bytedance/sdxl-lightning-4step 🔢📝❓✓ → 🖼️

▶️ 1027.9M runs 📅 Feb 2024 ⚙️ Cog 0.14.1 🔗 GitHub 📄 Paper ⚖️ License
sdxl text-to-image

About

SDXL-Lightning by ByteDance: a fast text-to-image model that makes high-quality images in 4 steps

Example Output

Prompt:

"self-portrait of a woman, lightning in the background"

Output

Example output

Performance Metrics

0.94s Prediction Time
0.95s Total Time
All Input Parameters
{
  "width": 1024,
  "height": 1024,
  "prompt": "self-portrait of a woman, lightning in the background",
  "scheduler": "K_EULER",
  "num_outputs": 1,
  "guidance_scale": 0,
  "negative_prompt": "worst quality, low quality",
  "num_inference_steps": 4
}
Input Parameters
seed Type: integerDefault: 0
Random seed. Leave blank to randomize the seed
width Type: integerDefault: 1024Range: 256 - 1280
Width of output image. Recommended 1024 or 1280
height Type: integerDefault: 1024Range: 256 - 1280
Height of output image. Recommended 1024 or 1280
prompt Type: stringDefault: self-portrait of a woman, lightning in the background
Input prompt
scheduler Default: K_EULER
scheduler
num_outputs Type: integerDefault: 1Range: 1 - 4
Number of images to output.
guidance_scale Type: numberDefault: 0Range: 0 - 50
Scale for classifier-free guidance
negative_prompt Type: stringDefault: worst quality, low quality
Negative Input prompt
num_inference_steps Type: integerDefault: 4Range: 1 - 10
Number of denoising steps. 4 for best results
disable_safety_checker Type: booleanDefault: false
Disable safety checker for generated images
Output Schema

Output

Type: arrayItems Type: stringItems Format: uri

Example Execution Logs
Using seed: 139922051
Prompt: self-portrait of a woman, lightning in the background
  0%|          | 0/4 [00:00<?, ?it/s]
 50%|█████     | 2/4 [00:00<00:00, 19.11it/s]
100%|██████████| 4/4 [00:00<00:00, 18.37it/s]
100%|██████████| 4/4 [00:00<00:00, 18.47it/s]
Version Details
Version ID
6f7a773af6fc3e8de9d5a3c00be77c17308914bf67772726aff83496ba1e3bbe
Version Created
March 17, 2025
Run on Replicate →