diff --git a/examples/community/clip_guided_stable_diffusion.py b/examples/community/clip_guided_stable_diffusion.py
index 5c34efee0970..fbb233dccd7a 100644
--- a/examples/community/clip_guided_stable_diffusion.py
+++ b/examples/community/clip_guided_stable_diffusion.py
@@ -11,6 +11,7 @@
     AutoencoderKL,
     DDIMScheduler,
     DiffusionPipeline,
+    DPMSolverMultistepScheduler,
     LMSDiscreteScheduler,
     PNDMScheduler,
     UNet2DConditionModel,
@@ -63,7 +64,7 @@ def __init__(
         clip_model: CLIPModel,
         tokenizer: CLIPTokenizer,
         unet: UNet2DConditionModel,
-        scheduler: Union[PNDMScheduler, LMSDiscreteScheduler, DDIMScheduler],
+        scheduler: Union[PNDMScheduler, LMSDiscreteScheduler, DDIMScheduler, DPMSolverMultistepScheduler],
         feature_extractor: CLIPImageProcessor,
     ):
         super().__init__()
@@ -125,17 +126,12 @@ def cond_fn(
     ):
         latents = latents.detach().requires_grad_()
 
-        if isinstance(self.scheduler, LMSDiscreteScheduler):
-            sigma = self.scheduler.sigmas[index]
-            # the model input needs to be scaled to match the continuous ODE formulation in K-LMS
-            latent_model_input = latents / ((sigma**2 + 1) ** 0.5)
-        else:
-            latent_model_input = latents
+        latent_model_input = self.scheduler.scale_model_input(latents, timestep)
 
         # predict the noise residual
         noise_pred = self.unet(latent_model_input, timestep, encoder_hidden_states=text_embeddings).sample
 
-        if isinstance(self.scheduler, (PNDMScheduler, DDIMScheduler)):
+        if isinstance(self.scheduler, (PNDMScheduler, DDIMScheduler, DPMSolverMultistepScheduler)):
             alpha_prod_t = self.scheduler.alphas_cumprod[timestep]
             beta_prod_t = 1 - alpha_prod_t
             # compute predicted original sample from predicted noise also called
diff --git a/examples/community/clip_guided_stable_diffusion_img2img.py b/examples/community/clip_guided_stable_diffusion_img2img.py
index c9d2bc6e5931..c3dee5aa9e9a 100644
--- a/examples/community/clip_guided_stable_diffusion_img2img.py
+++ b/examples/community/clip_guided_stable_diffusion_img2img.py
@@ -13,6 +13,7 @@
     AutoencoderKL,
     DDIMScheduler,
     DiffusionPipeline,
+    DPMSolverMultistepScheduler,
     LMSDiscreteScheduler,
     PNDMScheduler,
     UNet2DConditionModel,
@@ -140,7 +141,7 @@ def __init__(
         clip_model: CLIPModel,
         tokenizer: CLIPTokenizer,
         unet: UNet2DConditionModel,
-        scheduler: Union[PNDMScheduler, LMSDiscreteScheduler, DDIMScheduler],
+        scheduler: Union[PNDMScheduler, LMSDiscreteScheduler, DDIMScheduler, DPMSolverMultistepScheduler],
         feature_extractor: CLIPFeatureExtractor,
     ):
         super().__init__()
@@ -263,17 +264,12 @@ def cond_fn(
     ):
         latents = latents.detach().requires_grad_()
 
-        if isinstance(self.scheduler, LMSDiscreteScheduler):
-            sigma = self.scheduler.sigmas[index]
-            # the model input needs to be scaled to match the continuous ODE formulation in K-LMS
-            latent_model_input = latents / ((sigma**2 + 1) ** 0.5)
-        else:
-            latent_model_input = latents
+        latent_model_input = self.scheduler.scale_model_input(latents, timestep)
 
         # predict the noise residual
         noise_pred = self.unet(latent_model_input, timestep, encoder_hidden_states=text_embeddings).sample
 
-        if isinstance(self.scheduler, (PNDMScheduler, DDIMScheduler)):
+        if isinstance(self.scheduler, (PNDMScheduler, DDIMScheduler, DPMSolverMultistepScheduler)):
             alpha_prod_t = self.scheduler.alphas_cumprod[timestep]
             beta_prod_t = 1 - alpha_prod_t
             # compute predicted original sample from predicted noise also called