Language code "no" being read as boolean

[nobodyP]

I don't write many github issues so please excuse me if I miss anything important.

When trying to binarize a model that features the language code "no" (Norwegian), binarize.py seems to read the language code as a boolean, causing this error.

Traceback (most recent call last): File "D:\DiffSinger-muon_lynxnet2\scripts\binarize.py", line 25, in <module> binarize() File "D:\DiffSinger-muon_lynxnet2\scripts\binarize.py", line 21, in binarize binarizer_cls().process() File "D:\DiffSinger-muon_lynxnet2\preprocessing\acoustic_binarizer.py", line 61, in __init__ super().__init__(data_attrs=ACOUSTIC_ITEM_ATTRIBUTES) File "D:\DiffSinger-muon_lynxnet2\basics\base_binarizer.py", line 64, in __init__ self.build_lang_map() File "D:\DiffSinger-muon_lynxnet2\basics\base_binarizer.py", line 104, in build_lang_map for lang_id, lang_name in enumerate(sorted(self.dictionaries.keys()), start=1): TypeError: '<' not supported between instances of 'bool' and 'str'

This was done on the muon-lynxnet2 branch as you can see, and I have not tested on main, so it might be already fixed.

[yqzhishen]

Some words can be recognized as booleans in the YAML specification. If you want to make sure it is a string, use quotation marks like "no".