Submit an image-to-image task - sd_img2img

更新时间:
复制 MD 格式

Description

This operation generates an image from an existing image using Stable Diffusion (SD).

Prerequisites

The AI Creation Workshop software development kit (SDK) is integrated. For more information, see SDK Reference for earlier versions.

Authorization

This operation uses an AccessKey pair (AccessKey ID and AccessKey secret) for authentication. For information about how to obtain an AccessKey pair, see API authorization.

Endpoint

https://openai.edu-aliyun.com

Request path

/scc/sd_img2img

Request headers

Name

Type

Description

X-SP-RESOURCE-CONFIG-ID

string

The ID of the resource pool configuration. This parameter is optional. For more information about the format, see Resource pool configuration parameters.

Request parameters

Name

Type

Description

stable_diffusion_model

string

The image generation model. This is the model file name on the platform.

sd_vae

string

The Variational Autoencoder (VAE) decoder. Common options include the following:

Automatic

"None": Disables the VAE.

"sdxl_vae.safetensors": The VAE used with XL models.

init_images

list

A list of initial images. For more information about the image format, see Image field format.

denoising_strength

double

Repaint Intensity

prompt

string

The positive prompt.

negtive_prompt

string

The negative prompt.

batch_size

integer

The number of images to generate in each batch.

steps

integer

The number of iteration steps.

cfg_scale

double

The prompt relevance.

width

integer

The image width.

height

integer

The image height.

sampler_name

string

The sampler.

seed

long

The random seed.

data parameters

Name

Type

Description

taskId

string

The ID of the image generation task.

status

string

The task status. A value of succeeded indicates that the task is successful. A value of failed indicates that the task failed. A value of running indicates that the task is in progress. A value of waiting indicates that the task is in a queue.