Prompt expansion - Img Open English

Request

Header Params

X-Request-req-accessKeyId

string

required

X-Request-req-accessKeySecret

string

required

Body Params application/json

businessTaskId

string

required

mode

integer

required

0：SD，1：SDXL

sourcePrompt

string

required

topK

integer

optional

Value range: [0, 250], precision is 1; when generating the next token, only the top k tokens with the highest probability can be considered; the larger the value, the greater the diversity of the generated text; the smaller the value, the diversity of the generated text The smaller; when K=0, this sampling mode is disabled

topP

number

optional

Value range: [0.01, 1.00], accuracy is 0.01; when generating the next token, only tokens with cumulative probability reaching top_p from high to low are truncated and selected for prediction; the larger the value, the higher the diversity.

temperature

number

optional

Value range: [0.1, 3.0], accuracy is 0.1; used to adjust the probability of predicting the next token; the larger the value, the higher the randomness and diversity; temperature=1 means no effect

repetitionPenalty

number

optional

Value range: [0.1, 5.0], precision is 0.1; repetition penalty coefficient, repetition_penalty=1 means no penalty

maxTokens

integer

optional

Value range: [32, 512], precision is 16; the maximum number of tokens allowed to be generated, regardless of the number of tokens in the input prompt; the larger the setting, the longer the prompt may be generated and the longer it will take

minTokens

integer

optional

Value range: [16, max_new_tokens], precision is 16; the minimum number of tokens generated does not take into account the number of tokens in the input prompt

sequencesNum

integer

optional

Value range: [1, 10], precision is 1; the larger the number of prompts generated at one time, the longer it takes. When using the gpt2 model, it is best to set it to less than 3. If there are too many prompts, multiple prompts will be generated repeatedly.

Example

{
    "mode": 0,
    "sourcePrompt": "a girl",
    "businessTaskId": "1008699601"
}

Request samples

Shell

JavaScript

Java

Swift

Go

PHP

Python

HTTP

C

C#

Objective-C

Ruby

OCaml

Dart

R

curl --location --request POST 'https://api.zhiqite.com/AIGCImgOpenServ/saas/tool/prompts/expand' \
--header 'X-Request-req-accessKeyId;' \
--header 'X-Request-req-accessKeySecret;' \
--header 'Content-Type: application/json' \
--data-raw '{
    "mode": 0,
    "sourcePrompt": "a girl",
    "businessTaskId": "1008699601"
}'

Responses

🟢200成功

application/json

Body

code

string

required

msg

string

required

data

object

optional

status

integer

required

0: Failure, 2: Success

prompts

array[string]

required

taskId

string

required

toolTask

string

required

businessTaskId

string

required

Example

{
    "code": "200",
    "msg": "success",
    "data": {
        "status": 0,
        "prompts": [
            "masterpiece, (photorealistic:1.4), best quality, beautiful lighting, 1 girl, solo, highly detailed face, detailed eyes, delicate features, looking at viewer, ultra-detailed, UHD, HDR"
        ],
        "taskId": "pre1715773856999364951840",
        "toolTask": "sd_prompts"
    }
}