@h-lumos/llm-sdk/dist-types/vyro/resources/images.d.ts

通用大模型 SDK

import { ReadableStream } from 'node:stream/web';
import { RequestOptions, Uploadable } from 'openai/core';
import { APIResource } from '../resource';

export declare class Images extends APIResource {
    protected models: Record<ImageModel, number>;
    /**
     * Creates a variation of a given image.
     */
    createVariation(params: ImageCreateVariationParams, options?: RequestOptions): Promise<ImagesResponse>;
    /**
     * Experience the magic of Imagine's Image Remix feature, designed to breathe new life into your existing images.
     */
    edit(params: ImageEditParams, options?: RequestOptions): Promise<ImagesResponse>;
    /**
     * Creates an image given a prompt.
     */
    generate(params: ImageGenerateParams, options?: RequestOptions): Promise<ImagesResponse>;
    /**
     * The image upscale feature provides a better image to the user by increasing its resolution.
     */
    upscale(params: ImageUpscaleParams, options?: RequestOptions): Promise<ImagesResponse>;
    /**
     * Inpaint is an advanced feature of the Text-to-Image Stable Diffusion pipeline.
     * It allows users to remove unwanted objects or elements from an image by intelligently filling in the missing areas.
     */
    restoration(params: ImageRestorationParams, options?: RequestOptions): Promise<ImagesResponse>;
}
export type ImageModel = 'imagine-v5' | 'anime-v5' | 'imagine-v4.1' | 'imagine-v4' | 'imagine-v3' | 'imagine-v1' | 'realistic' | 'anime' | 'portrait' | 'sdxl-1.0';
export interface ImageRestorationParams {
    /**
     * The image to use as the basis for the variation(s). Must be a valid PNG file,
     * less than 4MB, and square.
     */
    image: Uploadable;
    /**
     * The mask indicating the areas to be inpainted.
     */
    mask: Uploadable;
    /**
     * The text guides the image generation.
     */
    prompt: string;
    /**
     * The model to use for image generation.
     */
    model?: 'vyro-inpaint' | null;
    /**
     * The negative_prompt parameter empowers you to provide additional
     * guidance to the AI by specifying what you don't want in the image.
     * It helps refine the creative direction, ensuring that the generated
     * content aligns with your intentions.
     */
    negative_prompt?: string | null;
    /**
     * Specifies the model to be used. Currently supports only 1 for realism.
     *
     * @defaultValue 1
     */
    style?: 1 | null;
    /**
     * Weightage to be given to text
     *
     * Range: 3 - 15
     *
     * @defaultValue 7.5
     */
    cfg?: number | null;
    /**
     * Weightage given to initial image. Greater this parameter more the output will be close to starting image and far from prompt.
     *
     * Range: 0 - 1
     *
     * @defaultValue 0.5
     */
    strength?: number | null;
    /**
     * 目前仅支持 binary 格式
     */
    response_format?: 'binary' | null;
}
export interface ImageCreateVariationParams {
    /**
     * The image to use as the basis for the variation(s). Must be a valid PNG file,
     * less than 4MB, and square.
     */
    image: Uploadable;
    /**
     * The text guides the image generation.
     */
    prompt: string;
    /**
     * The model to use for image generation.
     */
    model?: ImageModel | null;
    /**
     * The negative_prompt parameter empowers you to provide additional
     * guidance to the AI by specifying what you don't want in the image.
     * It helps refine the creative direction, ensuring that the generated
     * content aligns with your intentions.
     */
    negative_prompt?: string | null;
    /**
     * The style_id parameter is like choosing an artistic palette for your image.
     * By selecting a style id, you guide the AI in crafting the image with a particular visual aesthetic.
     * Style IDs range from 1 to N, each representing a unique artistic style.
     *
     * @defaultValue 30
     */
    style?: number | null;
    /**
     * The steps parameter defines the number of operations or iterations that the
     * generator will perform during image creation. It can impact the complexity
     * and detail of the generated image.
     *
     * Range: 30-50
     *
     * @defaultValue 30
     */
    steps?: number | null;
    /**
     * The cfg parameter acts as a creative control knob.
     * You can adjust it to fine-tune the level of artistic innovation in the image.
     * Lower values encourage faithful execution of the prompt,
     * while higher values introduce more creative and imaginative variations.
     *
     * Range: 3 - 15
     *
     * @defaultValue 7.5
     */
    cfg?: number | null;
    /**
     * The seed parameter serves as the initial value for the random number generator.
     * By setting a specific seed value, you can ensure that the AI generates the same
     * image or outcome each time you use that exact seed.
     *
     * range: 1-Infinity
     */
    seed?: number | null;
    /**
     * Influences the impact of the control image on output.
     *
     * Range: 0 - 1
     *
     * @defaultValue 0
     */
    strength?: number | null;
    /**
     * 目前仅支持 binary 格式
     */
    response_format?: 'binary' | null;
}
export interface ImageEditParams {
    /**
     * The image to use as the basis for the variation(s). Must be a valid PNG file,
     * less than 4MB, and square.
     */
    image: Uploadable;
    /**
     * The text guides the image generation.
     */
    prompt: string;
    /**
     * The model to use for image generation.
     */
    model?: ImageModel | null;
    /**
     * The negative_prompt parameter empowers you to provide additional
     * guidance to the AI by specifying what you don't want in the image.
     * It helps refine the creative direction, ensuring that the generated
     * content aligns with your intentions.
     */
    negative_prompt?: string | null;
    /**
     * The style_id parameter is like choosing an artistic palette for your image.
     * By selecting a style id, you guide the AI in crafting the image with a particular visual aesthetic.
     * Style IDs range from 1 to N, each representing a unique artistic style.
     *
     * @defaultValue 29
     */
    style?: number | null;
    /**
     * The steps parameter defines the number of operations or iterations that the
     * generator will perform during image creation. It can impact the complexity
     * and detail of the generated image.
     *
     * Range: 30-50
     *
     * @defaultValue 30
     */
    steps?: number | null;
    /**
     * The cfg parameter acts as a creative control knob.
     * You can adjust it to fine-tune the level of artistic innovation in the image.
     * Lower values encourage faithful execution of the prompt,
     * while higher values introduce more creative and imaginative variations.
     *
     * Range: 3 - 15
     *
     * @defaultValue 7.5
     */
    cfg?: number | null;
    /**
     * The seed parameter serves as the initial value for the random number generator.
     * By setting a specific seed value, you can ensure that the AI generates the same
     * image or outcome each time you use that exact seed.
     *
     * range: 1-Infinity
     */
    seed?: number | null;
    /**
     * Influences the impact of the control image on output.
     *
     * Range: 0 - 1
     *
     * @defaultValue 0
     */
    strength?: number | null;
    /**
     * The method/control used to guide image generation.
     *
     * @defaultValue openpose
     */
    control?: 'openpose' | 'scribble' | 'canny' | 'lineart' | 'depth' | null;
    /**
     * 目前仅支持 binary 格式
     */
    response_format?: 'binary' | null;
}
export interface ImageGenerateParams {
    /**
     * A prompt is the text input that guides the AI in generating visual content.
     * It defines the textual description or concept for the image you wish to generate.
     * Think of it as the creative vision you want the AI to bring to life.
     * Crafting clear and creative prompts is crucial for achieving the desired results with Imagine's API.
     * For example, A serene forest with a river under the moonlight, can be a prompt.
     */
    prompt: string;
    /**
     * The model to use for image generation.
     */
    model?: ImageModel | null;
    /**
     * The negative_prompt parameter empowers you to provide additional
     * guidance to the AI by specifying what you don't want in the image.
     * It helps refine the creative direction, ensuring that the generated
     * content aligns with your intentions.
     */
    negative_prompt?: string | null;
    /**
     * The aspect_ratio parameter allows you to specify the proportions and dimensions of the generated image.
     * You can set it to different ratios like 1:1 for square images, 16:9 for widescreen, or 3:4 for vertical formats,
     * shaping the visual composition to your liking.
     *
     * @defaultValue 1:1
     */
    aspect_ratio?: '1:1' | '3:2' | '4:3' | '3:4' | '16:9' | '9:16' | null;
    /**
     * The quality parameter is a flag that, when set to hd,
     * requests high-resolution results from the AI.
     *
     * @defaultValue standard
     */
    quality?: 'standard' | 'hd';
    /**
     * The style_id parameter is like choosing an artistic palette for your image.
     * By selecting a style id, you guide the AI in crafting the image with a particular visual aesthetic.
     * Style IDs range from 1 to N, each representing a unique artistic style.
     *
     * @defaultValue 30
     */
    style?: number | null;
    /**
     * The steps parameter defines the number of operations or iterations that the
     * generator will perform during image creation. It can impact the complexity
     * and detail of the generated image.
     *
     * Range: 30-50
     *
     * @defaultValue 30
     */
    steps?: number | null;
    /**
     * The cfg parameter acts as a creative control knob.
     * You can adjust it to fine-tune the level of artistic innovation in the image.
     * Lower values encourage faithful execution of the prompt,
     * while higher values introduce more creative and imaginative variations.
     *
     * Range: 3 - 15
     *
     * @defaultValue 7.5
     */
    cfg?: number | null;
    /**
     * The seed parameter serves as the initial value for the random number generator.
     * By setting a specific seed value, you can ensure that the AI generates the same
     * image or outcome each time you use that exact seed.
     *
     * range: 1-Infinity
     */
    seed?: number | null;
    /**
     * 目前仅支持 binary 格式
     */
    response_format?: 'binary' | null;
}
export interface ImageUpscaleParams {
    /**
     * The image to use as the basis for the variation(s). Must be a valid PNG file,
     * less than 4MB, and square.
     */
    image: Uploadable;
}
export interface Image {
    /**
     * The binary of the generated image.
     */
    binary?: ReadableStream;
    /**
     * The base64-encoded JSON of the generated image, if `response_format` is
     * `b64_json`.
     */
    b64_json?: string;
    /**
     * The prompt that was used to generate the image, if there was any revision to the
     * prompt.
     */
    revised_prompt?: string;
    /**
     * The URL of the generated image, if `response_format` is `url` (default).
     */
    url?: string;
}
export interface ImagesResponse {
    /**
     * When the request was made.
     */
    created: number;
    /**
     * The generated images.
     */
    data: Image[];
}