Minchul Lee

Hi @juneMJ Thank you for your suggestion. As you said, current version of tomotopy doesn't provide API for accessing weights value of each term of vocabularies or for adding custom...

Thank you for good suggestion, @ZechyW After looking at conda-forge, I'll add distributions for conda.

Hi @batmanscode , It seems that there is an empty document in your `df['clean_text']`. Could you check the value of `df['clean_text']` to make sure there are no blank documents?

@batmanscode `df.isnull()` tests only if the value is `NA` or not. Because an empty str `''` is not `NA`, it doesn't show any empty strings. Try following: ```python df['clean_text'].apply(lambda x:bool(x.strip())).value_counts()...

@batmanscode Currently, `add_doc` has no such feature. But I think it's a good idea to add the option to ignore empty docs.

Hello @juneMJ, thank you for your interest to my package. The first and second warning are related to numerical stability of tomotopy's CTM implementation. Currently, it seems to be a...

Hi @stephangreene , Currently, tomotopy don't support Apple Silicon M1. I am working on M1 support, but this may takes some time. Please use Rosetta until M1 is fully supported.

@alonjacovi Sorry for the late reply. Oh, I accidentally forgot to answer it. You can get topic assignment of each word in a trained or an inferred document. (https://bab2min.github.io/tomotopy/v0.11.0/en/#tomotopy.Document.topics) Get...

Hi @erip Yes, the model saved with `full=True` argument has all parameters related to whole train, thus it may take a long time to re-load it. If you save the...

@erip Actually, the package doesn't use numpy binarization for loading & saving, but uses a custom serialization functions. And it is true that there are many features related to backward...