600% memory footprint reduction on Vector Embeddings Quantizationgithub.com4 points·betto··1 commentOpen articleSaveView on HN