G
GetLLMs

DreamO by Bytedance

Discover DreamO by Bytedance, a unified model for advanced image customization. Seamlessly handle identity preservation and virtual try-ons.

Platform: Replicate
Image CustomizationFace Identity PreservationVirtual Try-OnMulti-Condition Generation
715 runs
A100 (80GB)
Commercial

🚀Function Overview

A unified framework for image customization tasks including identity preservation, virtual try-on, and style transfer using reference images and prompts.

Key Features

  • Identity preservation for objects/characters (IP) and facial features (ID)
  • Virtual try-on capabilities for clothing/accessories
  • Style transfer between images
  • Multi-condition generation handling (e.g., IP + ID combinations)
  • Feature routing constraint to resolve condition conflicts

Use Cases

  • Creating personalized avatars with specific facial identities
  • Generating objects/characters in new scenes
  • Virtual clothing/accessory try-ons
  • Applying artistic styles to generated images
  • Combining multiple subjects/styles in single images

⚙️Input Parameters

prompt

string

Prompt for image generation

ref_image1

string

Reference image 1 (optional)

ref_task1

string

Task for reference image 1 ('ip': object/character, 'id': face identity, 'style': preserve style/background)

ref_image2

string

Reference image 2 (optional)

ref_task2

string

Task for reference image 2 ('ip': object/character, 'id': face identity, 'style': preserve style/background)

width

integer

Width of the output image (must be multiple of 16)

height

integer

Height of the output image (must be multiple of 16)

num_steps

integer

Number of inference steps

guidance

number

Guidance scale. Lower for less intensity/more realism (e.g., faces), higher for stronger prompt adherence.

seed

integer

Random seed. Leave blank or set to -1 for random.

ref_res

integer

Resolution for non-ID reference image preprocessing (target pixel area)

neg_prompt

string

Negative prompt

neg_guidance

number

Negative guidance scale

true_cfg

number

True CFG scale (advanced, requires distilled CFG LoRA)

cfg_start_step

integer

CFG start step (advanced)

cfg_end_step

integer

CFG end step (advanced)

first_step_guidance

number

First step guidance scale override (advanced, 0 uses main guidance)

output_format

string

Format of the output image

output_quality

integer

Output quality for lossy formats (jpg, webp)

💡Usage Examples

Example 1

Input Parameters

{
  "seed": 7698454872441022000,
  "width": 1024,
  "height": 1024,
  "prompt": "the woman wearing a dress, In the banquet hall",
  "ref_res": 512,
  "guidance": 3.5,
  "true_cfg": 1,
  "num_steps": 12,
  "ref_task1": "id",
  "ref_task2": "ip",
  "neg_prompt": "",
  "ref_image1": "https://replicate.delivery/pbxt/MzZo0OcsWW6NBl10nsCZoIRcRP9aGSczwDwVSLW5QRn8zD42/0_1.webp",
  "ref_image2": "https://raw.githubusercontent.com/bytedance/DreamO/main/example_inputs/dress.png",
  "cfg_end_step": 0,
  "neg_guidance": 3.5,
  "output_format": "webp",
  "cfg_start_step": 0,
  "output_quality": 90,
  "first_step_guidance": 0
}

Output Results

https://replicate.delivery/yhqm/h78bCImZ6HrtBVM0pE0EEXW4naneIvVBkdviQJML2rC85erUA/output.webp