Interface EmbeddingProviderModelInfo

The specific models that the provider supports.

May include an endpoint-defined-model for some providers, such as huggingfaceDedicated, where the model may be truly arbitrary.

Example

// nvidia.models[0]
{
  name: 'NV-Embed-QA',
  vectorDimension: 1024,
  parameters: [],
}

Field

name - The name of the model to use

Field

vectorDimension - The preset, exact vector dimension to be used (if applicable)

Field

parameters - Any additional parameters the model may take in

See

EmbeddingProviderInfo

interface EmbeddingProviderModelInfo {
    name: string;
    parameters: EmbeddingProviderModelParameterInfo[];
    vectorDimension: null | number;
}

Properties

name: string

The name of the model to use.

May be endpoint-defined-model for some providers, such as huggingfaceDedicated, where the model may be truly arbitrary.

Example

// openai.models[0].name
'text-embedding-3-small'

// huggingfaceDedicated.models[0].name
'endpoint-defined-model'

Any additional, arbitrary parameters the modem may take in. May or may not be required.

Passed into the parameters block in VectorizeServiceOptions (except for vectorDimension).

Example

// openai.models[0].parameters[0] (text-embedding-3-small)
{
  name: 'vectorDimension',
  type: 'number',
  required: true,
  defaultValue: '1536',
  validation: { numericRange: [2, 1536] },
  help: 'Vector dimension to use in the database and when calling OpenAI.',
}
vectorDimension: null | number

The preset, exact vector dimension to be used (if applicable).

If not present, a vectorDimension parameter will be present in the model.parameters block.

Example

// openai.models[3].vectorDimension (text-embedding-ada-002)
1536

// huggingfaceDedicated.models[0].vectorDimension (endpoint-defined-model)
null