👐 Import Vectorized Data
In the context of this workshop, you might not have enough time to vectorize your own data. Therefore, we have provided a vectorized version of the data for you to use.
The data has been pre-encoded using various providers, so you can use the one that you are most comfortable with.
Keep in mind that your queries will need to be encoded with the same provider as the data.
The model that was used for each data set is listed below:
| Data Set | Model | Documentation |
|---|---|---|
| Google Cloud Vertex AI | multimodalembedding@001 | Docs |
| OpenAI | text-embedding-ada-002 | Docs |
Import the embeddings
To import the dataset with the pre-created embeddings, you can use the import tool that you used earlier.
Enter your connection string in the text box at the top, and pick your provider from the dropdown at the bottom of the screen.
If unsure, or don't want to create your own embeddings, you can use the Serverless Endpoint provider.
Once you've picked your provider, click on the Add Embeddings button.
This process will drop your existing books collection to replace it with a new one. Any existing indexes or changes to your collection will be lost.
Once the import is complete, you will see a confirmation message.