x2rgb/pipeline_x2rgb.py

import inspect
from dataclasses import dataclass
from typing import Callable, List, Optional, Union

import numpy as np
import PIL
import torch
import torch.nn.functional as F
from diffusers.configuration_utils import register_to_config
from diffusers.image_processor import VaeImageProcessor
from diffusers.loaders import LoraLoaderMixin, TextualInversionLoaderMixin
from diffusers.models import AutoencoderKL, UNet2DConditionModel
from diffusers.pipelines.pipeline_utils import DiffusionPipeline
from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion import (
    rescale_noise_cfg,
)
from diffusers.schedulers import KarrasDiffusionSchedulers
from diffusers.utils import CONFIG_NAME, BaseOutput, deprecate, logging, randn_tensor
from transformers import CLIPTextModel, CLIPTokenizer

logger = logging.get_logger(__name__)


class VaeImageProcrssorAOV(VaeImageProcessor):
    """
    Image processor for VAE AOV.

    Args:
        do_resize (`bool`, *optional*, defaults to `True`):
            Whether to downscale the image's (height, width) dimensions to multiples of `vae_scale_factor`.
        vae_scale_factor (`int`, *optional*, defaults to `8`):
            VAE scale factor. If `do_resize` is `True`, the image is automatically resized to multiples of this factor.
        resample (`str`, *optional*, defaults to `lanczos`):
            Resampling filter to use when resizing the image.
        do_normalize (`bool`, *optional*, defaults to `True`):
            Whether to normalize the image to [-1,1].
    """

    config_name = CONFIG_NAME

    @register_to_config
    def __init__(
        self,
        do_resize: bool = True,
        vae_scale_factor: int = 8,
        resample: str = "lanczos",
        do_normalize: bool = True,
    ):
        super().__init__()

    def postprocess(
        self,
        image: torch.FloatTensor,
        output_type: str = "pil",
        do_denormalize: Optional[List[bool]] = None,
        do_gamma_correction: bool = True,
    ):
        if not isinstance(image, torch.Tensor):
            raise ValueError(
                f"Input for postprocessing is in incorrect format: {type(image)}. We only support pytorch tensor"
            )
        if output_type not in ["latent", "pt", "np", "pil"]:
            deprecation_message = (
                f"the output_type {output_type} is outdated and has been set to `np`. Please make sure to set it to one of these instead: "
                "`pil`, `np`, `pt`, `latent`"
            )
            deprecate(
                "Unsupported output_type",
                "1.0.0",
                deprecation_message,
                standard_warn=False,
            )
            output_type = "np"

        if output_type == "latent":
            return image

        if do_denormalize is None:
            do_denormalize = [self.config.do_normalize] * image.shape[0]

        image = torch.stack(
            [
                self.denormalize(image[i]) if do_denormalize[i] else image[i]
                for i in range(image.shape[0])
            ]
        )

        # Gamma correction
        if do_gamma_correction:
            image = torch.pow(image, 1.0 / 2.2)

        if output_type == "pt":
            return image

        image = self.pt_to_numpy(image)

        if output_type == "np":
            return image

        if output_type == "pil":
            return self.numpy_to_pil(image)

    def preprocess_normal(
        self,
        image: Union[torch.FloatTensor, PIL.Image.Image, np.ndarray],
        height: Optional[int] = None,
        width: Optional[int] = None,
    ) -> torch.Tensor:
        image = torch.stack([image], axis=0)
        return image


@dataclass
class StableDiffusionAOVPipelineOutput(BaseOutput):
    """
    Output class for Stable Diffusion AOV 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]
    predicted_x0_images: Optional[Union[List[PIL.Image.Image], np.ndarray]] = None


class StableDiffusionAOVDropoutPipeline(
    DiffusionPipeline, TextualInversionLoaderMixin, LoraLoaderMixin
):
    r"""
    Pipeline for AOVs.

    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 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. Can be one of
            [`DDIMScheduler`], [`LMSDiscreteScheduler`], or [`PNDMScheduler`].
    """

    def __init__(
        self,
        vae: AutoencoderKL,
        text_encoder: CLIPTextModel,
        tokenizer: CLIPTokenizer,
        unet: UNet2DConditionModel,
        scheduler: KarrasDiffusionSchedulers,
    ):
        super().__init__()

        self.register_modules(
            vae=vae,
            text_encoder=text_encoder,
            tokenizer=tokenizer,
            unet=unet,
            scheduler=scheduler,
        )
        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
        self.image_processor = VaeImageProcrssorAOV(
            vae_scale_factor=self.vae_scale_factor
        )
        self.register_to_config()

    def _encode_prompt(
        self,
        prompt,
        device,
        num_images_per_prompt,
        do_classifier_free_guidance,
        negative_prompt=None,
        prompt_embeds: Optional[torch.FloatTensor] = None,
        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
    ):
        r"""
        Encodes the prompt into text encoder hidden states.

        Args:
             prompt (`str` or `List[str]`, *optional*):
                prompt to be encoded
            device: (`torch.device`):
                torch device
            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 (`torch.FloatTensor`, *optional*):
                Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                provided, text embeddings will be generated from `prompt` input argument.
            negative_prompt_embeds (`torch.FloatTensor`, *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.
        """
        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: procecss 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="pt",
            )
            text_input_ids = text_inputs.input_ids
            untruncated_ids = self.tokenizer(
                prompt, padding="longest", return_tensors="pt"
            ).input_ids

            if untruncated_ids.shape[-1] >= text_input_ids.shape[
                -1
            ] and not torch.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 = text_inputs.attention_mask.to(device)
            else:
                attention_mask = None

            prompt_embeds = self.text_encoder(
                text_input_ids.to(device),
                attention_mask=attention_mask,
            )
            prompt_embeds = prompt_embeds[0]

        prompt_embeds = prompt_embeds.to(dtype=self.text_encoder.dtype, device=device)

        bs_embed, seq_len, _ = prompt_embeds.shape
        # duplicate text embeddings for each generation per prompt, using mps friendly method
        prompt_embeds = prompt_embeds.repeat(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 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: procecss 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="pt",
            )

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

            negative_prompt_embeds = self.text_encoder(
                uncond_input.input_ids.to(device),
                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=self.text_encoder.dtype, device=device
            )

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

            # 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
            # pix2pix has two  negative embeddings, and unlike in other pipelines latents are ordered [prompt_embeds, negative_prompt_embeds, negative_prompt_embeds]
            prompt_embeds = torch.cat(
                [prompt_embeds, negative_prompt_embeds, negative_prompt_embeds]
            )

        return prompt_embeds

    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

    def check_inputs(
        self,
        prompt,
        callback_steps,
        negative_prompt=None,
        prompt_embeds=None,
        negative_prompt_embeds=None,
    ):
        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 prepare_latents(
        self,
        batch_size,
        num_channels_latents,
        height,
        width,
        dtype,
        device,
        generator,
        latents=None,
    ):
        shape = (
            batch_size,
            num_channels_latents,
            height // self.vae_scale_factor,
            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, device=device, dtype=dtype
            )
        else:
            latents = latents.to(device)

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

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

        image = image.to(device=device, dtype=dtype)

        batch_size = batch_size * num_images_per_prompt

        if image.shape[1] == 4:
            image_latents = image
        else:
            if isinstance(generator, list) and len(generator) != batch_size:
                raise ValueError(
                    f"You have passed a list of generators of length {len(generator)}, but requested an effective batch"
                    f" size of {batch_size}. Make sure the batch size matches the length of the generators."
                )

            if isinstance(generator, list):
                image_latents = [
                    self.vae.encode(image[i : i + 1]).latent_dist.mode()
                    for i in range(batch_size)
                ]
                image_latents = torch.cat(image_latents, dim=0)
            else:
                image_latents = self.vae.encode(image).latent_dist.mode()

        if (
            batch_size > image_latents.shape[0]
            and batch_size % image_latents.shape[0] == 0
        ):
            # expand image_latents for batch_size
            deprecation_message = (
                f"You have passed {batch_size} text prompts (`prompt`), but only {image_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_image_per_prompt = batch_size // image_latents.shape[0]
            image_latents = torch.cat(
                [image_latents] * additional_image_per_prompt, dim=0
            )
        elif (
            batch_size > image_latents.shape[0]
            and batch_size % image_latents.shape[0] != 0
        ):
            raise ValueError(
                f"Cannot duplicate `image` of batch size {image_latents.shape[0]} to {batch_size} text prompts."
            )
        else:
            image_latents = torch.cat([image_latents], dim=0)

        if do_classifier_free_guidance:
            uncond_image_latents = torch.zeros_like(image_latents)
            image_latents = torch.cat(
                [image_latents, image_latents, uncond_image_latents], dim=0
            )

        return image_latents

    @torch.no_grad()
    def __call__(
        self,
        height: int,
        width: int,
        prompt: Union[str, List[str]] = None,
        albedo: Optional[
            Union[
                torch.FloatTensor,
                PIL.Image.Image,
                np.ndarray,
                List[torch.FloatTensor],
                List[PIL.Image.Image],
                List[np.ndarray],
            ]
        ] = None,
        normal: Optional[
            Union[
                torch.FloatTensor,
                PIL.Image.Image,
                np.ndarray,
                List[torch.FloatTensor],
                List[PIL.Image.Image],
                List[np.ndarray],
            ]
        ] = None,
        roughness: Optional[
            Union[
                torch.FloatTensor,
                PIL.Image.Image,
                np.ndarray,
                List[torch.FloatTensor],
                List[PIL.Image.Image],
                List[np.ndarray],
            ]
        ] = None,
        metallic: Optional[
            Union[
                torch.FloatTensor,
                PIL.Image.Image,
                np.ndarray,
                List[torch.FloatTensor],
                List[PIL.Image.Image],
                List[np.ndarray],
            ]
        ] = None,
        irradiance: Optional[
            Union[
                torch.FloatTensor,
                PIL.Image.Image,
                np.ndarray,
                List[torch.FloatTensor],
                List[PIL.Image.Image],
                List[np.ndarray],
            ]
        ] = None,
        guidance_scale: float = 0.0,
        image_guidance_scale: float = 0.0,
        guidance_rescale: float = 0.0,
        num_inference_steps: int = 100,
        required_aovs: List[str] = ["albedo"],
        return_predicted_x0s: bool = False,
        negative_prompt: Optional[Union[str, List[str]]] = None,
        num_images_per_prompt: Optional[int] = 1,
        eta: float = 0.0,
        generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
        latents: Optional[torch.FloatTensor] = None,
        prompt_embeds: Optional[torch.FloatTensor] = None,
        negative_prompt_embeds: Optional[torch.FloatTensor] = None,
        output_type: Optional[str] = "pil",
        return_dict: bool = True,
        callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
        callback_steps: int = 1,
    ):
        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`.
            image (`torch.FloatTensor` `np.ndarray`, `PIL.Image.Image`, `List[torch.FloatTensor]`, `List[PIL.Image.Image]`, or `List[np.ndarray]`):
                `Image` or tensor representing an image batch to be repainted according to `prompt`. Can also accept
                image latents as `image`, but if passing latents directly it is not encoded again.
            num_inference_steps (`int`, *optional*, defaults to 100):
                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`.
            image_guidance_scale (`float`, *optional*, defaults to 1.5):
                Push the generated image towards the inital `image`. Image guidance scale is enabled by setting
                `image_guidance_scale > 1`. Higher image guidance scale encourages generated images that are closely
                linked to the source `image`, usually at the expense of lower image quality. This pipeline requires a
                value of at least `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 (`torch.Generator`, *optional*):
                A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
                generation deterministic.
            latents (`torch.FloatTensor`, *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 (`torch.FloatTensor`, *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 (`torch.FloatTensor`, *optional*):
                Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
                not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
            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: torch.FloatTensor)`.
            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.

        Examples:

        ```py
        >>> import PIL
        >>> import requests
        >>> import torch
        >>> from io import BytesIO

        >>> from diffusers import StableDiffusionInstructPix2PixPipeline


        >>> def download_image(url):
        ...     response = requests.get(url)
        ...     return PIL.Image.open(BytesIO(response.content)).convert("RGB")


        >>> img_url = "https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/mountain.png"

        >>> image = download_image(img_url).resize((512, 512))

        >>> pipe = StableDiffusionInstructPix2PixPipeline.from_pretrained(
        ...     "timbrooks/instruct-pix2pix", torch_dtype=torch.float16
        ... )
        >>> pipe = pipe.to("cuda")

        >>> prompt = "make the mountains snowy"
        >>> image = pipe(prompt=prompt, image=image).images[0]
        ```

        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.
        """
        # 0. Check inputs
        self.check_inputs(
            prompt,
            callback_steps,
            negative_prompt,
            prompt_embeds,
            negative_prompt_embeds,
        )

        # 1. 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]

        device = self._execution_device
        do_classifier_free_guidance = (
            guidance_scale >= 1.0 and image_guidance_scale >= 1.0
        )
        # check if scheduler is in sigmas space
        scheduler_is_in_sigma_space = hasattr(self.scheduler, "sigmas")

        # 2. Encode input prompt
        prompt_embeds = self._encode_prompt(
            prompt,
            device,
            num_images_per_prompt,
            do_classifier_free_guidance,
            negative_prompt,
            prompt_embeds=prompt_embeds,
            negative_prompt_embeds=negative_prompt_embeds,
        )

        # 3. Preprocess image
        # For normal, the preprocessing does nothing
        # For others, the preprocessing remap the values to [-1, 1]
        preprocessed_aovs = {}
        for aov_name in required_aovs:
            if aov_name == "albedo":
                if albedo is not None:
                    preprocessed_aovs[aov_name] = self.image_processor.preprocess(
                        albedo
                    )
                else:
                    preprocessed_aovs[aov_name] = None

            if aov_name == "normal":
                if normal is not None:
                    preprocessed_aovs[aov_name] = (
                        self.image_processor.preprocess_normal(normal)
                    )
                else:
                    preprocessed_aovs[aov_name] = None

            if aov_name == "roughness":
                if roughness is not None:
                    preprocessed_aovs[aov_name] = self.image_processor.preprocess(
                        roughness
                    )
                else:
                    preprocessed_aovs[aov_name] = None
            if aov_name == "metallic":
                if metallic is not None:
                    preprocessed_aovs[aov_name] = self.image_processor.preprocess(
                        metallic
                    )
                else:
                    preprocessed_aovs[aov_name] = None
            if aov_name == "irradiance":
                if irradiance is not None:
                    preprocessed_aovs[aov_name] = self.image_processor.preprocess(
                        irradiance
                    )
                else:
                    preprocessed_aovs[aov_name] = None

        # 4. set timesteps
        self.scheduler.set_timesteps(num_inference_steps, device=device)
        timesteps = self.scheduler.timesteps

        # 5. Prepare latent variables
        num_channels_latents = self.vae.config.latent_channels
        latents = self.prepare_latents(
            batch_size * num_images_per_prompt,
            num_channels_latents,
            height,
            width,
            prompt_embeds.dtype,
            device,
            generator,
            latents,
        )

        height_latent, width_latent = latents.shape[-2:]

        # 6. Prepare Image latents
        image_latents = []
        # Magicial scaling factors for each AOV (calculated from the training data)
        scaling_factors = {
            "albedo": 0.17301377137652138,
            "normal": 0.17483895473058078,
            "roughness": 0.1680724853626448,
            "metallic": 0.13135013390855135,
        }
        for aov_name, aov in preprocessed_aovs.items():
            if aov is None:
                image_latent = torch.zeros(
                    batch_size,
                    num_channels_latents,
                    height_latent,
                    width_latent,
                    dtype=prompt_embeds.dtype,
                    device=device,
                )
                if aov_name == "irradiance":
                    image_latent = image_latent[:, 0:3]
                if do_classifier_free_guidance:
                    image_latents.append(
                        torch.cat([image_latent, image_latent, image_latent], dim=0)
                    )
                else:
                    image_latents.append(image_latent)
            else:
                if aov_name == "irradiance":
                    image_latent = F.interpolate(
                        aov.to(device=device, dtype=prompt_embeds.dtype),
                        size=(height_latent, width_latent),
                        mode="bilinear",
                        align_corners=False,
                        antialias=True,
                    )
                    if do_classifier_free_guidance:
                        uncond_image_latent = torch.zeros_like(image_latent)
                        image_latent = torch.cat(
                            [image_latent, image_latent, uncond_image_latent], dim=0
                        )
                else:
                    scaling_factor = scaling_factors[aov_name]
                    image_latent = (
                        self.prepare_image_latents(
                            aov,
                            batch_size,
                            num_images_per_prompt,
                            prompt_embeds.dtype,
                            device,
                            do_classifier_free_guidance,
                            generator,
                        )
                        * scaling_factor
                    )
                image_latents.append(image_latent)
        image_latents = torch.cat(image_latents, dim=1)

        # 7. Check that shapes of latents and image match the UNet channels
        num_channels_image = image_latents.shape[1]
        if num_channels_latents + num_channels_image != self.unet.config.in_channels:
            raise ValueError(
                f"Incorrect configuration settings! The config of `pipeline.unet`: {self.unet.config} expects"
                f" {self.unet.config.in_channels} but received `num_channels_latents`: {num_channels_latents} +"
                f" `num_channels_image`: {num_channels_image} "
                f" = {num_channels_latents+num_channels_image}. Please verify the config of"
                " `pipeline.unet` or your `image` input."
            )

        # 8. 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)

        predicted_x0s = []

        # 9. Denoising loop
        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.
                # The latents are expanded 3 times because for pix2pix the guidance\
                # is applied for both the text and the input image.
                latent_model_input = (
                    torch.cat([latents] * 3) if do_classifier_free_guidance else latents
                )

                # concat latents, image_latents in the channel dimension
                scaled_latent_model_input = self.scheduler.scale_model_input(
                    latent_model_input, t
                )
                scaled_latent_model_input = torch.cat(
                    [scaled_latent_model_input, image_latents], dim=1
                )

                # predict the noise residual
                noise_pred = self.unet(
                    scaled_latent_model_input,
                    t,
                    encoder_hidden_states=prompt_embeds,
                    return_dict=False,
                )[0]

                # perform guidance
                if do_classifier_free_guidance:
                    (
                        noise_pred_text,
                        noise_pred_image,
                        noise_pred_uncond,
                    ) = noise_pred.chunk(3)
                    noise_pred = (
                        noise_pred_uncond
                        + guidance_scale * (noise_pred_text - noise_pred_image)
                        + image_guidance_scale * (noise_pred_image - noise_pred_uncond)
                    )

                if do_classifier_free_guidance and guidance_rescale > 0.0:
                    # Based on 3.4. in https://arxiv.org/pdf/2305.08891.pdf
                    noise_pred = rescale_noise_cfg(
                        noise_pred, noise_pred_text, guidance_rescale=guidance_rescale
                    )

                # compute the previous noisy sample x_t -> x_t-1
                output = self.scheduler.step(
                    noise_pred, t, latents, **extra_step_kwargs, return_dict=True
                )

                latents = output[0]

                if return_predicted_x0s:
                    predicted_x0s.append(output[1])

                # 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:
                        callback(i, t, latents)

        if not output_type == "latent":
            image = self.vae.decode(
                latents / self.vae.config.scaling_factor, return_dict=False
            )[0]

            if return_predicted_x0s:
                predicted_x0_images = [
                    self.vae.decode(
                        predicted_x0 / self.vae.config.scaling_factor, return_dict=False
                    )[0]
                    for predicted_x0 in predicted_x0s
                ]
        else:
            image = latents
            predicted_x0_images = predicted_x0s

        do_denormalize = [True] * image.shape[0]

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

        if return_predicted_x0s:
            predicted_x0_images = [
                self.image_processor.postprocess(
                    predicted_x0_image,
                    output_type=output_type,
                    do_denormalize=do_denormalize,
                )
                for predicted_x0_image in predicted_x0_images
            ]

        # Offload last model to CPU
        if hasattr(self, "final_offload_hook") and self.final_offload_hook is not None:
            self.final_offload_hook.offload()

        if not return_dict:
            return image

        if return_predicted_x0s:
            return StableDiffusionAOVPipelineOutput(
                images=image, predicted_x0_images=predicted_x0_images
            )
        else:
            return StableDiffusionAOVPipelineOutput(images=image)