep_model
self.select_and_save_best()
File "C:\Users\ZHENGJ\AppData\Local\Programs\Python\Python39\lib\site-packages\supervised\base_automl.py", line 1315, in select_and_save_best
self._best_model = min(
TypeError: '<' not supported between instances of 'numpy.ndarray' and 'str'

[Benjamin-Frost]

Same issue here. Did you manage to resolve it?

[pplonski]

Hi @jiaqizheng2000, @Benjamin-Frost,
Could you please provide code to reproduce the issue? Thank you!

[jiaqizheng2000]

automl = AutoML(
                ml_task="regression",
                train_ensemble=True,
                fairness_threshold=0.8,
                results_path=resultpath,
                model_time_limit=30 * 60,
                start_random_models=10,
                top_models_to_improve=3,
                hill_climbing_steps=3,
                golden_features=True,
                features_selection=False,
                stack_models=True,
                explain_level=2,
                validation_strategy={
                    "validation_type": "kfold",
                    "k_folds": 4,
                    "shuffle": False,
                    "stratify": True,}
     )
automl.fit(x_train, y_train)

[pplonski]

Could you please share dataset as well? or maybe data sample, do you have this error on synthetic data as well?

[jiaqizheng2000]

example.xlsx

[pplonski]

Thank you! How do you load code and prepare X_train and y_train variables?

[jiaqizheng2000]

Simply choose the first column as y_train, and the rest as X_train

[Rohan581]

@pplonski I Would like to work on this issue

[pplonski]

Sure @Rohan581, thanks!