Sentence Transformers v6.0 adds native ColBERT-style multi-vector retrieval, including ColPali
antoine_chaffin · x · 2026-08-18
Hugging Face shipped Sentence Transformers v6.0, adding a fourth model type, MultiVectorEncoder, for ColBERT-style late interaction (multi-vector) retrieval.
- Unlike dense embeddings that compress text into one vector, multi-vector models keep one vector per token and score via the MaxSim operator, preserving token-level matching that usually yields stronger retrieval at the cost of a bigger index.
- Any PyLate or Stanford ColBERT checkpoint loads directly; since ST now supports multimodal models, ColPali and other ColX visual document retrieval models can be trained and used with the same API — text queries match page images directly, no OCR needed.
- The blog covers loading checkpoints, encoding and MaxSim scoring, plugging into a search stack, running on page images, and keeping the index affordable — all on a plain pip install -U sentence-transformers.
Related event: Sentence Transformers v6.0 ships with first-class late interaction models(33 posts)→
More from coding & agent
- Using AI Agents to draft release reports from evidence collections — CodeByPoonam · 2026-08-19
- Netlify integrates OpenRouter to enable model swapping without code changes — thisiskp_ · 2026-08-19
- Dev runs three Codex accounts plus Claude to parallelize coding agents — ChanceKelch · 2026-08-19
- DeepSeek open sources 'deepseek-harness' agent framework with 130k+ stars — alex_verem · 2026-08-19
- Cloud agents save hours vs local setups: a dev's parallel-workflow experience — brandon_galang · 2026-08-19
- IBM Bob resolves 60.6% of SWE-bench Verified at a median $1.05 per task — TejasKumar_ · 2026-08-19