Text-to-Image
Diffusers
stable-diffusion
PeterL1n commited on
Commit
b468d42
1 Parent(s): df6d2ec

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +3 -0
README.md CHANGED
@@ -70,6 +70,7 @@ pipe("A girl smiling", num_inference_steps=4, guidance_scale=0).images[0].save("
70
  ```
71
 
72
  ### 1-Step UNet
 
73
 
74
  The 1-step model uses "sample" prediction instead of "epsilon" prediction! The scheduler needs to be configured correctly.
75
 
@@ -118,6 +119,8 @@ Please use Euler sampler with sgm_uniform scheduler.
118
 
119
  ### 1-Step UNet
120
 
 
 
121
  1. Update your ComfyUI to the latest version.
122
  1. Download the full checkpoint (`sdxl_lightning_1step_x0.safetensors`) to `/ComfyUI/models/checkpoints`.
123
  1. Download our [ComfyUI full 1-step workflow](comfyui/sdxl_lightning_workflow_full_1step.json).
 
70
  ```
71
 
72
  ### 1-Step UNet
73
+ The 1-step model is only experimental and the quality is much less stable. Consider using the 2-step model for much better quality.
74
 
75
  The 1-step model uses "sample" prediction instead of "epsilon" prediction! The scheduler needs to be configured correctly.
76
 
 
119
 
120
  ### 1-Step UNet
121
 
122
+ The 1-step model is only experimental and the quality is much less stable. Consider using the 2-step model for much better quality.
123
+
124
  1. Update your ComfyUI to the latest version.
125
  1. Download the full checkpoint (`sdxl_lightning_1step_x0.safetensors`) to `/ComfyUI/models/checkpoints`.
126
  1. Download our [ComfyUI full 1-step workflow](comfyui/sdxl_lightning_workflow_full_1step.json).