🔍 Semantic Search Engine (Nexus)
This repository contains model artifacts for a hybrid semantic search engine.
📦 Files
faiss_index.bin → Dense vector index (FAISS)
corpus.pkl → Document corpus (text, title, URL)
bm25.pkl → Sparse BM25 index
⚙️ Architecture
- Dense Retrieval: msmarco-bert-base-dot-v5
- Sparse Retrieval: BM25
- Reranker: cross-encoder/ms-marco-MiniLM-L-12-v2
📊 Dataset
- Wikipedia (100k articles)
🚀 Use Case
Hybrid semantic + keyword search (Google-like engine)