Bug: (conversion) --mistral-format doesn't work with latest mistral-common

[ngxson]

The current workaround is to use mistral-common version 1.8.3

With latest version, the conversion stops with this error:

Traceback (most recent call last):
  File "D:\Llama.cpp\llama.cpp\convert_hf_to_gguf.py", line 10503, in <module>
    main()
  File "D:\Llama.cpp\llama.cpp\convert_hf_to_gguf.py", line 10497, in main
    model_instance.write()
  File "D:\Llama.cpp\llama.cpp\convert_hf_to_gguf.py", line 661, in write
    self.prepare_metadata(vocab_only=False)
  File "D:\Llama.cpp\llama.cpp\convert_hf_to_gguf.py", line 782, in prepare_metadata
    self.set_vocab()
  File "D:\Llama.cpp\llama.cpp\convert_hf_to_gguf.py", line 2360, in set_vocab
    return self._set_vocab_mistral()
           ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\Llama.cpp\llama.cpp\convert_hf_to_gguf.py", line 2301, in _set_vocab_mistral
    vocab = MistralVocab(self.dir_model)
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\Llama.cpp\llama.cpp\gguf-py\gguf\vocab.py", line 693, in __init__
    base_path / tokenizer_file
    ~~~~~~~~~~^~~~~~~~~~~~~~~~
TypeError: unsupported operand type(s) for /: 'WindowsPath' and 'tuple'

[CISC]

Looks like they changed _filter_valid_tokenizer_files.

[CISC]

The offending PR: mistralai/mistral-common#142