Skip to main content
The Step Schedule determines how much freedom the model has when generating outputs. It controls the mix between:
  • Input Structure: Your original image (Reality)
  • Diffusion Noise: Random static (Creativity)
A diffusion model generates images by adding noise to “damage” an input. The model then tries to “repair” the damage by converting that noise into image details.

​
Single Step Schedule

Choose a single step between 1–50. Where you place the step determines the mixture of noise vs. structure:
  • Low steps (1–15): More emphasis on Diffusion Noise → more creativity, looser structure, less adherence to the input
  • High steps (35–50): More emphasis on Input Structure → outputs that strongly match your input
Step Schedule single step schedule moving from 1-49; prompt: blueberries

​
Multi-Step Schedule

You can also use a multi-step schedule with up to four steps in order from lowest to highest.
Multi-step schedules require more compute power.
A Multi-Step Schedule runs several refinement passes within a single frame. Instead of jumping directly from one noise level to the final output, the model moves through multiple steps in sequence (e.g., 1 → 24 → 49). Each step refines the result of the previous one. Step Scheduler 1 step schedule vs 3 step schedule; prompt: blueberries Stacking steps creates a smoother progression from “creative” to “accurate,” resulting in outputs that are more coherent and stable than a single-step schedule.

​
Usage

Set the step schedule using the t_index_list parameter: 
{
  "pipeline": "streamdiffusion",
  "params": {
    "model_id": "stabilityai/sdxl-turbo",
    "prompt": "blueberries",
    "t_index_list": [5, 15, 32]
  }
}

​
Examples

t_index_listEffect
[5]Single step, very creative, loose structure
[35]Single step, accurate, follows input closely
[5, 15, 32]Multi-step, balanced creativity and accuracy
[1, 24, 49]Multi-step, full range refinement