BM25 Part 3: Benchmarking Python BM25 Libraries
Benchmarking BM25 implementations: bm25-sparse, rankbm25, numpy vectorization, scipy, and plain Python. Find the fastest BM25 for your search use case.
نماذج التضمين
مقالات وتجارب وملاحظات حول نماذج التضمين.
BM25 Explained Part 2 | Qdrant hybrid search for real estate
Step-by-step guide to building a bilingual hybrid search system using Qdrant and BM25. Learn how to optimize real estate search with n-grams and RRF fusion.
BM25 Search Algorithm: Python Implementation from Scratch
Learn how BM25 improves search relevance over TF-IDF. This guide covers the math and Python code for implementing BM25 ranking from scratch with clear examples.
Diversifying Search Results with Pyversity and Qdrant
Learn how to diversify search results in a RAG system using Qdrant and Pyversity. Covers MMR, DPP, and other diversity algorithms with practical Python examples.
Bojji & Zarra Arabic Embeddings: MTEB vs all-MiniLM-L6-v2
Analyze Bojji and Zarra model2vec embeddings for Arabic, compare MTEB average score with all-MiniLM-L6-v2 and other static models.
Late Interaction & ColPali: Efficient Semantic Search
Explore Late Interaction models like ColBERT and ColPali for efficient token-level semantic search balancing speed and accuracy.
MTEB Benchmark: Tasks, Datasets & nDCG@10 Explained
Overview of MTEB, covering 58 datasets across 8 tasks, benchmarking models and explaining the nDCG@10 retrieval metric.
tiny-gte: Efficient Transformer for Semantic Search
tiny-gte is a 45MB distilled transformer from gte-small that produces 384D embeddings for efficient semantic search and vector DBs.