ALucek/QuicKB

Optimize Document Retrieval with Fine-Tuned KnowledgeBases

51
/ 100
Established

Implements an end-to-end ML pipeline combining semantic and token-based document chunking strategies, synthetic QA pair generation with deduplication, and Sentence Transformers-based embedding fine-tuning with Matryoshka dimension reduction. Supports multiple LLM providers via LiteLLM, parallel processing, and cross-device training (CUDA/MPS/CPU), with optional integration to Hugging Face Hub for dataset and model publishing.

183 stars.

No Package No Dependents
Maintenance 6 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 19 / 25

How are scores calculated?

Stars

183

Forks

32

Language

Python

License

MIT

Last pushed

Nov 05, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/embeddings/ALucek/QuicKB"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.