search

MilaNLProc / contextualized-topic-models

A python package to run contextualized topic modeling. CTMs combine contextualized embeddings (e.g., BERT) with topic models to get coherent topics. Published at EACL and ACL 2021 (Bianchi et al.).
MIT License
1.19k stars 143 forks source link

[help] Required versions #135

Open RodrigoAPJ opened 1 year ago

RodrigoAPJ commented 1 year ago

Hello, I'm trying to use:

but always get one or another error because incompatibilities, could you specify in detail wich versions and dependencies work for them? (python version, torch version, scikit-learn, scipy, numpy, pandas etc. etc. etc.? requirements.txt doesn't specify scikit-learn version for example, versions after 1.0 cause errors :(

vinid commented 1 year ago

Hello!

Where are you running this? On colab?

If you are on colab, did you restart the notebook after "pip install"?

Chen-X666 commented 1 year ago

Hello, I'm trying to use:

  • Combined TM on Wikipedia Data (Preproc+Saving+Viz) (stable v2.3.0)
  • Zero-Shot Cross-lingual Topic Modeling (Preproc+Viz) (stable v2.3.0)

but always get one or another error because incompatibilities, could you specify in detail wich versions and dependencies work for them? (python version, torch version, scikit-learn, scipy, numpy, pandas etc. etc. etc.? requirements.txt doesn't specify scikit-learn version for example, versions after 1.0 cause errors :(

I think you got 'numpy.core.multiarray failed to import' as me in colab?

vinid commented 1 year ago

Should be same to the issue described here: https://github.com/MilaNLProc/contextualized-topic-models/issues/126#issuecomment-1531232589

Restarting the colab after instal should fix the imports

RodrigoAPJ commented 1 year ago

(I was doing it on local machine) My problem was related on some point to that, but later (or before can't remember) the problem(s) was about compatibilities of dependencies that pip was messing up, at some point my ipykernel died but now everything is ok, I managed to get it working :)

I'll try to comment the versions that I'm using of ... everything? I guess (or at least the more important libraries, python, cuda and cuDNN versions, etc etc.) (I'm a litle busy so I'll do it later :P)

Thank you for your fast answer and sorry for the delay on mine.