Generate Embeddings

The Lantern Extras Postgres extension enables generating embeddings using SQL with the functions text_embedding and image_embedding.

Note that generating embeddings is a CPU-intensive task and large scale embedding generation processes. For large scale embedding generation, the Lantern CLI provides a separate process.

Run Embedding Generation

To generate one-off text embeddings, use the text_embedding function. For example, to generate an embedding for the text My text input using the embedding model BAAI/bge-small-en, run

sql

SELECT text_embedding('BAAI/bge-small-en', 'My text input');

To generate image embeddings, use the image_embedding function. For example, to generate an embedding for the image https://lantern.dev/images/home/footer.png using the embedding model clip/ViT-B-32-visual, run

sql

SELECT image_embedding('clip/ViT-B-32-visual', 'https://lantern.dev/images/home/footer.png');

The above mentioned functions will use local models using ort runtime.

If you want to generate embeddings using OpenAI or Cohere APIs you can use the following functions:

sql

SET lantern_extras.openai_token='xxxxxxxxxxxxx';
SET lantern_extras.openai_azure_api_token='xxxxxxxxxxxxx'; -- For Azure deployment with API Key authentication
SET lantern_extras.openai_azure_entra_token='xxxxxxxxxxxxx'; -- For Azure deployment with Microsoft Entra ID authentication
SET lantern_extras.openai_deployment_url='https://YOUR_RESOURCE_NAME.openai.azure.com/openai/deployments/YOUR_DEPLOYMENT_NAME/embeddings?api-version=2023-05-15' -- You can set this GUC or pass via arguments

SELECT openai_embedding('openai/text-embedding-ada-002', 'My text input');
SELECT openai_embedding('openai/text-embedding-ada-002', 'My text input', 'https://YOUR_RESOURCE_NAME.openai.azure.com/openai/deployments/YOUR_DEPLOYMENT_NAME/embeddings?api-version=2023-05-15'); -- Use azure deployment
SELECT openai_embedding('openai/text-embedding-v3-small', 'My text input', '', 768); -- Provide dimensions for new models
SELECT openai_embedding('openai/text-embedding-v3-large', 'My text input', '', 3072); -- Provide dimensions for new models

For more info about azure_api_token and azure_entra_token variables refer to Azure Docs

Cohere embeddings

sql

SET lantern_extras.cohere_token='xxxxxxxxxxxxx';
SELECT cohere_embedding('cohere/embed-multilingual-v3.0 ', 'My text input');
SELECT cohere_embedding('cohere/embed-multilingual-v3.0 ', 'My text input', 'search_query'); -- This is the default type for embedding. Use this when doing queries
SELECT cohere_embedding('cohere/embed-multilingual-v3.0 ', 'My text input', 'search_document'); -- This is type for embedding when you want to store in database

For more info about embedding type refer to Cohere Docs

Supported Models

The following embedding models are currently supported:

Model Name	Dimensions	Max Tokens	Data Type	Runtime
`clip/ViT-B-32-visual`	512	224	Image	ort
`clip/ViT-B-32-textual`	512	77	Text	ort
`microsoft/all-mpnet-base-v2`	768	128	Text	ort
`microsoft/all-MiniLM-L12-v2`	384	128	Text	ort
`transformers/multi-qa-mpnet-base-dot-v1`	768	250	Text	ort
`thenlper/gte-base`	768	128	Text	ort
`thenlper/gte-large`	1024	128	Text	ort
`llmrails/ember-v1`	1024	512	Text	ort
`intfloat/e5-base-v2`	768	512	Text	ort
`intfloat/e5-large-v2`	1024	512	Text	ort
`BAAI/bge-small-en`	384	512	Text	ort
`BAAI/bge-base-en`	768	512	Text	ort
`BAAI/bge-large-en`	1024	512	Text	ort
`BAAI/bge-m3`	1024	8192	Text	ort
`jinaai/jina-embeddings-v2-small-en`	512	8192	Text	ort
`jinaai/jina-embeddings-v2-base-en`	768	8192	Text	ort
`openai/text-embedding-ada-002`	1536	8192	Text	openai
`openai/text-embedding-3-small`	512 - 1536	8192	Text	openai
`openai/text-embedding-3-large`	256 - 3072	8192	Text	openai
`cohere/embed-english-v3.0`	1024	512	Text	cohere
`cohere/embed-multilingual-v3.0`	1024	512	Text	cohere
`cohere/embed-english-v2.0`	4096	512	Text	cohere
`cohere/embed-english-light-v2.0`	1024	512	Text	cohere
`cohere/embed-multilingual-v2.0`	768	512	Text	cohere
`cohere/embed-english-light-v3.0`	384	512	Text	cohere
`cohere/embed-multilingual-light-v3.0`	384	512	Text	cohere

Edit this page

On this page

Run Embedding Generation
Supported Models