Skip to content

Appendix A: FAQ & Supplementary Information

English | 中文

Reference materials and supplementary information.

Complete Parameter Reference

To avoid duplicated and drifting parameter definitions, the canonical parameter reference is maintained in:

  • advanced/hyperparameters.md (recommended)
  • api/run-pipeline.md (CLI-oriented reference)

Directory Structure

./
├── ETM/
│   ├── main.py
│   ├── run_pipeline.py
│   ├── prepare_data.py
│   └── src/
├── data/
│   └── {dataset}/
│       └── {dataset}_cleaned.csv
├── result/
│   ├── 0.6B/
│   ├── 4B/
│   ├── 8B/
│   └── baseline/
└── embedding_models/

Hardware Requirements

Setup CPU RAM GPU CUDA Storage
Minimum 4 cores 8GB 4GB VRAM 11.8+ 20GB
Recommended 8 cores 16GB 12GB VRAM 12.1+ 50GB SSD
High-Performance 16+ cores 32GB+ A100 40GB 12.1+ 200GB NVMe

FAQ

Q: What makes THETA different?
A: THETA uses Qwen embeddings and neural variational inference for better semantic understanding than LDA or ETM.

Q: Which model size to use?
A: 0.6B for prototyping, 4B for production, 8B for maximum quality.

Q: Minimum dataset size?
A: 500+ documents with 50+ words average recommended.

Q: Training time?
A: 5K docs with 0.6B on V100: ~25 min. 4B: ~50 min.

Q: GPU required?
A: Traditional models can run on CPU. Local neural embeddings can be slow on CPU; use compatible GPU hardware when needed. Desktop installers include CPU dependencies.

Citation

@misc{codesoul2026theta,
  author = {{CodeSoul-co}},
  title = {THETA: Local Topic Modeling and Research Analysis},
  year = {2026},
  howpublished = {GitHub repository},
  url = {https://github.com/CodeSoul-co/THETA},
  note = {Version 0.3.6}
}

Contact