ColBERT GitHub repository · inference-research

Fields

id

sources/niche-colbert-repo

URL

https://github.com/stanford-futuredata/ColBERT

Publisher

Stanford Future Data Systems

Kind

repo

Accessed

2026-10-03

notes

ColBERT README. Late-interaction retrieval: passages are encoded into matrices of token-level embeddings, MaxSim scoring at query time. Ships the ColBERTv2.0 checkpoint trained on MS MARCO Passage Ranking plus the LoTTE benchmark. Install is `pip install colbert-ai[torch,faiss-gpu]` with the README warning conda is more reliable for faiss/torch conflicts; a separate conda_env_cpu.yml exists for CPU-only. API metadata 2026-10-03: created 2020-05-25, last push 2025-10-14, v0.2.22 (2025-08-11), MIT, 290 open issues, 3944 stars.

Notes

ColBERT README. Late-interaction retrieval: passages are encoded into matrices of token-level embeddings, MaxSim scoring at query time. Ships the ColBERTv2.0 checkpoint trained on MS MARCO Passage Ranking plus the LoTTE benchmark. Install is `pip install colbert-ai[torch,faiss-gpu]` with the README warning conda is more reliable for faiss/torch conflicts; a separate conda_env_cpu.yml exists for CPU-only. API metadata 2026-10-03: created 2020-05-25, last push 2025-10-14, v0.2.22 (2025-08-11), MIT, 290 open issues, 3944 stars.

Cross-references

This record references no other records.