diffusers 0.32.2__py3-none-any.whl → 0.33.0__py3-none-any.whl

diffusers/pipelines/stable_diffusion_k_diffusion/pipeline_stable_diffusion_xl_k_diffusion.py

@@ -170,10 +170,14 @@ class StableDiffusionXLKDiffusionPipeline(
             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.default_sample_size = (
+            self.unet.config.sample_size
+            if hasattr(self, "unet") and self.unet is not None and hasattr(self.unet.config, "sample_size")
+            else 128
+        )
 
         model = ModelWrapper(unet, scheduler.alphas_cumprod)
         if scheduler.config.prediction_type == "v_prediction":
@@ -321,7 +325,9 @@ class StableDiffusionXLKDiffusionPipeline(
                 prompt_embeds = text_encoder(text_input_ids.to(device), output_hidden_states=True)
 
                 # We are only ALWAYS interested in the pooled output of the final text encoder
-                pooled_prompt_embeds = prompt_embeds[0]
+                if pooled_prompt_embeds is None and prompt_embeds[0].ndim == 2:
+                    pooled_prompt_embeds = prompt_embeds[0]
+
                 if clip_skip is None:
                     prompt_embeds = prompt_embeds.hidden_states[-2]
                 else:
@@ -380,8 +386,10 @@ class StableDiffusionXLKDiffusionPipeline(
                     uncond_input.input_ids.to(device),
                     output_hidden_states=True,
                 )
+
                 # We are only ALWAYS interested in the pooled output of the final text encoder
-                negative_pooled_prompt_embeds = negative_prompt_embeds[0]
+                if negative_pooled_prompt_embeds is None and negative_prompt_embeds[0].ndim == 2:
+                    negative_pooled_prompt_embeds = negative_prompt_embeds[0]
                 negative_prompt_embeds = negative_prompt_embeds.hidden_states[-2]
 
                 negative_prompt_embeds_list.append(negative_prompt_embeds)

diffusers/pipelines/stable_diffusion_ldm3d/pipeline_stable_diffusion_ldm3d.py

@@ -30,6 +30,7 @@ from ...utils import (
     USE_PEFT_BACKEND,
     BaseOutput,
     deprecate,
+    is_torch_xla_available,
     logging,
     replace_example_docstring,
     scale_lora_layers,
@@ -40,8 +41,16 @@ from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
 from ..stable_diffusion.safety_checker import StableDiffusionSafetyChecker
 
 
+if is_torch_xla_available():
+    import torch_xla.core.xla_model as xm
+
+    XLA_AVAILABLE = True
+else:
+    XLA_AVAILABLE = False
+
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 
+
 EXAMPLE_DOC_STRING = """
     Examples:
         ```python
@@ -203,8 +212,8 @@ class StableDiffusionLDM3DPipeline(
             [`DDIMScheduler`], [`LMSDiscreteScheduler`], or [`PNDMScheduler`].
         safety_checker ([`StableDiffusionSafetyChecker`]):
             Classification module that estimates whether generated images could be considered offensive or harmful.
-            Please refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for more details
-            about a model's potential harms.
+            Please refer to the [model card](https://huggingface.co/stable-diffusion-v1-5/stable-diffusion-v1-5) for
+            more details about a model's potential harms.
         feature_extractor ([`~transformers.CLIPImageProcessor`]):
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
@@ -254,7 +263,7 @@ class StableDiffusionLDM3DPipeline(
             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 = VaeImageProcessorLDM3D(vae_scale_factor=self.vae_scale_factor)
         self.register_to_config(requires_safety_checker=requires_safety_checker)
 
@@ -1002,6 +1011,9 @@ class StableDiffusionLDM3DPipeline(
                         step_idx = i // getattr(self.scheduler, "order", 1)
                         callback(step_idx, t, latents)
 
+                if XLA_AVAILABLE:
+                    xm.mark_step()
+
         if not output_type == "latent":
             image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False)[0]
             image, has_nsfw_concept = self.run_safety_checker(image, device, prompt_embeds.dtype)

diffusers/pipelines/stable_diffusion_panorama/pipeline_stable_diffusion_panorama.py

@@ -26,6 +26,7 @@ from ...schedulers import DDIMScheduler
 from ...utils import (
     USE_PEFT_BACKEND,
     deprecate,
+    is_torch_xla_available,
     logging,
     replace_example_docstring,
     scale_lora_layers,
@@ -37,8 +38,16 @@ from ..stable_diffusion import StableDiffusionPipelineOutput
 from ..stable_diffusion.safety_checker import StableDiffusionSafetyChecker
 
 
+if is_torch_xla_available():
+    import torch_xla.core.xla_model as xm
+
+    XLA_AVAILABLE = True
+else:
+    XLA_AVAILABLE = False
+
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 
+
 EXAMPLE_DOC_STRING = """
     Examples:
         ```py
@@ -179,8 +188,8 @@ class StableDiffusionPanoramaPipeline(
             [`DDIMScheduler`], [`LMSDiscreteScheduler`], or [`PNDMScheduler`].
         safety_checker ([`StableDiffusionSafetyChecker`]):
             Classification module that estimates whether generated images could be considered offensive or harmful.
-            Please refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for more details
-            about a model's potential harms.
+            Please refer to the [model card](https://huggingface.co/stable-diffusion-v1-5/stable-diffusion-v1-5) for
+            more details about a model's potential harms.
         feature_extractor ([`~transformers.CLIPImageProcessor`]):
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
@@ -230,7 +239,7 @@ 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)
 
@@ -1155,6 +1164,9 @@ class StableDiffusionPanoramaPipeline(
                         step_idx = i // getattr(self.scheduler, "order", 1)
                         callback(step_idx, t, latents)
 
+                if XLA_AVAILABLE:
+                    xm.mark_step()
+
         if output_type != "latent":
             if circular_padding:
                 image = self.decode_latents_with_padding(latents)

diffusers/pipelines/stable_diffusion_safe/pipeline_stable_diffusion_safe.py

@@ -12,13 +12,20 @@ from ...image_processor import PipelineImageInput
 from ...loaders import IPAdapterMixin
 from ...models import AutoencoderKL, ImageProjection, UNet2DConditionModel
 from ...schedulers import KarrasDiffusionSchedulers
-from ...utils import deprecate, logging
+from ...utils import deprecate, is_torch_xla_available, logging
 from ...utils.torch_utils import randn_tensor
 from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
 from . import StableDiffusionSafePipelineOutput
 from .safety_checker import SafeStableDiffusionSafetyChecker
 
 
+if is_torch_xla_available():
+    import torch_xla.core.xla_model as xm
+
+    XLA_AVAILABLE = True
+else:
+    XLA_AVAILABLE = False
+
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 
 
@@ -46,8 +53,8 @@ class StableDiffusionPipelineSafe(DiffusionPipeline, StableDiffusionMixin, IPAda
             [`DDIMScheduler`], [`LMSDiscreteScheduler`], or [`PNDMScheduler`].
         safety_checker ([`StableDiffusionSafetyChecker`]):
             Classification module that estimates whether generated images could be considered offensive or harmful.
-            Please refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for more details
-            about a model's potential harms.
+            Please refer to the [model card](https://huggingface.co/stable-diffusion-v1-5/stable-diffusion-v1-5) for
+            more details about a model's potential harms.
         feature_extractor ([`~transformers.CLIPImageProcessor`]):
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
@@ -74,7 +81,7 @@ class StableDiffusionPipelineSafe(DiffusionPipeline, StableDiffusionMixin, IPAda
             " abuse, brutality, cruelty"
         )
 
-        if hasattr(scheduler.config, "steps_offset") and scheduler.config.steps_offset != 1:
+        if scheduler is not None and getattr(scheduler.config, "steps_offset", 1) != 1:
             deprecation_message = (
                 f"The configuration file of this scheduler: {scheduler} is outdated. `steps_offset`"
                 f" should be set to 1 instead of {scheduler.config.steps_offset}. Please make sure "
@@ -88,7 +95,7 @@ class StableDiffusionPipelineSafe(DiffusionPipeline, StableDiffusionMixin, IPAda
             new_config["steps_offset"] = 1
             scheduler._internal_dict = FrozenDict(new_config)
 
-        if hasattr(scheduler.config, "clip_sample") and scheduler.config.clip_sample is True:
+        if scheduler is not None and getattr(scheduler.config, "clip_sample", False) is True:
             deprecation_message = (
                 f"The configuration file of this scheduler: {scheduler} has not set the configuration `clip_sample`."
                 " `clip_sample` should be set to False in the configuration file. Please make sure to update the"
@@ -117,17 +124,21 @@ class StableDiffusionPipelineSafe(DiffusionPipeline, StableDiffusionMixin, IPAda
                 " checker. If you do not want to use the safety checker, you can pass `'safety_checker=None'` instead."
             )
 
-        is_unet_version_less_0_9_0 = hasattr(unet.config, "_diffusers_version") and version.parse(
-            version.parse(unet.config._diffusers_version).base_version
-        ) < version.parse("0.9.0.dev0")
-        is_unet_sample_size_less_64 = hasattr(unet.config, "sample_size") and unet.config.sample_size < 64
+        is_unet_version_less_0_9_0 = (
+            unet is not None
+            and hasattr(unet.config, "_diffusers_version")
+            and version.parse(version.parse(unet.config._diffusers_version).base_version) < version.parse("0.9.0.dev0")
+        )
+        is_unet_sample_size_less_64 = (
+            unet is not None and hasattr(unet.config, "sample_size") and unet.config.sample_size < 64
+        )
         if is_unet_version_less_0_9_0 and is_unet_sample_size_less_64:
             deprecation_message = (
                 "The configuration file of the unet has set the default `sample_size` to smaller than"
                 " 64 which seems highly unlikely .If you're checkpoint is a fine-tuned version of any of the"
                 " following: \n- CompVis/stable-diffusion-v1-4 \n- CompVis/stable-diffusion-v1-3 \n-"
-                " CompVis/stable-diffusion-v1-2 \n- CompVis/stable-diffusion-v1-1 \n- runwayml/stable-diffusion-v1-5"
-                " \n- runwayml/stable-diffusion-inpainting \n you should change 'sample_size' to 64 in the"
+                " CompVis/stable-diffusion-v1-2 \n- CompVis/stable-diffusion-v1-1 \n- stable-diffusion-v1-5/stable-diffusion-v1-5"
+                " \n- stable-diffusion-v1-5/stable-diffusion-inpainting \n you should change 'sample_size' to 64 in the"
                 " configuration file. Please make sure to update the config accordingly as leaving `sample_size=32`"
                 " in the config might lead to incorrect results in future versions. If you have downloaded this"
                 " checkpoint from the Hugging Face Hub, it would be very nice if you could open a Pull request for"
@@ -149,7 +160,7 @@ 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
@@ -739,6 +750,9 @@ class StableDiffusionPipelineSafe(DiffusionPipeline, StableDiffusionMixin, IPAda
                         step_idx = i // getattr(self.scheduler, "order", 1)
                         callback(step_idx, t, latents)
 
+                if XLA_AVAILABLE:
+                    xm.mark_step()
+
         # 8. Post-processing
         image = self.decode_latents(latents)
 

diffusers/pipelines/stable_diffusion_sag/pipeline_stable_diffusion_sag.py

@@ -27,6 +27,7 @@ from ...schedulers import KarrasDiffusionSchedulers
 from ...utils import (
     USE_PEFT_BACKEND,
     deprecate,
+    is_torch_xla_available,
     logging,
     replace_example_docstring,
     scale_lora_layers,
@@ -38,8 +39,16 @@ from ..stable_diffusion import StableDiffusionPipelineOutput
 from ..stable_diffusion.safety_checker import StableDiffusionSafetyChecker
 
 
+if is_torch_xla_available():
+    import torch_xla.core.xla_model as xm
+
+    XLA_AVAILABLE = True
+else:
+    XLA_AVAILABLE = False
+
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 
+
 EXAMPLE_DOC_STRING = """
     Examples:
         ```py
@@ -47,7 +56,7 @@ EXAMPLE_DOC_STRING = """
         >>> from diffusers import StableDiffusionSAGPipeline
 
         >>> pipe = StableDiffusionSAGPipeline.from_pretrained(
-        ...     "runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16
+        ...     "stable-diffusion-v1-5/stable-diffusion-v1-5", torch_dtype=torch.float16
         ... )
         >>> pipe = pipe.to("cuda")
 
@@ -123,8 +132,8 @@ class StableDiffusionSAGPipeline(DiffusionPipeline, StableDiffusionMixin, Textua
             [`DDIMScheduler`], [`LMSDiscreteScheduler`], or [`PNDMScheduler`].
         safety_checker ([`StableDiffusionSafetyChecker`]):
             Classification module that estimates whether generated images could be considered offensive or harmful.
-            Please refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for more details
-            about a model's potential harms.
+            Please refer to the [model card](https://huggingface.co/stable-diffusion-v1-5/stable-diffusion-v1-5) for
+            more details about a model's potential harms.
         feature_extractor ([`~transformers.CLIPImageProcessor`]):
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
@@ -157,7 +166,7 @@ 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)
 
@@ -840,6 +849,9 @@ class StableDiffusionSAGPipeline(DiffusionPipeline, StableDiffusionMixin, Textua
                             step_idx = i // getattr(self.scheduler, "order", 1)
                             callback(step_idx, t, latents)
 
+                    if XLA_AVAILABLE:
+                        xm.mark_step()
+
         if not output_type == "latent":
             image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False)[0]
             image, has_nsfw_concept = self.run_safety_checker(image, device, prompt_embeds.dtype)

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)):

diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl.py

@@ -269,10 +269,14 @@ 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
+        self.default_sample_size = (
+            self.unet.config.sample_size
+            if hasattr(self, "unet") and self.unet is not None and hasattr(self.unet.config, "sample_size")
+            else 128
+        )
 
         add_watermarker = add_watermarker if add_watermarker is not None else is_invisible_watermark_available()
 
@@ -406,7 +410,9 @@ class StableDiffusionXLPipeline(
                 prompt_embeds = text_encoder(text_input_ids.to(device), output_hidden_states=True)
 
                 # We are only ALWAYS interested in the pooled output of the final text encoder
-                pooled_prompt_embeds = prompt_embeds[0]
+                if pooled_prompt_embeds is None and prompt_embeds[0].ndim == 2:
+                    pooled_prompt_embeds = prompt_embeds[0]
+
                 if clip_skip is None:
                     prompt_embeds = prompt_embeds.hidden_states[-2]
                 else:
@@ -465,8 +471,10 @@ class StableDiffusionXLPipeline(
                     uncond_input.input_ids.to(device),
                     output_hidden_states=True,
                 )
+
                 # We are only ALWAYS interested in the pooled output of the final text encoder
-                negative_pooled_prompt_embeds = negative_prompt_embeds[0]
+                if negative_pooled_prompt_embeds is None and negative_prompt_embeds[0].ndim == 2:
+                    negative_pooled_prompt_embeds = negative_prompt_embeds[0]
                 negative_prompt_embeds = negative_prompt_embeds.hidden_states[-2]
 
                 negative_prompt_embeds_list.append(negative_prompt_embeds)

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()
@@ -427,7 +427,9 @@ class StableDiffusionXLImg2ImgPipeline(
                 prompt_embeds = text_encoder(text_input_ids.to(device), output_hidden_states=True)
 
                 # We are only ALWAYS interested in the pooled output of the final text encoder
-                pooled_prompt_embeds = prompt_embeds[0]
+                if pooled_prompt_embeds is None and prompt_embeds[0].ndim == 2:
+                    pooled_prompt_embeds = prompt_embeds[0]
+
                 if clip_skip is None:
                     prompt_embeds = prompt_embeds.hidden_states[-2]
                 else:
@@ -486,8 +488,10 @@ class StableDiffusionXLImg2ImgPipeline(
                     uncond_input.input_ids.to(device),
                     output_hidden_states=True,
                 )
+
                 # We are only ALWAYS interested in the pooled output of the final text encoder
-                negative_pooled_prompt_embeds = negative_prompt_embeds[0]
+                if negative_pooled_prompt_embeds is None and negative_prompt_embeds[0].ndim == 2:
+                    negative_pooled_prompt_embeds = negative_prompt_embeds[0]
                 negative_prompt_embeds = negative_prompt_embeds.hidden_states[-2]
 
                 negative_prompt_embeds_list.append(negative_prompt_embeds)

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
@@ -531,7 +531,9 @@ class StableDiffusionXLInpaintPipeline(
                 prompt_embeds = text_encoder(text_input_ids.to(device), output_hidden_states=True)
 
                 # We are only ALWAYS interested in the pooled output of the final text encoder
-                pooled_prompt_embeds = prompt_embeds[0]
+                if pooled_prompt_embeds is None and prompt_embeds[0].ndim == 2:
+                    pooled_prompt_embeds = prompt_embeds[0]
+
                 if clip_skip is None:
                     prompt_embeds = prompt_embeds.hidden_states[-2]
                 else:
@@ -590,8 +592,10 @@ class StableDiffusionXLInpaintPipeline(
                     uncond_input.input_ids.to(device),
                     output_hidden_states=True,
                 )
+
                 # We are only ALWAYS interested in the pooled output of the final text encoder
-                negative_pooled_prompt_embeds = negative_prompt_embeds[0]
+                if negative_pooled_prompt_embeds is None and negative_prompt_embeds[0].ndim == 2:
+                    negative_pooled_prompt_embeds = negative_prompt_embeds[0]
                 negative_prompt_embeds = negative_prompt_embeds.hidden_states[-2]
 
                 negative_prompt_embeds_list.append(negative_prompt_embeds)
@@ -737,7 +741,7 @@ class StableDiffusionXLInpaintPipeline(
         if padding_mask_crop is not None:
             if not isinstance(image, PIL.Image.Image):
                 raise ValueError(
-                    f"The image should be a PIL image when inpainting mask crop, but is of type" f" {type(image)}."
+                    f"The image should be a PIL image when inpainting mask crop, but is of type {type(image)}."
                 )
             if not isinstance(mask_image, PIL.Image.Image):
                 raise ValueError(
@@ -745,7 +749,7 @@ class StableDiffusionXLInpaintPipeline(
                     f" {type(mask_image)}."
                 )
             if output_type != "pil":
-                raise ValueError(f"The output type should be PIL when inpainting mask crop, but is" f" {output_type}.")
+                raise ValueError(f"The output type should be PIL when inpainting mask crop, but is {output_type}.")
 
         if ip_adapter_image is not None and ip_adapter_image_embeds is not None:
             raise ValueError(
@@ -1505,7 +1509,7 @@ class StableDiffusionXLInpaintPipeline(
                     f"Incorrect configuration settings! The config of `pipeline.unet`: {self.unet.config} expects"
                     f" {self.unet.config.in_channels} but received `num_channels_latents`: {num_channels_latents} +"
                     f" `num_channels_mask`: {num_channels_mask} + `num_channels_masked_image`: {num_channels_masked_image}"
-                    f" = {num_channels_latents+num_channels_masked_image+num_channels_mask}. Please verify the config of"
+                    f" = {num_channels_latents + num_channels_masked_image + num_channels_mask}. Please verify the config of"
                     " `pipeline.unet` or your `mask_image` or `image` input."
                 )
         elif num_channels_unet != 4:

diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_instruct_pix2pix.py

@@ -199,9 +199,13 @@ 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.default_sample_size = (
+            self.unet.config.sample_size
+            if hasattr(self, "unet") and self.unet is not None and hasattr(self.unet.config, "sample_size")
+            else 128
+        )
         self.is_cosxl_edit = is_cosxl_edit
 
         add_watermarker = add_watermarker if add_watermarker is not None else is_invisible_watermark_available()
@@ -333,7 +337,9 @@ class StableDiffusionXLInstructPix2PixPipeline(
                 )
 
                 # We are only ALWAYS interested in the pooled output of the final text encoder
-                pooled_prompt_embeds = prompt_embeds[0]
+                if pooled_prompt_embeds is None and prompt_embeds[0].ndim == 2:
+                    pooled_prompt_embeds = prompt_embeds[0]
+
                 prompt_embeds = prompt_embeds.hidden_states[-2]
 
                 prompt_embeds_list.append(prompt_embeds)
@@ -385,7 +391,8 @@ class StableDiffusionXLInstructPix2PixPipeline(
                     output_hidden_states=True,
                 )
                 # We are only ALWAYS interested in the pooled output of the final text encoder
-                negative_pooled_prompt_embeds = negative_prompt_embeds[0]
+                if negative_pooled_prompt_embeds is None and negative_prompt_embeds[0].ndim == 2:
+                    negative_pooled_prompt_embeds = negative_prompt_embeds[0]
                 negative_prompt_embeds = negative_prompt_embeds.hidden_states[-2]
 
                 negative_prompt_embeds_list.append(negative_prompt_embeds)

diffusers/pipelines/stable_video_diffusion/pipeline_stable_video_diffusion.py

@@ -24,14 +24,22 @@ from transformers import CLIPImageProcessor, CLIPVisionModelWithProjection
 from ...image_processor import PipelineImageInput
 from ...models import AutoencoderKLTemporalDecoder, UNetSpatioTemporalConditionModel
 from ...schedulers import EulerDiscreteScheduler
-from ...utils import BaseOutput, logging, replace_example_docstring
+from ...utils import BaseOutput, is_torch_xla_available, logging, replace_example_docstring
 from ...utils.torch_utils import is_compiled_module, randn_tensor
 from ...video_processor import VideoProcessor
 from ..pipeline_utils import DiffusionPipeline
 
 
+if is_torch_xla_available():
+    import torch_xla.core.xla_model as xm
+
+    XLA_AVAILABLE = True
+else:
+    XLA_AVAILABLE = False
+
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 
+
 EXAMPLE_DOC_STRING = """
     Examples:
         ```py
@@ -177,7 +185,7 @@ 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(
@@ -600,6 +608,9 @@ class StableVideoDiffusionPipeline(DiffusionPipeline):
                 if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0):
                     progress_bar.update()
 
+                if XLA_AVAILABLE:
+                    xm.mark_step()
+
         if not output_type == "latent":
             # cast back to fp16 if needed
             if needs_upcasting:

diffusers/pipelines/t2i_adapter/pipeline_stable_diffusion_adapter.py

@@ -22,7 +22,7 @@ import torch
 from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer
 
 from ...image_processor import VaeImageProcessor
-from ...loaders import StableDiffusionLoraLoaderMixin, TextualInversionLoaderMixin
+from ...loaders import FromSingleFileMixin, StableDiffusionLoraLoaderMixin, TextualInversionLoaderMixin
 from ...models import AutoencoderKL, MultiAdapter, T2IAdapter, UNet2DConditionModel
 from ...models.lora import adjust_lora_scale_text_encoder
 from ...schedulers import KarrasDiffusionSchedulers
@@ -31,6 +31,7 @@ from ...utils import (
     USE_PEFT_BACKEND,
     BaseOutput,
     deprecate,
+    is_torch_xla_available,
     logging,
     replace_example_docstring,
     scale_lora_layers,
@@ -41,6 +42,14 @@ from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
 from ..stable_diffusion.safety_checker import StableDiffusionSafetyChecker
 
 
+if is_torch_xla_available():
+    import torch_xla.core.xla_model as xm
+
+    XLA_AVAILABLE = True
+else:
+    XLA_AVAILABLE = False
+
+
 @dataclass
 class StableDiffusionAdapterPipelineOutput(BaseOutput):
     """
@@ -59,6 +68,7 @@ class StableDiffusionAdapterPipelineOutput(BaseOutput):
 
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 
+
 EXAMPLE_DOC_STRING = """
     Examples:
         ```py
@@ -178,7 +188,7 @@ def retrieve_timesteps(
     return timesteps, num_inference_steps
 
 
-class StableDiffusionAdapterPipeline(DiffusionPipeline, StableDiffusionMixin):
+class StableDiffusionAdapterPipeline(DiffusionPipeline, StableDiffusionMixin, FromSingleFileMixin):
     r"""
     Pipeline for text-to-image generation using Stable Diffusion augmented with T2I-Adapter
     https://arxiv.org/abs/2302.08453
@@ -208,7 +218,8 @@ class StableDiffusionAdapterPipeline(DiffusionPipeline, StableDiffusionMixin):
             [`DDIMScheduler`], [`LMSDiscreteScheduler`], or [`PNDMScheduler`].
         safety_checker ([`StableDiffusionSafetyChecker`]):
             Classification module that estimates whether generated images could be considered offensive or harmful.
-            Please, refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for details.
+            Please, refer to the [model card](https://huggingface.co/stable-diffusion-v1-5/stable-diffusion-v1-5) for
+            details.
         feature_extractor ([`CLIPImageProcessor`]):
             Model that extracts features from generated images to be used as inputs for the `safety_checker`.
     """
@@ -259,7 +270,7 @@ 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)
 
@@ -914,6 +925,9 @@ class StableDiffusionAdapterPipeline(DiffusionPipeline, StableDiffusionMixin):
                         step_idx = i // getattr(self.scheduler, "order", 1)
                         callback(step_idx, t, latents)
 
+                if XLA_AVAILABLE:
+                    xm.mark_step()
+
         if output_type == "latent":
             image = latents
             has_nsfw_concept = None