Embedding Quantization: 25-45x retrieval speedup, 32x or 4x less memory usagehuggingface.co4 points·cubie··0 commentsOpen articleSaveView on HN