diffusers 0.23.0__py3-none-any.whl → 0.24.0__py3-none-any.whl

diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_img2img.py

@@ -19,11 +19,11 @@ import numpy as np
 import PIL.Image
 import torch
 from packaging import version
-from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer
+from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer, CLIPVisionModelWithProjection
 
 from ...configuration_utils import FrozenDict
 from ...image_processor import PipelineImageInput, VaeImageProcessor
-from ...loaders import FromSingleFileMixin, LoraLoaderMixin, TextualInversionLoaderMixin
+from ...loaders import FromSingleFileMixin, IPAdapterMixin, LoraLoaderMixin, TextualInversionLoaderMixin
 from ...models import AutoencoderKL, UNet2DConditionModel
 from ...models.lora import adjust_lora_scale_text_encoder
 from ...schedulers import KarrasDiffusionSchedulers
@@ -73,9 +73,13 @@ EXAMPLE_DOC_STRING = """
 """
 
 
-def retrieve_latents(encoder_output, generator):
-    if hasattr(encoder_output, "latent_dist"):
+def retrieve_latents(
+    encoder_output: torch.Tensor, generator: Optional[torch.Generator] = None, sample_mode: str = "sample"
+):
+    if hasattr(encoder_output, "latent_dist") and sample_mode == "sample":
         return encoder_output.latent_dist.sample(generator)
+    elif hasattr(encoder_output, "latent_dist") and sample_mode == "argmax":
+        return encoder_output.latent_dist.mode()
     elif hasattr(encoder_output, "latents"):
         return encoder_output.latents
     else:
@@ -105,8 +109,53 @@ def preprocess(image):
     return image
 
 
+# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.retrieve_timesteps
+def retrieve_timesteps(
+    scheduler,
+    num_inference_steps: Optional[int] = None,
+    device: Optional[Union[str, torch.device]] = None,
+    timesteps: Optional[List[int]] = None,
+    **kwargs,
+):
+    """
+    Calls the scheduler's `set_timesteps` method and retrieves timesteps from the scheduler after the call. Handles
+    custom timesteps. Any kwargs will be supplied to `scheduler.set_timesteps`.
+
+    Args:
+        scheduler (`SchedulerMixin`):
+            The scheduler to get timesteps from.
+        num_inference_steps (`int`):
+            The number of diffusion steps used when generating samples with a pre-trained model. If used,
+            `timesteps` must be `None`.
+        device (`str` or `torch.device`, *optional*):
+            The device to which the timesteps should be moved to. If `None`, the timesteps are not moved.
+        timesteps (`List[int]`, *optional*):
+                Custom timesteps used to support arbitrary spacing between timesteps. If `None`, then the default
+                timestep spacing strategy of the scheduler is used. If `timesteps` is passed, `num_inference_steps`
+                must be `None`.
+
+    Returns:
+        `Tuple[torch.Tensor, int]`: A tuple where the first element is the timestep schedule from the scheduler and the
+        second element is the number of inference steps.
+    """
+    if timesteps is not None:
+        accepts_timesteps = "timesteps" in set(inspect.signature(scheduler.set_timesteps).parameters.keys())
+        if not accepts_timesteps:
+            raise ValueError(
+                f"The current scheduler class {scheduler.__class__}'s `set_timesteps` does not support custom"
+                f" timestep schedules. Please check whether you are using the correct scheduler."
+            )
+        scheduler.set_timesteps(timesteps=timesteps, device=device, **kwargs)
+        timesteps = scheduler.timesteps
+        num_inference_steps = len(timesteps)
+    else:
+        scheduler.set_timesteps(num_inference_steps, device=device, **kwargs)
+        timesteps = scheduler.timesteps
+    return timesteps, num_inference_steps
+
+
 class StableDiffusionImg2ImgPipeline(
-    DiffusionPipeline, TextualInversionLoaderMixin, LoraLoaderMixin, FromSingleFileMixin
+    DiffusionPipeline, TextualInversionLoaderMixin, IPAdapterMixin, LoraLoaderMixin, FromSingleFileMixin
 ):
     r"""
     Pipeline for text-guided image-to-image generation using Stable Diffusion.
@@ -119,6 +168,7 @@ class StableDiffusionImg2ImgPipeline(
         - [`~loaders.LoraLoaderMixin.load_lora_weights`] for loading LoRA weights
         - [`~loaders.LoraLoaderMixin.save_lora_weights`] for saving LoRA weights
         - [`~loaders.FromSingleFileMixin.from_single_file`] for loading `.ckpt` files
+        - [`~loaders.IPAdapterMixin.load_ip_adapter`] for loading IP Adapters
 
     Args:
         vae ([`AutoencoderKL`]):
@@ -139,8 +189,9 @@ class StableDiffusionImg2ImgPipeline(
         feature_extractor ([`~transformers.CLIPImageProcessor`]):
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
+
     model_cpu_offload_seq = "text_encoder->unet->vae"
-    _optional_components = ["safety_checker", "feature_extractor"]
+    _optional_components = ["safety_checker", "feature_extractor", "image_encoder"]
     _exclude_from_cpu_offload = ["safety_checker"]
     _callback_tensor_inputs = ["latents", "prompt_embeds", "negative_prompt_embeds"]
 
@@ -153,6 +204,7 @@ class StableDiffusionImg2ImgPipeline(
         scheduler: KarrasDiffusionSchedulers,
         safety_checker: StableDiffusionSafetyChecker,
         feature_extractor: CLIPImageProcessor,
+        image_encoder: CLIPVisionModelWithProjection = None,
         requires_safety_checker: bool = True,
     ):
         super().__init__()
@@ -229,6 +281,7 @@ class StableDiffusionImg2ImgPipeline(
             scheduler=scheduler,
             safety_checker=safety_checker,
             feature_extractor=feature_extractor,
+            image_encoder=image_encoder,
         )
         self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
         self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
@@ -449,6 +502,20 @@ class StableDiffusionImg2ImgPipeline(
 
         return prompt_embeds, negative_prompt_embeds
 
+    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.encode_image
+    def encode_image(self, image, device, num_images_per_prompt):
+        dtype = next(self.image_encoder.parameters()).dtype
+
+        if not isinstance(image, torch.Tensor):
+            image = self.feature_extractor(image, return_tensors="pt").pixel_values
+
+        image = image.to(device=device, dtype=dtype)
+        image_embeds = self.image_encoder(image).image_embeds
+        image_embeds = image_embeds.repeat_interleave(num_images_per_prompt, dim=0)
+
+        uncond_image_embeds = torch.zeros_like(image_embeds)
+        return image_embeds, uncond_image_embeds
+
     # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.run_safety_checker
     def run_safety_checker(self, image, device, dtype):
         if self.safety_checker is None:
@@ -700,6 +767,7 @@ class StableDiffusionImg2ImgPipeline(
         image: PipelineImageInput = None,
         strength: float = 0.8,
         num_inference_steps: Optional[int] = 50,
+        timesteps: List[int] = None,
         guidance_scale: Optional[float] = 7.5,
         negative_prompt: Optional[Union[str, List[str]]] = None,
         num_images_per_prompt: Optional[int] = 1,
@@ -707,6 +775,7 @@ class StableDiffusionImg2ImgPipeline(
         generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
         prompt_embeds: Optional[torch.FloatTensor] = None,
         negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        ip_adapter_image: Optional[PipelineImageInput] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
@@ -736,6 +805,10 @@ class StableDiffusionImg2ImgPipeline(
             num_inference_steps (`int`, *optional*, defaults to 50):
                 The number of denoising steps. More denoising steps usually lead to a higher quality image at the
                 expense of slower inference. This parameter is modulated by `strength`.
+            timesteps (`List[int]`, *optional*):
+                Custom timesteps to use for the denoising process with schedulers which support a `timesteps` argument
+                in their `set_timesteps` method. If not defined, the default behavior when `num_inference_steps` is
+                passed will be used. Must be in descending order.
             guidance_scale (`float`, *optional*, defaults to 7.5):
                 A higher guidance scale value encourages the model to generate images closely linked to the text
                 `prompt` at the expense of lower image quality. Guidance scale is enabled when `guidance_scale > 1`.
@@ -756,6 +829,7 @@ class StableDiffusionImg2ImgPipeline(
             negative_prompt_embeds (`torch.FloatTensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
                 not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
+            ip_adapter_image: (`PipelineImageInput`, *optional*): Optional image input to work with IP Adapters.
             output_type (`str`, *optional*, defaults to `"pil"`):
                 The output format of the generated image. Choose between `PIL.Image` or `np.array`.
             return_dict (`bool`, *optional*, defaults to `True`):
@@ -775,7 +849,7 @@ class StableDiffusionImg2ImgPipeline(
             callback_on_step_end_tensor_inputs (`List`, *optional*):
                 The list of tensor inputs for the `callback_on_step_end` function. The tensors specified in the list
                 will be passed as `callback_kwargs` argument. You will only be able to include variables listed in the
-                `._callback_tensor_inputs` attribute of your pipeine class.
+                `._callback_tensor_inputs` attribute of your pipeline class.
         Examples:
 
         Returns:
@@ -848,11 +922,16 @@ class StableDiffusionImg2ImgPipeline(
         if self.do_classifier_free_guidance:
             prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds])
 
+        if ip_adapter_image is not None:
+            image_embeds, negative_image_embeds = self.encode_image(ip_adapter_image, device, num_images_per_prompt)
+            if self.do_classifier_free_guidance:
+                image_embeds = torch.cat([negative_image_embeds, image_embeds])
+
         # 4. Preprocess image
         image = self.image_processor.preprocess(image)
 
         # 5. set timesteps
-        self.scheduler.set_timesteps(num_inference_steps, device=device)
+        timesteps, num_inference_steps = retrieve_timesteps(self.scheduler, num_inference_steps, device, timesteps)
         timesteps, num_inference_steps = self.get_timesteps(num_inference_steps, strength, device)
         latent_timestep = timesteps[:1].repeat(batch_size * num_images_per_prompt)
 
@@ -870,7 +949,10 @@ class StableDiffusionImg2ImgPipeline(
         # 7. Prepare extra step kwargs. TODO: Logic should ideally just be moved out of the pipeline
         extra_step_kwargs = self.prepare_extra_step_kwargs(generator, eta)
 
-        # 7.5 Optionally get Guidance Scale Embedding
+        # 7.1 Add image embeds for IP-Adapter
+        added_cond_kwargs = {"image_embeds": image_embeds} if ip_adapter_image is not None else None
+
+        # 7.2 Optionally get Guidance Scale Embedding
         timestep_cond = None
         if self.unet.config.time_cond_proj_dim is not None:
             guidance_scale_tensor = torch.tensor(self.guidance_scale - 1).repeat(batch_size * num_images_per_prompt)
@@ -894,6 +976,7 @@ class StableDiffusionImg2ImgPipeline(
                     encoder_hidden_states=prompt_embeds,
                     timestep_cond=timestep_cond,
                     cross_attention_kwargs=self.cross_attention_kwargs,
+                    added_cond_kwargs=added_cond_kwargs,
                     return_dict=False,
                 )[0]
 

diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_inpaint.py

@@ -19,11 +19,11 @@ import numpy as np
 import PIL.Image
 import torch
 from packaging import version
-from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer
+from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer, CLIPVisionModelWithProjection
 
 from ...configuration_utils import FrozenDict
 from ...image_processor import PipelineImageInput, VaeImageProcessor
-from ...loaders import FromSingleFileMixin, LoraLoaderMixin, TextualInversionLoaderMixin
+from ...loaders import FromSingleFileMixin, IPAdapterMixin, LoraLoaderMixin, TextualInversionLoaderMixin
 from ...models import AsymmetricAutoencoderKL, AutoencoderKL, UNet2DConditionModel
 from ...models.lora import adjust_lora_scale_text_encoder
 from ...schedulers import KarrasDiffusionSchedulers
@@ -160,17 +160,66 @@ def prepare_mask_and_masked_image(image, mask, height, width, return_image: bool
 
 
 # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion_img2img.retrieve_latents
-def retrieve_latents(encoder_output, generator):
-    if hasattr(encoder_output, "latent_dist"):
+def retrieve_latents(
+    encoder_output: torch.Tensor, generator: Optional[torch.Generator] = None, sample_mode: str = "sample"
+):
+    if hasattr(encoder_output, "latent_dist") and sample_mode == "sample":
         return encoder_output.latent_dist.sample(generator)
+    elif hasattr(encoder_output, "latent_dist") and sample_mode == "argmax":
+        return encoder_output.latent_dist.mode()
     elif hasattr(encoder_output, "latents"):
         return encoder_output.latents
     else:
         raise AttributeError("Could not access latents of provided encoder_output")
 
 
+# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.retrieve_timesteps
+def retrieve_timesteps(
+    scheduler,
+    num_inference_steps: Optional[int] = None,
+    device: Optional[Union[str, torch.device]] = None,
+    timesteps: Optional[List[int]] = None,
+    **kwargs,
+):
+    """
+    Calls the scheduler's `set_timesteps` method and retrieves timesteps from the scheduler after the call. Handles
+    custom timesteps. Any kwargs will be supplied to `scheduler.set_timesteps`.
+
+    Args:
+        scheduler (`SchedulerMixin`):
+            The scheduler to get timesteps from.
+        num_inference_steps (`int`):
+            The number of diffusion steps used when generating samples with a pre-trained model. If used,
+            `timesteps` must be `None`.
+        device (`str` or `torch.device`, *optional*):
+            The device to which the timesteps should be moved to. If `None`, the timesteps are not moved.
+        timesteps (`List[int]`, *optional*):
+                Custom timesteps used to support arbitrary spacing between timesteps. If `None`, then the default
+                timestep spacing strategy of the scheduler is used. If `timesteps` is passed, `num_inference_steps`
+                must be `None`.
+
+    Returns:
+        `Tuple[torch.Tensor, int]`: A tuple where the first element is the timestep schedule from the scheduler and the
+        second element is the number of inference steps.
+    """
+    if timesteps is not None:
+        accepts_timesteps = "timesteps" in set(inspect.signature(scheduler.set_timesteps).parameters.keys())
+        if not accepts_timesteps:
+            raise ValueError(
+                f"The current scheduler class {scheduler.__class__}'s `set_timesteps` does not support custom"
+                f" timestep schedules. Please check whether you are using the correct scheduler."
+            )
+        scheduler.set_timesteps(timesteps=timesteps, device=device, **kwargs)
+        timesteps = scheduler.timesteps
+        num_inference_steps = len(timesteps)
+    else:
+        scheduler.set_timesteps(num_inference_steps, device=device, **kwargs)
+        timesteps = scheduler.timesteps
+    return timesteps, num_inference_steps
+
+
 class StableDiffusionInpaintPipeline(
-    DiffusionPipeline, TextualInversionLoaderMixin, LoraLoaderMixin, FromSingleFileMixin
+    DiffusionPipeline, TextualInversionLoaderMixin, IPAdapterMixin, LoraLoaderMixin, FromSingleFileMixin
 ):
     r"""
     Pipeline for text-guided image inpainting using Stable Diffusion.
@@ -182,6 +231,7 @@ class StableDiffusionInpaintPipeline(
         - [`~loaders.TextualInversionLoaderMixin.load_textual_inversion`] for loading textual inversion embeddings
         - [`~loaders.LoraLoaderMixin.load_lora_weights`] for loading LoRA weights
         - [`~loaders.LoraLoaderMixin.save_lora_weights`] for saving LoRA weights
+        - [`~loaders.IPAdapterMixin.load_ip_adapter`] for loading IP Adapters
 
     Args:
         vae ([`AutoencoderKL`, `AsymmetricAutoencoderKL`]):
@@ -202,8 +252,9 @@ class StableDiffusionInpaintPipeline(
         feature_extractor ([`~transformers.CLIPImageProcessor`]):
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
+
     model_cpu_offload_seq = "text_encoder->unet->vae"
-    _optional_components = ["safety_checker", "feature_extractor"]
+    _optional_components = ["safety_checker", "feature_extractor", "image_encoder"]
     _exclude_from_cpu_offload = ["safety_checker"]
     _callback_tensor_inputs = ["latents", "prompt_embeds", "negative_prompt_embeds", "mask", "masked_image_latents"]
 
@@ -216,6 +267,7 @@ class StableDiffusionInpaintPipeline(
         scheduler: KarrasDiffusionSchedulers,
         safety_checker: StableDiffusionSafetyChecker,
         feature_extractor: CLIPImageProcessor,
+        image_encoder: CLIPVisionModelWithProjection = None,
         requires_safety_checker: bool = True,
     ):
         super().__init__()
@@ -297,6 +349,7 @@ class StableDiffusionInpaintPipeline(
             scheduler=scheduler,
             safety_checker=safety_checker,
             feature_extractor=feature_extractor,
+            image_encoder=image_encoder,
         )
         self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
         self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
@@ -520,6 +573,20 @@ class StableDiffusionInpaintPipeline(
 
         return prompt_embeds, negative_prompt_embeds
 
+    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.encode_image
+    def encode_image(self, image, device, num_images_per_prompt):
+        dtype = next(self.image_encoder.parameters()).dtype
+
+        if not isinstance(image, torch.Tensor):
+            image = self.feature_extractor(image, return_tensors="pt").pixel_values
+
+        image = image.to(device=device, dtype=dtype)
+        image_embeds = self.image_encoder(image).image_embeds
+        image_embeds = image_embeds.repeat_interleave(num_images_per_prompt, dim=0)
+
+        uncond_image_embeds = torch.zeros_like(image_embeds)
+        return image_embeds, uncond_image_embeds
+
     # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.run_safety_checker
     def run_safety_checker(self, image, device, dtype):
         if self.safety_checker is None:
@@ -828,6 +895,7 @@ class StableDiffusionInpaintPipeline(
         width: Optional[int] = None,
         strength: float = 1.0,
         num_inference_steps: int = 50,
+        timesteps: List[int] = None,
         guidance_scale: float = 7.5,
         negative_prompt: Optional[Union[str, List[str]]] = None,
         num_images_per_prompt: Optional[int] = 1,
@@ -836,6 +904,7 @@ class StableDiffusionInpaintPipeline(
         latents: Optional[torch.FloatTensor] = None,
         prompt_embeds: Optional[torch.FloatTensor] = None,
         negative_prompt_embeds: Optional[torch.FloatTensor] = None,
+        ip_adapter_image: Optional[PipelineImageInput] = None,
         output_type: Optional[str] = "pil",
         return_dict: bool = True,
         cross_attention_kwargs: Optional[Dict[str, Any]] = None,
@@ -877,6 +946,10 @@ class StableDiffusionInpaintPipeline(
             num_inference_steps (`int`, *optional*, defaults to 50):
                 The number of denoising steps. More denoising steps usually lead to a higher quality image at the
                 expense of slower inference. This parameter is modulated by `strength`.
+            timesteps (`List[int]`, *optional*):
+                Custom timesteps to use for the denoising process with schedulers which support a `timesteps` argument
+                in their `set_timesteps` method. If not defined, the default behavior when `num_inference_steps` is
+                passed will be used. Must be in descending order.
             guidance_scale (`float`, *optional*, defaults to 7.5):
                 A higher guidance scale value encourages the model to generate images closely linked to the text
                 `prompt` at the expense of lower image quality. Guidance scale is enabled when `guidance_scale > 1`.
@@ -901,6 +974,7 @@ class StableDiffusionInpaintPipeline(
             negative_prompt_embeds (`torch.FloatTensor`, *optional*):
                 Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
                 not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
+            ip_adapter_image: (`PipelineImageInput`, *optional*): Optional image input to work with IP Adapters.
             output_type (`str`, *optional*, defaults to `"pil"`):
                 The output format of the generated image. Choose between `PIL.Image` or `np.array`.
             return_dict (`bool`, *optional*, defaults to `True`):
@@ -920,7 +994,7 @@ class StableDiffusionInpaintPipeline(
             callback_on_step_end_tensor_inputs (`List`, *optional*):
                 The list of tensor inputs for the `callback_on_step_end` function. The tensors specified in the list
                 will be passed as `callback_kwargs` argument. You will only be able to include variables listed in the
-                `._callback_tensor_inputs` attribute of your pipeine class.
+                `._callback_tensor_inputs` attribute of your pipeline class.
         Examples:
 
         ```py
@@ -1028,8 +1102,13 @@ class StableDiffusionInpaintPipeline(
         if self.do_classifier_free_guidance:
             prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds])
 
+        if ip_adapter_image is not None:
+            image_embeds, negative_image_embeds = self.encode_image(ip_adapter_image, device, num_images_per_prompt)
+            if self.do_classifier_free_guidance:
+                image_embeds = torch.cat([negative_image_embeds, image_embeds])
+
         # 4. set timesteps
-        self.scheduler.set_timesteps(num_inference_steps, device=device)
+        timesteps, num_inference_steps = retrieve_timesteps(self.scheduler, num_inference_steps, device, timesteps)
         timesteps, num_inference_steps = self.get_timesteps(
             num_inference_steps=num_inference_steps, strength=strength, device=device
         )
@@ -1116,7 +1195,10 @@ class StableDiffusionInpaintPipeline(
         # 9. Prepare extra step kwargs. TODO: Logic should ideally just be moved out of the pipeline
         extra_step_kwargs = self.prepare_extra_step_kwargs(generator, eta)
 
-        # 9.5 Optionally get Guidance Scale Embedding
+        # 9.1 Add image embeds for IP-Adapter
+        added_cond_kwargs = {"image_embeds": image_embeds} if ip_adapter_image is not None else None
+
+        # 9.2 Optionally get Guidance Scale Embedding
         timestep_cond = None
         if self.unet.config.time_cond_proj_dim is not None:
             guidance_scale_tensor = torch.tensor(self.guidance_scale - 1).repeat(batch_size * num_images_per_prompt)
@@ -1145,6 +1227,7 @@ class StableDiffusionInpaintPipeline(
                     encoder_hidden_states=prompt_embeds,
                     timestep_cond=timestep_cond,
                     cross_attention_kwargs=self.cross_attention_kwargs,
+                    added_cond_kwargs=added_cond_kwargs,
                     return_dict=False,
                 )[0]
 

diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_inpaint_legacy.py

@@ -115,6 +115,7 @@ class StableDiffusionInpaintPipelineLegacy(
         feature_extractor ([`CLIPImageProcessor`]):
             Model that extracts features from generated images to be used as inputs for the `safety_checker`.
     """
+
     model_cpu_offload_seq = "text_encoder->unet->vae"
     _optional_components = ["feature_extractor"]
     _exclude_from_cpu_offload = ["safety_checker"]

diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_instruct_pix2pix.py

@@ -58,6 +58,20 @@ def preprocess(image):
     return image
 
 
+# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion_img2img.retrieve_latents
+def retrieve_latents(
+    encoder_output: torch.Tensor, generator: Optional[torch.Generator] = None, sample_mode: str = "sample"
+):
+    if hasattr(encoder_output, "latent_dist") and sample_mode == "sample":
+        return encoder_output.latent_dist.sample(generator)
+    elif hasattr(encoder_output, "latent_dist") and sample_mode == "argmax":
+        return encoder_output.latent_dist.mode()
+    elif hasattr(encoder_output, "latents"):
+        return encoder_output.latents
+    else:
+        raise AttributeError("Could not access latents of provided encoder_output")
+
+
 class StableDiffusionInstructPix2PixPipeline(DiffusionPipeline, TextualInversionLoaderMixin, LoraLoaderMixin):
     r"""
     Pipeline for pixel-level image editing by following text instructions (based on Stable Diffusion).
@@ -89,6 +103,7 @@ class StableDiffusionInstructPix2PixPipeline(DiffusionPipeline, TextualInversion
         feature_extractor ([`~transformers.CLIPImageProcessor`]):
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
+
     model_cpu_offload_seq = "text_encoder->unet->vae"
     _optional_components = ["safety_checker", "feature_extractor"]
     _exclude_from_cpu_offload = ["safety_checker"]
@@ -211,7 +226,7 @@ class StableDiffusionInstructPix2PixPipeline(DiffusionPipeline, TextualInversion
             callback_on_step_end_tensor_inputs (`List`, *optional*):
                 The list of tensor inputs for the `callback_on_step_end` function. The tensors specified in the list
                 will be passed as `callback_kwargs` argument. You will only be able to include variables listed in the
-                `._callback_tensor_inputs` attribute of your pipeine class.
+                `._callback_tensor_inputs` attribute of your pipeline class.
 
         Examples:
 
@@ -319,7 +334,6 @@ class StableDiffusionInstructPix2PixPipeline(DiffusionPipeline, TextualInversion
             prompt_embeds.dtype,
             device,
             self.do_classifier_free_guidance,
-            generator,
         )
 
         height, width = image_latents.shape[-2:]
@@ -715,17 +729,7 @@ class StableDiffusionInstructPix2PixPipeline(DiffusionPipeline, TextualInversion
         if image.shape[1] == 4:
             image_latents = image
         else:
-            if isinstance(generator, list) and len(generator) != batch_size:
-                raise ValueError(
-                    f"You have passed a list of generators of length {len(generator)}, but requested an effective batch"
-                    f" size of {batch_size}. Make sure the batch size matches the length of the generators."
-                )
-
-            if isinstance(generator, list):
-                image_latents = [self.vae.encode(image[i : i + 1]).latent_dist.mode() for i in range(batch_size)]
-                image_latents = torch.cat(image_latents, dim=0)
-            else:
-                image_latents = self.vae.encode(image).latent_dist.mode()
+            image_latents = retrieve_latents(self.vae.encode(image), sample_mode="argmax")
 
         if batch_size > image_latents.shape[0] and batch_size % image_latents.shape[0] == 0:
             # expand image_latents for batch_size

diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_k_diffusion.py

@@ -80,6 +80,7 @@ class StableDiffusionKDiffusionPipeline(DiffusionPipeline, TextualInversionLoade
         feature_extractor ([`CLIPImageProcessor`]):
             Model that extracts features from generated images to be used as inputs for the `safety_checker`.
     """
+
     model_cpu_offload_seq = "text_encoder->unet->vae"
     _optional_components = ["safety_checker", "feature_extractor"]
     _exclude_from_cpu_offload = ["safety_checker"]

diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_latent_upscale.py

@@ -79,6 +79,7 @@ class StableDiffusionLatentUpscalePipeline(DiffusionPipeline, FromSingleFileMixi
         scheduler ([`SchedulerMixin`]):
             A [`EulerDiscreteScheduler`] to be used in combination with `unet` to denoise the encoded image latents.
     """
+
     model_cpu_offload_seq = "text_encoder->unet->vae"
 
     def __init__(

diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_ldm3d.py

@@ -115,6 +115,7 @@ class StableDiffusionLDM3DPipeline(
         feature_extractor ([`~transformers.CLIPImageProcessor`]):
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
+
     model_cpu_offload_seq = "text_encoder->unet->vae"
     _optional_components = ["safety_checker", "feature_extractor"]
     _exclude_from_cpu_offload = ["safety_checker"]

diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_model_editing.py

@@ -66,6 +66,7 @@ class StableDiffusionModelEditingPipeline(DiffusionPipeline, TextualInversionLoa
         with_augs ([`list`]):
             Textual augmentations to apply while editing the text-to-image model. Set to `[]` for no augmentations.
     """
+
     model_cpu_offload_seq = "text_encoder->unet->vae"
     _optional_components = ["safety_checker", "feature_extractor"]
     _exclude_from_cpu_offload = ["safety_checker"]

diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_panorama.py

@@ -85,6 +85,7 @@ class StableDiffusionPanoramaPipeline(DiffusionPipeline, TextualInversionLoaderM
         feature_extractor ([`~transformers.CLIPImageProcessor`]):
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
+
     model_cpu_offload_seq = "text_encoder->unet->vae"
     _optional_components = ["safety_checker", "feature_extractor"]
     _exclude_from_cpu_offload = ["safety_checker"]

diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_paradigms.py

@@ -96,6 +96,7 @@ class StableDiffusionParadigmsPipeline(
         feature_extractor ([`~transformers.CLIPImageProcessor`]):
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
+
     model_cpu_offload_seq = "text_encoder->unet->vae"
     _optional_components = ["safety_checker", "feature_extractor"]
     _exclude_from_cpu_offload = ["safety_checker"]

diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_pix2pix_zero.py

@@ -310,6 +310,7 @@ class StableDiffusionPix2PixZeroPipeline(DiffusionPipeline):
             Whether the pipeline requires a safety checker. We recommend setting it to True if you're using the
             pipeline publicly.
     """
+
     model_cpu_offload_seq = "text_encoder->unet->vae"
     _optional_components = [
         "safety_checker",

diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_sag.py

@@ -124,6 +124,7 @@ class StableDiffusionSAGPipeline(DiffusionPipeline, TextualInversionLoaderMixin)
         feature_extractor ([`~transformers.CLIPImageProcessor`]):
             A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
     """
+
     model_cpu_offload_seq = "text_encoder->unet->vae"
     _optional_components = ["safety_checker", "feature_extractor"]
     _exclude_from_cpu_offload = ["safety_checker"]

diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_upscale.py

@@ -92,6 +92,7 @@ class StableDiffusionUpscalePipeline(
             A scheduler to be used in combination with `unet` to denoise the encoded image latents. Can be one of
             [`DDIMScheduler`], [`LMSDiscreteScheduler`], or [`PNDMScheduler`].
     """
+
     model_cpu_offload_seq = "text_encoder->unet->vae"
     _optional_components = ["watermarker", "safety_checker", "feature_extractor"]
     _exclude_from_cpu_offload = ["safety_checker"]