600% memory footprint reduction on Vector Embeddings Quantization github.com 3 points by betto 14 hours ago
Reduced by 600%?