dls = TabularDataLoaders.from_df(df, \'.\', procs=procs, cat_names=cat_names, cont_names=cont_names, y_names="won",valid_idx=list(range(int(len(df)*0.8),len(df))), bs=
