Скачать или смотреть Partitioned S3 Vector Indexes for Faster, Scalable Search | Spice.ai Demo

In this demo, we introduce Partitioned S3 Vector Indexes, a new feature in Spice.ai that makes large-scale vector search faster and more efficient.

Partitioning allows you to divide an S3 vector index into multiple buckets based on a column or expression, enabling scatter-gather queries across multiple partitions. This helps bypass per-index ingestion limits, improves query latency, and scales seamlessly to billions of embeddings.

In this walkthrough, we show how to:

Add partitioning to a vector dataset in your spicepod.yaml
Embed and index data with a local embedding model
Perform vector searches with SQL and inspect the explain plan
See how scatter-gather queries operate across multiple S3 vector indices

📖 Check out Spice's Amazon S3 Vectors documentation for more details:
https://spiceai.org/docs/components/v...

Explore the docs:

OSS Docs: https://spiceai.org/docs
Cloud Docs: https://docs.spice.ai/

💬 Questions? Reach out to us on Discord:   / discord