tokenized = df[0].apply((lambda x: tokenizer.encode(x, add_special_tokens=True)))