diffusers 0.32.1__py3-none-any.whl → 0.33.0__py3-none-any.whl

diffusers/pipelines/stable_diffusion_3/pipeline_stable_diffusion_3_inpaint.py

@@ -13,19 +13,21 @@
 # limitations under the License.
 
 import inspect
-from typing import Callable, Dict, List, Optional, Union
+from typing import Any, Callable, Dict, List, Optional, Union
 
 import torch
 from transformers import (
     CLIPTextModelWithProjection,
     CLIPTokenizer,
+    SiglipImageProcessor,
+    SiglipVisionModel,
     T5EncoderModel,
     T5TokenizerFast,
 )
 
 from ...callbacks import MultiPipelineCallbacks, PipelineCallback
 from ...image_processor import PipelineImageInput, VaeImageProcessor
-from ...loaders import FromSingleFileMixin, SD3LoraLoaderMixin
+from ...loaders import FromSingleFileMixin, SD3IPAdapterMixin, SD3LoraLoaderMixin
 from ...models.autoencoders import AutoencoderKL
 from ...models.transformers import SD3Transformer2DModel
 from ...schedulers import FlowMatchEulerDiscreteScheduler
@@ -80,7 +82,7 @@ def calculate_shift(
     base_seq_len: int = 256,
     max_seq_len: int = 4096,
     base_shift: float = 0.5,
-    max_shift: float = 1.16,
+    max_shift: float = 1.15,
 ):
     m = (max_shift - base_shift) / (max_seq_len - base_seq_len)
     b = base_shift - m * base_seq_len
@@ -162,7 +164,7 @@ def retrieve_timesteps(
     return timesteps, num_inference_steps
 
 
-class StableDiffusion3InpaintPipeline(DiffusionPipeline, SD3LoraLoaderMixin, FromSingleFileMixin):
+class StableDiffusion3InpaintPipeline(DiffusionPipeline, SD3LoraLoaderMixin, FromSingleFileMixin, SD3IPAdapterMixin):
     r"""
     Args:
         transformer ([`SD3Transformer2DModel`]):
@@ -194,10 +196,14 @@ class StableDiffusion3InpaintPipeline(DiffusionPipeline, SD3LoraLoaderMixin, Fro
         tokenizer_3 (`T5TokenizerFast`):
             Tokenizer of class
             [T5Tokenizer](https://huggingface.co/docs/transformers/model_doc/t5#transformers.T5Tokenizer).
+        image_encoder (`SiglipVisionModel`, *optional*):
+            Pre-trained Vision Model for IP Adapter.
+        feature_extractor (`SiglipImageProcessor`, *optional*):
+            Image processor for IP Adapter.
     """
 
-    model_cpu_offload_seq = "text_encoder->text_encoder_2->text_encoder_3->transformer->vae"
-    _optional_components = []
+    model_cpu_offload_seq = "text_encoder->text_encoder_2->text_encoder_3->image_encoder->transformer->vae"
+    _optional_components = ["image_encoder", "feature_extractor"]
     _callback_tensor_inputs = ["latents", "prompt_embeds", "negative_prompt_embeds", "negative_pooled_prompt_embeds"]
 
     def __init__(
@@ -211,6 +217,8 @@ class StableDiffusion3InpaintPipeline(DiffusionPipeline, SD3LoraLoaderMixin, Fro
         tokenizer_2: CLIPTokenizer,
         text_encoder_3: T5EncoderModel,
         tokenizer_3: T5TokenizerFast,
+        image_encoder: Optional[SiglipVisionModel] = None,
+        feature_extractor: Optional[SiglipImageProcessor] = None,
     ):
         super().__init__()
 
@@ -224,20 +232,29 @@ class StableDiffusion3InpaintPipeline(DiffusionPipeline, SD3LoraLoaderMixin, Fro
             tokenizer_3=tokenizer_3,
             transformer=transformer,
             scheduler=scheduler,
+            image_encoder=image_encoder,
+            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
+        latent_channels = self.vae.config.latent_channels if getattr(self, "vae", None) else 16
         self.image_processor = VaeImageProcessor(
-            vae_scale_factor=self.vae_scale_factor, vae_latent_channels=self.vae.config.latent_channels
+            vae_scale_factor=self.vae_scale_factor, vae_latent_channels=latent_channels
         )
         self.mask_processor = VaeImageProcessor(
             vae_scale_factor=self.vae_scale_factor,
-            vae_latent_channels=self.vae.config.latent_channels,
+            vae_latent_channels=latent_channels,
             do_normalize=False,
             do_binarize=True,
             do_convert_grayscale=True,
         )
-        self.tokenizer_max_length = self.tokenizer.model_max_length
-        self.default_sample_size = self.transformer.config.sample_size
+        self.tokenizer_max_length = (
+            self.tokenizer.model_max_length if hasattr(self, "tokenizer") and self.tokenizer is not None else 77
+        )
+        self.default_sample_size = (
+            self.transformer.config.sample_size
+            if hasattr(self, "transformer") and self.transformer is not None
+            else 128
+        )
         self.patch_size = (
             self.transformer.config.patch_size if hasattr(self, "transformer") and self.transformer is not None else 2
         )
@@ -399,9 +416,9 @@ class StableDiffusion3InpaintPipeline(DiffusionPipeline, SD3LoraLoaderMixin, Fro
             negative_prompt_2 (`str` or `List[str]`, *optional*):
                 The prompt or prompts not to guide the image generation to be sent to `tokenizer_2` and
                 `text_encoder_2`. If not defined, `negative_prompt` is used in all the text-encoders.
-            negative_prompt_2 (`str` or `List[str]`, *optional*):
+            negative_prompt_3 (`str` or `List[str]`, *optional*):
                 The prompt or prompts not to guide the image generation to be sent to `tokenizer_3` and
-                `text_encoder_3`. If not defined, `negative_prompt` is used in both text-encoders
+                `text_encoder_3`. If not defined, `negative_prompt` is used in all the text-encoders.
             prompt_embeds (`torch.FloatTensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
@@ -811,6 +828,10 @@ class StableDiffusion3InpaintPipeline(DiffusionPipeline, SD3LoraLoaderMixin, Fro
     def do_classifier_free_guidance(self):
         return self._guidance_scale > 1
 
+    @property
+    def joint_attention_kwargs(self):
+        return self._joint_attention_kwargs
+
     @property
     def num_timesteps(self):
         return self._num_timesteps
@@ -819,6 +840,84 @@ class StableDiffusion3InpaintPipeline(DiffusionPipeline, SD3LoraLoaderMixin, Fro
     def interrupt(self):
         return self._interrupt
 
+    # Copied from diffusers.pipelines.stable_diffusion_3.pipeline_stable_diffusion_3.StableDiffusion3Pipeline.encode_image
+    def encode_image(self, image: PipelineImageInput, device: torch.device) -> torch.Tensor:
+        """Encodes the given image into a feature representation using a pre-trained image encoder.
+
+        Args:
+            image (`PipelineImageInput`):
+                Input image to be encoded.
+            device: (`torch.device`):
+                Torch device.
+
+        Returns:
+            `torch.Tensor`: The encoded image feature representation.
+        """
+        if not isinstance(image, torch.Tensor):
+            image = self.feature_extractor(image, return_tensors="pt").pixel_values
+
+        image = image.to(device=device, dtype=self.dtype)
+
+        return self.image_encoder(image, output_hidden_states=True).hidden_states[-2]
+
+    # Copied from diffusers.pipelines.stable_diffusion_3.pipeline_stable_diffusion_3.StableDiffusion3Pipeline.prepare_ip_adapter_image_embeds
+    def prepare_ip_adapter_image_embeds(
+        self,
+        ip_adapter_image: Optional[PipelineImageInput] = None,
+        ip_adapter_image_embeds: Optional[torch.Tensor] = None,
+        device: Optional[torch.device] = None,
+        num_images_per_prompt: int = 1,
+        do_classifier_free_guidance: bool = True,
+    ) -> torch.Tensor:
+        """Prepares image embeddings for use in the IP-Adapter.
+
+        Either `ip_adapter_image` or `ip_adapter_image_embeds` must be passed.
+
+        Args:
+            ip_adapter_image (`PipelineImageInput`, *optional*):
+                The input image to extract features from for IP-Adapter.
+            ip_adapter_image_embeds (`torch.Tensor`, *optional*):
+                Precomputed image embeddings.
+            device: (`torch.device`, *optional*):
+                Torch device.
+            num_images_per_prompt (`int`, defaults to 1):
+                Number of images that should be generated per prompt.
+            do_classifier_free_guidance (`bool`, defaults to True):
+                Whether to use classifier free guidance or not.
+        """
+        device = device or self._execution_device
+
+        if ip_adapter_image_embeds is not None:
+            if do_classifier_free_guidance:
+                single_negative_image_embeds, single_image_embeds = ip_adapter_image_embeds.chunk(2)
+            else:
+                single_image_embeds = ip_adapter_image_embeds
+        elif ip_adapter_image is not None:
+            single_image_embeds = self.encode_image(ip_adapter_image, device)
+            if do_classifier_free_guidance:
+                single_negative_image_embeds = torch.zeros_like(single_image_embeds)
+        else:
+            raise ValueError("Neither `ip_adapter_image_embeds` or `ip_adapter_image_embeds` were provided.")
+
+        image_embeds = torch.cat([single_image_embeds] * num_images_per_prompt, dim=0)
+
+        if do_classifier_free_guidance:
+            negative_image_embeds = torch.cat([single_negative_image_embeds] * num_images_per_prompt, dim=0)
+            image_embeds = torch.cat([negative_image_embeds, image_embeds], dim=0)
+
+        return image_embeds.to(device=device)
+
+    # Copied from diffusers.pipelines.stable_diffusion_3.pipeline_stable_diffusion_3.StableDiffusion3Pipeline.enable_sequential_cpu_offload
+    def enable_sequential_cpu_offload(self, *args, **kwargs):
+        if self.image_encoder is not None and "image_encoder" not in self._exclude_from_cpu_offload:
+            logger.warning(
+                "`pipe.enable_sequential_cpu_offload()` might fail for `image_encoder` if it uses "
+                "`torch.nn.MultiheadAttention`. You can exclude `image_encoder` from CPU offloading by calling "
+                "`pipe._exclude_from_cpu_offload.append('image_encoder')` before `pipe.enable_sequential_cpu_offload()`."
+            )
+
+        super().enable_sequential_cpu_offload(*args, **kwargs)
+
     @torch.no_grad()
     @replace_example_docstring(EXAMPLE_DOC_STRING)
     def __call__(
@@ -846,8 +945,11 @@ class StableDiffusion3InpaintPipeline(DiffusionPipeline, SD3LoraLoaderMixin, Fro
         negative_prompt_embeds: Optional[torch.Tensor] = None,
         pooled_prompt_embeds: Optional[torch.Tensor] = None,
         negative_pooled_prompt_embeds: Optional[torch.Tensor] = None,
+        ip_adapter_image: Optional[PipelineImageInput] = None,
+        ip_adapter_image_embeds: Optional[torch.Tensor] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
+        joint_attention_kwargs: Optional[Dict[str, Any]] = None,
         clip_skip: Optional[int] = None,
         callback_on_step_end: Optional[Callable[[int, int, Dict], None]] = None,
         callback_on_step_end_tensor_inputs: List[str] = ["latents"],
@@ -883,9 +985,9 @@ class StableDiffusion3InpaintPipeline(DiffusionPipeline, SD3LoraLoaderMixin, Fro
             mask_image_latent (`torch.Tensor`, `List[torch.Tensor]`):
                 `Tensor` representing an image batch to mask `image` generated by VAE. If not provided, the mask
                 latents tensor will ge generated by `mask_image`.
-            height (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):
+            height (`int`, *optional*, defaults to self.transformer.config.sample_size * self.vae_scale_factor):
                 The height in pixels of the generated image. This is set to 1024 by default for the best results.
-            width (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):
+            width (`int`, *optional*, defaults to self.transformer.config.sample_size * self.vae_scale_factor):
                 The width in pixels of the generated image. This is set to 1024 by default for the best results.
             padding_mask_crop (`int`, *optional*, defaults to `None`):
                 The size of margin in the crop to be applied to the image and masking. If `None`, no crop is applied to
@@ -946,12 +1048,22 @@ class StableDiffusion3InpaintPipeline(DiffusionPipeline, SD3LoraLoaderMixin, Fro
                 Pre-generated negative pooled text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                 weighting. If not provided, pooled negative_prompt_embeds will be generated from `negative_prompt`
                 input argument.
+            ip_adapter_image (`PipelineImageInput`, *optional*):
+                Optional image input to work with IP Adapters.
+            ip_adapter_image_embeds (`torch.Tensor`, *optional*):
+                Pre-generated image embeddings for IP-Adapter. Should be a tensor of shape `(batch_size, num_images,
+                emb_dim)`. It should contain the negative image embedding if `do_classifier_free_guidance` is set to
+                `True`. If not provided, embeddings are computed from the `ip_adapter_image` input argument.
             output_type (`str`, *optional*, defaults to `"pil"`):
                 The output format of the generate image. Choose between
                 [PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
             return_dict (`bool`, *optional*, defaults to `True`):
                 Whether or not to return a [`~pipelines.stable_diffusion_3.StableDiffusion3PipelineOutput`] instead of
                 a plain tuple.
+            joint_attention_kwargs (`dict`, *optional*):
+                A kwargs dictionary that if specified is passed along to the `AttentionProcessor` as defined under
+                `self.processor` in
+                [diffusers.models.attention_processor](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).
             callback_on_step_end (`Callable`, *optional*):
                 A function that calls at the end of each denoising steps during the inference. The function is called
                 with the following arguments: `callback_on_step_end(self: DiffusionPipeline, step: int, timestep: int,
@@ -999,6 +1111,7 @@ class StableDiffusion3InpaintPipeline(DiffusionPipeline, SD3LoraLoaderMixin, Fro
 
         self._guidance_scale = guidance_scale
         self._clip_skip = clip_skip
+        self._joint_attention_kwargs = joint_attention_kwargs
         self._interrupt = False
 
         # 2. Define call parameters
@@ -1046,10 +1159,10 @@ class StableDiffusion3InpaintPipeline(DiffusionPipeline, SD3LoraLoaderMixin, Fro
             )
             mu = calculate_shift(
                 image_seq_len,
-                self.scheduler.config.base_image_seq_len,
-                self.scheduler.config.max_image_seq_len,
-                self.scheduler.config.base_shift,
-                self.scheduler.config.max_shift,
+                self.scheduler.config.get("base_image_seq_len", 256),
+                self.scheduler.config.get("max_image_seq_len", 4096),
+                self.scheduler.config.get("base_shift", 0.5),
+                self.scheduler.config.get("max_shift", 1.16),
             )
             scheduler_kwargs["mu"] = mu
         elif mu is not None:
@@ -1145,7 +1258,7 @@ class StableDiffusion3InpaintPipeline(DiffusionPipeline, SD3LoraLoaderMixin, Fro
                     f"Incorrect configuration settings! The config of `pipeline.transformer`: {self.transformer.config} expects"
                     f" {self.transformer.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.transformer` or your `mask_image` or `image` input."
                 )
         elif num_channels_transformer != 16:
@@ -1153,7 +1266,22 @@ class StableDiffusion3InpaintPipeline(DiffusionPipeline, SD3LoraLoaderMixin, Fro
                 f"The transformer {self.transformer.__class__} should have 16 input channels or 33 input channels, not {self.transformer.config.in_channels}."
             )
 
-        # 7. Denoising loop
+        # 7. Prepare image embeddings
+        if (ip_adapter_image is not None and self.is_ip_adapter_active) or ip_adapter_image_embeds is not None:
+            ip_adapter_image_embeds = self.prepare_ip_adapter_image_embeds(
+                ip_adapter_image,
+                ip_adapter_image_embeds,
+                device,
+                batch_size * num_images_per_prompt,
+                self.do_classifier_free_guidance,
+            )
+
+            if self.joint_attention_kwargs is None:
+                self._joint_attention_kwargs = {"ip_adapter_image_embeds": ip_adapter_image_embeds}
+            else:
+                self._joint_attention_kwargs.update(ip_adapter_image_embeds=ip_adapter_image_embeds)
+
+        # 8. Denoising loop
         num_warmup_steps = max(len(timesteps) - num_inference_steps * self.scheduler.order, 0)
         self._num_timesteps = len(timesteps)
         with self.progress_bar(total=num_inference_steps) as progress_bar:
@@ -1174,6 +1302,7 @@ class StableDiffusion3InpaintPipeline(DiffusionPipeline, SD3LoraLoaderMixin, Fro
                     timestep=timestep,
                     encoder_hidden_states=prompt_embeds,
                     pooled_projections=pooled_prompt_embeds,
+                    joint_attention_kwargs=self.joint_attention_kwargs,
                     return_dict=False,
                 )[0]
 

diffusers/pipelines/stable_diffusion_attend_and_excite/pipeline_stable_diffusion_attend_and_excite.py

@@ -30,6 +30,7 @@ from ...schedulers import KarrasDiffusionSchedulers
 from ...utils import (
     USE_PEFT_BACKEND,
     deprecate,
+    is_torch_xla_available,
     logging,
     replace_example_docstring,
     scale_lora_layers,
@@ -41,6 +42,14 @@ 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__)
 
 EXAMPLE_DOC_STRING = """
@@ -194,8 +203,8 @@ class StableDiffusionAttendAndExcitePipeline(DiffusionPipeline, StableDiffusionM
             [`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`.
     """
@@ -242,7 +251,7 @@ class StableDiffusionAttendAndExcitePipeline(DiffusionPipeline, StableDiffusionM
             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)
 
@@ -1008,6 +1017,9 @@ class StableDiffusionAttendAndExcitePipeline(DiffusionPipeline, StableDiffusionM
                         step_idx = i // getattr(self.scheduler, "order", 1)
                         callback(step_idx, t, latents)
 
+                if XLA_AVAILABLE:
+                    xm.mark_step()
+
         # 8. Post-processing
         if not output_type == "latent":
             image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False)[0]

diffusers/pipelines/stable_diffusion_diffedit/pipeline_stable_diffusion_diffedit.py

@@ -33,6 +33,7 @@ from ...utils import (
     USE_PEFT_BACKEND,
     BaseOutput,
     deprecate,
+    is_torch_xla_available,
     logging,
     replace_example_docstring,
     scale_lora_layers,
@@ -44,6 +45,13 @@ 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
 
 
@@ -268,8 +276,8 @@ class StableDiffusionDiffEditPipeline(
             A scheduler to be used in combination with `unet` to fill in the unmasked part of the input latents.
         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`.
     """
@@ -292,7 +300,7 @@ class StableDiffusionDiffEditPipeline(
     ):
         super().__init__()
 
-        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 "
@@ -306,7 +314,7 @@ class StableDiffusionDiffEditPipeline(
             new_config["steps_offset"] = 1
             scheduler._internal_dict = FrozenDict(new_config)
 
-        if hasattr(scheduler.config, "skip_prk_steps") and scheduler.config.skip_prk_steps is False:
+        if scheduler is not None and getattr(scheduler.config, "skip_prk_steps", True) is False:
             deprecation_message = (
                 f"The configuration file of this scheduler: {scheduler} has not set the configuration"
                 " `skip_prk_steps`. `skip_prk_steps` should be set to True in the configuration file. Please make"
@@ -336,17 +344,21 @@ class StableDiffusionDiffEditPipeline(
                 " 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"
@@ -367,7 +379,7 @@ class StableDiffusionDiffEditPipeline(
             feature_extractor=feature_extractor,
             inverse_scheduler=inverse_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.register_to_config(requires_safety_checker=requires_safety_checker)
 
@@ -1508,6 +1520,9 @@ class StableDiffusionDiffEditPipeline(
                         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_gligen/pipeline_stable_diffusion_gligen.py

@@ -29,6 +29,7 @@ from ...schedulers import KarrasDiffusionSchedulers
 from ...utils import (
     USE_PEFT_BACKEND,
     deprecate,
+    is_torch_xla_available,
     logging,
     replace_example_docstring,
     scale_lora_layers,
@@ -40,8 +41,16 @@ 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
@@ -120,8 +129,8 @@ class StableDiffusionGLIGENPipeline(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 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`.
     """
@@ -168,7 +177,7 @@ class StableDiffusionGLIGENPipeline(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, do_convert_rgb=True)
         self.register_to_config(requires_safety_checker=requires_safety_checker)
 
@@ -828,6 +837,9 @@ class StableDiffusionGLIGENPipeline(DiffusionPipeline, StableDiffusionMixin):
                         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_gligen/pipeline_stable_diffusion_gligen_text_image.py

@@ -32,7 +32,14 @@ from ...models import AutoencoderKL, UNet2DConditionModel
 from ...models.attention import GatedSelfAttentionDense
 from ...models.lora import adjust_lora_scale_text_encoder
 from ...schedulers import KarrasDiffusionSchedulers
-from ...utils import USE_PEFT_BACKEND, logging, replace_example_docstring, scale_lora_layers, unscale_lora_layers
+from ...utils import (
+    USE_PEFT_BACKEND,
+    is_torch_xla_available,
+    logging,
+    replace_example_docstring,
+    scale_lora_layers,
+    unscale_lora_layers,
+)
 from ...utils.torch_utils import randn_tensor
 from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
 from ..stable_diffusion import StableDiffusionPipelineOutput
@@ -40,8 +47,16 @@ from ..stable_diffusion.clip_image_project_model import CLIPImageProjection
 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
@@ -172,8 +187,8 @@ class StableDiffusionGLIGENTextImagePipeline(DiffusionPipeline, StableDiffusionM
             [`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`.
     """
@@ -226,7 +241,7 @@ class StableDiffusionGLIGENTextImagePipeline(DiffusionPipeline, StableDiffusionM
             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, do_convert_rgb=True)
         self.register_to_config(requires_safety_checker=requires_safety_checker)
 
@@ -1010,6 +1025,9 @@ class StableDiffusionGLIGENTextImagePipeline(DiffusionPipeline, StableDiffusionM
                         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_k_diffusion/pipeline_stable_diffusion_k_diffusion.py

@@ -19,15 +19,31 @@ from typing import Callable, List, Optional, Union
 import torch
 from k_diffusion.external import CompVisDenoiser, CompVisVDenoiser
 from k_diffusion.sampling import BrownianTreeNoiseSampler, get_sigmas_karras
+from transformers import (
+    CLIPImageProcessor,
+    CLIPTextModel,
+    CLIPTokenizer,
+    CLIPTokenizerFast,
+)
 
 from ...image_processor import VaeImageProcessor
-from ...loaders import StableDiffusionLoraLoaderMixin, TextualInversionLoaderMixin
+from ...loaders import (
+    StableDiffusionLoraLoaderMixin,
+    TextualInversionLoaderMixin,
+)
+from ...models import AutoencoderKL, UNet2DConditionModel
 from ...models.lora import adjust_lora_scale_text_encoder
-from ...schedulers import LMSDiscreteScheduler
-from ...utils import USE_PEFT_BACKEND, deprecate, logging, scale_lora_layers, unscale_lora_layers
+from ...schedulers import KarrasDiffusionSchedulers, LMSDiscreteScheduler
+from ...utils import (
+    USE_PEFT_BACKEND,
+    deprecate,
+    logging,
+    scale_lora_layers,
+    unscale_lora_layers,
+)
 from ...utils.torch_utils import randn_tensor
 from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
-from ..stable_diffusion import StableDiffusionPipelineOutput
+from ..stable_diffusion import StableDiffusionPipelineOutput, StableDiffusionSafetyChecker
 
 
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
@@ -83,7 +99,8 @@ class StableDiffusionKDiffusionPipeline(
             [`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`.
     """
@@ -94,13 +111,13 @@ class StableDiffusionKDiffusionPipeline(
 
     def __init__(
         self,
-        vae,
-        text_encoder,
-        tokenizer,
-        unet,
-        scheduler,
-        safety_checker,
-        feature_extractor,
+        vae: AutoencoderKL,
+        text_encoder: CLIPTextModel,
+        tokenizer: Union[CLIPTokenizer, CLIPTokenizerFast],
+        unet: UNet2DConditionModel,
+        scheduler: KarrasDiffusionSchedulers,
+        safety_checker: StableDiffusionSafetyChecker,
+        feature_extractor: CLIPImageProcessor,
         requires_safety_checker: bool = True,
     ):
         super().__init__()
@@ -124,7 +141,7 @@ class StableDiffusionKDiffusionPipeline(
             feature_extractor=feature_extractor,
         )
         self.register_to_config(requires_safety_checker=requires_safety_checker)
-        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)
 
         model = ModelWrapper(unet, scheduler.alphas_cumprod)