Hi,
I’m trying to load a bunch of text to fine-tune a language model. I’ve tried a few different ways, but so far to no success.
I have a .txt file with a lot of text.
I’ve tried two ways:
TextList.from_folder(path)
and also loading it as a pandas dataframe with only one column and then
TextList.from_df(df)
but both of them throw the same error:
~/anaconda3/envs/fastai/lib/python3.6/site-packages/fastai/text/data.py in get(self, i)
325 def get(self, i):
326 o = super().get(i)
--> 327 return Text(o, self.vocab.textify(o))
328
329 def label_for_lm(self, **kwargs):
AttributeError: 'NoneType' object has no attribute 'textify'
Even if I create a minimal text file with only 10 lines of random text, I get the same error.
I tried with fastai 1.0.49 and 1.0.50.
Thank you.