IDK-ab0ut commited on
Commit
a4808bf
·
verified ·
1 Parent(s): ab2d4cc

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +39 -1
README.md CHANGED
@@ -38,4 +38,42 @@ image = pipeline(prompt=prompt, negative_prompt=negative_prompt,
38
  image
39
  ```
40
 
41
- Feel free to edit the image's configuration with your desire.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
38
  image
39
  ```
40
 
41
+ Feel free to edit the image's configuration with your desire.
42
+
43
+ ### Scheduler's Customization ⚙️
44
+ #### (for Diffusers)
45
+ You can see all available schedulers [here](https://huggingface.co/docs/diffusers/v0.11.0/en/api/schedulers/overview).
46
+
47
+ To use scheduler other than DPM++ 2M Karras for this repo, make sure to import the
48
+ corresponding pipeline for the scheduler you want to use. For example, we want to use Euler. First, import [EulerDiscreteScheduler](https://huggingface.co/docs/diffusers/v0.29.2/en/api/schedulers/euler#diffusers.EulerDiscreteScheduler) from Diffusers by adding this line of code.
49
+ ```py
50
+ from diffusers import StableDiffusionXLPipeline, EulerDiscreteScheduler
51
+ ```
52
+
53
+ Next step is to load the scheduler.
54
+ ```py
55
+ model = "IDK-ab0ut/Yiffymix_v51"
56
+ euler = EulerDiscreteScheduler.from_pretrained(
57
+ model, subfolder="scheduler")
58
+ pipeline = StableDiffusionXLPipeline.from_pretrained(
59
+ model, scheduler=euler, torch.dtype=torch.float16
60
+ ).to("cuda")
61
+ ```
62
+ Now you can generate any images using the scheduler you want.
63
+
64
+ Another example is using DPM++ 2M SDE Karras. We want to import [DPMSolverMultistepScheduler](https://huggingface.co/docs/diffusers/v0.29.2/api/schedulers/multistep_dpm_solver) from Diffusers first.
65
+ ```py
66
+ from diffusers import StableDiffusionXLPipeline, DPMSolverMultistepScheduler
67
+ ```
68
+ Next, load the scheduler into the model.
69
+ ```py
70
+ model = "IDK-ab0ut/Yiffymix_v51"
71
+ dpmsolver = DPMSolverMultistepScheduler.from_pretrained(
72
+ model, subfolder="scheduler", use_karras_sigmas=True,
73
+ algorithm_type="sde-dpmsolver++").to("cuda")
74
+ # 'use_karras_sigmas' is called to make the scheduler
75
+ # use Karras sigmas during sampling.
76
+ pipeline = StableDiffusionXLPipeline.from_pretrained(
77
+ model, scheduler=dpmsolver, torch.dtype=torch.float16,
78
+ ).to("cuda")
79
+ ```