When importing objects into Weaviate, you can choose to:
- Provide object vectors to Weaviate,
- Configure a vectorizer to obtain vectors at import time, or
- Combine these options to configure a vectorizer while explicitly providing vectors.
This 5-minute video provides an overview of all three workflows, with code examples shown in Python.
We recommend that you go through the Quickstart tutorial before watching this video, or to make sure you are familiar with the basic of a schema, import, and query in Weaviate.
Jupyter notebook & data
You can find the accompanying Jupyter notebook and data at this repository.
Video (4mins 46secs)
- 0:00: Introduction
- 0:15: Overview
- 0:30: Instantiate Weaviate client & load data
- 1:18: Import data with Weaviate-obtained vectors
- 2:17: Import data with user-specified vectors
- 3:14: Combined option - import with user-specified vectors and specify a vectorizer
- 4:20: Wrap-up