Skip to main content
Weaviate Docs (migrated from docs.weaviate.io) Docs

Search documentation

Type to search this documentation.

On this pageOverview

Multimodal Embeddings

Weaviate's integration with the Meta ImageBind library allows you to access its capabilities directly from Weaviate. The ImageBind model supports multiple modalities (text, image, audio, video, thermal, IMU and depth).

Configure a Weaviate vector index to use the ImageBind integration, and configure the Weaviate instance with a model image, and Weaviate will generate embeddings for various operations using the specified model in the ImageBind inference container. This feature is called the vectorizer.

At import time, Weaviate generates multimodal object embeddings and saves them into the index. For vector and hybrid search operations, Weaviate converts queries of one or more modalities into embeddings. Multimodal search operations are also supported.

Embedding integration illustration

Requirements

Section titled “Requirements”

Weaviate configuration

Section titled “Weaviate configuration”

Your Weaviate instance must be configured with the ImageBind multimodal vectorizer integration (multi2vec-bind) module.

For Weaviate Cloud (WCD) users

This integration is not available for Weaviate Cloud (WCD) instances, as it requires spinning up a container with the ImageBind model.

Enable the integration module

Section titled “Enable the integration module”

Configure the integration

Section titled “Configure the integration”

To use this integration, you must configure the container image of the ImageBind model, and the inference endpoint of the containerized model.

The following example shows how to configure the ImageBind integration in Weaviate:

Docker Option 1: Use a pre-configured docker-compose.yml file

Follow the instructions on the Weaviate Docker installation configurator to download a pre-configured docker-compose.yml file with a selected model

Docker Option 2: Add the configuration manually

Alternatively, add the configuration to the docker-compose.yml file manually as in the example below.

YAML
services:
  weaviate:
    # Other Weaviate configuration
    environment:
      BIND_INFERENCE_API: http://multi2vec-bind:8080  # Set the inference API endpoint
  multi2vec-bind:  # Set the name of the inference container
    mem_limit: 12g
    image: cr.weaviate.io/semitechnologies/multi2vec-bind:imagebind
    environment:
      ENABLE_CUDA: 0  # Set to 1 to enable
  • BIND_INFERENCE_API environment variable sets the inference API endpoint
  • multi2vec-bind is the name of the inference container
  • image is the container image
  • ENABLE_CUDA environment variable enables GPU usage

Configure the ImageBind integration in Weaviate by adding or updating the multi2vec-bind module in the modules section of the Weaviate Helm chart values file. For example, modify the values.yaml file as follows:

YAML
modules:

  multi2vec-bind:

    enabled: true
    tag: imagebind
    repo: semitechnologies/multi2vec-bind
    registry: cr.weaviate.io
    envconfig:
      enable_cuda: true

See the Weaviate Helm chart for an example of the values.yaml file including more configuration options.

Credentials

Section titled “Credentials”

As this integration runs a local container with the ImageBind model, no additional credentials (e.g. API key) are required. Connect to Weaviate as usual, such as in the examples below.

Python
JavaScript/TypeScript

Configure the vectorizer

Section titled “Configure the vectorizer”

Configure a Weaviate index as follows to use an ImageBind embedding model:

Python
from weaviate.classes.config import Configure, DataType, Multi2VecField, Propertyclient.collections.create(    "DemoCollection",    properties=[        Property(name="title", data_type=DataType.TEXT),        Property(name="poster", data_type=DataType.BLOB),    ],    vector_config=[        Configure.Vectors.multi2vec_bind(            name="title_vector",            # Define the fields to be used for the vectorization - using image_fields, text_fields, video_fields            image_fields=[                Multi2VecField(name="poster", weight=0.9)            ],            text_fields=[                Multi2VecField(name="title", weight=0.1)            ]        )    ],    # Additional parameters not shown)
JavaScript/TypeScript
await client.collections.create({  name: 'DemoCollection',  properties: [    {      name: 'title',      dataType: 'text' as const,    },    {      name: 'poster',      dataType: 'blob' as const,    },  ],  vectorizers: [    weaviate.configure.vectors.multi2VecClip({      name: 'title_vector',      imageFields: [        {          name: 'poster',          weight: 0.9,        },      ],      textFields: [        {          name: 'title',          weight: 0.1,        },      ],    }),  ],  // Additional parameters not shown});

There is only one ImageBind model available.

Vectorization behavior

Weaviate follows the collection configuration and a set of predetermined rules to vectorize objects.

Unless specified otherwise in the collection definition, the default behavior is to:

  • Only vectorize properties that use the text or text[] data type (unless skipped)
  • Sort properties in alphabetical (a-z) order before concatenating values
  • If vectorizePropertyName is true (false by default) prepend the property name to each property value
  • Join the (prepended) property values with spaces
  • Prepend the class name (unless vectorizeClassName is false)
  • Convert the produced string to lowercase

Vectorizer parameters

Section titled “Vectorizer parameters”

The ImageBind vectorizer supports multiple modalities (text, image, audio, video, thermal, IMU and depth). One or more of these can be specified in the vectorizer configuration as shown.

Python
from weaviate.classes.config import Configure, DataType, Multi2VecField, Propertyclient.collections.create(    "DemoCollection",    properties=[        Property(name="title", data_type=DataType.TEXT),        Property(name="poster", data_type=DataType.BLOB),        Property(name="sound", data_type=DataType.BLOB),        Property(name="video", data_type=DataType.BLOB),    ],    vector_config=[        Configure.Vectors.multi2vec_bind(            name="title_vector",            # Define the fields to be used for the vectorization            image_fields=[                Multi2VecField(name="poster", weight=0.7)            ],            text_fields=[                Multi2VecField(name="title", weight=0.1)            ],            audio_fields=[                Multi2VecField(name="sound", weight=0.1)            ],            video_fields=[                Multi2VecField(name="video", weight=0.1)            ],            # depth, IMU and thermal fields are also available        )    ],    # Additional parameters not shown)
JavaScript/TypeScript
await client.collections.create({  name: 'DemoCollection',  properties: [    {      name: 'title',      dataType: 'text' as const,    },    {      name: 'poster',      dataType: 'blob' as const,    },    {      name: 'sound',      dataType: 'blob' as const,    },    {      name: 'video',      dataType: 'blob' as const,    },  ],  vectorizers: [    weaviate.configure.vectors.multi2VecBind({      name: 'title_vector',      imageFields: [        {          name: 'poster',          weight: 0.7,        },      ],      textFields: [        {          name: 'title',          weight: 0.1,        },      ],      audioFields: [        {          name: 'sound',          weight: 0.1,        },      ],      videoFields: [        {          name: 'video',          weight: 0.1,        },      ],      // depth, IMU and thermal fields are also available    }),  ],});

Data import

Section titled “Data import”

After configuring the vectorizer, import data into Weaviate. Weaviate generates embeddings for the objects using the specified model.

Python
collection = client.collections.use("DemoCollection")with collection.batch.fixed_size(batch_size=200) as batch:    for src_obj in source_objects:        poster_b64 = url_to_base64(src_obj["poster_path"])        weaviate_obj = {            "title": src_obj["title"],            "poster": poster_b64  # Add the image in base64 encoding        }        # The model provider integration will automatically vectorize the object        batch.add_object(            properties=weaviate_obj,            # vector=vector  # Optionally provide a pre-obtained vector        )
JavaScript/TypeScript
const collectionName = 'DemoCollection'
const myCollection = client.collections.use(collectionName)

Searches

Section titled “Searches”

Once the vectorizer is configured, Weaviate will perform vector and hybrid search operations using the specified ImageBind model.

Embedding integration at search illustration

Vector (near text) search

Section titled “Vector (near text) search”

When you perform a vector search, Weaviate converts the text query into an embedding using the specified model and returns the most similar objects from the database.

The query below returns the n most similar objects from the database, set by limit.

Python
collection = client.collections.use("DemoCollection")response = collection.query.near_text(    query="A holiday film",  # The model provider integration will automatically vectorize the query    limit=2)for obj in response.objects:    print(obj.properties["title"])
JavaScript/TypeScript
const collectionName = 'DemoCollection'
const myCollection = client.collections.use(collectionName)
Section titled “Hybrid search”

When you perform a hybrid search, Weaviate converts the text query into an embedding using the specified model and returns the best scoring objects from the database.

The query below returns the n best scoring objects from the database, set by limit.

Python
collection = client.collections.use("DemoCollection")response = collection.query.hybrid(    query="A holiday film",  # The model provider integration will automatically vectorize the query    limit=2)for obj in response.objects:    print(obj.properties["title"])
JavaScript/TypeScript
const collectionName = 'DemoCollection'
const myCollection = client.collections.use(collectionName)
Section titled “Vector (near media) search”

When you perform a media search such as a near image search, Weaviate converts the query into an embedding using the specified model and returns the most similar objects from the database.

To perform a near media search such as near image search, convert the media query into a base64 string and pass it to the search query.

The query below returns the n most similar objects to the input image from the database, set by limit.

Python
def url_to_base64(url):
    import requests
    import base64

    image_response = requests.get(url)
    content = image_response.content
    return base64.b64encode(content).decode("utf-8")
JavaScript/TypeScript
const base64String = 'SOME_BASE_64_REPRESENTATION';

result = await myCollection.query.nearImage(
  base64String,  // The model provider integration will automatically vectorize the query
  {
    limit: 2,
  }
)

console.log(JSON.stringify(result.objects, null, 2));

You can perform similar searches for other media types such as audio, video, thermal, IMU, and depth, by using an equivalent search query for the respective media type.

References

Section titled “References”

Available models

Section titled “Available models”

There is only one ImageBind model available.

Further resources

Section titled “Further resources”

Code examples

Section titled “Code examples”

Once the integrations are configured at the collection, the data management and search operations in Weaviate work identically to any other collection. See the following model-agnostic examples:

Model licenses

Section titled “Model licenses”

Review the license for the model on the ImageBind page.

It is your responsibility to evaluate whether the terms of its license(s), if any, are appropriate for your intended use.

External resources

Section titled “External resources”

Questions and feedback

Section titled “Questions and feedback”

Have a question or feedback? Here's how to reach us.

Community Forum

Ask questions and connect with other developers on our Community forum.

Support

Weaviate Cloud user or customer? Find the right channel on the Support page.

Suggest an edit

Propose a replacement for this page. The site team reviews it before applying any changes.

Export
Documentation menu