Hey everyone, I’m messing around with the CHAMPS molecular properties dataset and I was wondering what is the best way for dealing with multiple files of tabular data? Should I put them each in their own dataframe and pass them into the databunch? I tried passing them as an array of dataframes but that didn’t work. Is it even possible to pass in multiple dataframes? I can’t append them into one csv file because the train.csv is organized by bond and the rest are organized by atom.
Update: I found this kernel on the kaggle page and copied what he did (under the label “pre-process data”). Turns out you totally can smush everything into one csv file.
1 Like