sambanova/resources/embeddings.d.mts

The official TypeScript library for the Samba Nova API

import { APIResource } from "../core/resource.mjs";
import { APIPromise } from "../core/api-promise.mjs";
import { RequestOptions } from "../internal/request-options.mjs";
export declare class Embeddings extends APIResource {
    /**
     * Create embeddings
     *
     * @example
     * ```ts
     * const embeddingsResponse = await client.embeddings.create({
     *   input: [
     *     'text to embed number 1',
     *     'text to embed number 2',
     *   ],
     *   model: 'E5-Mistral-7B-Instruct',
     * });
     * ```
     */
    create(body: EmbeddingCreateParams, options?: RequestOptions): APIPromise<EmbeddingsResponse>;
}
/**
 * Embeddings response returned by the model
 */
export interface EmbeddingsResponse {
    /**
     * The list of embeddings generated by the model.
     */
    data: Array<EmbeddingsResponse.Data>;
    /**
     * The name of the model used to generate the embedding.
     */
    model: string;
    /**
     * The object type, which is always "list".
     */
    object: 'list';
    /**
     * Usage metrics for the completion, embeddings,transcription or translation
     * request
     */
    usage: EmbeddingsResponse.Usage | null;
}
export declare namespace EmbeddingsResponse {
    /**
     * Represents an embedding vector returned by embeddings endpoint.
     */
    interface Data {
        /**
         * List of floats containing the embedding vector.
         */
        embedding: Array<number> | null;
        /**
         * The index of the embedding in the list of embeddings.
         */
        index: number;
        /**
         * Object type, always embedding.
         */
        object: 'embedding';
    }
    /**
     * Usage metrics for the completion, embeddings,transcription or translation
     * request
     */
    interface Usage {
        /**
         * acceptance rate
         */
        acceptance_rate?: number;
        /**
         * number of tokens generated in completion
         */
        completion_tokens?: number;
        /**
         * completion tokens per second after first token generation
         */
        completion_tokens_after_first_per_sec?: number;
        /**
         * completion tokens per second after first token generation first ten
         */
        completion_tokens_after_first_per_sec_first_ten?: number;
        /**
         * completion tokens per second after first token generation
         */
        completion_tokens_after_first_per_sec_graph?: number;
        /**
         * completion tokens per second
         */
        completion_tokens_per_sec?: number;
        /**
         * The Unix timestamp (in seconds) of when the generation finished.
         */
        end_time?: number;
        /**
         * whether or not is last response, always true for non streaming response
         */
        is_last_response?: true;
        /**
         * number of tokens used in the prompt sent
         */
        prompt_tokens?: number;
        /**
         * Extra tokens details
         */
        prompt_tokens_details?: Usage.PromptTokensDetails;
        /**
         * The Unix timestamp (in seconds) of when the generation started.
         */
        start_time?: number;
        /**
         * also TTF, time (in seconds) taken to generate the first token
         */
        time_to_first_token?: number;
        /**
         * total time (in seconds) taken to generate the full generation
         */
        total_latency?: number;
        /**
         * prompt tokens + completion tokens
         */
        total_tokens?: number;
        /**
         * tokens per second including prompt and completion
         */
        total_tokens_per_sec?: number;
        [k: string]: unknown;
    }
    namespace Usage {
        /**
         * Extra tokens details
         */
        interface PromptTokensDetails {
            /**
             * amount of cached tokens
             */
            cached_tokens?: number;
            [k: string]: unknown;
        }
    }
}
export interface EmbeddingCreateParams {
    /**
     * Input text to embed. to embed multiple inputs in a single request, pass an array
     * of strings. The input must not exceed the max input tokens for the model
     */
    input: string | Array<string>;
    /**
     * The model ID to use See available
     * [models](https://docs.sambanova.ai/cloud/docs/get-started/supported-models)
     */
    model: (string & {}) | 'E5-Mistral-7B-Instruct';
    [k: string]: unknown;
}
export declare namespace Embeddings {
    export { type EmbeddingsResponse as EmbeddingsResponse, type EmbeddingCreateParams as EmbeddingCreateParams, };
}
//# sourceMappingURL=embeddings.d.mts.map