lpw_stable_diffusion.py

import inspect
import re
from typing import Any, Callable, Dict, List, Optional, Union

import numpy as np
import PIL.Image
import torch
from packaging import version
from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer

from diffusers import DiffusionPipeline
from diffusers.configuration_utils import FrozenDict
from diffusers.image_processor import VaeImageProcessor
from diffusers.loaders import FromSingleFileMixin, StableDiffusionLoraLoaderMixin, TextualInversionLoaderMixin
from diffusers.models import AutoencoderKL, UNet2DConditionModel
from diffusers.models.lora import adjust_lora_scale_text_encoder
from diffusers.pipelines.pipeline_utils import StableDiffusionMixin
from diffusers.pipelines.stable_diffusion import StableDiffusionPipelineOutput, StableDiffusionSafetyChecker
from diffusers.schedulers import KarrasDiffusionSchedulers
from diffusers.utils import (
    PIL_INTERPOLATION,
    USE_PEFT_BACKEND,
    deprecate,
    logging,
    scale_lora_layers,
    unscale_lora_layers,
)
from diffusers.utils.torch_utils import randn_tensor


# ------------------------------------------------------------------------------

logger = logging.get_logger(__name__)  # pylint: disable=invalid-name

re_attention = re.compile(
    r"""
\\\(|
\\\)|
\\\[|
\\]|
\\\\|
\\|
\(|
\[|
:([+-]?[.\d]+)\)|
\)|
]|
[^\\()\[\]:]+|
:
""",
    re.X,
)


def parse_prompt_attention(text):
    """
    Parses a string with attention tokens and returns a list of pairs: text and its associated weight.
    Accepted tokens are:
      (abc) - increases attention to abc by a multiplier of 1.1
      (abc:3.12) - increases attention to abc by a multiplier of 3.12
      [abc] - decreases attention to abc by a multiplier of 1.1
      \\( - literal character '('
      \\[ - literal character '['
      \\) - literal character ')'
      \\] - literal character ']'
      \\ - literal character '\'
      anything else - just text
    >>> parse_prompt_attention('normal text')
    [['normal text', 1.0]]
    >>> parse_prompt_attention('an (important) word')
    [['an ', 1.0], ['important', 1.1], [' word', 1.0]]
    >>> parse_prompt_attention('(unbalanced')
    [['unbalanced', 1.1]]
    >>> parse_prompt_attention('\\(literal\\]')
    [['(literal]', 1.0]]
    >>> parse_prompt_attention('(unnecessary)(parens)')
    [['unnecessaryparens', 1.1]]
    >>> parse_prompt_attention('a (((house:1.3)) [on] a (hill:0.5), sun, (((sky))).')
    [['a ', 1.0],
     ['house', 1.5730000000000004],
     [' ', 1.1],
     ['on', 1.0],
     [' a ', 1.1],
     ['hill', 0.55],
     [', sun, ', 1.1],
     ['sky', 1.4641000000000006],
     ['.', 1.1]]
    """

    res = []
    round_brackets = []
    square_brackets = []

    round_bracket_multiplier = 1.1
    square_bracket_multiplier = 1 / 1.1

    def multiply_range(start_position, multiplier):
        for p in range(start_position, len(res)):
            res[p][1] *= multiplier

    for m in re_attention.finditer(text):
        text = m.group(0)
        weight = m.group(1)

        if text.startswith("\\"):
            res.append([text[1:], 1.0])
        elif text == "(":
            round_brackets.append(len(res))
        elif text == "[":
            square_brackets.append(len(res))
        elif weight is not None and len(round_brackets) > 0:
            multiply_range(round_brackets.pop(), float(weight))
        elif text == ")" and len(round_brackets) > 0:
            multiply_range(round_brackets.pop(), round_bracket_multiplier)
        elif text == "]" and len(square_brackets) > 0:
            multiply_range(square_brackets.pop(), square_bracket_multiplier)
        else:
            res.append([text, 1.0])

    for pos in round_brackets:
        multiply_range(pos, round_bracket_multiplier)

    for pos in square_brackets:
        multiply_range(pos, square_bracket_multiplier)

    if len(res) == 0:
        res = [["", 1.0]]

    # merge runs of identical weights
    i = 0
    while i + 1 < len(res):
        if res[i][1] == res[i + 1][1]:
            res[i][0] += res[i + 1][0]
            res.pop(i + 1)
        else:
            i += 1

    return res


def get_prompts_with_weights(pipe: DiffusionPipeline, prompt: List[str], max_length: int):
    r"""
    Tokenize a list of prompts and return its tokens with weights of each token.

    No padding, starting or ending token is included.
    """
    tokens = []
    weights = []
    truncated = False
    for text in prompt:
        texts_and_weights = parse_prompt_attention(text)
        text_token = []
        text_weight = []
        for word, weight in texts_and_weights:
            # tokenize and discard the starting and the ending token
            token = pipe.tokenizer(word).input_ids[1:-1]
            text_token += token
            # copy the weight by length of token
            text_weight += [weight] * len(token)
            # stop if the text is too long (longer than truncation limit)
            if len(text_token) > max_length:
                truncated = True
                break
        # truncate
        if len(text_token) > max_length:
            truncated = True
            text_token = text_token[:max_length]
            text_weight = text_weight[:max_length]
        tokens.append(text_token)
        weights.append(text_weight)
    if truncated:
        logger.warning("Prompt was truncated. Try to shorten the prompt or increase max_embeddings_multiples")
    return tokens, weights


def pad_tokens_and_weights(tokens, weights, max_length, bos, eos, pad, no_boseos_middle=True, chunk_length=77):
    r"""
    Pad the tokens (with starting and ending tokens) and weights (with 1.0) to max_length.
    """
    max_embeddings_multiples = (max_length - 2) // (chunk_length - 2)
    weights_length = max_length if no_boseos_middle else max_embeddings_multiples * chunk_length
    for i in range(len(tokens)):
        tokens[i] = [bos] + tokens[i] + [pad] * (max_length - 1 - len(tokens[i]) - 1) + [eos]
        if no_boseos_middle:
            weights[i] = [1.0] + weights[i] + [1.0] * (max_length - 1 - len(weights[i]))
        else:
            w = []
            if len(weights[i]) == 0:
                w = [1.0] * weights_length
            else:
                for j in range(max_embeddings_multiples):
                    w.append(1.0)  # weight for starting token in this chunk
                    w += weights[i][j * (chunk_length - 2) : min(len(weights[i]), (j + 1) * (chunk_length - 2))]
                    w.append(1.0)  # weight for ending token in this chunk
                w += [1.0] * (weights_length - len(w))
            weights[i] = w[:]

    return tokens, weights


def get_unweighted_text_embeddings(
    pipe: DiffusionPipeline,
    text_input: torch.Tensor,
    chunk_length: int,
    no_boseos_middle: Optional[bool] = True,
    clip_skip: Optional[int] = None,
):
    """
    When the length of tokens is a multiple of the capacity of the text encoder,
    it should be split into chunks and sent to the text encoder individually.
    """
    max_embeddings_multiples = (text_input.shape[1] - 2) // (chunk_length - 2)
    if max_embeddings_multiples > 1:
        text_embeddings = []
        for i in range(max_embeddings_multiples):
            # extract the i-th chunk
            text_input_chunk = text_input[:, i * (chunk_length - 2) : (i + 1) * (chunk_length - 2) + 2].clone()

            # cover the head and the tail by the starting and the ending tokens
            text_input_chunk[:, 0] = text_input[0, 0]
            text_input_chunk[:, -1] = text_input[0, -1]
            if clip_skip is None:
                prompt_embeds = pipe.text_encoder(text_input_chunk.to(pipe.device))
                text_embedding = prompt_embeds[0]
            else:
                prompt_embeds = pipe.text_encoder(text_input_chunk.to(pipe.device), 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.
                text_embedding = pipe.text_encoder.text_model.final_layer_norm(prompt_embeds)

            if no_boseos_middle:
                if i == 0:
                    # discard the ending token
                    text_embedding = text_embedding[:, :-1]
                elif i == max_embeddings_multiples - 1:
                    # discard the starting token
                    text_embedding = text_embedding[:, 1:]
                else:
                    # discard both starting and ending tokens
                    text_embedding = text_embedding[:, 1:-1]

            text_embeddings.append(text_embedding)
        text_embeddings = torch.concat(text_embeddings, axis=1)
    else:
        if clip_skip is None:
            clip_skip = 0
        prompt_embeds = pipe.text_encoder(text_input, output_hidden_states=True)[-1][-(clip_skip + 1)]
        text_embeddings = pipe.text_encoder.text_model.final_layer_norm(prompt_embeds)
    return text_embeddings


def get_weighted_text_embeddings(
    pipe: DiffusionPipeline,
    prompt: Union[str, List[str]],
    uncond_prompt: Optional[Union[str, List[str]]] = None,
    max_embeddings_multiples: Optional[int] = 3,
    no_boseos_middle: Optional[bool] = False,
    skip_parsing: Optional[bool] = False,
    skip_weighting: Optional[bool] = False,
    clip_skip=None,
    lora_scale=None,
):
    r"""
    Prompts can be assigned with local weights using brackets. For example,
    prompt 'A (very beautiful) masterpiece' highlights the words 'very beautiful',
    and the embedding tokens corresponding to the words get multiplied by a constant, 1.1.

    Also, to regularize of the embedding, the weighted embedding would be scaled to preserve the original mean.

    Args:
        pipe (`DiffusionPipeline`):
            Pipe to provide access to the tokenizer and the text encoder.
        prompt (`str` or `List[str]`):
            The prompt or prompts to guide the image generation.
        uncond_prompt (`str` or `List[str]`):
            The unconditional prompt or prompts for guide the image generation. If unconditional prompt
            is provided, the embeddings of prompt and uncond_prompt are concatenated.
        max_embeddings_multiples (`int`, *optional*, defaults to `3`):
            The max multiple length of prompt embeddings compared to the max output length of text encoder.
        no_boseos_middle (`bool`, *optional*, defaults to `False`):
            If the length of text token is multiples of the capacity of text encoder, whether reserve the starting and
            ending token in each of the chunk in the middle.
        skip_parsing (`bool`, *optional*, defaults to `False`):
            Skip the parsing of brackets.
        skip_weighting (`bool`, *optional*, defaults to `False`):
            Skip the weighting. When the parsing is skipped, it is forced True.
    """
    # set lora scale so that monkey patched LoRA
    # function of text encoder can correctly access it
    if lora_scale is not None and isinstance(pipe, StableDiffusionLoraLoaderMixin):
        pipe._lora_scale = lora_scale

        # dynamically adjust the LoRA scale
        if not USE_PEFT_BACKEND:
            adjust_lora_scale_text_encoder(pipe.text_encoder, lora_scale)
        else:
            scale_lora_layers(pipe.text_encoder, lora_scale)
    max_length = (pipe.tokenizer.model_max_length - 2) * max_embeddings_multiples + 2
    if isinstance(prompt, str):
        prompt = [prompt]

    if not skip_parsing:
        prompt_tokens, prompt_weights = get_prompts_with_weights(pipe, prompt, max_length - 2)
        if uncond_prompt is not None:
            if isinstance(uncond_prompt, str):
                uncond_prompt = [uncond_prompt]
            uncond_tokens, uncond_weights = get_prompts_with_weights(pipe, uncond_prompt, max_length - 2)
    else:
        prompt_tokens = [
            token[1:-1] for token in pipe.tokenizer(prompt, max_length=max_length, truncation=True).input_ids
        ]
        prompt_weights = [[1.0] * len(token) for token in prompt_tokens]
        if uncond_prompt is not None:
            if isinstance(uncond_prompt, str):
                uncond_prompt = [uncond_prompt]
            uncond_tokens = [
                token[1:-1]
                for token in pipe.tokenizer(uncond_prompt, max_length=max_length, truncation=True).input_ids
            ]
            uncond_weights = [[1.0] * len(token) for token in uncond_tokens]

    # round up the longest length of tokens to a multiple of (model_max_length - 2)
    max_length = max([len(token) for token in prompt_tokens])
    if uncond_prompt is not None:
        max_length = max(max_length, max([len(token) for token in uncond_tokens]))

    max_embeddings_multiples = min(
        max_embeddings_multiples,
        (max_length - 1) // (pipe.tokenizer.model_max_length - 2) + 1,
    )
    max_embeddings_multiples = max(1, max_embeddings_multiples)
    max_length = (pipe.tokenizer.model_max_length - 2) * max_embeddings_multiples + 2

    # pad the length of tokens and weights
    bos = pipe.tokenizer.bos_token_id
    eos = pipe.tokenizer.eos_token_id
    pad = getattr(pipe.tokenizer, "pad_token_id", eos)
    prompt_tokens, prompt_weights = pad_tokens_and_weights(
        prompt_tokens,
        prompt_weights,
        max_length,
        bos,
        eos,
        pad,
        no_boseos_middle=no_boseos_middle,
        chunk_length=pipe.tokenizer.model_max_length,
    )
    prompt_tokens = torch.tensor(prompt_tokens, dtype=torch.long, device=pipe.device)
    if uncond_prompt is not None:
        uncond_tokens, uncond_weights = pad_tokens_and_weights(
            uncond_tokens,
            uncond_weights,
            max_length,
            bos,
            eos,
            pad,
            no_boseos_middle=no_boseos_middle,
            chunk_length=pipe.tokenizer.model_max_length,
        )
        uncond_tokens = torch.tensor(uncond_tokens, dtype=torch.long, device=pipe.device)

    # get the embeddings
    text_embeddings = get_unweighted_text_embeddings(
        pipe, prompt_tokens, pipe.tokenizer.model_max_length, no_boseos_middle=no_boseos_middle, clip_skip=clip_skip
    )
    prompt_weights = torch.tensor(prompt_weights, dtype=text_embeddings.dtype, device=text_embeddings.device)
    if uncond_prompt is not None:
        uncond_embeddings = get_unweighted_text_embeddings(
            pipe,
            uncond_tokens,
            pipe.tokenizer.model_max_length,
            no_boseos_middle=no_boseos_middle,
            clip_skip=clip_skip,
        )
        uncond_weights = torch.tensor(uncond_weights, dtype=uncond_embeddings.dtype, device=uncond_embeddings.device)

    # assign weights to the prompts and normalize in the sense of mean
    # TODO: should we normalize by chunk or in a whole (current implementation)?
    if (not skip_parsing) and (not skip_weighting):
        previous_mean = text_embeddings.float().mean(axis=[-2, -1]).to(text_embeddings.dtype)
        text_embeddings *= prompt_weights.unsqueeze(-1)
        current_mean = text_embeddings.float().mean(axis=[-2, -1]).to(text_embeddings.dtype)
        text_embeddings *= (previous_mean / current_mean).unsqueeze(-1).unsqueeze(-1)
        if uncond_prompt is not None:
            previous_mean = uncond_embeddings.float().mean(axis=[-2, -1]).to(uncond_embeddings.dtype)
            uncond_embeddings *= uncond_weights.unsqueeze(-1)
            current_mean = uncond_embeddings.float().mean(axis=[-2, -1]).to(uncond_embeddings.dtype)
            uncond_embeddings *= (previous_mean / current_mean).unsqueeze(-1).unsqueeze(-1)

    if pipe.text_encoder is not None:
        if isinstance(pipe, StableDiffusionLoraLoaderMixin) and USE_PEFT_BACKEND:
            # Retrieve the original scale by scaling back the LoRA layers
            unscale_lora_layers(pipe.text_encoder, lora_scale)

    if uncond_prompt is not None:
        return text_embeddings, uncond_embeddings
    return text_embeddings, None


def preprocess_image(image, batch_size):
    w, h = image.size
    w, h = (x - x % 8 for x in (w, h))  # resize to integer multiple of 8
    image = image.resize((w, h), resample=PIL_INTERPOLATION["lanczos"])
    image = np.array(image).astype(np.float32) / 255.0
    image = np.vstack([image[None].transpose(0, 3, 1, 2)] * batch_size)
    image = torch.from_numpy(image)
    return 2.0 * image - 1.0


def preprocess_mask(mask, batch_size, scale_factor=8):
    if not isinstance(mask, torch.Tensor):
        mask = mask.convert("L")
        w, h = mask.size
        w, h = (x - x % 8 for x in (w, h))  # resize to integer multiple of 8
        mask = mask.resize((w // scale_factor, h // scale_factor), resample=PIL_INTERPOLATION["nearest"])
        mask = np.array(mask).astype(np.float32) / 255.0
        mask = np.tile(mask, (4, 1, 1))
        mask = np.vstack([mask[None]] * batch_size)
        mask = 1 - mask  # repaint white, keep black
        mask = torch.from_numpy(mask)
        return mask

    else:
        valid_mask_channel_sizes = [1, 3]
        # if mask channel is fourth tensor dimension, permute dimensions to pytorch standard (B, C, H, W)
        if mask.shape[3] in valid_mask_channel_sizes:
            mask = mask.permute(0, 3, 1, 2)
        elif mask.shape[1] not in valid_mask_channel_sizes:
            raise ValueError(
                f"Mask channel dimension of size in {valid_mask_channel_sizes} should be second or fourth dimension,"
                f" but received mask of shape {tuple(mask.shape)}"
            )
        # (potentially) reduce mask channel dimension from 3 to 1 for broadcasting to latent shape
        mask = mask.mean(dim=1, keepdim=True)
        h, w = mask.shape[-2:]
        h, w = (x - x % 8 for x in (h, w))  # resize to integer multiple of 8
        mask = torch.nn.functional.interpolate(mask, (h // scale_factor, w // scale_factor))
        return mask


class StableDiffusionLongPromptWeightingPipeline(
    DiffusionPipeline,
    StableDiffusionMixin,
    TextualInversionLoaderMixin,
    StableDiffusionLoraLoaderMixin,
    FromSingleFileMixin,
):
    r"""
    Pipeline for text-to-image generation using Stable Diffusion without tokens length limit, and support parsing
    weighting in prompt.

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

    Args:
        vae ([`AutoencoderKL`]):
            Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations.
        text_encoder ([`CLIPTextModel`]):
            Frozen text-encoder. Stable Diffusion uses the text portion of
            [CLIP](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTextModel), specifically
            the [clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14) variant.
        tokenizer (`CLIPTokenizer`):
            Tokenizer of class
            [CLIPTokenizer](https://huggingface.co/docs/transformers/v4.21.0/en/model_doc/clip#transformers.CLIPTokenizer).
        unet ([`UNet2DConditionModel`]): Conditional U-Net architecture 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`].
        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/CompVis/stable-diffusion-v1-4) for details.
        feature_extractor ([`CLIPImageProcessor`]):
            Model that extracts features from generated images to be used as inputs for the `safety_checker`.
    """

    model_cpu_offload_seq = "text_encoder-->unet->vae"
    _optional_components = ["safety_checker", "feature_extractor"]
    _exclude_from_cpu_offload = ["safety_checker"]

    def __init__(
        self,
        vae: AutoencoderKL,
        text_encoder: CLIPTextModel,
        tokenizer: CLIPTokenizer,
        unet: UNet2DConditionModel,
        scheduler: KarrasDiffusionSchedulers,
        safety_checker: StableDiffusionSafetyChecker,
        feature_extractor: CLIPImageProcessor,
        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, "clip_sample") and scheduler.config.clip_sample is True:
            deprecation_message = (
                f"The configuration file of this scheduler: {scheduler} has not set the configuration `clip_sample`."
                " `clip_sample` should be set to False in the configuration file. Please make sure to update the"
                " config accordingly as not setting `clip_sample` 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("clip_sample not set", "1.0.0", deprecation_message, standard_warn=False)
            new_config = dict(scheduler.config)
            new_config["clip_sample"] = False
            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 your 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,
        )
        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,
        )

    def _encode_prompt(
        self,
        prompt,
        device,
        num_images_per_prompt,
        do_classifier_free_guidance,
        negative_prompt=None,
        max_embeddings_multiples=3,
        prompt_embeds: Optional[torch.Tensor] = None,
        negative_prompt_embeds: Optional[torch.Tensor] = None,
        clip_skip: Optional[int] = None,
        lora_scale: Optional[float] = None,
    ):
        r"""
        Encodes the prompt into text encoder hidden states.

        Args:
            prompt (`str` or `list(int)`):
                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]`):
                The prompt or prompts not to guide the image generation. Ignored when not using guidance (i.e., ignored
                if `guidance_scale` is less than `1`).
            max_embeddings_multiples (`int`, *optional*, defaults to `3`):
                The max multiple length of prompt embeddings compared to the max output length of text encoder.
        """
        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 negative_prompt_embeds is None:
            if negative_prompt is None:
                negative_prompt = [""] * batch_size
            elif isinstance(negative_prompt, str):
                negative_prompt = [negative_prompt] * batch_size
            if 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`."
                )
        if prompt_embeds is None or negative_prompt_embeds is None:
            if isinstance(self, TextualInversionLoaderMixin):
                prompt = self.maybe_convert_prompt(prompt, self.tokenizer)
                if do_classifier_free_guidance and negative_prompt_embeds is None:
                    negative_prompt = self.maybe_convert_prompt(negative_prompt, self.tokenizer)

            prompt_embeds1, negative_prompt_embeds1 = get_weighted_text_embeddings(
                pipe=self,
                prompt=prompt,
                uncond_prompt=negative_prompt if do_classifier_free_guidance else None,
                max_embeddings_multiples=max_embeddings_multiples,
                clip_skip=clip_skip,
                lora_scale=lora_scale,
            )
            if prompt_embeds is None:
                prompt_embeds = prompt_embeds1
            if negative_prompt_embeds is None:
                negative_prompt_embeds = negative_prompt_embeds1

        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)

        if do_classifier_free_guidance:
            bs_embed, seq_len, _ = negative_prompt_embeds.shape
            negative_prompt_embeds = negative_prompt_embeds.repeat(1, num_images_per_prompt, 1)
            negative_prompt_embeds = negative_prompt_embeds.view(bs_embed * num_images_per_prompt, seq_len, -1)
            prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds])

        return prompt_embeds

    def check_inputs(
        self,
        prompt,
        height,
        width,
        strength,
        callback_steps,
        negative_prompt=None,
        prompt_embeds=None,
        negative_prompt_embeds=None,
    ):
        if height % 8 != 0 or width % 8 != 0:
            raise ValueError(f"`height` and `width` have to be divisible by 8 but are {height} and {width}.")

        if strength < 0 or strength > 1:
            raise ValueError(f"The value of strength should in [0.0, 1.0] but is {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 get_timesteps(self, num_inference_steps, strength, device, is_text2img):
        if is_text2img:
            return self.scheduler.timesteps.to(device), num_inference_steps
        else:
            # 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 run_safety_checker(self, image, device, dtype):
        if self.safety_checker is not None:
            safety_checker_input = self.feature_extractor(self.numpy_to_pil(image), return_tensors="pt").to(device)
            image, has_nsfw_concept = self.safety_checker(
                images=image, clip_input=safety_checker_input.pixel_values.to(dtype)
            )
        else:
            has_nsfw_concept = None
        return image, has_nsfw_concept

    def decode_latents(self, latents):
        latents = 1 / self.vae.config.scaling_factor * latents
        image = self.vae.decode(latents).sample
        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.cpu().permute(0, 2, 3, 1).float().numpy()
        return image

    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 prepare_latents(
        self,
        image,
        timestep,
        num_images_per_prompt,
        batch_size,
        num_channels_latents,
        height,
        width,
        dtype,
        device,
        generator,
        latents=None,
    ):
        if image is None:
            batch_size = batch_size * num_images_per_prompt
            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, 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, None, None
        else:
            image = image.to(device=self.device, dtype=dtype)
            init_latent_dist = self.vae.encode(image).latent_dist
            init_latents = init_latent_dist.sample(generator=generator)
            init_latents = self.vae.config.scaling_factor * init_latents

            # Expand init_latents for batch_size and num_images_per_prompt
            init_latents = torch.cat([init_latents] * num_images_per_prompt, dim=0)
            init_latents_orig = init_latents

            # add noise to latents using the timesteps
            noise = randn_tensor(init_latents.shape, generator=generator, device=self.device, dtype=dtype)
            init_latents = self.scheduler.add_noise(init_latents, noise, timestep)
            latents = init_latents
            return latents, init_latents_orig, noise

    @torch.no_grad()
    def __call__(
        self,
        prompt: Union[str, List[str]],
        negative_prompt: Optional[Union[str, List[str]]] = None,
        image: Union[torch.Tensor, PIL.Image.Image] = None,
        mask_image: Union[torch.Tensor, PIL.Image.Image] = None,
        height: int = 512,
        width: int = 512,
        num_inference_steps: int = 50,
        guidance_scale: float = 7.5,
        strength: float = 0.8,
        num_images_per_prompt: Optional[int] = 1,
        add_predicted_noise: Optional[bool] = False,
        eta: float = 0.0,
        generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
        latents: Optional[torch.Tensor] = None,
        prompt_embeds: Optional[torch.Tensor] = None,
        negative_prompt_embeds: Optional[torch.Tensor] = None,
        max_embeddings_multiples: Optional[int] = 3,
        output_type: Optional[str] = "pil",
        return_dict: bool = True,
        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
        is_cancelled_callback: Optional[Callable[[], bool]] = None,
        clip_skip: Optional[int] = None,
        callback_steps: int = 1,
        cross_attention_kwargs: Optional[Dict[str, Any]] = None,
    ):
        r"""
        Function invoked when calling the pipeline for generation.

        Args:
            prompt (`str` or `List[str]`):
                The prompt or prompts to guide the image generation.
            negative_prompt (`str` or `List[str]`, *optional*):
                The prompt or prompts not to guide the image generation. Ignored when not using guidance (i.e., ignored
                if `guidance_scale` is less than `1`).
            image (`torch.Tensor` or `PIL.Image.Image`):
                `Image`, or tensor representing an image batch, that will be used as the starting point for the
                process.
            mask_image (`torch.Tensor` or `PIL.Image.Image`):
                `Image`, or tensor representing an image batch, to mask `image`. White pixels in the mask will be
                replaced by noise and therefore repainted, while black pixels will be preserved. If `mask_image` is a
                PIL image, it will be 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, H, W, 1)`.
            height (`int`, *optional*, defaults to 512):
                The height in pixels of the generated image.
            width (`int`, *optional*, defaults to 512):
                The width in pixels of the generated image.
            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):
                Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
                `guidance_scale` is defined as `w` of equation 2. of [Imagen
                Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
                1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
                usually at the expense of lower image quality.
            strength (`float`, *optional*, defaults to 0.8):
                Conceptually, indicates how much to transform the reference `image`. Must be between 0 and 1.
                `image` will be used as a starting point, adding more noise to it the larger the `strength`. The
                number of denoising steps depends on the amount of noise initially added. When `strength` is 1, added
                noise will be maximum and the denoising process will run for the full number of iterations specified in
                `num_inference_steps`. A value of 1, therefore, essentially ignores `image`.
            num_images_per_prompt (`int`, *optional*, defaults to 1):
                The number of images to generate per prompt.
            add_predicted_noise (`bool`, *optional*, defaults to True):
                Use predicted noise instead of random noise when constructing noisy versions of the original image in
                the reverse diffusion process
            eta (`float`, *optional*, defaults to 0.0):
                Corresponds to parameter eta (η) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
                [`schedulers.DDIMScheduler`], will be ignored for others.
            generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
                to make generation deterministic.
            latents (`torch.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 will ge generated by sampling using the supplied random `generator`.
            prompt_embeds (`torch.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 (`torch.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.
            max_embeddings_multiples (`int`, *optional*, defaults to `3`):
                The max multiple length of prompt embeddings compared to the max output length of text encoder.
            output_type (`str`, *optional*, defaults to `"pil"`):
                The output format of the generate image. Choose between
                [PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
            return_dict (`bool`, *optional*, defaults to `True`):
                Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] instead of a
                plain tuple.
            callback (`Callable`, *optional*):
                A function that will be called every `callback_steps` steps during inference. The function will be
                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
            is_cancelled_callback (`Callable`, *optional*):
                A function that will be called every `callback_steps` steps during inference. If the function returns
                `True`, the inference will be cancelled.
            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.
            callback_steps (`int`, *optional*, defaults to 1):
                The frequency at which the `callback` function will be called. If not specified, the callback will be
                called at every step.
            cross_attention_kwargs (`dict`, *optional*):
                A kwargs dictionary that if specified is passed along to the `AttentionProcessor` as defined under
                `self.processor` in
                [diffusers.models.attention_processor](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).

        Returns:
            `None` if cancelled by `is_cancelled_callback`,
            [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] or `tuple`:
            [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] if `return_dict` is True, otherwise a `tuple.
            When returning a tuple, the first element is a list with the generated images, and the second element is a
            list of `bool`s denoting whether the corresponding generated image likely represents "not-safe-for-work"
            (nsfw) content, according to the `safety_checker`.
        """
        # 0. Default height and width to unet
        height = height or self.unet.config.sample_size * self.vae_scale_factor
        width = width or self.unet.config.sample_size * self.vae_scale_factor

        # 1. Check inputs. Raise error if not correct
        self.check_inputs(
            prompt, height, width, strength, callback_steps, negative_prompt, prompt_embeds, negative_prompt_embeds
        )

        # 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]

        device = self._execution_device
        # 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
        lora_scale = cross_attention_kwargs.get("scale", None) if cross_attention_kwargs is not None else None

        # 3. Encode input prompt
        prompt_embeds = self._encode_prompt(
            prompt,
            device,
            num_images_per_prompt,
            do_classifier_free_guidance,
            negative_prompt,
            max_embeddings_multiples,
            prompt_embeds=prompt_embeds,
            negative_prompt_embeds=negative_prompt_embeds,
            clip_skip=clip_skip,
            lora_scale=lora_scale,
        )
        dtype = prompt_embeds.dtype

        # 4. Preprocess image and mask
        if isinstance(image, PIL.Image.Image):
            image = preprocess_image(image, batch_size)
        if image is not None:
            image = image.to(device=self.device, dtype=dtype)
        if isinstance(mask_image, PIL.Image.Image):
            mask_image = preprocess_mask(mask_image, batch_size, self.vae_scale_factor)
        if mask_image is not None:
            mask = mask_image.to(device=self.device, dtype=dtype)
            mask = torch.cat([mask] * num_images_per_prompt)
        else:
            mask = None

        # 5. set timesteps
        self.scheduler.set_timesteps(num_inference_steps, device=device)
        timesteps, num_inference_steps = self.get_timesteps(num_inference_steps, strength, device, image is None)
        latent_timestep = timesteps[:1].repeat(batch_size * num_images_per_prompt)

        # 6. Prepare latent variables
        latents, init_latents_orig, noise = self.prepare_latents(
            image,
            latent_timestep,
            num_images_per_prompt,
            batch_size,
            self.unet.config.in_channels,
            height,
            width,
            dtype,
            device,
            generator,
            latents,
        )

        # 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
        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 = torch.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,
                ).sample

                # 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).prev_sample

                if mask is not None:
                    # masking
                    if add_predicted_noise:
                        init_latents_proper = self.scheduler.add_noise(
                            init_latents_orig, noise_pred_uncond, torch.tensor([t])
                        )
                    else:
                        init_latents_proper = self.scheduler.add_noise(init_latents_orig, noise, torch.tensor([t]))
                    latents = (init_latents_proper * mask) + (latents * (1 - mask))

                # 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 i % callback_steps == 0:
                        if callback is not None:
                            step_idx = i // getattr(self.scheduler, "order", 1)
                            callback(step_idx, t, latents)
                        if is_cancelled_callback is not None and is_cancelled_callback():
                            return None

        if output_type == "latent":
            image = latents
            has_nsfw_concept = None
        elif output_type == "pil":
            # 9. Post-processing
            image = self.decode_latents(latents)

            # 10. Run safety checker
            image, has_nsfw_concept = self.run_safety_checker(image, device, prompt_embeds.dtype)

            # 11. Convert to PIL
            image = self.numpy_to_pil(image)
        else:
            # 9. Post-processing
            image = self.decode_latents(latents)

            # 10. Run safety checker
            image, has_nsfw_concept = self.run_safety_checker(image, device, prompt_embeds.dtype)

        # 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, has_nsfw_concept

        return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=has_nsfw_concept)

    def text2img(
        self,
        prompt: Union[str, List[str]],
        negative_prompt: Optional[Union[str, List[str]]] = None,
        height: int = 512,
        width: int = 512,
        num_inference_steps: int = 50,
        guidance_scale: float = 7.5,
        num_images_per_prompt: Optional[int] = 1,
        eta: float = 0.0,
        generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
        latents: Optional[torch.Tensor] = None,
        prompt_embeds: Optional[torch.Tensor] = None,
        negative_prompt_embeds: Optional[torch.Tensor] = None,
        max_embeddings_multiples: Optional[int] = 3,
        output_type: Optional[str] = "pil",
        return_dict: bool = True,
        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
        is_cancelled_callback: Optional[Callable[[], bool]] = None,
        clip_skip=None,
        callback_steps: int = 1,
        cross_attention_kwargs: Optional[Dict[str, Any]] = None,
    ):
        r"""
        Function for text-to-image generation.
        Args:
            prompt (`str` or `List[str]`):
                The prompt or prompts to guide the image generation.
            negative_prompt (`str` or `List[str]`, *optional*):
                The prompt or prompts not to guide the image generation. Ignored when not using guidance (i.e., ignored
                if `guidance_scale` is less than `1`).
            height (`int`, *optional*, defaults to 512):
                The height in pixels of the generated image.
            width (`int`, *optional*, defaults to 512):
                The width in pixels of the generated image.
            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):
                Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
                `guidance_scale` is defined as `w` of equation 2. of [Imagen
                Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
                1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
                usually at the expense of lower image quality.
            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) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
                [`schedulers.DDIMScheduler`], will be ignored for others.
            generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
                to make generation deterministic.
            latents (`torch.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 will ge generated by sampling using the supplied random `generator`.
            prompt_embeds (`torch.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 (`torch.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.
            max_embeddings_multiples (`int`, *optional*, defaults to `3`):
                The max multiple length of prompt embeddings compared to the max output length of text encoder.
            output_type (`str`, *optional*, defaults to `"pil"`):
                The output format of the generate image. Choose between
                [PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
            return_dict (`bool`, *optional*, defaults to `True`):
                Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] instead of a
                plain tuple.
            callback (`Callable`, *optional*):
                A function that will be called every `callback_steps` steps during inference. The function will be
                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
            is_cancelled_callback (`Callable`, *optional*):
                A function that will be called every `callback_steps` steps during inference. If the function returns
                `True`, the inference will be cancelled.
            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.
            callback_steps (`int`, *optional*, defaults to 1):
                The frequency at which the `callback` function will be called. If not specified, the callback will be
                called at every step.
            cross_attention_kwargs (`dict`, *optional*):
                A kwargs dictionary that if specified is passed along to the `AttentionProcessor` as defined under
                `self.processor` in
                [diffusers.models.attention_processor](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).

        Returns:
            `None` if cancelled by `is_cancelled_callback`,
            [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] or `tuple`:
            [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] if `return_dict` is True, otherwise a `tuple.
            When returning a tuple, the first element is a list with the generated images, and the second element is a
            list of `bool`s denoting whether the corresponding generated image likely represents "not-safe-for-work"
            (nsfw) content, according to the `safety_checker`.
        """
        return self.__call__(
            prompt=prompt,
            negative_prompt=negative_prompt,
            height=height,
            width=width,
            num_inference_steps=num_inference_steps,
            guidance_scale=guidance_scale,
            num_images_per_prompt=num_images_per_prompt,
            eta=eta,
            generator=generator,
            latents=latents,
            prompt_embeds=prompt_embeds,
            negative_prompt_embeds=negative_prompt_embeds,
            max_embeddings_multiples=max_embeddings_multiples,
            output_type=output_type,
            return_dict=return_dict,
            callback=callback,
            is_cancelled_callback=is_cancelled_callback,
            clip_skip=clip_skip,
            callback_steps=callback_steps,
            cross_attention_kwargs=cross_attention_kwargs,
        )

    def img2img(
        self,
        image: Union[torch.Tensor, PIL.Image.Image],
        prompt: Union[str, List[str]],
        negative_prompt: Optional[Union[str, List[str]]] = None,
        strength: float = 0.8,
        num_inference_steps: Optional[int] = 50,
        guidance_scale: Optional[float] = 7.5,
        num_images_per_prompt: Optional[int] = 1,
        eta: Optional[float] = 0.0,
        generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
        prompt_embeds: Optional[torch.Tensor] = None,
        negative_prompt_embeds: Optional[torch.Tensor] = None,
        max_embeddings_multiples: Optional[int] = 3,
        output_type: Optional[str] = "pil",
        return_dict: bool = True,
        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
        is_cancelled_callback: Optional[Callable[[], bool]] = None,
        callback_steps: int = 1,
        cross_attention_kwargs: Optional[Dict[str, Any]] = None,
    ):
        r"""
        Function for image-to-image generation.
        Args:
            image (`torch.Tensor` or `PIL.Image.Image`):
                `Image`, or tensor representing an image batch, that will be used as the starting point for the
                process.
            prompt (`str` or `List[str]`):
                The prompt or prompts to guide the image generation.
            negative_prompt (`str` or `List[str]`, *optional*):
                The prompt or prompts not to guide the image generation. Ignored when not using guidance (i.e., ignored
                if `guidance_scale` is less than `1`).
            strength (`float`, *optional*, defaults to 0.8):
                Conceptually, indicates how much to transform the reference `image`. Must be between 0 and 1.
                `image` will be used as a starting point, adding more noise to it the larger the `strength`. The
                number of denoising steps depends on the amount of noise initially added. When `strength` is 1, added
                noise will be maximum and the denoising process will run for the full number of iterations specified in
                `num_inference_steps`. A value of 1, therefore, essentially ignores `image`.
            num_inference_steps (`int`, *optional*, defaults to 50):
                The number of denoising steps. More denoising steps usually lead to a higher quality image at the
                expense of slower inference. This parameter will be modulated by `strength`.
            guidance_scale (`float`, *optional*, defaults to 7.5):
                Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
                `guidance_scale` is defined as `w` of equation 2. of [Imagen
                Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
                1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
                usually at the expense of lower image quality.
            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) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
                [`schedulers.DDIMScheduler`], will be ignored for others.
            generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
                to make generation deterministic.
            prompt_embeds (`torch.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 (`torch.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.
            max_embeddings_multiples (`int`, *optional*, defaults to `3`):
                The max multiple length of prompt embeddings compared to the max output length of text encoder.
            output_type (`str`, *optional*, defaults to `"pil"`):
                The output format of the generate image. Choose between
                [PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
            return_dict (`bool`, *optional*, defaults to `True`):
                Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] instead of a
                plain tuple.
            callback (`Callable`, *optional*):
                A function that will be called every `callback_steps` steps during inference. The function will be
                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
            is_cancelled_callback (`Callable`, *optional*):
                A function that will be called every `callback_steps` steps during inference. If the function returns
                `True`, the inference will be cancelled.
            callback_steps (`int`, *optional*, defaults to 1):
                The frequency at which the `callback` function will be called. If not specified, the callback will be
                called at every step.
            cross_attention_kwargs (`dict`, *optional*):
                A kwargs dictionary that if specified is passed along to the `AttentionProcessor` as defined under
                `self.processor` in
                [diffusers.models.attention_processor](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).

        Returns:
            `None` if cancelled by `is_cancelled_callback`,
            [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] if `return_dict` is True, otherwise a `tuple.
            When returning a tuple, the first element is a list with the generated images, and the second element is a
            list of `bool`s denoting whether the corresponding generated image likely represents "not-safe-for-work"
            (nsfw) content, according to the `safety_checker`.
        """
        return self.__call__(
            prompt=prompt,
            negative_prompt=negative_prompt,
            image=image,
            num_inference_steps=num_inference_steps,
            guidance_scale=guidance_scale,
            strength=strength,
            num_images_per_prompt=num_images_per_prompt,
            eta=eta,
            generator=generator,
            prompt_embeds=prompt_embeds,
            negative_prompt_embeds=negative_prompt_embeds,
            max_embeddings_multiples=max_embeddings_multiples,
            output_type=output_type,
            return_dict=return_dict,
            callback=callback,
            is_cancelled_callback=is_cancelled_callback,
            callback_steps=callback_steps,
            cross_attention_kwargs=cross_attention_kwargs,
        )

    def inpaint(
        self,
        image: Union[torch.Tensor, PIL.Image.Image],
        mask_image: Union[torch.Tensor, PIL.Image.Image],
        prompt: Union[str, List[str]],
        negative_prompt: Optional[Union[str, List[str]]] = None,
        strength: float = 0.8,
        num_inference_steps: Optional[int] = 50,
        guidance_scale: Optional[float] = 7.5,
        num_images_per_prompt: Optional[int] = 1,
        add_predicted_noise: Optional[bool] = False,
        eta: Optional[float] = 0.0,
        generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
        prompt_embeds: Optional[torch.Tensor] = None,
        negative_prompt_embeds: Optional[torch.Tensor] = None,
        max_embeddings_multiples: Optional[int] = 3,
        output_type: Optional[str] = "pil",
        return_dict: bool = True,
        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,
        is_cancelled_callback: Optional[Callable[[], bool]] = None,
        callback_steps: int = 1,
        cross_attention_kwargs: Optional[Dict[str, Any]] = None,
    ):
        r"""
        Function for inpaint.
        Args:
            image (`torch.Tensor` or `PIL.Image.Image`):
                `Image`, or tensor representing an image batch, that will be used as the starting point for the
                process. This is the image whose masked region will be inpainted.
            mask_image (`torch.Tensor` or `PIL.Image.Image`):
                `Image`, or tensor representing an image batch, to mask `image`. White pixels in the mask will be
                replaced by noise and therefore repainted, while black pixels will be preserved. If `mask_image` is a
                PIL image, it will be 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, H, W, 1)`.
            prompt (`str` or `List[str]`):
                The prompt or prompts to guide the image generation.
            negative_prompt (`str` or `List[str]`, *optional*):
                The prompt or prompts not to guide the image generation. Ignored when not using guidance (i.e., ignored
                if `guidance_scale` is less than `1`).
            strength (`float`, *optional*, defaults to 0.8):
                Conceptually, indicates how much to inpaint the masked area. Must be between 0 and 1. When `strength`
                is 1, the denoising process will be run on the masked area for the full number of iterations specified
                in `num_inference_steps`. `image` will be used as a reference for the masked area, adding more
                noise to that region the larger the `strength`. If `strength` is 0, no inpainting will occur.
            num_inference_steps (`int`, *optional*, defaults to 50):
                The reference number of denoising steps. More denoising steps usually lead to a higher quality image at
                the expense of slower inference. This parameter will be modulated by `strength`, as explained above.
            guidance_scale (`float`, *optional*, defaults to 7.5):
                Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
                `guidance_scale` is defined as `w` of equation 2. of [Imagen
                Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
                1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
                usually at the expense of lower image quality.
            num_images_per_prompt (`int`, *optional*, defaults to 1):
                The number of images to generate per prompt.
            add_predicted_noise (`bool`, *optional*, defaults to True):
                Use predicted noise instead of random noise when constructing noisy versions of the original image in
                the reverse diffusion process
            eta (`float`, *optional*, defaults to 0.0):
                Corresponds to parameter eta (η) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
                [`schedulers.DDIMScheduler`], will be ignored for others.
            generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
                One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
                to make generation deterministic.
            prompt_embeds (`torch.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 (`torch.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.
            max_embeddings_multiples (`int`, *optional*, defaults to `3`):
                The max multiple length of prompt embeddings compared to the max output length of text encoder.
            output_type (`str`, *optional*, defaults to `"pil"`):
                The output format of the generate image. Choose between
                [PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
            return_dict (`bool`, *optional*, defaults to `True`):
                Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] instead of a
                plain tuple.
            callback (`Callable`, *optional*):
                A function that will be called every `callback_steps` steps during inference. The function will be
                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.
            is_cancelled_callback (`Callable`, *optional*):
                A function that will be called every `callback_steps` steps during inference. If the function returns
                `True`, the inference will be cancelled.
            callback_steps (`int`, *optional*, defaults to 1):
                The frequency at which the `callback` function will be called. If not specified, the callback will be
                called at every step.
            cross_attention_kwargs (`dict`, *optional*):
                A kwargs dictionary that if specified is passed along to the `AttentionProcessor` as defined under
                `self.processor` in
                [diffusers.models.attention_processor](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).

        Returns:
            `None` if cancelled by `is_cancelled_callback`,
            [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] if `return_dict` is True, otherwise a `tuple.
            When returning a tuple, the first element is a list with the generated images, and the second element is a
            list of `bool`s denoting whether the corresponding generated image likely represents "not-safe-for-work"
            (nsfw) content, according to the `safety_checker`.
        """
        return self.__call__(
            prompt=prompt,
            negative_prompt=negative_prompt,
            image=image,
            mask_image=mask_image,
            num_inference_steps=num_inference_steps,
            guidance_scale=guidance_scale,
            strength=strength,
            num_images_per_prompt=num_images_per_prompt,
            add_predicted_noise=add_predicted_noise,
            eta=eta,
            generator=generator,
            prompt_embeds=prompt_embeds,
            negative_prompt_embeds=negative_prompt_embeds,
            max_embeddings_multiples=max_embeddings_multiples,
            output_type=output_type,
            return_dict=return_dict,
            callback=callback,
            is_cancelled_callback=is_cancelled_callback,
            callback_steps=callback_steps,
            cross_attention_kwargs=cross_attention_kwargs,
        )