Upload combined_stable_diffusion.py with huggingface_hub

combined_stable_diffusion.py

@@ -1,10 +1,9 @@
 import torch
-from diffusers import DiffusionPipeline, DDPMScheduler
-from diffusers.pipelines.stable_diffusion import StableDiffusionSafetyChecker
+from diffusers import DiffusionPipeline, DDPMScheduler, StableDiffusionPipeline
 from diffusers.pipelines.stable_diffusion_xl.pipeline_output import StableDiffusionXLPipelineOutput
 from diffusers.image_processor import VaeImageProcessor
 from huggingface_hub import PyTorchModelHubMixin
-from transformers import CLIPTextModel, CLIPImageProcessor, CLIPTextModelWithProjection
+from transformers import CLIPTextModel, CLIPTextModelWithProjection
 from diffusers.models.attention_processor import (
     AttnProcessor2_0,
     FusedAttnProcessor2_0,
@@ -358,7 +357,6 @@ class CombinedStableDiffusionXL(
             guidance_scale=5,
             output_type: str = "pil",
             return_dict: bool = True,
-            generator=None,
     ):
         self.guidance_scale = guidance_scale
         self.resolution = resolution
@@ -440,20 +438,7 @@ class CombinedStableDiffusionXL(
                     # some platforms (eg. apple mps) misbehave due to a pytorch bug: https://github.com/pytorch/pytorch/pull/99272
                     self.vae = self.vae.to(latents.dtype)
 
-            # unscale/denormalize the latents
-            # denormalize with the mean and std if available and not None
-            has_latents_mean = hasattr(self.vae.config, "latents_mean") and self.vae.config.latents_mean is not None
-            has_latents_std = hasattr(self.vae.config, "latents_std") and self.vae.config.latents_std is not None
-            if has_latents_mean and has_latents_std:
-                latents_mean = (
-                    torch.tensor(self.vae.config.latents_mean).view(1, 4, 1, 1).to(latents.device, latents.dtype)
-                )
-                latents_std = (
-                    torch.tensor(self.vae.config.latents_std).view(1, 4, 1, 1).to(latents.device, latents.dtype)
-                )
-                latents = latents * latents_std / self.vae.config.scaling_factor + latents_mean
-            else:
-                latents = latents / self.vae.config.scaling_factor
+            latents = latents / self.vae.config.scaling_factor
 
             image = self.vae.decode(latents, return_dict=False)[0]