PrithivirajDamodaran/FlashRank

Lite & Super-fast re-ranking for your search & retrieval pipelines. Supports SoTA Listwise and Pairwise reranking based on LLMs and cross-encoders and more. Created by Prithivi Da, open for PRs & Collaborations.

70
/ 100
Verified

Operates without PyTorch or Transformers dependencies, running efficiently on CPU with models as small as 4MB. Supports flexible model selection ranging from ultra-lightweight cross-encoders (TinyBERT) to larger LLM-based listwise rerankers (Zephyr 7B with 4-bit quantization), with multilingual options and domain-specific variants. Integrates seamlessly into existing RAG pipelines through a simple Python API, with configurable max_length parameters optimized for inference speed based on passage token counts.

948 stars and 256,714 monthly downloads. Used by 2 other packages. Available on PyPI.

Maintenance 6 / 25
Adoption 22 / 25
Maturity 25 / 25
Community 17 / 25

How are scores calculated?

Stars

948

Forks

67

Language

Python

License

Apache-2.0

Last pushed

Jan 01, 2026

Monthly downloads

256,714

Commits (30d)

0

Dependencies

5

Reverse dependents

2

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/vector-db/PrithivirajDamodaran/FlashRank"

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