Error on NLP Subword Permission denied: ""tmp/spm".model": No such file or directory Error #2

I was going over the NLP chapter of the book and when I was getting to subword tokenizations I encounter an issue that I don’t know why its being cause

The error is happening after the function definition of subword in the 10_nlp chapter of the book I’m hoping someone can help me figure out whats going on or what I’m doing wrong.

I’ve installed sentencepiece and tried to reinstall but not working. Currently working on paperspace gradient.

Thanks a lot for your help

def subword(sz):
    sp = SubwordTokenizer(vocab_sz=sz)
    sp.setup(txts)
    return ' '.join(first(sp([txt]))[:40])

---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
<ipython-input-17-d9dfabdbc5e1> in <module>
----> 1 subword(116)

<ipython-input-15-8b64e195bd76> in subword(sz)
      1 def subword(sz):
      2     sp = SubwordTokenizer(vocab_sz=sz)
----> 3     sp.setup(txts)
      4     return ' '.join(first(sp([txt]))[:40])

/opt/conda/envs/fastai/lib/python3.7/site-packages/fastai2/text/core.py in setup(self, items, rules)
    356             for t in progress_bar(maps(*rules, items), total=len(items), leave=False):
    357                 f.write(f'{t}\n')
--> 358         sp_model = self.train(raw_text_path)
    359         self.tok = SentencePieceProcessor()
    360         self.tok.Load(str(sp_model))

/opt/conda/envs/fastai/lib/python3.7/site-packages/fastai2/text/core.py in train(self, raw_text_path)
    344             f"--character_coverage={self.char_coverage} --model_type={self.model_type}",
    345             f"--unk_id={len(spec_tokens)} --pad_id=-1 --bos_id=-1 --eos_id=-1",
--> 346             f"--user_defined_symbols={','.join(spec_tokens)}"]))
    347         raw_text_path.unlink()
    348         return self.cache_dir/'spm.model'

/opt/conda/envs/fastai/lib/python3.7/site-packages/sentencepiece.py in Train(arg, **kwargs)
    402       """Train Sentencepiece model. Accept both kwargs and legacy string arg."""
    403       if arg is not None and type(arg) is str:
--> 404         return SentencePieceTrainer._TrainFromString(arg)
    405 
    406       def _encode(value):

/opt/conda/envs/fastai/lib/python3.7/site-packages/sentencepiece.py in _TrainFromString(arg)
    380     @staticmethod
    381     def _TrainFromString(arg):
--> 382         return _sentencepiece.SentencePieceTrainer__TrainFromString(arg)
    383 
    384     @staticmethod

RuntimeError: Permission denied: ""tmp/spm".model": No such file or directory Error #2

I am getting the same issue. Anyone figured out the issue?

Can you confirm you are using "sentencepiece<0.1.90" ??