TimeSeriesPredictor - after load - EOFError: Ran out of input

[5xcor]

Hi,

i ahve versio 1.1.0 on Python 3.10 apple sillicon M2

a) first step - I create models with simple timeseries and after filt() a call the save() / i read in the manual is not neccessary

b) predictor=None
c) predictor=TimeSeriesPredictor.load("str")

d) predictor.info() ... work
....
{'path': './ag-tplM15',
'version': '1.1.0',
'time_fit_training': 1240.153370141983,
'time_limit': 2398.856751680374,
'best_model': 'WeightedEnsemble',
'best_model_score_val': -0.13336230268863725,
'num_models_trained': 12,........

e) if I call predict, or evaluate ... every time I get error "Ran out input"

File ~/miniforge3/envs/ag/lib/python3.10/site-packages/autogluon/common/loaders/load_pkl.py:44, in load(path, format, verbose, **kwargs)
42 if compression_fn in compression_fn_map:
43 with compression_fn_map[compression_fn]["open"](validated_path, "rb", **compression_fn_kwargs) as fin:
---> 44 object = pickle.load(fin)
45 else:
46 raise ValueError(
47 f"compression_fn={compression_fn} or compression_fn_kwargs={compression_fn_kwargs} are not valid."
48 f" Valid function values: {compression_fn_map.keys()}"
49 )

EOFError: Ran out of input

......
Please, do you anybody knew, where I have a bug? Mayby wrong pickle or perrmision?

Thank's Milan

[5xcor]

So,

I found where is bug ( on my side )

My input have two columns - one for prediction, second I don't used. But is in this case not work import properly.