I’m using a custom datablock to load a dataframe from a .csv.
phrase_bank = DataBlock(blocks=(TextBlock.from_df(‘text’, seq_len=80), CategoryBlock), get_x=ColReader(‘text’), get_y=ColReader(‘label’), splitter=RandomSplitter(valid_pct=0.1, seed=1))
dls = phrase_bank.dataloaders(df, bs=128)
KeyError: “Label ‘nan’ was not included in the training dataset”
This seemed strange to me so I’m sure it has something to do with how I’m calling the DataBlock but why is a common label asked for by the library? And, if I can’t add it manually how do I get around this?