Closed franchuterivera closed 3 years ago
This problem can be easily reproduced by removing this line
The problem is that https://github.com/LMZimmer/Auto-PyTorch_refactor/blob/3683897e3587971d50a2fcf82897c0a65b45e927/autoPyTorch/pipeline/tabular_classification.py#L114 assumes that the output shape is unidimensional or not a tuple.
The goal is to enable predict probability in this testcase.
This problem can be easily reproduced by removing this line
The problem is that https://github.com/LMZimmer/Auto-PyTorch_refactor/blob/3683897e3587971d50a2fcf82897c0a65b45e927/autoPyTorch/pipeline/tabular_classification.py#L114 assumes that the output shape is unidimensional or not a tuple.
The goal is to enable predict probability in this testcase.