Use pipelines without vae (#10441)

* Use pipelines without vae

* getattr

* vqvae

---------
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>

src/diffusers/pipelines/stable_diffusion_panorama/pipeline_stable_diffusion_panorama.py

@@ -230,7 +230,7 @@ class StableDiffusionPanoramaPipeline(
             feature_extractor=feature_extractor,
             image_encoder=image_encoder,
         )
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
         self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
         self.register_to_config(requires_safety_checker=requires_safety_checker)
 

src/diffusers/pipelines/stable_diffusion_safe/pipeline_stable_diffusion_safe.py

@@ -149,7 +149,7 @@ class StableDiffusionPipelineSafe(DiffusionPipeline, StableDiffusionMixin, IPAda
             image_encoder=image_encoder,
         )
         self._safety_text_concept = safety_concept
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
         self.register_to_config(requires_safety_checker=requires_safety_checker)
 
     @property

src/diffusers/pipelines/stable_diffusion_sag/pipeline_stable_diffusion_sag.py

@@ -157,7 +157,7 @@ class StableDiffusionSAGPipeline(DiffusionPipeline, StableDiffusionMixin, Textua
             feature_extractor=feature_extractor,
             image_encoder=image_encoder,
         )
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
         self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
         self.register_to_config(requires_safety_checker=requires_safety_checker)
 

src/diffusers/pipelines/stable_diffusion_xl/pipeline_flax_stable_diffusion_xl.py

@@ -65,7 +65,7 @@ class FlaxStableDiffusionXLPipeline(FlaxDiffusionPipeline):
             unet=unet,
             scheduler=scheduler,
         )
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
 
     def prepare_inputs(self, prompt: Union[str, List[str]]):
         if not isinstance(prompt, (str, list)):

src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl.py

@@ -269,7 +269,7 @@ class StableDiffusionXLPipeline(
             feature_extractor=feature_extractor,
         )
         self.register_to_config(force_zeros_for_empty_prompt=force_zeros_for_empty_prompt)
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
         self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
 
         self.default_sample_size = self.unet.config.sample_size

src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_img2img.py

@@ -291,7 +291,7 @@ class StableDiffusionXLImg2ImgPipeline(
         )
         self.register_to_config(force_zeros_for_empty_prompt=force_zeros_for_empty_prompt)
         self.register_to_config(requires_aesthetics_score=requires_aesthetics_score)
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
         self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
 
         add_watermarker = add_watermarker if add_watermarker is not None else is_invisible_watermark_available()

src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_inpaint.py

@@ -321,7 +321,7 @@ class StableDiffusionXLInpaintPipeline(
         )
         self.register_to_config(force_zeros_for_empty_prompt=force_zeros_for_empty_prompt)
         self.register_to_config(requires_aesthetics_score=requires_aesthetics_score)
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
         self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
         self.mask_processor = VaeImageProcessor(
             vae_scale_factor=self.vae_scale_factor, do_normalize=False, do_binarize=True, do_convert_grayscale=True

src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_instruct_pix2pix.py

@@ -199,7 +199,7 @@ class StableDiffusionXLInstructPix2PixPipeline(
             scheduler=scheduler,
         )
         self.register_to_config(force_zeros_for_empty_prompt=force_zeros_for_empty_prompt)
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
         self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
         self.default_sample_size = self.unet.config.sample_size
         self.is_cosxl_edit = is_cosxl_edit

src/diffusers/pipelines/stable_video_diffusion/pipeline_stable_video_diffusion.py

@@ -177,7 +177,7 @@ class StableVideoDiffusionPipeline(DiffusionPipeline):
             scheduler=scheduler,
             feature_extractor=feature_extractor,
         )
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
         self.video_processor = VideoProcessor(do_resize=True, vae_scale_factor=self.vae_scale_factor)
 
     def _encode_image(

src/diffusers/pipelines/t2i_adapter/pipeline_stable_diffusion_adapter.py

@@ -260,7 +260,7 @@ class StableDiffusionAdapterPipeline(DiffusionPipeline, StableDiffusionMixin):
             safety_checker=safety_checker,
             feature_extractor=feature_extractor,
         )
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
         self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
         self.register_to_config(requires_safety_checker=requires_safety_checker)
 

src/diffusers/pipelines/t2i_adapter/pipeline_stable_diffusion_xl_adapter.py

@@ -293,7 +293,7 @@ class StableDiffusionXLAdapterPipeline(
             image_encoder=image_encoder,
         )
         self.register_to_config(force_zeros_for_empty_prompt=force_zeros_for_empty_prompt)
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
         self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
         self.default_sample_size = self.unet.config.sample_size
 

src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth.py

@@ -105,7 +105,7 @@ class TextToVideoSDPipeline(
             unet=unet,
             scheduler=scheduler,
         )
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
         self.video_processor = VideoProcessor(do_resize=False, vae_scale_factor=self.vae_scale_factor)
 
     # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline._encode_prompt

src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth_img2img.py

@@ -140,7 +140,7 @@ class VideoToVideoSDPipeline(
             unet=unet,
             scheduler=scheduler,
         )
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
         self.video_processor = VideoProcessor(do_resize=False, vae_scale_factor=self.vae_scale_factor)
 
     # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline._encode_prompt

src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_zero.py

@@ -358,7 +358,7 @@ class TextToVideoZeroPipeline(
                 " it only for use-cases that involve analyzing network behavior or auditing its results. For more"
                 " information, please have a look at https://github.com/huggingface/diffusers/pull/254 ."
             )
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
         self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
 
     def forward_loop(self, x_t0, t0, t1, generator):

src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_zero_sdxl.py

@@ -409,7 +409,7 @@ class TextToVideoZeroSDXLPipeline(
             feature_extractor=feature_extractor,
         )
         self.register_to_config(force_zeros_for_empty_prompt=force_zeros_for_empty_prompt)
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
         self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
 
         self.default_sample_size = self.unet.config.sample_size

src/diffusers/pipelines/unidiffuser/pipeline_unidiffuser.py

@@ -117,7 +117,7 @@ class UniDiffuserPipeline(DiffusionPipeline):
             scheduler=scheduler,
         )
 
-        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
+        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
         self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
 
         self.num_channels_latents = vae.config.latent_channels