Type alias Multi2VecNvidiaConfig

Multi2VecNvidiaConfig: {
    baseURL?: string;
    imageFields?: string[];
    model?: string;
    output_encoding?: string;
    textFields?: string[];
    truncation?: boolean;
    weights?: {
        imageFields?: number[];
        textFields?: number[];
    };
}

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

See the documentation for detailed usage.

Type declaration

  • Optional baseURL?: string

    The base URL where API requests should go.

  • Optional imageFields?: string[]

    The image fields used when vectorizing.

  • Optional model?: string

    The model to use. Defaults to None, which uses the server-defined default.

  • Optional output_encoding?: string

    Format in which the embeddings are encoded. Defaults to None, so the embeddings are represented as a list of floating-point numbers.

  • Optional textFields?: string[]

    The text fields used when vectorizing.

  • Optional truncation?: boolean

    Whether to apply truncation.

  • Optional weights?: {
        imageFields?: number[];
        textFields?: number[];
    }

    The weights of the fields used for vectorization.

    • Optional imageFields?: number[]

      The weights of the image fields.

    • Optional textFields?: number[]

      The weights of the text fields.