MIND-Lab/OCTIS
A Python library for optimizing and evaluating topic models including LDA, NMF, and neural topic models using Bayesian hyperparameter search.
Not currently ranked — collecting fresh signals.
star history
OCTIS is a Python package for training, optimizing, and evaluating topic models. It supports classical approaches like Latent Dirichlet Allocation, Non-negative Matrix Factorization, and neural topic models with Bayesian optimization for hyperparameter tuning. The library provides evaluation metrics for comparing model performance and includes pre-built benchmark datasets. It was presented at the EACL 2021 demo track.
Frequently asked
- What is MIND-Lab/OCTIS?
- A Python library for optimizing and evaluating topic models including LDA, NMF, and neural topic models using Bayesian hyperparameter search.
- Is OCTIS open source?
- Yes — MIND-Lab/OCTIS is open source, released under the MIT license.
- What language is OCTIS written in?
- MIND-Lab/OCTIS is primarily written in Python.
- How popular is OCTIS?
- MIND-Lab/OCTIS has 803 stars on GitHub.
- Where can I find OCTIS?
- MIND-Lab/OCTIS is on GitHub at https://github.com/MIND-Lab/OCTIS.