Skip to main content

hunyuan-image Generate images model parameter specifications

TCB AI's hunyuan-image image generation model supports generating images from text descriptions. This document describes the parameters supported by the model.

Supported Parameters​

Parameter NameTypeRequiredDescription
promptstringRequiredThe text description used to generate images
modelstringRequiredModel name, use hunyuan-image
sizestringNot requiredImage dimensions, defaults to 1024x1024
revisebooleanNot requiredwhether to revise the prompt, defaults to true
footnotestringNot requiredCustom watermark content, limited to 16 characters
seednumberNot requiredGeneration seed, range [1, 4294967295]

Parameter Details​

size Image Dimensions​

  • Default value: 1024x1024
  • Supports custom dimensions: width and height must be between 768-1280 and must be multiples of 64.
  • If the input number is not a multiple of 64, it will be rounded up to the nearest multiple of 64, and then cropped to the target size from the center point.

Common Dimensions:

  • 1024x1024 - Square (default)
  • 1024x768 - Landscape
  • 768x1024 - Portrait

revise Prompt Rewriting​

  • Default value: true
  • After enabling, the model will automatically optimize your prompt, which can improve generation results in most scenarios.
  • If you wish to strictly follow the original prompt for generation, set this to false

footnote Watermark​

  • Limit to 16 characters (both Chinese and English characters count as one character each).
  • The watermark is generated at the bottom-right corner of the image.
  • Can be used to add brand logos or copyright information.

seed generation seed​

  • Only takes effect when the number of generated images is 1.
  • Range: 1 to 4294967295
  • Random when not passed
  • Using the same seed + prompt can generate similar images, which can be used to reproduce results.

Invocation Example​

Mini Program Cloud Function Invocation​

const cloud = require('wx-server-sdk');
cloud.init({ env: cloud.DYNAMIC_CURRENT_ENV });

exports.main = async (event) => {
  const ai = cloud.extend.AI;
  
  const result = await ai.createImage({
    model: 'hunyuan-image',
    prompt: 'a cute orange cat napping in the sunlight',
    size: '1024x1024',
    revise: true,
    footnote: 'TCB generation'
  });
  
  return result;
};

Node.js SDK Invocation​

const cloudbase = require('@cloudbase/node-sdk');
const app = cloudbase.init({ env: 'your-env-id' });

const result = await app.ai.createImage({
  model: 'hunyuan-image',
  prompt: 'a cute orange cat napping in the sunlight',
  size: '1024x1024',
  revise: true
});

Common Errors​

400 Bad Request​

A 400 error will be returned when unsupported parameters are passed. Please make sure to only use the parameters listed in the table above.

Note: The style parameter is currently not supported. Do not pass it.

Similar Issues​

  1. What parameters does hunyuan-image support?
  2. How to set image dimensions for the image generation model?
  3. What does the revise parameter mean in hunyuan-image?
  4. How to add a watermark to generated images?
  5. What is the use of the seed parameter in the image generation model?
  6. How to resolve a 400 error reported by hunyuan-image?
  7. What sizes does TCB AI image generation support?
  8. How to reproduce identical image generation results?
  9. hunyuan-image image generation model invocation example
  10. Why is there an error with the style parameter in the image generation model?