search

hochschule-darmstadt / MetaAutoML

MIT License
8 stars 4 forks source link

[Bug] Different x input in prediction in EvalML #384

Open piafriehe opened 1 year ago

piafriehe commented 1 year ago

Describe the bug When training EvalML with https://www.kaggle.com/datasets/falgunipatel19/biomedical-text-publication-classification an error occurs image

To Reproduce Steps to reproduce the behavior ((if applicable, add screenshots)):

  1. Upload https://www.kaggle.com/datasets/falgunipatel19/biomedical-text-publication-classification
  2. Start Training text classification with target 0 and Adapter EvalMl
  3. See error

Expected behavior Run without problems

kehuydietlatoi commented 1 year ago

i can fix that later :) woodwork causes that problem, wood work schema from training and testing data set have to be identical ( solution: add following code snippet to predict.ji template https://github.com/hochschule-darmstadt/MetaAutoML/blob/57058cf3867b88251ab04d4d50d5d52187178c23/adapters/EvalML/AutoMLs/EvalMLAdapterManager.py#L118C5-L120C5 )