DiffEdit

DiffEdit: Diffusion-based semantic image editing with mask guidance is by Guillaume Couairon, Jakob Verbeek, Holger Schwenk, and Matthieu Cord.

The abstract from the paper is:

Image generation has recently seen tremendous advances, with diffusion models allowing to synthesize convincing images for a large variety of text prompts. In this article, we propose DiffEdit, a method to take advantage of text-conditioned diffusion models for the task of semantic image editing, where the goal is to edit an image based on a text query. Semantic image editing is an extension of image generation, with the additional constraint that the generated image should be as similar as possible to a given input image. Current editing methods based on diffusion models usually require to provide a mask, making the task much easier by treating it as a conditional inpainting task. In contrast, our main contribution is able to automatically generate a mask highlighting regions of the input image that need to be edited, by contrasting predictions of a diffusion model conditioned on different text prompts. Moreover, we rely on latent inference to preserve content in those regions of interest and show excellent synergies with mask-based diffusion. DiffEdit achieves state-of-the-art editing performance on ImageNet. In addition, we evaluate semantic image editing in more challenging settings, using images from the COCO dataset as well as text-based generated images.

The original codebase can be found at Xiang-cd/DiffEdit-stable-diffusion, and you can try it out in this demo.

This pipeline was contributed by clarencechen. ❤️

Tips

• The pipeline can generate masks that can be fed into other inpainting pipelines.
• In order to generate an image using this pipeline, both an image mask (source and target prompts can be manually specified or generated, and passed to StableDiffusionDiffEditPipeline.generate_mask) and a set of partially inverted latents (generated using StableDiffusionDiffEditPipeline.invert) must be provided as arguments when calling the pipeline to generate the final edited image.
• The function StableDiffusionDiffEditPipeline.generate_mask exposes two prompt arguments, source_prompt and target_prompt that let you control the locations of the semantic edits in the final image to be generated. Let's say, you wanted to translate from "cat" to "dog". In this case, the edit direction will be "cat -> dog". To reflect this in the generated mask, you simply have to set the embeddings related to the phrases including "cat" to source_prompt and "dog" to target_prompt.
• When generating partially inverted latents using invert, assign a caption or text embedding describing the overall image to the prompt argument to help guide the inverse latent sampling process. In most cases, the source concept is sufficiently descriptive to yield good results, but feel free to explore alternatives.
• When calling the pipeline to generate the final edited image, assign the source concept to negative_prompt and the target concept to prompt. Taking the above example, you simply have to set the embeddings related to the phrases including "cat" to negative_prompt and "dog" to prompt.
• If you wanted to reverse the direction in the example above, i.e., "dog -> cat", then it's recommended to:
  • Swap the source_prompt and target_prompt in the arguments to generate_mask.
  • Change the input prompt in StableDiffusionDiffEditPipeline.invert to include "dog".
  • Swap the prompt and negative_prompt in the arguments to call the pipeline to generate the final edited image.
• The source and target prompts, or their corresponding embeddings, can also be automatically generated. Please refer to the DiffEdit guide for more details.

mindone.diffusers.StableDiffusionDiffEditPipeline

Bases: DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin, LoraLoaderMixin

This is an experimental feature!

Pipeline for text-guided image inpainting using Stable Diffusion and DiffEdit.

This model inherits from [DiffusionPipeline]. Check the superclass documentation for the generic methods implemented for all pipelines (downloading, saving, running on a particular device, etc.).

The pipeline also inherits the following loading and saving methods

• [~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

PARAMETER	DESCRIPTION
vae	Variational Auto-Encoder (VAE) model to encode and decode images to and from latent representations. TYPE: [`AutoencoderKL`]
text_encoder	Frozen text-encoder (clip-vit-large-patch14). TYPE: [`~transformers.CLIPTextModel`]
tokenizer	A CLIPTokenizer to tokenize text. TYPE: [`~transformers.CLIPTokenizer`]
unet	A UNet2DConditionModel to denoise the encoded image latents. TYPE: [`UNet2DConditionModel`]
scheduler	A scheduler to be used in combination with unet to denoise the encoded image latents. TYPE: [`SchedulerMixin`]
inverse_scheduler	A scheduler to be used in combination with unet to fill in the unmasked part of the input latents. TYPE: [`DDIMInverseScheduler`]
safety_checker	Classification module that estimates whether generated images could be considered offensive or harmful. Please refer to the model card for more details about a model's potential harms. TYPE: [`StableDiffusionSafetyChecker`]
feature_extractor	A CLIPImageProcessor to extract features from generated images; used as inputs to the safety_checker. TYPE: [`~transformers.CLIPImageProcessor`]

Source code in mindone/diffusers/pipelines/stable_diffusion_diffedit/pipeline_stable_diffusion_diffedit.py

class StableDiffusionDiffEditPipeline(
    DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin, LoraLoaderMixin
):
    r"""
    <Tip warning={true}>

    This is an experimental feature!

    </Tip>

    Pipeline for text-guided image inpainting using Stable Diffusion and DiffEdit.

    This model inherits from [`DiffusionPipeline`]. Check the superclass documentation for the generic methods
    implemented for all pipelines (downloading, saving, running on a particular device, etc.).

    The pipeline also inherits the following loading and saving methods:
        - [`~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

    Args:
        vae ([`AutoencoderKL`]):
            Variational Auto-Encoder (VAE) model to encode and decode images to and from latent representations.
        text_encoder ([`~transformers.CLIPTextModel`]):
            Frozen text-encoder ([clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14)).
        tokenizer ([`~transformers.CLIPTokenizer`]):
            A `CLIPTokenizer` to tokenize text.
        unet ([`UNet2DConditionModel`]):
            A `UNet2DConditionModel` to denoise the encoded image latents.
        scheduler ([`SchedulerMixin`]):
            A scheduler to be used in combination with `unet` to denoise the encoded image latents.
        inverse_scheduler ([`DDIMInverseScheduler`]):
            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.
        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", "inverse_scheduler"]
    _exclude_from_cpu_offload = ["safety_checker"]

    def __init__(
        self,
        vae: AutoencoderKL,
        text_encoder: CLIPTextModel,
        tokenizer: CLIPTokenizer,
        unet: UNet2DConditionModel,
        scheduler: KarrasDiffusionSchedulers,
        safety_checker: StableDiffusionSafetyChecker,
        feature_extractor: CLIPImageProcessor,
        inverse_scheduler: DDIMInverseScheduler,
        requires_safety_checker: bool = True,
    ):
        super().__init__()

        if hasattr(scheduler.config, "steps_offset") and scheduler.config.steps_offset != 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 "
                "to update the config accordingly as leaving `steps_offset` might led 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 the `scheduler/scheduler_config.json`"
                " file"
            )
            deprecate("steps_offset!=1", "1.0.0", deprecation_message, standard_warn=False)
            new_config = dict(scheduler.config)
            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:
            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"
                " sure to update the config accordingly as not setting `skip_prk_steps` 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 the"
                " `scheduler/scheduler_config.json` file"
            )
            deprecate("skip_prk_steps not set", "1.0.0", deprecation_message, standard_warn=False)
            new_config = dict(scheduler.config)
            new_config["skip_prk_steps"] = True
            scheduler._internal_dict = FrozenDict(new_config)

        if safety_checker is None and requires_safety_checker:
            logger.warning(
                f"You have disabled the safety checker for {self.__class__} by passing `safety_checker=None`. Ensure"
                " that you abide to the conditions of the Stable Diffusion license and do not expose unfiltered"
                " results in services or applications open to the public. Both the diffusers team and Hugging Face"
                " strongly recommend to keep the safety filter enabled in all public facing circumstances, disabling"
                " it only for use-cases that involve analyzing network behavior or auditing its results. For more"
                " information, please have a look at https://github.com/huggingface/diffusers/pull/254 ."
            )

        if safety_checker is not None and feature_extractor is None:
            raise ValueError(
                "Make sure to define a feature extractor when loading {self.__class__} if you want to use the safety"
                " 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
        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"
                " 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"
                " the `unet/config.json` file"
            )
            deprecate("sample_size<64", "1.0.0", deprecation_message, standard_warn=False)
            new_config = dict(unet.config)
            new_config["sample_size"] = 64
            unet._internal_dict = FrozenDict(new_config)

        self.register_modules(
            vae=vae,
            text_encoder=text_encoder,
            tokenizer=tokenizer,
            unet=unet,
            scheduler=scheduler,
            safety_checker=safety_checker,
            feature_extractor=feature_extractor,
            inverse_scheduler=inverse_scheduler,
        )
        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
        self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
        self.register_to_config(requires_safety_checker=requires_safety_checker)

    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline._encode_prompt
    def _encode_prompt(
        self,
        prompt,
        num_images_per_prompt,
        do_classifier_free_guidance,
        negative_prompt=None,
        prompt_embeds: Optional[ms.Tensor] = None,
        negative_prompt_embeds: Optional[ms.Tensor] = None,
        lora_scale: Optional[float] = None,
        **kwargs,
    ):
        deprecation_message = (
            "`_encode_prompt()` is deprecated and it will be removed in a future version. Use"
            "`encode_prompt()` instead. Also, be aware that the output format changed from a"
            "concatenated tensor to a tuple."
        )
        deprecate("_encode_prompt()", "1.0.0", deprecation_message, standard_warn=False)

        prompt_embeds_tuple = self.encode_prompt(
            prompt=prompt,
            num_images_per_prompt=num_images_per_prompt,
            do_classifier_free_guidance=do_classifier_free_guidance,
            negative_prompt=negative_prompt,
            prompt_embeds=prompt_embeds,
            negative_prompt_embeds=negative_prompt_embeds,
            lora_scale=lora_scale,
            **kwargs,
        )

        # concatenate for backwards comp
        prompt_embeds = ops.cat([prompt_embeds_tuple[1], prompt_embeds_tuple[0]])

        return prompt_embeds

    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.encode_prompt
    def encode_prompt(
        self,
        prompt,
        num_images_per_prompt,
        do_classifier_free_guidance,
        negative_prompt=None,
        prompt_embeds: Optional[ms.Tensor] = None,
        negative_prompt_embeds: Optional[ms.Tensor] = None,
        lora_scale: Optional[float] = None,
        clip_skip: Optional[int] = None,
    ):
        r"""
        Encodes the prompt into text encoder hidden states.

        Args:
            prompt (`str` or `List[str]`, *optional*):
                prompt to be encoded
            num_images_per_prompt (`int`):
                number of images that should be generated per prompt
            do_classifier_free_guidance (`bool`):
                whether to use classifier free guidance or not
            negative_prompt (`str` or `List[str]`, *optional*):
                The prompt or prompts not to guide the image generation. If not defined, one has to pass
                `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is
                less than `1`).
            prompt_embeds (`ms.Tensor`, *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.
            negative_prompt_embeds (`ms.Tensor`, *optional*):
                Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
                weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
                argument.
            lora_scale (`float`, *optional*):
                A LoRA scale that will be applied to all LoRA layers of the text encoder if LoRA layers are loaded.
            clip_skip (`int`, *optional*):
                Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that
                the output of the pre-final layer will be used for computing the prompt embeddings.
        """
        # set lora scale so that monkey patched LoRA
        # function of text encoder can correctly access it
        if lora_scale is not None and isinstance(self, LoraLoaderMixin):
            self._lora_scale = lora_scale

            scale_lora_layers(self.text_encoder, lora_scale)

        if prompt is not None and isinstance(prompt, str):
            batch_size = 1
        elif prompt is not None and isinstance(prompt, list):
            batch_size = len(prompt)
        else:
            batch_size = prompt_embeds.shape[0]

        if prompt_embeds is None:
            # textual inversion: process multi-vector tokens if necessary
            if isinstance(self, TextualInversionLoaderMixin):
                prompt = self.maybe_convert_prompt(prompt, self.tokenizer)

            text_inputs = self.tokenizer(
                prompt,
                padding="max_length",
                max_length=self.tokenizer.model_max_length,
                truncation=True,
                return_tensors="np",
            )
            text_input_ids = text_inputs.input_ids
            untruncated_ids = self.tokenizer(prompt, padding="longest", return_tensors="np").input_ids

            if untruncated_ids.shape[-1] >= text_input_ids.shape[-1] and not np.array_equal(
                text_input_ids, untruncated_ids
            ):
                removed_text = self.tokenizer.batch_decode(untruncated_ids[:, self.tokenizer.model_max_length - 1 : -1])
                logger.warning(
                    "The following part of your input was truncated because CLIP can only handle sequences up to"
                    f" {self.tokenizer.model_max_length} tokens: {removed_text}"
                )

            if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
                attention_mask = ms.tensor(text_inputs.attention_mask)
            else:
                attention_mask = None

            if clip_skip is None:
                prompt_embeds = self.text_encoder(ms.tensor(text_input_ids), attention_mask=attention_mask)
                prompt_embeds = prompt_embeds[0]
            else:
                prompt_embeds = self.text_encoder(
                    ms.tensor(text_input_ids), attention_mask=attention_mask, output_hidden_states=True
                )
                # Access the `hidden_states` first, that contains a tuple of
                # all the hidden states from the encoder layers. Then index into
                # the tuple to access the hidden states from the desired layer.
                prompt_embeds = prompt_embeds[-1][-(clip_skip + 1)]
                # We also need to apply the final LayerNorm here to not mess with the
                # representations. The `last_hidden_states` that we typically use for
                # obtaining the final prompt representations passes through the LayerNorm
                # layer.
                prompt_embeds = self.text_encoder.text_model.final_layer_norm(prompt_embeds)

        if self.text_encoder is not None:
            prompt_embeds_dtype = self.text_encoder.dtype
        elif self.unet is not None:
            prompt_embeds_dtype = self.unet.dtype
        else:
            prompt_embeds_dtype = prompt_embeds.dtype

        prompt_embeds = prompt_embeds.to(dtype=prompt_embeds_dtype)

        bs_embed, seq_len, _ = prompt_embeds.shape
        # duplicate text embeddings for each generation per prompt, using mps friendly method
        prompt_embeds = prompt_embeds.tile((1, num_images_per_prompt, 1))
        prompt_embeds = prompt_embeds.view(bs_embed * num_images_per_prompt, seq_len, -1)

        # get unconditional embeddings for classifier free guidance
        if do_classifier_free_guidance and negative_prompt_embeds is None:
            uncond_tokens: List[str]
            if negative_prompt is None:
                uncond_tokens = [""] * batch_size
            elif prompt is not None and type(prompt) is not type(negative_prompt):
                raise TypeError(
                    f"`negative_prompt` should be the same type to `prompt`, but got {type(negative_prompt)} !="
                    f" {type(prompt)}."
                )
            elif isinstance(negative_prompt, str):
                uncond_tokens = [negative_prompt]
            elif batch_size != len(negative_prompt):
                raise ValueError(
                    f"`negative_prompt`: {negative_prompt} has batch size {len(negative_prompt)}, but `prompt`:"
                    f" {prompt} has batch size {batch_size}. Please make sure that passed `negative_prompt` matches"
                    " the batch size of `prompt`."
                )
            else:
                uncond_tokens = negative_prompt

            # textual inversion: process multi-vector tokens if necessary
            if isinstance(self, TextualInversionLoaderMixin):
                uncond_tokens = self.maybe_convert_prompt(uncond_tokens, self.tokenizer)

            max_length = prompt_embeds.shape[1]
            uncond_input = self.tokenizer(
                uncond_tokens,
                padding="max_length",
                max_length=max_length,
                truncation=True,
                return_tensors="np",
            )

            if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
                attention_mask = ms.Tensor(uncond_input.attention_mask)
            else:
                attention_mask = None

            negative_prompt_embeds = self.text_encoder(
                ms.Tensor(uncond_input.input_ids),
                attention_mask=attention_mask,
            )
            negative_prompt_embeds = negative_prompt_embeds[0]

        if do_classifier_free_guidance:
            # duplicate unconditional embeddings for each generation per prompt, using mps friendly method
            seq_len = negative_prompt_embeds.shape[1]

            negative_prompt_embeds = negative_prompt_embeds.to(dtype=prompt_embeds_dtype)

            negative_prompt_embeds = negative_prompt_embeds.tile((1, num_images_per_prompt, 1))
            negative_prompt_embeds = negative_prompt_embeds.view(batch_size * num_images_per_prompt, seq_len, -1)

        if self.text_encoder is not None:
            if isinstance(self, LoraLoaderMixin):
                # Retrieve the original scale by scaling back the LoRA layers
                unscale_lora_layers(self.text_encoder, lora_scale)

        return prompt_embeds, negative_prompt_embeds

    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.run_safety_checker
    def run_safety_checker(self, image, dtype):
        if self.safety_checker is None:
            has_nsfw_concept = None
        else:
            if ops.is_tensor(image):
                feature_extractor_input = self.image_processor.postprocess(image, output_type="pil")
            else:
                feature_extractor_input = self.image_processor.numpy_to_pil(image)
            safety_checker_input = self.feature_extractor(feature_extractor_input, return_tensors="pt")
            image, has_nsfw_concept = self.safety_checker(
                images=image, clip_input=safety_checker_input.pixel_values.to(dtype)
            )
        return image, has_nsfw_concept

    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_extra_step_kwargs
    def prepare_extra_step_kwargs(self, generator, eta):
        # prepare extra kwargs for the scheduler step, since not all schedulers have the same signature
        # eta (η) is only used with the DDIMScheduler, it will be ignored for other schedulers.
        # eta corresponds to η in DDIM paper: https://arxiv.org/abs/2010.02502
        # and should be between [0, 1]

        accepts_eta = "eta" in set(inspect.signature(self.scheduler.step).parameters.keys())
        extra_step_kwargs = {}
        if accepts_eta:
            extra_step_kwargs["eta"] = eta

        # check if the scheduler accepts generator
        accepts_generator = "generator" in set(inspect.signature(self.scheduler.step).parameters.keys())
        if accepts_generator:
            extra_step_kwargs["generator"] = generator
        return extra_step_kwargs

    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.decode_latents
    def decode_latents(self, latents):
        deprecation_message = "The decode_latents method is deprecated and will be removed in 1.0.0. Please use VaeImageProcessor.postprocess(...) instead"
        deprecate("decode_latents", "1.0.0", deprecation_message, standard_warn=False)

        latents = 1 / self.vae.config.scaling_factor * latents
        image = self.vae.decode(latents, return_dict=False)[0]
        image = (image / 2 + 0.5).clamp(0, 1)
        # we always cast to float32 as this does not cause significant overhead and is compatible with bfloat16
        image = image.permute(0, 2, 3, 1).float().numpy()
        return image

    def check_inputs(
        self,
        prompt,
        strength,
        callback_steps,
        negative_prompt=None,
        prompt_embeds=None,
        negative_prompt_embeds=None,
    ):
        if (strength is None) or (strength is not None and (strength < 0 or strength > 1)):
            raise ValueError(
                f"The value of `strength` should in [0.0, 1.0] but is, but is {strength} of type {type(strength)}."
            )

        if (callback_steps is None) or (
            callback_steps is not None and (not isinstance(callback_steps, int) or callback_steps <= 0)
        ):
            raise ValueError(
                f"`callback_steps` has to be a positive integer but is {callback_steps} of type"
                f" {type(callback_steps)}."
            )

        if prompt is not None and prompt_embeds is not None:
            raise ValueError(
                f"Cannot forward both `prompt`: {prompt} and `prompt_embeds`: {prompt_embeds}. Please make sure to"
                " only forward one of the two."
            )
        elif prompt is None and prompt_embeds is None:
            raise ValueError(
                "Provide either `prompt` or `prompt_embeds`. Cannot leave both `prompt` and `prompt_embeds` undefined."
            )
        elif prompt is not None and (not isinstance(prompt, str) and not isinstance(prompt, list)):
            raise ValueError(f"`prompt` has to be of type `str` or `list` but is {type(prompt)}")

        if negative_prompt is not None and negative_prompt_embeds is not None:
            raise ValueError(
                f"Cannot forward both `negative_prompt`: {negative_prompt} and `negative_prompt_embeds`:"
                f" {negative_prompt_embeds}. Please make sure to only forward one of the two."
            )

        if prompt_embeds is not None and negative_prompt_embeds is not None:
            if prompt_embeds.shape != negative_prompt_embeds.shape:
                raise ValueError(
                    "`prompt_embeds` and `negative_prompt_embeds` must have the same shape when passed directly, but"
                    f" got: `prompt_embeds` {prompt_embeds.shape} != `negative_prompt_embeds`"
                    f" {negative_prompt_embeds.shape}."
                )

    def check_source_inputs(
        self,
        source_prompt=None,
        source_negative_prompt=None,
        source_prompt_embeds=None,
        source_negative_prompt_embeds=None,
    ):
        if source_prompt is not None and source_prompt_embeds is not None:
            raise ValueError(
                f"Cannot forward both `source_prompt`: {source_prompt} and `source_prompt_embeds`: {source_prompt_embeds}."
                "  Please make sure to only forward one of the two."
            )
        elif source_prompt is None and source_prompt_embeds is None:
            raise ValueError(
                "Provide either `source_image` or `source_prompt_embeds`. Cannot leave all both of the arguments undefined."
            )
        elif source_prompt is not None and (not isinstance(source_prompt, str) and not isinstance(source_prompt, list)):
            raise ValueError(f"`source_prompt` has to be of type `str` or `list` but is {type(source_prompt)}")

        if source_negative_prompt is not None and source_negative_prompt_embeds is not None:
            raise ValueError(
                f"Cannot forward both `source_negative_prompt`: {source_negative_prompt} and `source_negative_prompt_embeds`:"
                f" {source_negative_prompt_embeds}. Please make sure to only forward one of the two."
            )

        if source_prompt_embeds is not None and source_negative_prompt_embeds is not None:
            if source_prompt_embeds.shape != source_negative_prompt_embeds.shape:
                raise ValueError(
                    "`source_prompt_embeds` and `source_negative_prompt_embeds` must have the same shape when passed"
                    f" directly, but got: `source_prompt_embeds` {source_prompt_embeds.shape} !="
                    f" `source_negative_prompt_embeds` {source_negative_prompt_embeds.shape}."
                )

    def get_timesteps(self, num_inference_steps, strength):
        # get the original timestep using init_timestep
        init_timestep = min(int(num_inference_steps * strength), num_inference_steps)

        t_start = max(num_inference_steps - init_timestep, 0)
        timesteps = self.scheduler.timesteps[t_start * self.scheduler.order :]

        return timesteps, num_inference_steps - t_start

    def get_inverse_timesteps(self, num_inference_steps, strength):
        # get the original timestep using init_timestep
        init_timestep = min(int(num_inference_steps * strength), num_inference_steps)

        t_start = max(num_inference_steps - init_timestep, 0)

        # safety for t_start overflow to prevent empty timsteps slice
        if t_start == 0:
            return self.inverse_scheduler.timesteps, num_inference_steps
        timesteps = self.inverse_scheduler.timesteps[:-t_start]

        return timesteps, num_inference_steps - t_start

    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_latents
    def prepare_latents(self, batch_size, num_channels_latents, height, width, dtype, generator, latents=None):
        shape = (
            batch_size,
            num_channels_latents,
            int(height) // self.vae_scale_factor,
            int(width) // self.vae_scale_factor,
        )
        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 latents is None:
            latents = randn_tensor(shape, generator=generator, dtype=dtype)

        # scale the initial noise by the standard deviation required by the scheduler
        latents = (latents * self.scheduler.init_noise_sigma).to(dtype)
        return latents

    def prepare_image_latents(self, image, batch_size, dtype, generator=None):
        if not isinstance(image, (ms.Tensor, PIL.Image.Image, list)):
            raise ValueError(f"`image` has to be of type `ms.Tensor`, `PIL.Image.Image` or list but is {type(image)}")

        image = image.to(dtype=dtype)

        if image.shape[1] == 4:
            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):
                latents = [
                    self.vae.encode(image[i : i + 1]).latent_dist.sample(generator[i], generator=generator)
                    for i in range(batch_size)
                ]
                latents = ops.cat(latents, axis=0)
            else:
                latents = self.vae.diag_gauss_dist.sample(self.vae.encode(image)[0], generator=generator)

            latents = self.vae.config.scaling_factor * latents

        if batch_size != latents.shape[0]:
            if batch_size % latents.shape[0] == 0:
                # expand image_latents for batch_size
                deprecation_message = (
                    f"You have passed {batch_size} text prompts (`prompt`), but only {latents.shape[0]} initial"
                    " images (`image`). Initial images are now duplicating to match the number of text prompts. Note"
                    " that this behavior is deprecated and will be removed in a version 1.0.0. Please make sure to update"
                    " your script to pass as many initial images as text prompts to suppress this warning."
                )
                deprecate("len(prompt) != len(image)", "1.0.0", deprecation_message, standard_warn=False)
                additional_latents_per_image = batch_size // latents.shape[0]
                latents = ops.cat([latents] * additional_latents_per_image, axis=0)
            else:
                raise ValueError(
                    f"Cannot duplicate `image` of batch size {latents.shape[0]} to {batch_size} text prompts."
                )
        else:
            latents = ops.cat([latents], axis=0)

        return latents

    def get_epsilon(self, model_output: ms.Tensor, sample: ms.Tensor, timestep: int):
        pred_type = self.inverse_scheduler.config.prediction_type
        alpha_prod_t = self.inverse_scheduler.alphas_cumprod[timestep]

        beta_prod_t = 1 - alpha_prod_t

        if pred_type == "epsilon":
            return model_output
        elif pred_type == "sample":
            return (sample - alpha_prod_t ** (0.5) * model_output) / beta_prod_t ** (0.5)
        elif pred_type == "v_prediction":
            return (alpha_prod_t**0.5) * model_output + (beta_prod_t**0.5) * sample
        else:
            raise ValueError(
                f"prediction_type given as {pred_type} must be one of `epsilon`, `sample`, or `v_prediction`"
            )

    def generate_mask(
        self,
        image: Union[ms.Tensor, PIL.Image.Image] = None,
        target_prompt: Optional[Union[str, List[str]]] = None,
        target_negative_prompt: Optional[Union[str, List[str]]] = None,
        target_prompt_embeds: Optional[ms.Tensor] = None,
        target_negative_prompt_embeds: Optional[ms.Tensor] = None,
        source_prompt: Optional[Union[str, List[str]]] = None,
        source_negative_prompt: Optional[Union[str, List[str]]] = None,
        source_prompt_embeds: Optional[ms.Tensor] = None,
        source_negative_prompt_embeds: Optional[ms.Tensor] = None,
        num_maps_per_mask: Optional[int] = 10,
        mask_encode_strength: Optional[float] = 0.5,
        mask_thresholding_ratio: Optional[float] = 3.0,
        num_inference_steps: int = 50,
        guidance_scale: float = 7.5,
        generator: Optional[Union[np.random.Generator, List[np.random.Generator]]] = None,
        output_type: Optional[str] = "np",
        cross_attention_kwargs: Optional[Dict[str, Any]] = None,
    ):
        r"""
        Generate a latent mask given a mask prompt, a target prompt, and an image.

        Args:
            image (`PIL.Image.Image`):
                `Image` or tensor representing an image batch to be used for computing the mask.
            target_prompt (`str` or `List[str]`, *optional*):
                The prompt or prompts to guide semantic mask generation. If not defined, you need to pass
                `prompt_embeds`.
            target_negative_prompt (`str` or `List[str]`, *optional*):
                The prompt or prompts to guide what to not include in image generation. If not defined, you need to
                pass `negative_prompt_embeds` instead. Ignored when not using guidance (`guidance_scale < 1`).
            target_prompt_embeds (`ms.Tensor`, *optional*):
                Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
                provided, text embeddings are generated from the `prompt` input argument.
            target_negative_prompt_embeds (`ms.Tensor`, *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.
            source_prompt (`str` or `List[str]`, *optional*):
                The prompt or prompts to guide semantic mask generation using DiffEdit. If not defined, you need to
                pass `source_prompt_embeds` or `source_image` instead.
            source_negative_prompt (`str` or `List[str]`, *optional*):
                The prompt or prompts to guide semantic mask generation away from using DiffEdit. If not defined, you
                need to pass `source_negative_prompt_embeds` or `source_image` instead.
            source_prompt_embeds (`ms.Tensor`, *optional*):
                Pre-generated text embeddings to guide the semantic mask generation. Can be used to easily tweak text
                inputs (prompt weighting). If not provided, text embeddings are generated from `source_prompt` input
                argument.
            source_negative_prompt_embeds (`ms.Tensor`, *optional*):
                Pre-generated text embeddings to negatively guide the semantic mask generation. Can be used to easily
                tweak text inputs (prompt weighting). If not provided, text embeddings are generated from
                `source_negative_prompt` input argument.
            num_maps_per_mask (`int`, *optional*, defaults to 10):
                The number of noise maps sampled to generate the semantic mask using DiffEdit.
            mask_encode_strength (`float`, *optional*, defaults to 0.5):
                The strength of the noise maps sampled to generate the semantic mask using DiffEdit. Must be between 0
                and 1.
            mask_thresholding_ratio (`float`, *optional*, defaults to 3.0):
                The maximum multiple of the mean absolute difference used to clamp the semantic guidance map before
                mask binarization.
            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.
            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`.
            generator (`np.random.Generator` or `List[np.random.Generator]`, *optional*):
                A [`np.random.Generator`](https://numpy.org/doc/stable/reference/random/generator.html) to make
                generation deterministic.
            output_type (`str`, *optional*, defaults to `"pil"`):
                The output format of the generated image. Choose between `PIL.Image` or `np.array`.
            cross_attention_kwargs (`dict`, *optional*):
                A kwargs dictionary that if specified is passed along to the
                [`~models.attention_processor.AttnProcessor`] as defined in
                [`self.processor`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).

        Examples:

        Returns:
            `List[PIL.Image.Image]` or `np.array`:
                When returning a `List[PIL.Image.Image]`, the list consists of a batch of single-channel binary images
                with dimensions `(height // self.vae_scale_factor, width // self.vae_scale_factor)`. If it's
                `np.array`, the shape is `(batch_size, height // self.vae_scale_factor, width //
                self.vae_scale_factor)`.
        """

        # 1. Check inputs (Provide dummy argument for callback_steps)
        self.check_inputs(
            target_prompt,
            mask_encode_strength,
            1,
            target_negative_prompt,
            target_prompt_embeds,
            target_negative_prompt_embeds,
        )

        self.check_source_inputs(
            source_prompt,
            source_negative_prompt,
            source_prompt_embeds,
            source_negative_prompt_embeds,
        )

        if (num_maps_per_mask is None) or (
            num_maps_per_mask is not None and (not isinstance(num_maps_per_mask, int) or num_maps_per_mask <= 0)
        ):
            raise ValueError(
                f"`num_maps_per_mask` has to be a positive integer but is {num_maps_per_mask} of type"
                f" {type(num_maps_per_mask)}."
            )

        if mask_thresholding_ratio is None or mask_thresholding_ratio <= 0:
            raise ValueError(
                f"`mask_thresholding_ratio` has to be positive but is {mask_thresholding_ratio} of type"
                f" {type(mask_thresholding_ratio)}."
            )

        # 2. Define call parameters
        if target_prompt is not None and isinstance(target_prompt, str):
            batch_size = 1
        elif target_prompt is not None and isinstance(target_prompt, list):
            batch_size = len(target_prompt)
        else:
            batch_size = target_prompt_embeds.shape[0]
        if cross_attention_kwargs is None:
            cross_attention_kwargs = {}

        # here `guidance_scale` is defined analog to the guidance weight `w` of equation (2)
        # of the Imagen paper: https://arxiv.org/pdf/2205.11487.pdf . `guidance_scale = 1`
        # corresponds to doing no classifier free guidance.
        do_classifier_free_guidance = guidance_scale > 1.0

        # 3. Encode input prompts
        (cross_attention_kwargs.get("scale", None) if cross_attention_kwargs is not None else None)
        target_negative_prompt_embeds, target_prompt_embeds = self.encode_prompt(
            target_prompt,
            num_maps_per_mask,
            do_classifier_free_guidance,
            target_negative_prompt,
            prompt_embeds=target_prompt_embeds,
            negative_prompt_embeds=target_negative_prompt_embeds,
        )
        # For classifier free guidance, we need to do two forward passes.
        # Here we concatenate the unconditional and text embeddings into a single batch
        # to avoid doing two forward passes
        if do_classifier_free_guidance:
            target_prompt_embeds = ops.cat([target_negative_prompt_embeds, target_prompt_embeds])

        source_negative_prompt_embeds, source_prompt_embeds = self.encode_prompt(
            source_prompt,
            num_maps_per_mask,
            do_classifier_free_guidance,
            source_negative_prompt,
            prompt_embeds=source_prompt_embeds,
            negative_prompt_embeds=source_negative_prompt_embeds,
        )
        if do_classifier_free_guidance:
            source_prompt_embeds = ops.cat([source_negative_prompt_embeds, source_prompt_embeds])

        # 4. Preprocess image
        image = self.image_processor.preprocess(image).repeat_interleave(num_maps_per_mask, dim=0)

        # 5. Set timesteps
        self.scheduler.set_timesteps(num_inference_steps)
        timesteps, _ = self.get_timesteps(num_inference_steps, mask_encode_strength)
        encode_timestep = timesteps[0]

        # 6. Prepare image latents and add noise with specified strength
        image_latents = self.prepare_image_latents(image, batch_size * num_maps_per_mask, self.vae.dtype, generator)
        noise = randn_tensor(image_latents.shape, generator=generator, dtype=self.vae.dtype)
        image_latents = self.scheduler.add_noise(image_latents, noise, encode_timestep)

        latent_model_input = ops.cat([image_latents] * (4 if do_classifier_free_guidance else 2))
        latent_model_input = self.scheduler.scale_model_input(latent_model_input, encode_timestep)

        # 7. Predict the noise residual
        prompt_embeds = ops.cat([source_prompt_embeds, target_prompt_embeds])
        noise_pred = self.unet(
            latent_model_input,
            encode_timestep,
            encoder_hidden_states=prompt_embeds,
            cross_attention_kwargs=cross_attention_kwargs,
        )[0]

        if do_classifier_free_guidance:
            noise_pred_neg_src, noise_pred_source, noise_pred_uncond, noise_pred_target = noise_pred.chunk(4)
            noise_pred_source = noise_pred_neg_src + guidance_scale * (noise_pred_source - noise_pred_neg_src)
            noise_pred_target = noise_pred_uncond + guidance_scale * (noise_pred_target - noise_pred_uncond)
        else:
            noise_pred_source, noise_pred_target = noise_pred.chunk(2)

        # 8. Compute the mask from the absolute difference of predicted noise residuals
        # TODO: Consider smoothing mask guidance map
        mask_guidance_map = (
            ops.abs(noise_pred_target - noise_pred_source)
            .reshape(batch_size, num_maps_per_mask, *noise_pred_target.shape[-3:])
            .mean([1, 2])
        )
        clamp_magnitude = mask_guidance_map.mean() * mask_thresholding_ratio
        semantic_mask_image = mask_guidance_map.clamp(ms.Tensor(0), clamp_magnitude) / clamp_magnitude
        semantic_mask_image = ops.where(semantic_mask_image <= 0.5, ms.Tensor(0), ms.Tensor(1))
        mask_image = semantic_mask_image.asnumpy()

        # 9. Convert to Numpy array or PIL.
        if output_type == "pil":
            mask_image = self.image_processor.numpy_to_pil(mask_image)

        return mask_image

    def invert(
        self,
        prompt: Optional[Union[str, List[str]]] = None,
        image: Union[ms.Tensor, PIL.Image.Image] = None,
        num_inference_steps: int = 50,
        inpaint_strength: float = 0.8,
        guidance_scale: float = 7.5,
        negative_prompt: Optional[Union[str, List[str]]] = None,
        generator: Optional[Union[np.random.Generator, List[np.random.Generator]]] = None,
        prompt_embeds: Optional[ms.Tensor] = None,
        negative_prompt_embeds: Optional[ms.Tensor] = None,
        decode_latents: bool = False,
        output_type: Optional[str] = "pil",
        return_dict: bool = False,
        callback: Optional[Callable[[int, int, ms.Tensor], None]] = None,
        callback_steps: Optional[int] = 1,
        cross_attention_kwargs: Optional[Dict[str, Any]] = None,
        lambda_auto_corr: float = 20.0,
        lambda_kl: float = 20.0,
        num_reg_steps: int = 0,
        num_auto_corr_rolls: int = 5,
    ):
        r"""
        Generate inverted latents given a prompt and image.

        Args:
            prompt (`str` or `List[str]`, *optional*):
                The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`.
            image (`PIL.Image.Image`):
                `Image` or tensor representing an image batch to produce the inverted latents guided by `prompt`.
            inpaint_strength (`float`, *optional*, defaults to 0.8):
                Indicates extent of the noising process to run latent inversion. Must be between 0 and 1. When
                `inpaint_strength` is 1, the inversion process is run for the full number of iterations specified in
                `num_inference_steps`. `image` is used as a reference for the inversion process, and adding more noise
                increases `inpaint_strength`. If `inpaint_strength` is 0, no inpainting occurs.
            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.
            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`.
            negative_prompt (`str` or `List[str]`, *optional*):
                The prompt or prompts to guide what to not include in image generation. If not defined, you need to
                pass `negative_prompt_embeds` instead. Ignored when not using guidance (`guidance_scale < 1`).
            generator (`np.random.Generator`, *optional*):
                A [`np.random.Generator`](https://numpy.org/doc/stable/reference/random/generator.html) to make
                generation deterministic.
            prompt_embeds (`ms.Tensor`, *optional*):
                Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
                provided, text embeddings are generated from the `prompt` input argument.
            negative_prompt_embeds (`ms.Tensor`, *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.
            decode_latents (`bool`, *optional*, defaults to `False`):
                Whether or not to decode the inverted latents into a generated image. Setting this argument to `True`
                decodes all inverted latents for each timestep into a list of generated images.
            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`):
                Whether or not to return a [`~pipelines.stable_diffusion.DiffEditInversionPipelineOutput`] instead of a
                plain tuple.
            callback (`Callable`, *optional*):
                A function that calls every `callback_steps` steps during inference. The function is called with the
                following arguments: `callback(step: int, timestep: int, latents: ms.Tensor)`.
            callback_steps (`int`, *optional*, defaults to 1):
                The frequency at which the `callback` function is called. If not specified, the callback is called at
                every step.
            cross_attention_kwargs (`dict`, *optional*):
                A kwargs dictionary that if specified is passed along to the
                [`~models.attention_processor.AttnProcessor`] as defined in
                [`self.processor`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).
            lambda_auto_corr (`float`, *optional*, defaults to 20.0):
                Lambda parameter to control auto correction.
            lambda_kl (`float`, *optional*, defaults to 20.0):
                Lambda parameter to control Kullback-Leibler divergence output.
            num_reg_steps (`int`, *optional*, defaults to 0):
                Number of regularization loss steps.
            num_auto_corr_rolls (`int`, *optional*, defaults to 5):
                Number of auto correction roll steps.

        Examples:

        Returns:
            [`~pipelines.stable_diffusion.pipeline_stable_diffusion_diffedit.DiffEditInversionPipelineOutput`] or
            `tuple`:
                If `return_dict` is `True`,
                [`~pipelines.stable_diffusion.pipeline_stable_diffusion_diffedit.DiffEditInversionPipelineOutput`] is
                returned, otherwise a `tuple` is returned where the first element is the inverted latents tensors
                ordered by increasing noise, and the second is the corresponding decoded images if `decode_latents` is
                `True`, otherwise `None`.
        """

        # 1. Check inputs
        self.check_inputs(
            prompt,
            inpaint_strength,
            callback_steps,
            negative_prompt,
            prompt_embeds,
            negative_prompt_embeds,
        )

        if image is None:
            raise ValueError("`image` input cannot be undefined.")

        # 2. Define call parameters
        if prompt is not None and isinstance(prompt, str):
            batch_size = 1
        elif prompt is not None and isinstance(prompt, list):
            batch_size = len(prompt)
        else:
            batch_size = prompt_embeds.shape[0]
        if cross_attention_kwargs is None:
            cross_attention_kwargs = {}

        # here `guidance_scale` is defined analog to the guidance weight `w` of equation (2)
        # of the Imagen paper: https://arxiv.org/pdf/2205.11487.pdf . `guidance_scale = 1`
        # corresponds to doing no classifier free guidance.
        do_classifier_free_guidance = guidance_scale > 1.0

        # 3. Preprocess image
        image = self.image_processor.preprocess(image)

        # 4. Prepare latent variables
        num_images_per_prompt = 1
        latents = self.prepare_image_latents(image, batch_size * num_images_per_prompt, self.vae.dtype, generator)

        # 5. Encode input prompt
        prompt_embeds, negative_prompt_embeds = self.encode_prompt(
            prompt,
            num_images_per_prompt,
            do_classifier_free_guidance,
            negative_prompt,
            prompt_embeds=prompt_embeds,
            negative_prompt_embeds=negative_prompt_embeds,
        )
        # For classifier free guidance, we need to do two forward passes.
        # Here we concatenate the unconditional and text embeddings into a single batch
        # to avoid doing two forward passes
        if do_classifier_free_guidance:
            prompt_embeds = ops.cat([negative_prompt_embeds, prompt_embeds])

        # 6. Prepare timesteps
        self.inverse_scheduler.set_timesteps(num_inference_steps)
        timesteps, num_inference_steps = self.get_inverse_timesteps(num_inference_steps, inpaint_strength)

        # 7. Noising loop where we obtain the intermediate noised latent image for each timestep.
        num_warmup_steps = len(timesteps) - num_inference_steps * self.inverse_scheduler.order
        inverted_latents = []
        with self.progress_bar(total=num_inference_steps) as progress_bar:
            for i, t in enumerate(timesteps):
                # expand the latents if we are doing classifier free guidance
                latent_model_input = ops.cat([latents] * 2) if do_classifier_free_guidance else latents
                latent_model_input = self.inverse_scheduler.scale_model_input(latent_model_input, t)

                # predict the noise residual
                noise_pred = self.unet(
                    latent_model_input,
                    t,
                    encoder_hidden_states=prompt_embeds,
                    cross_attention_kwargs=cross_attention_kwargs,
                )[0]

                # perform guidance
                if do_classifier_free_guidance:
                    noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
                    noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)

                # regularization of the noise prediction (not in original code or paper but borrowed from Pix2PixZero)
                if num_reg_steps > 0:
                    raise NotImplementedError("Regularization of the noise prediction is not yet supported.")

                # compute the previous noisy sample x_t -> x_t-1
                latents = self.inverse_scheduler.step(noise_pred, t, latents)[0]
                inverted_latents.append(latents)

                # call the callback, if provided
                if i == len(timesteps) - 1 or (
                    (i + 1) > num_warmup_steps and (i + 1) % self.inverse_scheduler.order == 0
                ):
                    progress_bar.update()
                    if callback is not None and i % callback_steps == 0:
                        step_idx = i // getattr(self.scheduler, "order", 1)
                        callback(step_idx, t, latents)

        assert len(inverted_latents) == len(timesteps)
        latents = ops.stack(list(reversed(inverted_latents)), 1)

        # 8. Post-processing
        image = None
        if decode_latents:
            image = self.decode_latents(latents.flatten(start_dim=0, end_dim=1))

        # 9. Convert to PIL.
        if decode_latents and output_type == "pil":
            image = self.image_processor.numpy_to_pil(image)

        if not return_dict:
            return (latents, image)

        return DiffEditInversionPipelineOutput(latents=latents, images=image)

    def __call__(
        self,
        prompt: Optional[Union[str, List[str]]] = None,
        mask_image: Union[ms.Tensor, PIL.Image.Image] = None,
        image_latents: Union[ms.Tensor, PIL.Image.Image] = None,
        inpaint_strength: Optional[float] = 0.8,
        num_inference_steps: int = 50,
        guidance_scale: float = 7.5,
        negative_prompt: Optional[Union[str, List[str]]] = None,
        num_images_per_prompt: Optional[int] = 1,
        eta: float = 0.0,
        generator: Optional[Union[np.random.Generator, List[np.random.Generator]]] = None,
        latents: Optional[ms.Tensor] = None,
        prompt_embeds: Optional[ms.Tensor] = None,
        negative_prompt_embeds: Optional[ms.Tensor] = None,
        output_type: Optional[str] = "pil",
        return_dict: bool = False,
        callback: Optional[Callable[[int, int, ms.Tensor], None]] = None,
        callback_steps: int = 1,
        cross_attention_kwargs: Optional[Dict[str, Any]] = None,
        clip_skip: int = None,
    ):
        r"""
        The call function to the pipeline for generation.

        Args:
            prompt (`str` or `List[str]`, *optional*):
                The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`.
            mask_image (`PIL.Image.Image`):
                `Image` or tensor representing an image batch to mask the generated image. White pixels in the mask are
                repainted, while black pixels are preserved. If `mask_image` is a PIL image, it is converted to a
                single channel (luminance) before use. If it's a tensor, it should contain one color channel (L)
                instead of 3, so the expected shape would be `(B, 1, H, W)`.
            image_latents (`PIL.Image.Image` or `ms.Tensor`):
                Partially noised image latents from the inversion process to be used as inputs for image generation.
            inpaint_strength (`float`, *optional*, defaults to 0.8):
                Indicates extent to inpaint the masked area. Must be between 0 and 1. When `inpaint_strength` is 1, the
                denoising process is run on the masked area for the full number of iterations specified in
                `num_inference_steps`. `image_latents` is used as a reference for the masked area, and adding more
                noise to a region increases `inpaint_strength`. If `inpaint_strength` is 0, no inpainting occurs.
            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.
            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`.
            negative_prompt (`str` or `List[str]`, *optional*):
                The prompt or prompts to guide what to not include in image generation. If not defined, you need to
                pass `negative_prompt_embeds` instead. Ignored when not using guidance (`guidance_scale < 1`).
            num_images_per_prompt (`int`, *optional*, defaults to 1):
                The number of images to generate per prompt.
            eta (`float`, *optional*, defaults to 0.0):
                Corresponds to parameter eta (η) from the [DDIM](https://arxiv.org/abs/2010.02502) paper. Only applies
                to the [`~schedulers.DDIMScheduler`], and is ignored in other schedulers.
            generator (`np.random.Generator`, *optional*):
                A [`np.random.Generator`](https://numpy.org/doc/stable/reference/random/generator.html) to make
                generation deterministic.
            latents (`ms.Tensor`, *optional*):
                Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image
                generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
                tensor is generated by sampling using the supplied random `generator`.
            prompt_embeds (`ms.Tensor`, *optional*):
                Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
                provided, text embeddings are generated from the `prompt` input argument.
            negative_prompt_embeds (`ms.Tensor`, *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.
            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`):
                Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] instead of a
                plain tuple.
            callback (`Callable`, *optional*):
                A function that calls every `callback_steps` steps during inference. The function is called with the
                following arguments: `callback(step: int, timestep: int, latents: ms.Tensor)`.
            callback_steps (`int`, *optional*, defaults to 1):
                The frequency at which the `callback` function is called. If not specified, the callback is called at
                every step.
            cross_attention_kwargs (`dict`, *optional*):
                A kwargs dictionary that if specified is passed along to the [`AttentionProcessor`] as defined in
                [`self.processor`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).
            clip_skip (`int`, *optional*):
                Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that
                the output of the pre-final layer will be used for computing the prompt embeddings.
        Examples:

        Returns:
            [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] or `tuple`:
                If `return_dict` is `True`, [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] is returned,
                otherwise a `tuple` is returned where the first element is a list with the generated images and the
                second element is a list of `bool`s indicating whether the corresponding generated image contains
                "not-safe-for-work" (nsfw) content.
        """

        # 1. Check inputs
        self.check_inputs(
            prompt,
            inpaint_strength,
            callback_steps,
            negative_prompt,
            prompt_embeds,
            negative_prompt_embeds,
        )

        if mask_image is None:
            raise ValueError(
                "`mask_image` input cannot be undefined. Use `generate_mask()` to compute `mask_image` from text prompts."
            )
        if image_latents is None:
            raise ValueError(
                "`image_latents` input cannot be undefined. Use `invert()` to compute `image_latents` from input images."
            )

        # 2. Define call parameters
        if prompt is not None and isinstance(prompt, str):
            batch_size = 1
        elif prompt is not None and isinstance(prompt, list):
            batch_size = len(prompt)
        else:
            batch_size = prompt_embeds.shape[0]
        if cross_attention_kwargs is None:
            cross_attention_kwargs = {}

        # here `guidance_scale` is defined analog to the guidance weight `w` of equation (2)
        # of the Imagen paper: https://arxiv.org/pdf/2205.11487.pdf . `guidance_scale = 1`
        # corresponds to doing no classifier free guidance.
        do_classifier_free_guidance = guidance_scale > 1.0

        # 3. Encode input prompt
        text_encoder_lora_scale = (
            cross_attention_kwargs.get("scale", None) if cross_attention_kwargs is not None else None
        )
        prompt_embeds, negative_prompt_embeds = self.encode_prompt(
            prompt,
            num_images_per_prompt,
            do_classifier_free_guidance,
            negative_prompt,
            prompt_embeds=prompt_embeds,
            negative_prompt_embeds=negative_prompt_embeds,
            lora_scale=text_encoder_lora_scale,
            clip_skip=clip_skip,
        )
        # For classifier free guidance, we need to do two forward passes.
        # Here we concatenate the unconditional and text embeddings into a single batch
        # to avoid doing two forward passes
        if do_classifier_free_guidance:
            prompt_embeds = ops.cat([negative_prompt_embeds, prompt_embeds])

        # 4. Preprocess mask
        mask_image = preprocess_mask(mask_image, batch_size)
        latent_height, latent_width = mask_image.shape[-2:]
        mask_image = ops.cat([mask_image] * num_images_per_prompt)
        mask_image = mask_image.to(prompt_embeds.dtype)

        # 5. Set timesteps
        self.scheduler.set_timesteps(num_inference_steps)
        timesteps, num_inference_steps = self.get_timesteps(num_inference_steps, inpaint_strength)

        # 6. Preprocess image latents
        if isinstance(image_latents, list) and any(isinstance(i, ms.Tensor) and i.ndim == 5 for i in image_latents):
            image_latents = ops.cat(image_latents)
        elif isinstance(image_latents, ms.Tensor) and image_latents.ndim == 5:
            image_latents = image_latents
        else:
            image_latents = self.image_processor.preprocess(image_latents)

        latent_shape = (self.vae.config.latent_channels, latent_height, latent_width)
        if image_latents.shape[-3:] != latent_shape:
            raise ValueError(
                f"Each latent image in `image_latents` must have shape {latent_shape}, "
                f"but has shape {image_latents.shape[-3:]}"
            )
        if image_latents.ndim == 4:
            image_latents = image_latents.reshape(batch_size, len(timesteps), *latent_shape)
        if image_latents.shape[:2] != (batch_size, len(timesteps)):
            raise ValueError(
                f"`image_latents` must have batch size {batch_size} with latent images from {len(timesteps)}"
                f" timesteps, but has batch size {image_latents.shape[0]} with latent images from"
                f" {image_latents.shape[1]} timesteps."
            )
        image_latents = image_latents.swapaxes(0, 1).repeat_interleave(num_images_per_prompt, dim=1)
        image_latents = image_latents.to(prompt_embeds.dtype)

        # 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)

        # 8. Denoising loop
        latents = image_latents[0]
        num_warmup_steps = len(timesteps) - num_inference_steps * self.scheduler.order
        with self.progress_bar(total=num_inference_steps) as progress_bar:
            for i, t in enumerate(timesteps):
                # expand the latents if we are doing classifier free guidance
                latent_model_input = ops.cat([latents] * 2) if do_classifier_free_guidance else latents
                latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)

                # predict the noise residual
                noise_pred = self.unet(
                    latent_model_input,
                    t,
                    encoder_hidden_states=prompt_embeds,
                    cross_attention_kwargs=cross_attention_kwargs,
                )[0]

                # perform guidance
                if do_classifier_free_guidance:
                    noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
                    noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)

                # compute the previous noisy sample x_t -> x_t-1
                latents = self.scheduler.step(noise_pred, t, latents, **extra_step_kwargs)[0]

                # mask with inverted latents from appropriate timestep - use original image latent for last step
                latents = latents * mask_image + image_latents[i] * (1 - mask_image)

                # call the callback, if provided
                if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0):
                    progress_bar.update()
                    if callback is not None and i % callback_steps == 0:
                        step_idx = i // getattr(self.scheduler, "order", 1)
                        callback(step_idx, t, latents)

        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, prompt_embeds.dtype)
        else:
            image = latents
            has_nsfw_concept = None

        if has_nsfw_concept is None:
            do_denormalize = [True] * image.shape[0]
        else:
            do_denormalize = [not has_nsfw for has_nsfw in has_nsfw_concept]

        image = self.image_processor.postprocess(image, output_type=output_type, do_denormalize=do_denormalize)

        if not return_dict:
            return (image, has_nsfw_concept)

        return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=has_nsfw_concept)

mindone.diffusers.StableDiffusionDiffEditPipeline.__call__(prompt=None, mask_image=None, image_latents=None, inpaint_strength=0.8, num_inference_steps=50, guidance_scale=7.5, negative_prompt=None, num_images_per_prompt=1, eta=0.0, generator=None, latents=None, prompt_embeds=None, negative_prompt_embeds=None, output_type='pil', return_dict=False, callback=None, callback_steps=1, cross_attention_kwargs=None, clip_skip=None)

The call function to the pipeline for generation.

PARAMETER	DESCRIPTION
prompt	The prompt or prompts to guide image generation. If not defined, you need to pass prompt_embeds. TYPE: `str` or `List[str]`, *optional* DEFAULT: None
mask_image	Image or tensor representing an image batch to mask the generated image. White pixels in the mask are repainted, while black pixels are preserved. If mask_image is a PIL image, it is converted to a single channel (luminance) before use. If it's a tensor, it should contain one color channel (L) instead of 3, so the expected shape would be (B, 1, H, W). TYPE: `PIL.Image.Image` DEFAULT: None
image_latents	Partially noised image latents from the inversion process to be used as inputs for image generation. TYPE: `PIL.Image.Image` or `ms.Tensor` DEFAULT: None
inpaint_strength	Indicates extent to inpaint the masked area. Must be between 0 and 1. When inpaint_strength is 1, the denoising process is run on the masked area for the full number of iterations specified in num_inference_steps. image_latents is used as a reference for the masked area, and adding more noise to a region increases inpaint_strength. If inpaint_strength is 0, no inpainting occurs. TYPE: `float`, *optional*, defaults to 0.8 DEFAULT: 0.8
num_inference_steps	The number of denoising steps. More denoising steps usually lead to a higher quality image at the expense of slower inference. TYPE: `int`, *optional*, defaults to 50 DEFAULT: 50
guidance_scale	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. TYPE: `float`, *optional*, defaults to 7.5 DEFAULT: 7.5
negative_prompt	The prompt or prompts to guide what to not include in image generation. If not defined, you need to pass negative_prompt_embeds instead. Ignored when not using guidance (guidance_scale < 1). TYPE: `str` or `List[str]`, *optional* DEFAULT: None
num_images_per_prompt	The number of images to generate per prompt. TYPE: `int`, *optional*, defaults to 1 DEFAULT: 1
eta	Corresponds to parameter eta (η) from the DDIM paper. Only applies to the [~schedulers.DDIMScheduler], and is ignored in other schedulers. TYPE: `float`, *optional*, defaults to 0.0 DEFAULT: 0.0
generator	A np.random.Generator to make generation deterministic. TYPE: `np.random.Generator`, *optional* DEFAULT: None
latents	Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image generation. Can be used to tweak the same generation with different prompts. If not provided, a latents tensor is generated by sampling using the supplied random generator. TYPE: `ms.Tensor`, *optional* DEFAULT: None
prompt_embeds	Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not provided, text embeddings are generated from the prompt input argument. TYPE: `ms.Tensor`, *optional* DEFAULT: None
negative_prompt_embeds	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. TYPE: `ms.Tensor`, *optional* DEFAULT: None
output_type	The output format of the generated image. Choose between PIL.Image or np.array. TYPE: `str`, *optional*, defaults to `"pil"` DEFAULT: 'pil'
return_dict	Whether or not to return a [~pipelines.stable_diffusion.StableDiffusionPipelineOutput] instead of a plain tuple. TYPE: `bool`, *optional*, defaults to `True` DEFAULT: False
callback	A function that calls every callback_steps steps during inference. The function is called with the following arguments: callback(step: int, timestep: int, latents: ms.Tensor). TYPE: `Callable`, *optional* DEFAULT: None
callback_steps	The frequency at which the callback function is called. If not specified, the callback is called at every step. TYPE: `int`, *optional*, defaults to 1 DEFAULT: 1
cross_attention_kwargs	A kwargs dictionary that if specified is passed along to the [AttentionProcessor] as defined in self.processor. TYPE: `dict`, *optional* DEFAULT: None
clip_skip	Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that the output of the pre-final layer will be used for computing the prompt embeddings. TYPE: `int`, *optional* DEFAULT: None

RETURNS	DESCRIPTION
	[~pipelines.stable_diffusion.StableDiffusionPipelineOutput] or tuple: If return_dict is True, [~pipelines.stable_diffusion.StableDiffusionPipelineOutput] is returned, otherwise a tuple is returned where the first element is a list with the generated images and the second element is a list of bools indicating whether the corresponding generated image contains "not-safe-for-work" (nsfw) content.

Source code in mindone/diffusers/pipelines/stable_diffusion_diffedit/pipeline_stable_diffusion_diffedit.py

def __call__(
    self,
    prompt: Optional[Union[str, List[str]]] = None,
    mask_image: Union[ms.Tensor, PIL.Image.Image] = None,
    image_latents: Union[ms.Tensor, PIL.Image.Image] = None,
    inpaint_strength: Optional[float] = 0.8,
    num_inference_steps: int = 50,
    guidance_scale: float = 7.5,
    negative_prompt: Optional[Union[str, List[str]]] = None,
    num_images_per_prompt: Optional[int] = 1,
    eta: float = 0.0,
    generator: Optional[Union[np.random.Generator, List[np.random.Generator]]] = None,
    latents: Optional[ms.Tensor] = None,
    prompt_embeds: Optional[ms.Tensor] = None,
    negative_prompt_embeds: Optional[ms.Tensor] = None,
    output_type: Optional[str] = "pil",
    return_dict: bool = False,
    callback: Optional[Callable[[int, int, ms.Tensor], None]] = None,
    callback_steps: int = 1,
    cross_attention_kwargs: Optional[Dict[str, Any]] = None,
    clip_skip: int = None,
):
    r"""
    The call function to the pipeline for generation.

    Args:
        prompt (`str` or `List[str]`, *optional*):
            The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`.
        mask_image (`PIL.Image.Image`):
            `Image` or tensor representing an image batch to mask the generated image. White pixels in the mask are
            repainted, while black pixels are preserved. If `mask_image` is a PIL image, it is converted to a
            single channel (luminance) before use. If it's a tensor, it should contain one color channel (L)
            instead of 3, so the expected shape would be `(B, 1, H, W)`.
        image_latents (`PIL.Image.Image` or `ms.Tensor`):
            Partially noised image latents from the inversion process to be used as inputs for image generation.
        inpaint_strength (`float`, *optional*, defaults to 0.8):
            Indicates extent to inpaint the masked area. Must be between 0 and 1. When `inpaint_strength` is 1, the
            denoising process is run on the masked area for the full number of iterations specified in
            `num_inference_steps`. `image_latents` is used as a reference for the masked area, and adding more
            noise to a region increases `inpaint_strength`. If `inpaint_strength` is 0, no inpainting occurs.
        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.
        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`.
        negative_prompt (`str` or `List[str]`, *optional*):
            The prompt or prompts to guide what to not include in image generation. If not defined, you need to
            pass `negative_prompt_embeds` instead. Ignored when not using guidance (`guidance_scale < 1`).
        num_images_per_prompt (`int`, *optional*, defaults to 1):
            The number of images to generate per prompt.
        eta (`float`, *optional*, defaults to 0.0):
            Corresponds to parameter eta (η) from the [DDIM](https://arxiv.org/abs/2010.02502) paper. Only applies
            to the [`~schedulers.DDIMScheduler`], and is ignored in other schedulers.
        generator (`np.random.Generator`, *optional*):
            A [`np.random.Generator`](https://numpy.org/doc/stable/reference/random/generator.html) to make
            generation deterministic.
        latents (`ms.Tensor`, *optional*):
            Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image
            generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
            tensor is generated by sampling using the supplied random `generator`.
        prompt_embeds (`ms.Tensor`, *optional*):
            Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
            provided, text embeddings are generated from the `prompt` input argument.
        negative_prompt_embeds (`ms.Tensor`, *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.
        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`):
            Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] instead of a
            plain tuple.
        callback (`Callable`, *optional*):
            A function that calls every `callback_steps` steps during inference. The function is called with the
            following arguments: `callback(step: int, timestep: int, latents: ms.Tensor)`.
        callback_steps (`int`, *optional*, defaults to 1):
            The frequency at which the `callback` function is called. If not specified, the callback is called at
            every step.
        cross_attention_kwargs (`dict`, *optional*):
            A kwargs dictionary that if specified is passed along to the [`AttentionProcessor`] as defined in
            [`self.processor`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).
        clip_skip (`int`, *optional*):
            Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that
            the output of the pre-final layer will be used for computing the prompt embeddings.
    Examples:

    Returns:
        [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] or `tuple`:
            If `return_dict` is `True`, [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] is returned,
            otherwise a `tuple` is returned where the first element is a list with the generated images and the
            second element is a list of `bool`s indicating whether the corresponding generated image contains
            "not-safe-for-work" (nsfw) content.
    """

    # 1. Check inputs
    self.check_inputs(
        prompt,
        inpaint_strength,
        callback_steps,
        negative_prompt,
        prompt_embeds,
        negative_prompt_embeds,
    )

    if mask_image is None:
        raise ValueError(
            "`mask_image` input cannot be undefined. Use `generate_mask()` to compute `mask_image` from text prompts."
        )
    if image_latents is None:
        raise ValueError(
            "`image_latents` input cannot be undefined. Use `invert()` to compute `image_latents` from input images."
        )

    # 2. Define call parameters
    if prompt is not None and isinstance(prompt, str):
        batch_size = 1
    elif prompt is not None and isinstance(prompt, list):
        batch_size = len(prompt)
    else:
        batch_size = prompt_embeds.shape[0]
    if cross_attention_kwargs is None:
        cross_attention_kwargs = {}

    # here `guidance_scale` is defined analog to the guidance weight `w` of equation (2)
    # of the Imagen paper: https://arxiv.org/pdf/2205.11487.pdf . `guidance_scale = 1`
    # corresponds to doing no classifier free guidance.
    do_classifier_free_guidance = guidance_scale > 1.0

    # 3. Encode input prompt
    text_encoder_lora_scale = (
        cross_attention_kwargs.get("scale", None) if cross_attention_kwargs is not None else None
    )
    prompt_embeds, negative_prompt_embeds = self.encode_prompt(
        prompt,
        num_images_per_prompt,
        do_classifier_free_guidance,
        negative_prompt,
        prompt_embeds=prompt_embeds,
        negative_prompt_embeds=negative_prompt_embeds,
        lora_scale=text_encoder_lora_scale,
        clip_skip=clip_skip,
    )
    # For classifier free guidance, we need to do two forward passes.
    # Here we concatenate the unconditional and text embeddings into a single batch
    # to avoid doing two forward passes
    if do_classifier_free_guidance:
        prompt_embeds = ops.cat([negative_prompt_embeds, prompt_embeds])

    # 4. Preprocess mask
    mask_image = preprocess_mask(mask_image, batch_size)
    latent_height, latent_width = mask_image.shape[-2:]
    mask_image = ops.cat([mask_image] * num_images_per_prompt)
    mask_image = mask_image.to(prompt_embeds.dtype)

    # 5. Set timesteps
    self.scheduler.set_timesteps(num_inference_steps)
    timesteps, num_inference_steps = self.get_timesteps(num_inference_steps, inpaint_strength)

    # 6. Preprocess image latents
    if isinstance(image_latents, list) and any(isinstance(i, ms.Tensor) and i.ndim == 5 for i in image_latents):
        image_latents = ops.cat(image_latents)
    elif isinstance(image_latents, ms.Tensor) and image_latents.ndim == 5:
        image_latents = image_latents
    else:
        image_latents = self.image_processor.preprocess(image_latents)

    latent_shape = (self.vae.config.latent_channels, latent_height, latent_width)
    if image_latents.shape[-3:] != latent_shape:
        raise ValueError(
            f"Each latent image in `image_latents` must have shape {latent_shape}, "
            f"but has shape {image_latents.shape[-3:]}"
        )
    if image_latents.ndim == 4:
        image_latents = image_latents.reshape(batch_size, len(timesteps), *latent_shape)
    if image_latents.shape[:2] != (batch_size, len(timesteps)):
        raise ValueError(
            f"`image_latents` must have batch size {batch_size} with latent images from {len(timesteps)}"
            f" timesteps, but has batch size {image_latents.shape[0]} with latent images from"
            f" {image_latents.shape[1]} timesteps."
        )
    image_latents = image_latents.swapaxes(0, 1).repeat_interleave(num_images_per_prompt, dim=1)
    image_latents = image_latents.to(prompt_embeds.dtype)

    # 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)

    # 8. Denoising loop
    latents = image_latents[0]
    num_warmup_steps = len(timesteps) - num_inference_steps * self.scheduler.order
    with self.progress_bar(total=num_inference_steps) as progress_bar:
        for i, t in enumerate(timesteps):
            # expand the latents if we are doing classifier free guidance
            latent_model_input = ops.cat([latents] * 2) if do_classifier_free_guidance else latents
            latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)

            # predict the noise residual
            noise_pred = self.unet(
                latent_model_input,
                t,
                encoder_hidden_states=prompt_embeds,
                cross_attention_kwargs=cross_attention_kwargs,
            )[0]

            # perform guidance
            if do_classifier_free_guidance:
                noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
                noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)

            # compute the previous noisy sample x_t -> x_t-1
            latents = self.scheduler.step(noise_pred, t, latents, **extra_step_kwargs)[0]

            # mask with inverted latents from appropriate timestep - use original image latent for last step
            latents = latents * mask_image + image_latents[i] * (1 - mask_image)

            # call the callback, if provided
            if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0):
                progress_bar.update()
                if callback is not None and i % callback_steps == 0:
                    step_idx = i // getattr(self.scheduler, "order", 1)
                    callback(step_idx, t, latents)

    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, prompt_embeds.dtype)
    else:
        image = latents
        has_nsfw_concept = None

    if has_nsfw_concept is None:
        do_denormalize = [True] * image.shape[0]
    else:
        do_denormalize = [not has_nsfw for has_nsfw in has_nsfw_concept]

    image = self.image_processor.postprocess(image, output_type=output_type, do_denormalize=do_denormalize)

    if not return_dict:
        return (image, has_nsfw_concept)

    return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=has_nsfw_concept)

mindone.diffusers.StableDiffusionDiffEditPipeline.encode_prompt(prompt, num_images_per_prompt, do_classifier_free_guidance, negative_prompt=None, prompt_embeds=None, negative_prompt_embeds=None, lora_scale=None, clip_skip=None)

Encodes the prompt into text encoder hidden states.

PARAMETER	DESCRIPTION
prompt	prompt to be encoded TYPE: `str` or `List[str]`, *optional*
num_images_per_prompt	number of images that should be generated per prompt TYPE: `int`
do_classifier_free_guidance	whether to use classifier free guidance or not TYPE: `bool`
negative_prompt	The prompt or prompts not to guide the image generation. If not defined, one has to pass negative_prompt_embeds instead. Ignored when not using guidance (i.e., ignored if guidance_scale is less than 1). TYPE: `str` or `List[str]`, *optional* DEFAULT: None
prompt_embeds	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. TYPE: `ms.Tensor`, *optional* DEFAULT: None
negative_prompt_embeds	Pre-generated negative text embeddings. Can be used to easily tweak text inputs, e.g. prompt weighting. If not provided, negative_prompt_embeds will be generated from negative_prompt input argument. TYPE: `ms.Tensor`, *optional* DEFAULT: None
lora_scale	A LoRA scale that will be applied to all LoRA layers of the text encoder if LoRA layers are loaded. TYPE: `float`, *optional* DEFAULT: None
clip_skip	Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that the output of the pre-final layer will be used for computing the prompt embeddings. TYPE: `int`, *optional* DEFAULT: None

Source code in mindone/diffusers/pipelines/stable_diffusion_diffedit/pipeline_stable_diffusion_diffedit.py

def encode_prompt(
    self,
    prompt,
    num_images_per_prompt,
    do_classifier_free_guidance,
    negative_prompt=None,
    prompt_embeds: Optional[ms.Tensor] = None,
    negative_prompt_embeds: Optional[ms.Tensor] = None,
    lora_scale: Optional[float] = None,
    clip_skip: Optional[int] = None,
):
    r"""
    Encodes the prompt into text encoder hidden states.

    Args:
        prompt (`str` or `List[str]`, *optional*):
            prompt to be encoded
        num_images_per_prompt (`int`):
            number of images that should be generated per prompt
        do_classifier_free_guidance (`bool`):
            whether to use classifier free guidance or not
        negative_prompt (`str` or `List[str]`, *optional*):
            The prompt or prompts not to guide the image generation. If not defined, one has to pass
            `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is
            less than `1`).
        prompt_embeds (`ms.Tensor`, *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.
        negative_prompt_embeds (`ms.Tensor`, *optional*):
            Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
            weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
            argument.
        lora_scale (`float`, *optional*):
            A LoRA scale that will be applied to all LoRA layers of the text encoder if LoRA layers are loaded.
        clip_skip (`int`, *optional*):
            Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that
            the output of the pre-final layer will be used for computing the prompt embeddings.
    """
    # set lora scale so that monkey patched LoRA
    # function of text encoder can correctly access it
    if lora_scale is not None and isinstance(self, LoraLoaderMixin):
        self._lora_scale = lora_scale

        scale_lora_layers(self.text_encoder, lora_scale)

    if prompt is not None and isinstance(prompt, str):
        batch_size = 1
    elif prompt is not None and isinstance(prompt, list):
        batch_size = len(prompt)
    else:
        batch_size = prompt_embeds.shape[0]

    if prompt_embeds is None:
        # textual inversion: process multi-vector tokens if necessary
        if isinstance(self, TextualInversionLoaderMixin):
            prompt = self.maybe_convert_prompt(prompt, self.tokenizer)

        text_inputs = self.tokenizer(
            prompt,
            padding="max_length",
            max_length=self.tokenizer.model_max_length,
            truncation=True,
            return_tensors="np",
        )
        text_input_ids = text_inputs.input_ids
        untruncated_ids = self.tokenizer(prompt, padding="longest", return_tensors="np").input_ids

        if untruncated_ids.shape[-1] >= text_input_ids.shape[-1] and not np.array_equal(
            text_input_ids, untruncated_ids
        ):
            removed_text = self.tokenizer.batch_decode(untruncated_ids[:, self.tokenizer.model_max_length - 1 : -1])
            logger.warning(
                "The following part of your input was truncated because CLIP can only handle sequences up to"
                f" {self.tokenizer.model_max_length} tokens: {removed_text}"
            )

        if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
            attention_mask = ms.tensor(text_inputs.attention_mask)
        else:
            attention_mask = None

        if clip_skip is None:
            prompt_embeds = self.text_encoder(ms.tensor(text_input_ids), attention_mask=attention_mask)
            prompt_embeds = prompt_embeds[0]
        else:
            prompt_embeds = self.text_encoder(
                ms.tensor(text_input_ids), attention_mask=attention_mask, output_hidden_states=True
            )
            # Access the `hidden_states` first, that contains a tuple of
            # all the hidden states from the encoder layers. Then index into
            # the tuple to access the hidden states from the desired layer.
            prompt_embeds = prompt_embeds[-1][-(clip_skip + 1)]
            # We also need to apply the final LayerNorm here to not mess with the
            # representations. The `last_hidden_states` that we typically use for
            # obtaining the final prompt representations passes through the LayerNorm
            # layer.
            prompt_embeds = self.text_encoder.text_model.final_layer_norm(prompt_embeds)

    if self.text_encoder is not None:
        prompt_embeds_dtype = self.text_encoder.dtype
    elif self.unet is not None:
        prompt_embeds_dtype = self.unet.dtype
    else:
        prompt_embeds_dtype = prompt_embeds.dtype

    prompt_embeds = prompt_embeds.to(dtype=prompt_embeds_dtype)

    bs_embed, seq_len, _ = prompt_embeds.shape
    # duplicate text embeddings for each generation per prompt, using mps friendly method
    prompt_embeds = prompt_embeds.tile((1, num_images_per_prompt, 1))
    prompt_embeds = prompt_embeds.view(bs_embed * num_images_per_prompt, seq_len, -1)

    # get unconditional embeddings for classifier free guidance
    if do_classifier_free_guidance and negative_prompt_embeds is None:
        uncond_tokens: List[str]
        if negative_prompt is None:
            uncond_tokens = [""] * batch_size
        elif prompt is not None and type(prompt) is not type(negative_prompt):
            raise TypeError(
                f"`negative_prompt` should be the same type to `prompt`, but got {type(negative_prompt)} !="
                f" {type(prompt)}."
            )
        elif isinstance(negative_prompt, str):
            uncond_tokens = [negative_prompt]
        elif batch_size != len(negative_prompt):
            raise ValueError(
                f"`negative_prompt`: {negative_prompt} has batch size {len(negative_prompt)}, but `prompt`:"
                f" {prompt} has batch size {batch_size}. Please make sure that passed `negative_prompt` matches"
                " the batch size of `prompt`."
            )
        else:
            uncond_tokens = negative_prompt

        # textual inversion: process multi-vector tokens if necessary
        if isinstance(self, TextualInversionLoaderMixin):
            uncond_tokens = self.maybe_convert_prompt(uncond_tokens, self.tokenizer)

        max_length = prompt_embeds.shape[1]
        uncond_input = self.tokenizer(
            uncond_tokens,
            padding="max_length",
            max_length=max_length,
            truncation=True,
            return_tensors="np",
        )

        if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
            attention_mask = ms.Tensor(uncond_input.attention_mask)
        else:
            attention_mask = None

        negative_prompt_embeds = self.text_encoder(
            ms.Tensor(uncond_input.input_ids),
            attention_mask=attention_mask,
        )
        negative_prompt_embeds = negative_prompt_embeds[0]

    if do_classifier_free_guidance:
        # duplicate unconditional embeddings for each generation per prompt, using mps friendly method
        seq_len = negative_prompt_embeds.shape[1]

        negative_prompt_embeds = negative_prompt_embeds.to(dtype=prompt_embeds_dtype)

        negative_prompt_embeds = negative_prompt_embeds.tile((1, num_images_per_prompt, 1))
        negative_prompt_embeds = negative_prompt_embeds.view(batch_size * num_images_per_prompt, seq_len, -1)

    if self.text_encoder is not None:
        if isinstance(self, LoraLoaderMixin):
            # Retrieve the original scale by scaling back the LoRA layers
            unscale_lora_layers(self.text_encoder, lora_scale)

    return prompt_embeds, negative_prompt_embeds

mindone.diffusers.StableDiffusionDiffEditPipeline.generate_mask(image=None, target_prompt=None, target_negative_prompt=None, target_prompt_embeds=None, target_negative_prompt_embeds=None, source_prompt=None, source_negative_prompt=None, source_prompt_embeds=None, source_negative_prompt_embeds=None, num_maps_per_mask=10, mask_encode_strength=0.5, mask_thresholding_ratio=3.0, num_inference_steps=50, guidance_scale=7.5, generator=None, output_type='np', cross_attention_kwargs=None)

Generate a latent mask given a mask prompt, a target prompt, and an image.

PARAMETER	DESCRIPTION
image	Image or tensor representing an image batch to be used for computing the mask. TYPE: `PIL.Image.Image` DEFAULT: None
target_prompt	The prompt or prompts to guide semantic mask generation. If not defined, you need to pass prompt_embeds. TYPE: `str` or `List[str]`, *optional* DEFAULT: None
target_negative_prompt	The prompt or prompts to guide what to not include in image generation. If not defined, you need to pass negative_prompt_embeds instead. Ignored when not using guidance (guidance_scale < 1). TYPE: `str` or `List[str]`, *optional* DEFAULT: None
target_prompt_embeds	Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not provided, text embeddings are generated from the prompt input argument. TYPE: `ms.Tensor`, *optional* DEFAULT: None
target_negative_prompt_embeds	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. TYPE: `ms.Tensor`, *optional* DEFAULT: None
source_prompt	The prompt or prompts to guide semantic mask generation using DiffEdit. If not defined, you need to pass source_prompt_embeds or source_image instead. TYPE: `str` or `List[str]`, *optional* DEFAULT: None
source_negative_prompt	The prompt or prompts to guide semantic mask generation away from using DiffEdit. If not defined, you need to pass source_negative_prompt_embeds or source_image instead. TYPE: `str` or `List[str]`, *optional* DEFAULT: None
source_prompt_embeds	Pre-generated text embeddings to guide the semantic mask generation. Can be used to easily tweak text inputs (prompt weighting). If not provided, text embeddings are generated from source_prompt input argument. TYPE: `ms.Tensor`, *optional* DEFAULT: None
source_negative_prompt_embeds	Pre-generated text embeddings to negatively guide the semantic mask generation. Can be used to easily tweak text inputs (prompt weighting). If not provided, text embeddings are generated from source_negative_prompt input argument. TYPE: `ms.Tensor`, *optional* DEFAULT: None
num_maps_per_mask	The number of noise maps sampled to generate the semantic mask using DiffEdit. TYPE: `int`, *optional*, defaults to 10 DEFAULT: 10
mask_encode_strength	The strength of the noise maps sampled to generate the semantic mask using DiffEdit. Must be between 0 and 1. TYPE: `float`, *optional*, defaults to 0.5 DEFAULT: 0.5
mask_thresholding_ratio	The maximum multiple of the mean absolute difference used to clamp the semantic guidance map before mask binarization. TYPE: `float`, *optional*, defaults to 3.0 DEFAULT: 3.0
num_inference_steps	The number of denoising steps. More denoising steps usually lead to a higher quality image at the expense of slower inference. TYPE: `int`, *optional*, defaults to 50 DEFAULT: 50
guidance_scale	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. TYPE: `float`, *optional*, defaults to 7.5 DEFAULT: 7.5
generator	A np.random.Generator to make generation deterministic. TYPE: `np.random.Generator` or `List[np.random.Generator]`, *optional* DEFAULT: None
output_type	The output format of the generated image. Choose between PIL.Image or np.array. TYPE: `str`, *optional*, defaults to `"pil"` DEFAULT: 'np'
cross_attention_kwargs	A kwargs dictionary that if specified is passed along to the [~models.attention_processor.AttnProcessor] as defined in self.processor. TYPE: `dict`, *optional* DEFAULT: None

RETURNS	DESCRIPTION
	List[PIL.Image.Image] or np.array: When returning a List[PIL.Image.Image], the list consists of a batch of single-channel binary images with dimensions (height // self.vae_scale_factor, width // self.vae_scale_factor). If it's np.array, the shape is (batch_size, height // self.vae_scale_factor, width // self.vae_scale_factor).

Source code in mindone/diffusers/pipelines/stable_diffusion_diffedit/pipeline_stable_diffusion_diffedit.py

def generate_mask(
    self,
    image: Union[ms.Tensor, PIL.Image.Image] = None,
    target_prompt: Optional[Union[str, List[str]]] = None,
    target_negative_prompt: Optional[Union[str, List[str]]] = None,
    target_prompt_embeds: Optional[ms.Tensor] = None,
    target_negative_prompt_embeds: Optional[ms.Tensor] = None,
    source_prompt: Optional[Union[str, List[str]]] = None,
    source_negative_prompt: Optional[Union[str, List[str]]] = None,
    source_prompt_embeds: Optional[ms.Tensor] = None,
    source_negative_prompt_embeds: Optional[ms.Tensor] = None,
    num_maps_per_mask: Optional[int] = 10,
    mask_encode_strength: Optional[float] = 0.5,
    mask_thresholding_ratio: Optional[float] = 3.0,
    num_inference_steps: int = 50,
    guidance_scale: float = 7.5,
    generator: Optional[Union[np.random.Generator, List[np.random.Generator]]] = None,
    output_type: Optional[str] = "np",
    cross_attention_kwargs: Optional[Dict[str, Any]] = None,
):
    r"""
    Generate a latent mask given a mask prompt, a target prompt, and an image.

    Args:
        image (`PIL.Image.Image`):
            `Image` or tensor representing an image batch to be used for computing the mask.
        target_prompt (`str` or `List[str]`, *optional*):
            The prompt or prompts to guide semantic mask generation. If not defined, you need to pass
            `prompt_embeds`.
        target_negative_prompt (`str` or `List[str]`, *optional*):
            The prompt or prompts to guide what to not include in image generation. If not defined, you need to
            pass `negative_prompt_embeds` instead. Ignored when not using guidance (`guidance_scale < 1`).
        target_prompt_embeds (`ms.Tensor`, *optional*):
            Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
            provided, text embeddings are generated from the `prompt` input argument.
        target_negative_prompt_embeds (`ms.Tensor`, *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.
        source_prompt (`str` or `List[str]`, *optional*):
            The prompt or prompts to guide semantic mask generation using DiffEdit. If not defined, you need to
            pass `source_prompt_embeds` or `source_image` instead.
        source_negative_prompt (`str` or `List[str]`, *optional*):
            The prompt or prompts to guide semantic mask generation away from using DiffEdit. If not defined, you
            need to pass `source_negative_prompt_embeds` or `source_image` instead.
        source_prompt_embeds (`ms.Tensor`, *optional*):
            Pre-generated text embeddings to guide the semantic mask generation. Can be used to easily tweak text
            inputs (prompt weighting). If not provided, text embeddings are generated from `source_prompt` input
            argument.
        source_negative_prompt_embeds (`ms.Tensor`, *optional*):
            Pre-generated text embeddings to negatively guide the semantic mask generation. Can be used to easily
            tweak text inputs (prompt weighting). If not provided, text embeddings are generated from
            `source_negative_prompt` input argument.
        num_maps_per_mask (`int`, *optional*, defaults to 10):
            The number of noise maps sampled to generate the semantic mask using DiffEdit.
        mask_encode_strength (`float`, *optional*, defaults to 0.5):
            The strength of the noise maps sampled to generate the semantic mask using DiffEdit. Must be between 0
            and 1.
        mask_thresholding_ratio (`float`, *optional*, defaults to 3.0):
            The maximum multiple of the mean absolute difference used to clamp the semantic guidance map before
            mask binarization.
        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.
        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`.
        generator (`np.random.Generator` or `List[np.random.Generator]`, *optional*):
            A [`np.random.Generator`](https://numpy.org/doc/stable/reference/random/generator.html) to make
            generation deterministic.
        output_type (`str`, *optional*, defaults to `"pil"`):
            The output format of the generated image. Choose between `PIL.Image` or `np.array`.
        cross_attention_kwargs (`dict`, *optional*):
            A kwargs dictionary that if specified is passed along to the
            [`~models.attention_processor.AttnProcessor`] as defined in
            [`self.processor`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).

    Examples:

    Returns:
        `List[PIL.Image.Image]` or `np.array`:
            When returning a `List[PIL.Image.Image]`, the list consists of a batch of single-channel binary images
            with dimensions `(height // self.vae_scale_factor, width // self.vae_scale_factor)`. If it's
            `np.array`, the shape is `(batch_size, height // self.vae_scale_factor, width //
            self.vae_scale_factor)`.
    """

    # 1. Check inputs (Provide dummy argument for callback_steps)
    self.check_inputs(
        target_prompt,
        mask_encode_strength,
        1,
        target_negative_prompt,
        target_prompt_embeds,
        target_negative_prompt_embeds,
    )

    self.check_source_inputs(
        source_prompt,
        source_negative_prompt,
        source_prompt_embeds,
        source_negative_prompt_embeds,
    )

    if (num_maps_per_mask is None) or (
        num_maps_per_mask is not None and (not isinstance(num_maps_per_mask, int) or num_maps_per_mask <= 0)
    ):
        raise ValueError(
            f"`num_maps_per_mask` has to be a positive integer but is {num_maps_per_mask} of type"
            f" {type(num_maps_per_mask)}."
        )

    if mask_thresholding_ratio is None or mask_thresholding_ratio <= 0:
        raise ValueError(
            f"`mask_thresholding_ratio` has to be positive but is {mask_thresholding_ratio} of type"
            f" {type(mask_thresholding_ratio)}."
        )

    # 2. Define call parameters
    if target_prompt is not None and isinstance(target_prompt, str):
        batch_size = 1
    elif target_prompt is not None and isinstance(target_prompt, list):
        batch_size = len(target_prompt)
    else:
        batch_size = target_prompt_embeds.shape[0]
    if cross_attention_kwargs is None:
        cross_attention_kwargs = {}

    # here `guidance_scale` is defined analog to the guidance weight `w` of equation (2)
    # of the Imagen paper: https://arxiv.org/pdf/2205.11487.pdf . `guidance_scale = 1`
    # corresponds to doing no classifier free guidance.
    do_classifier_free_guidance = guidance_scale > 1.0

    # 3. Encode input prompts
    (cross_attention_kwargs.get("scale", None) if cross_attention_kwargs is not None else None)
    target_negative_prompt_embeds, target_prompt_embeds = self.encode_prompt(
        target_prompt,
        num_maps_per_mask,
        do_classifier_free_guidance,
        target_negative_prompt,
        prompt_embeds=target_prompt_embeds,
        negative_prompt_embeds=target_negative_prompt_embeds,
    )
    # For classifier free guidance, we need to do two forward passes.
    # Here we concatenate the unconditional and text embeddings into a single batch
    # to avoid doing two forward passes
    if do_classifier_free_guidance:
        target_prompt_embeds = ops.cat([target_negative_prompt_embeds, target_prompt_embeds])

    source_negative_prompt_embeds, source_prompt_embeds = self.encode_prompt(
        source_prompt,
        num_maps_per_mask,
        do_classifier_free_guidance,
        source_negative_prompt,
        prompt_embeds=source_prompt_embeds,
        negative_prompt_embeds=source_negative_prompt_embeds,
    )
    if do_classifier_free_guidance:
        source_prompt_embeds = ops.cat([source_negative_prompt_embeds, source_prompt_embeds])

    # 4. Preprocess image
    image = self.image_processor.preprocess(image).repeat_interleave(num_maps_per_mask, dim=0)

    # 5. Set timesteps
    self.scheduler.set_timesteps(num_inference_steps)
    timesteps, _ = self.get_timesteps(num_inference_steps, mask_encode_strength)
    encode_timestep = timesteps[0]

    # 6. Prepare image latents and add noise with specified strength
    image_latents = self.prepare_image_latents(image, batch_size * num_maps_per_mask, self.vae.dtype, generator)
    noise = randn_tensor(image_latents.shape, generator=generator, dtype=self.vae.dtype)
    image_latents = self.scheduler.add_noise(image_latents, noise, encode_timestep)

    latent_model_input = ops.cat([image_latents] * (4 if do_classifier_free_guidance else 2))
    latent_model_input = self.scheduler.scale_model_input(latent_model_input, encode_timestep)

    # 7. Predict the noise residual
    prompt_embeds = ops.cat([source_prompt_embeds, target_prompt_embeds])
    noise_pred = self.unet(
        latent_model_input,
        encode_timestep,
        encoder_hidden_states=prompt_embeds,
        cross_attention_kwargs=cross_attention_kwargs,
    )[0]

    if do_classifier_free_guidance:
        noise_pred_neg_src, noise_pred_source, noise_pred_uncond, noise_pred_target = noise_pred.chunk(4)
        noise_pred_source = noise_pred_neg_src + guidance_scale * (noise_pred_source - noise_pred_neg_src)
        noise_pred_target = noise_pred_uncond + guidance_scale * (noise_pred_target - noise_pred_uncond)
    else:
        noise_pred_source, noise_pred_target = noise_pred.chunk(2)

    # 8. Compute the mask from the absolute difference of predicted noise residuals
    # TODO: Consider smoothing mask guidance map
    mask_guidance_map = (
        ops.abs(noise_pred_target - noise_pred_source)
        .reshape(batch_size, num_maps_per_mask, *noise_pred_target.shape[-3:])
        .mean([1, 2])
    )
    clamp_magnitude = mask_guidance_map.mean() * mask_thresholding_ratio
    semantic_mask_image = mask_guidance_map.clamp(ms.Tensor(0), clamp_magnitude) / clamp_magnitude
    semantic_mask_image = ops.where(semantic_mask_image <= 0.5, ms.Tensor(0), ms.Tensor(1))
    mask_image = semantic_mask_image.asnumpy()

    # 9. Convert to Numpy array or PIL.
    if output_type == "pil":
        mask_image = self.image_processor.numpy_to_pil(mask_image)

    return mask_image

mindone.diffusers.StableDiffusionDiffEditPipeline.invert(prompt=None, image=None, num_inference_steps=50, inpaint_strength=0.8, guidance_scale=7.5, negative_prompt=None, generator=None, prompt_embeds=None, negative_prompt_embeds=None, decode_latents=False, output_type='pil', return_dict=False, callback=None, callback_steps=1, cross_attention_kwargs=None, lambda_auto_corr=20.0, lambda_kl=20.0, num_reg_steps=0, num_auto_corr_rolls=5)

Generate inverted latents given a prompt and image.

PARAMETER	DESCRIPTION
prompt	The prompt or prompts to guide image generation. If not defined, you need to pass prompt_embeds. TYPE: `str` or `List[str]`, *optional* DEFAULT: None
image	Image or tensor representing an image batch to produce the inverted latents guided by prompt. TYPE: `PIL.Image.Image` DEFAULT: None
inpaint_strength	Indicates extent of the noising process to run latent inversion. Must be between 0 and 1. When inpaint_strength is 1, the inversion process is run for the full number of iterations specified in num_inference_steps. image is used as a reference for the inversion process, and adding more noise increases inpaint_strength. If inpaint_strength is 0, no inpainting occurs. TYPE: `float`, *optional*, defaults to 0.8 DEFAULT: 0.8
num_inference_steps	The number of denoising steps. More denoising steps usually lead to a higher quality image at the expense of slower inference. TYPE: `int`, *optional*, defaults to 50 DEFAULT: 50
guidance_scale	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. TYPE: `float`, *optional*, defaults to 7.5 DEFAULT: 7.5
negative_prompt	The prompt or prompts to guide what to not include in image generation. If not defined, you need to pass negative_prompt_embeds instead. Ignored when not using guidance (guidance_scale < 1). TYPE: `str` or `List[str]`, *optional* DEFAULT: None
generator	A np.random.Generator to make generation deterministic. TYPE: `np.random.Generator`, *optional* DEFAULT: None
prompt_embeds	Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not provided, text embeddings are generated from the prompt input argument. TYPE: `ms.Tensor`, *optional* DEFAULT: None
negative_prompt_embeds	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. TYPE: `ms.Tensor`, *optional* DEFAULT: None
decode_latents	Whether or not to decode the inverted latents into a generated image. Setting this argument to True decodes all inverted latents for each timestep into a list of generated images. TYPE: `bool`, *optional*, defaults to `False` DEFAULT: False
output_type	The output format of the generated image. Choose between PIL.Image or np.array. TYPE: `str`, *optional*, defaults to `"pil"` DEFAULT: 'pil'
return_dict	Whether or not to return a [~pipelines.stable_diffusion.DiffEditInversionPipelineOutput] instead of a plain tuple. TYPE: `bool`, *optional*, defaults to `True` DEFAULT: False
callback	A function that calls every callback_steps steps during inference. The function is called with the following arguments: callback(step: int, timestep: int, latents: ms.Tensor). TYPE: `Callable`, *optional* DEFAULT: None
callback_steps	The frequency at which the callback function is called. If not specified, the callback is called at every step. TYPE: `int`, *optional*, defaults to 1 DEFAULT: 1
cross_attention_kwargs	A kwargs dictionary that if specified is passed along to the [~models.attention_processor.AttnProcessor] as defined in self.processor. TYPE: `dict`, *optional* DEFAULT: None
lambda_auto_corr	Lambda parameter to control auto correction. TYPE: `float`, *optional*, defaults to 20.0 DEFAULT: 20.0
lambda_kl	Lambda parameter to control Kullback-Leibler divergence output. TYPE: `float`, *optional*, defaults to 20.0 DEFAULT: 20.0
num_reg_steps	Number of regularization loss steps. TYPE: `int`, *optional*, defaults to 0 DEFAULT: 0
num_auto_corr_rolls	Number of auto correction roll steps. TYPE: `int`, *optional*, defaults to 5 DEFAULT: 5

RETURNS	DESCRIPTION
	[~pipelines.stable_diffusion.pipeline_stable_diffusion_diffedit.DiffEditInversionPipelineOutput] or
	tuple: If return_dict is True, [~pipelines.stable_diffusion.pipeline_stable_diffusion_diffedit.DiffEditInversionPipelineOutput] is returned, otherwise a tuple is returned where the first element is the inverted latents tensors ordered by increasing noise, and the second is the corresponding decoded images if decode_latents is True, otherwise None.

Source code in mindone/diffusers/pipelines/stable_diffusion_diffedit/pipeline_stable_diffusion_diffedit.py

def invert(
    self,
    prompt: Optional[Union[str, List[str]]] = None,
    image: Union[ms.Tensor, PIL.Image.Image] = None,
    num_inference_steps: int = 50,
    inpaint_strength: float = 0.8,
    guidance_scale: float = 7.5,
    negative_prompt: Optional[Union[str, List[str]]] = None,
    generator: Optional[Union[np.random.Generator, List[np.random.Generator]]] = None,
    prompt_embeds: Optional[ms.Tensor] = None,
    negative_prompt_embeds: Optional[ms.Tensor] = None,
    decode_latents: bool = False,
    output_type: Optional[str] = "pil",
    return_dict: bool = False,
    callback: Optional[Callable[[int, int, ms.Tensor], None]] = None,
    callback_steps: Optional[int] = 1,
    cross_attention_kwargs: Optional[Dict[str, Any]] = None,
    lambda_auto_corr: float = 20.0,
    lambda_kl: float = 20.0,
    num_reg_steps: int = 0,
    num_auto_corr_rolls: int = 5,
):
    r"""
    Generate inverted latents given a prompt and image.

    Args:
        prompt (`str` or `List[str]`, *optional*):
            The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`.
        image (`PIL.Image.Image`):
            `Image` or tensor representing an image batch to produce the inverted latents guided by `prompt`.
        inpaint_strength (`float`, *optional*, defaults to 0.8):
            Indicates extent of the noising process to run latent inversion. Must be between 0 and 1. When
            `inpaint_strength` is 1, the inversion process is run for the full number of iterations specified in
            `num_inference_steps`. `image` is used as a reference for the inversion process, and adding more noise
            increases `inpaint_strength`. If `inpaint_strength` is 0, no inpainting occurs.
        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.
        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`.
        negative_prompt (`str` or `List[str]`, *optional*):
            The prompt or prompts to guide what to not include in image generation. If not defined, you need to
            pass `negative_prompt_embeds` instead. Ignored when not using guidance (`guidance_scale < 1`).
        generator (`np.random.Generator`, *optional*):
            A [`np.random.Generator`](https://numpy.org/doc/stable/reference/random/generator.html) to make
            generation deterministic.
        prompt_embeds (`ms.Tensor`, *optional*):
            Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
            provided, text embeddings are generated from the `prompt` input argument.
        negative_prompt_embeds (`ms.Tensor`, *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.
        decode_latents (`bool`, *optional*, defaults to `False`):
            Whether or not to decode the inverted latents into a generated image. Setting this argument to `True`
            decodes all inverted latents for each timestep into a list of generated images.
        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`):
            Whether or not to return a [`~pipelines.stable_diffusion.DiffEditInversionPipelineOutput`] instead of a
            plain tuple.
        callback (`Callable`, *optional*):
            A function that calls every `callback_steps` steps during inference. The function is called with the
            following arguments: `callback(step: int, timestep: int, latents: ms.Tensor)`.
        callback_steps (`int`, *optional*, defaults to 1):
            The frequency at which the `callback` function is called. If not specified, the callback is called at
            every step.
        cross_attention_kwargs (`dict`, *optional*):
            A kwargs dictionary that if specified is passed along to the
            [`~models.attention_processor.AttnProcessor`] as defined in
            [`self.processor`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).
        lambda_auto_corr (`float`, *optional*, defaults to 20.0):
            Lambda parameter to control auto correction.
        lambda_kl (`float`, *optional*, defaults to 20.0):
            Lambda parameter to control Kullback-Leibler divergence output.
        num_reg_steps (`int`, *optional*, defaults to 0):
            Number of regularization loss steps.
        num_auto_corr_rolls (`int`, *optional*, defaults to 5):
            Number of auto correction roll steps.

    Examples:

    Returns:
        [`~pipelines.stable_diffusion.pipeline_stable_diffusion_diffedit.DiffEditInversionPipelineOutput`] or
        `tuple`:
            If `return_dict` is `True`,
            [`~pipelines.stable_diffusion.pipeline_stable_diffusion_diffedit.DiffEditInversionPipelineOutput`] is
            returned, otherwise a `tuple` is returned where the first element is the inverted latents tensors
            ordered by increasing noise, and the second is the corresponding decoded images if `decode_latents` is
            `True`, otherwise `None`.
    """

    # 1. Check inputs
    self.check_inputs(
        prompt,
        inpaint_strength,
        callback_steps,
        negative_prompt,
        prompt_embeds,
        negative_prompt_embeds,
    )

    if image is None:
        raise ValueError("`image` input cannot be undefined.")

    # 2. Define call parameters
    if prompt is not None and isinstance(prompt, str):
        batch_size = 1
    elif prompt is not None and isinstance(prompt, list):
        batch_size = len(prompt)
    else:
        batch_size = prompt_embeds.shape[0]
    if cross_attention_kwargs is None:
        cross_attention_kwargs = {}

    # here `guidance_scale` is defined analog to the guidance weight `w` of equation (2)
    # of the Imagen paper: https://arxiv.org/pdf/2205.11487.pdf . `guidance_scale = 1`
    # corresponds to doing no classifier free guidance.
    do_classifier_free_guidance = guidance_scale > 1.0

    # 3. Preprocess image
    image = self.image_processor.preprocess(image)

    # 4. Prepare latent variables
    num_images_per_prompt = 1
    latents = self.prepare_image_latents(image, batch_size * num_images_per_prompt, self.vae.dtype, generator)

    # 5. Encode input prompt
    prompt_embeds, negative_prompt_embeds = self.encode_prompt(
        prompt,
        num_images_per_prompt,
        do_classifier_free_guidance,
        negative_prompt,
        prompt_embeds=prompt_embeds,
        negative_prompt_embeds=negative_prompt_embeds,
    )
    # For classifier free guidance, we need to do two forward passes.
    # Here we concatenate the unconditional and text embeddings into a single batch
    # to avoid doing two forward passes
    if do_classifier_free_guidance:
        prompt_embeds = ops.cat([negative_prompt_embeds, prompt_embeds])

    # 6. Prepare timesteps
    self.inverse_scheduler.set_timesteps(num_inference_steps)
    timesteps, num_inference_steps = self.get_inverse_timesteps(num_inference_steps, inpaint_strength)

    # 7. Noising loop where we obtain the intermediate noised latent image for each timestep.
    num_warmup_steps = len(timesteps) - num_inference_steps * self.inverse_scheduler.order
    inverted_latents = []
    with self.progress_bar(total=num_inference_steps) as progress_bar:
        for i, t in enumerate(timesteps):
            # expand the latents if we are doing classifier free guidance
            latent_model_input = ops.cat([latents] * 2) if do_classifier_free_guidance else latents
            latent_model_input = self.inverse_scheduler.scale_model_input(latent_model_input, t)

            # predict the noise residual
            noise_pred = self.unet(
                latent_model_input,
                t,
                encoder_hidden_states=prompt_embeds,
                cross_attention_kwargs=cross_attention_kwargs,
            )[0]

            # perform guidance
            if do_classifier_free_guidance:
                noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
                noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)

            # regularization of the noise prediction (not in original code or paper but borrowed from Pix2PixZero)
            if num_reg_steps > 0:
                raise NotImplementedError("Regularization of the noise prediction is not yet supported.")

            # compute the previous noisy sample x_t -> x_t-1
            latents = self.inverse_scheduler.step(noise_pred, t, latents)[0]
            inverted_latents.append(latents)

            # call the callback, if provided
            if i == len(timesteps) - 1 or (
                (i + 1) > num_warmup_steps and (i + 1) % self.inverse_scheduler.order == 0
            ):
                progress_bar.update()
                if callback is not None and i % callback_steps == 0:
                    step_idx = i // getattr(self.scheduler, "order", 1)
                    callback(step_idx, t, latents)

    assert len(inverted_latents) == len(timesteps)
    latents = ops.stack(list(reversed(inverted_latents)), 1)

    # 8. Post-processing
    image = None
    if decode_latents:
        image = self.decode_latents(latents.flatten(start_dim=0, end_dim=1))

    # 9. Convert to PIL.
    if decode_latents and output_type == "pil":
        image = self.image_processor.numpy_to_pil(image)

    if not return_dict:
        return (latents, image)

    return DiffEditInversionPipelineOutput(latents=latents, images=image)

mindone.diffusers.pipelines.stable_diffusion.StableDiffusionPipelineOutput dataclass

Bases: BaseOutput

Output class for Stable Diffusion pipelines.

Source code in mindone/diffusers/pipelines/stable_diffusion/pipeline_output.py

@dataclass
class StableDiffusionPipelineOutput(BaseOutput):
    """
    Output class for Stable Diffusion pipelines.

    Args:
        images (`List[PIL.Image.Image]` or `np.ndarray`)
            List of denoised PIL images of length `batch_size` or NumPy array of shape `(batch_size, height, width,
            num_channels)`.
        nsfw_content_detected (`List[bool]`)
            List indicating whether the corresponding generated image contains "not-safe-for-work" (nsfw) content or
            `None` if safety checking could not be performed.
    """

    images: Union[List[PIL.Image.Image], np.ndarray]
    nsfw_content_detected: Optional[List[bool]]