Error in /tutorial/basic.py

[KyrieCui]

When I run the code in /tutorial/basic.py:
tokenized_example = wrapped_t5tokenizer.tokenize_one_example(wrapped_example, teacher_forcing=False)
There is an AttributeError:

/usr/local/lib/python3.7/dist-packages/openprompt/plms/seq2seq.py in tokenize_one_example(self, wrapped_example, teacher_forcing)
89 encode_text = [0] # can be replace by any token, since these token will use their own embeddings
90 else:
---> 91 encode_text = self.tokenizer.encode(piece['text'], add_special_tokens=False)
92
93 encoding_length = len(encode_text)

AttributeError: 'NoneType' object has no attribute 'encode'

Could you tell me how to fix this problem?

[dingn42]

When I run the code in /tutorial/basic.py: tokenized_example = wrapped_t5tokenizer.tokenize_one_example(wrapped_example, teacher_forcing=False) There is an AttributeError:

/usr/local/lib/python3.7/dist-packages/openprompt/plms/seq2seq.py in tokenize_one_example(self, wrapped_example, teacher_forcing) 89 encode_text = [0] # can be replace by any token, since these token will use their own embeddings 90 else: ---> 91 encode_text = self.tokenizer.encode(piece['text'], add_special_tokens=False) 92 93 encoding_length = len(encode_text)

AttributeError: 'NoneType' object has no attribute 'encode'

Could you tell me how to fix this problem?

It seems like a version error of huggingface transformers, or you don't install SentencePiece in your system. See if this link will help you fix the problem.

[KyrieCui]

It seems like a version error of huggingface transformers, or you don't install SentencePiece in your system. See if this link will help you fix the problem.

Thanks for your help! It can run after I restart the kernel on Colab.