Optimizing ColPali for Retrieval at Scale - From Theory to Practice
Qdrant - Vector Database & Search Engine via YouTube
Overview
Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Join a technical webinar exploring the optimization of ColPali's multivector approach for document retrieval at scale. Learn how mean-pooling ColPali multivectors for first-stage retrieval, followed by reranking with original multivectors, achieves 12x faster search speeds while maintaining performance. Discover the implementation of ColPali in improving document retrieval, leveraging Qdrant's Binary Quantization, and applying ColPali pooling optimization techniques. Explore practical applications in RAG and Vision RAG systems through live demonstrations available on GitHub. Hosted by industry experts Sabrina Aquino, Evgeniya Sukhodolskaya, and Atita Arora, gain insights into addressing the computational challenges of building HNSW indices with multivectors and implementing efficient solutions for handling visually rich documents.
Syllabus
Optimizing ColPali for Retrieval at Scale, from Theory to Practice
Taught by
Qdrant - Vector Database & Search Engine