Qdrant 1.20 makes search cheaper and faster. Cheaper: turbo8 and turbo16 store vectors in a quarter...
Qdrant(@qdrant_engine) · 人工智能
Qdrant 1.20 makes search cheaper and faster. Cheaper: turbo8 and turbo16 store vectors in a quarter or half the space of float32, with recall holding steady. Faster: nested filters get up to 4x more queries per second. Check the blog in the first comment👇 💬 2 🔄 2 ❤️ 6 👀 328 📊 4 ⚡ Powered by xgo.ing