Turn PostgreSQL into your search engine in a Pythonic way.
-
Updated
Aug 29, 2025 - Python
Turn PostgreSQL into your search engine in a Pythonic way.
Fused Triton kernels for late-interaction (MaxSim) scoring
Composable candidate generation and optional late-interaction (MaxSim) reranking, in Rust with Python bindings
Embedding and similarity scoring algorithm of Clair news API
SIMD-accelerated vector similarity primitives
ColPali-style multimodal late-interaction retrieval (text → document-image patches via MaxSim) — numpy reference implementation
A building docker file for Maxine VM and ZSim
Fast JavaScript MaxSim implementation for ColBERT retrieval. Supports pure JS and WebAssembly modes. Works in browsers and Node.js.
Code, data and run outputs for the article Negation Scope in MaxSim-Family Scorers: exact limits of max-pooled token matching, NegPairs permutation minimal pairs, 18 released systems on NegPairs and NevIR, late-interaction fine-tuning and a prespecified 718-run study. Every number in the article is rebuilt from the outputs.
MCP server exposing late-interaction (ColBERT-style MaxSim) document retrieval over a local folder — pluggable encoder, runs offline by default.
Local late-interaction retrieval: a cheap first stage, then an exact MaxSim rerank
ColBERT-style late-interaction MaxSim scoring for multi-vector retrieval — tiny, zero-dependency Rust crate
To associate your repository with the maxsim topic, visit your repo's landing page and select "manage topics."