Closed ArlindKadra closed 3 years ago
The evaluation measure chosen by the user in not passed to fit in the trainer and consequentially only accuracy is used in the debugging logs.
@ravinkohli took an initial look at the problem and it seems to be located here: https://github.com/automl/Auto-PyTorch/blob/7f9305d4f3ae8e866e8e318ba82bc93721d2175c/autoPyTorch/pipeline/components/training/trainer/base_trainer_choice.py#L329
Addressed in #194
ArlindKadra
commented
3 years ago ArlindKadra
commented
3 years ago
The evaluation measure chosen by the user in not passed to fit in the trainer and consequentially only accuracy is used in the debugging logs.
@ravinkohli took an initial look at the problem and it seems to be located here: https://github.com/automl/Auto-PyTorch/blob/7f9305d4f3ae8e866e8e318ba82bc93721d2175c/autoPyTorch/pipeline/components/training/trainer/base_trainer_choice.py#L329