weaviate-client
    Preparing search index...

    Type Alias Multi2VecVoyageAIConfig

    The configuration for multi-media vectorization using the VoyageAI module.

    See the documentation for detailed usage.

    type Multi2VecVoyageAIConfig = {
        baseURL?: string;
        dimensions?: number;
        imageFields?: string[];
        model?: "voyage-multimodal-3" | "voyage-multimodal-3.5" | string;
        outputEncoding?: string;
        textFields?: string[];
        truncate?: boolean;
        vectorizeCollectionName?: boolean;
        videoFields?: string[];
        weights?: {
            imageFields?: number[];
            textFields?: number[];
            videoFields?: number[];
        };
    }
    Index

    Properties

    baseURL?: string

    The base URL to use where API requests should go.

    dimensions?: number

    The dimensionality of the vector once embedded. Supported values: 256, 512, 1024 (default), 2048.

    imageFields?: string[]

    The image fields used when vectorizing.

    model?: "voyage-multimodal-3" | "voyage-multimodal-3.5" | string

    The model to use.

    outputEncoding?: string

    How the output from the model should be encoded on return.

    textFields?: string[]

    The text fields used when vectorizing.

    truncate?: boolean

    Whether the input should be truncated to fit in the context window.

    vectorizeCollectionName?: boolean

    Whether the collection name is vectorized.

    videoFields?: string[]

    The video fields used when vectorizing.

    weights?: {
        imageFields?: number[];
        textFields?: number[];
        videoFields?: number[];
    }

    The weights of the fields used for vectorization.

    Type Declaration

    • OptionalimageFields?: number[]

      The weights of the image fields.

    • OptionaltextFields?: number[]

      The weights of the text fields.

    • OptionalvideoFields?: number[]

      The weights of the video fields.