This project contain two Jupyter Notebooks:
wiki-large-setup
– shows how to configure a collection withtext2vec-cohere
andgenerative-cohere
modules. Then loads pre-vectorized Wikipedia data in 10 different languageswiki-large-query
– examples for Semantic Search and Generative Search queries
Notes:
- collection name:
Wikipedia
- languages included:
en
,de
,fr
,es
,it
,ja
,ar
,zh
,ko
,hi
- source: Cohere/wikipedia-22-12-(lang)-embeddings
To make this notebook work on your own Weaviate instance, you need to update or provide:
access_token
– auth token used to authenticate with your Weaviate instance. Note, if your Weaviate instance is configured without authentication, then you can remove the wholeauth_config
.url
– provide the url for your Weaviate instanceX-Cohere-Api-Key
– this is your API key you need to get from dashboard.cohere.ai/api-keys
Weaviate instances created with WCS provide all the required modules. No action required.
Make sure that your docker configuration includes the following modules:
text2vec-cohere
generative-cohere