diffusers/community-pipelines-mirror

1# Copyright 2024 The HuggingFace Team. All rights reserved.2#3# Licensed under the Apache License, Version 2.0 (the "License");4# you may not use this file except in compliance with the License.5# You may obtain a copy of the License at6#7#     http://www.apache.org/licenses/LICENSE-2.08#9# Unless required by applicable law or agreed to in writing, software10# distributed under the License is distributed on an "AS IS" BASIS,11# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.12# See the License for the specific language governing permissions and13# limitations under the License.14 15import inspect16from typing import Any, Callable, Dict, List, Optional, Union17 18import intel_extension_for_pytorch as ipex19import torch20from packaging import version21from transformers import CLIPFeatureExtractor, CLIPTextModel, CLIPTokenizer22 23from diffusers.configuration_utils import FrozenDict24from diffusers.loaders import LoraLoaderMixin, TextualInversionLoaderMixin25from diffusers.models import AutoencoderKL, UNet2DConditionModel26from diffusers.pipelines.pipeline_utils import DiffusionPipeline, StableDiffusionMixin27from diffusers.pipelines.stable_diffusion import StableDiffusionPipelineOutput28from diffusers.pipelines.stable_diffusion.safety_checker import StableDiffusionSafetyChecker29from diffusers.schedulers import KarrasDiffusionSchedulers30from diffusers.utils import (31    deprecate,32    logging,33    replace_example_docstring,34)35from diffusers.utils.torch_utils import randn_tensor36 37 38logger = logging.get_logger(__name__)  # pylint: disable=invalid-name39 40EXAMPLE_DOC_STRING = """41    Examples:42        ```py43        >>> import torch44        >>> from diffusers import StableDiffusionPipeline45 46        >>> pipe = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", custom_pipeline="stable_diffusion_ipex")47 48        >>> # For Float3249        >>> pipe.prepare_for_ipex(prompt, dtype=torch.float32, height=512, width=512) #value of image height/width should be consistent with the pipeline inference50        >>> # For BFloat1651        >>> pipe.prepare_for_ipex(prompt, dtype=torch.bfloat16, height=512, width=512) #value of image height/width should be consistent with the pipeline inference52 53        >>> prompt = "a photo of an astronaut riding a horse on mars"54        >>> # For Float3255        >>> image = pipe(prompt, num_inference_steps=num_inference_steps, height=512, width=512).images[0] #value of image height/width should be consistent with 'prepare_for_ipex()'56        >>> # For BFloat1657        >>> with torch.cpu.amp.autocast(enabled=True, dtype=torch.bfloat16):58        >>>     image = pipe(prompt, num_inference_steps=num_inference_steps, height=512, width=512).images[0] #value of image height/width should be consistent with 'prepare_for_ipex()'59        ```60"""61 62 63class StableDiffusionIPEXPipeline(64    DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin, LoraLoaderMixin65):66    r"""67    Pipeline for text-to-image generation using Stable Diffusion on IPEX.68 69    This model inherits from [`DiffusionPipeline`]. Check the superclass documentation for the generic methods the70    library implements for all the pipelines (such as downloading or saving, running on a particular device, etc.)71 72    Args:73        vae ([`AutoencoderKL`]):74            Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations.75        text_encoder ([`CLIPTextModel`]):76            Frozen text-encoder. Stable Diffusion uses the text portion of77            [CLIP](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTextModel), specifically78            the [clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14) variant.79        tokenizer (`CLIPTokenizer`):80            Tokenizer of class81            [CLIPTokenizer](https://huggingface.co/docs/transformers/v4.21.0/en/model_doc/clip#transformers.CLIPTokenizer).82        unet ([`UNet2DConditionModel`]): Conditional U-Net architecture to denoise the encoded image latents.83        scheduler ([`SchedulerMixin`]):84            A scheduler to be used in combination with `unet` to denoise the encoded image latents. Can be one of85            [`DDIMScheduler`], [`LMSDiscreteScheduler`], or [`PNDMScheduler`].86        safety_checker ([`StableDiffusionSafetyChecker`]):87            Classification module that estimates whether generated images could be considered offensive or harmful.88            Please, refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for details.89        feature_extractor ([`CLIPFeatureExtractor`]):90            Model that extracts features from generated images to be used as inputs for the `safety_checker`.91    """92 93    _optional_components = ["safety_checker", "feature_extractor"]94 95    def __init__(96        self,97        vae: AutoencoderKL,98        text_encoder: CLIPTextModel,99        tokenizer: CLIPTokenizer,100        unet: UNet2DConditionModel,101        scheduler: KarrasDiffusionSchedulers,102        safety_checker: StableDiffusionSafetyChecker,103        feature_extractor: CLIPFeatureExtractor,104        requires_safety_checker: bool = True,105    ):106        super().__init__()107 108        if hasattr(scheduler.config, "steps_offset") and scheduler.config.steps_offset != 1:109            deprecation_message = (110                f"The configuration file of this scheduler: {scheduler} is outdated. `steps_offset`"111                f" should be set to 1 instead of {scheduler.config.steps_offset}. Please make sure "112                "to update the config accordingly as leaving `steps_offset` might led to incorrect results"113                " in future versions. If you have downloaded this checkpoint from the Hugging Face Hub,"114                " it would be very nice if you could open a Pull request for the `scheduler/scheduler_config.json`"115                " file"116            )117            deprecate("steps_offset!=1", "1.0.0", deprecation_message, standard_warn=False)118            new_config = dict(scheduler.config)119            new_config["steps_offset"] = 1120            scheduler._internal_dict = FrozenDict(new_config)121 122        if hasattr(scheduler.config, "clip_sample") and scheduler.config.clip_sample is True:123            deprecation_message = (124                f"The configuration file of this scheduler: {scheduler} has not set the configuration `clip_sample`."125                " `clip_sample` should be set to False in the configuration file. Please make sure to update the"126                " config accordingly as not setting `clip_sample` in the config might lead to incorrect results in"127                " future versions. If you have downloaded this checkpoint from the Hugging Face Hub, it would be very"128                " nice if you could open a Pull request for the `scheduler/scheduler_config.json` file"129            )130            deprecate("clip_sample not set", "1.0.0", deprecation_message, standard_warn=False)131            new_config = dict(scheduler.config)132            new_config["clip_sample"] = False133            scheduler._internal_dict = FrozenDict(new_config)134 135        if safety_checker is None and requires_safety_checker:136            logger.warning(137                f"You have disabled the safety checker for {self.__class__} by passing `safety_checker=None`. Ensure"138                " that you abide to the conditions of the Stable Diffusion license and do not expose unfiltered"139                " results in services or applications open to the public. Both the diffusers team and Hugging Face"140                " strongly recommend to keep the safety filter enabled in all public facing circumstances, disabling"141                " it only for use-cases that involve analyzing network behavior or auditing its results. For more"142                " information, please have a look at https://github.com/huggingface/diffusers/pull/254 ."143            )144 145        if safety_checker is not None and feature_extractor is None:146            raise ValueError(147                "Make sure to define a feature extractor when loading {self.__class__} if you want to use the safety"148                " checker. If you do not want to use the safety checker, you can pass `'safety_checker=None'` instead."149            )150 151        is_unet_version_less_0_9_0 = hasattr(unet.config, "_diffusers_version") and version.parse(152            version.parse(unet.config._diffusers_version).base_version153        ) < version.parse("0.9.0.dev0")154        is_unet_sample_size_less_64 = hasattr(unet.config, "sample_size") and unet.config.sample_size < 64155        if is_unet_version_less_0_9_0 and is_unet_sample_size_less_64:156            deprecation_message = (157                "The configuration file of the unet has set the default `sample_size` to smaller than"158                " 64 which seems highly unlikely. If your checkpoint is a fine-tuned version of any of the"159                " following: \n- CompVis/stable-diffusion-v1-4 \n- CompVis/stable-diffusion-v1-3 \n-"160                " CompVis/stable-diffusion-v1-2 \n- CompVis/stable-diffusion-v1-1 \n- runwayml/stable-diffusion-v1-5"161                " \n- runwayml/stable-diffusion-inpainting \n you should change 'sample_size' to 64 in the"162                " configuration file. Please make sure to update the config accordingly as leaving `sample_size=32`"163                " in the config might lead to incorrect results in future versions. If you have downloaded this"164                " checkpoint from the Hugging Face Hub, it would be very nice if you could open a Pull request for"165                " the `unet/config.json` file"166            )167            deprecate("sample_size<64", "1.0.0", deprecation_message, standard_warn=False)168            new_config = dict(unet.config)169            new_config["sample_size"] = 64170            unet._internal_dict = FrozenDict(new_config)171 172        self.register_modules(173            vae=vae,174            text_encoder=text_encoder,175            tokenizer=tokenizer,176            unet=unet,177            scheduler=scheduler,178            safety_checker=safety_checker,179            feature_extractor=feature_extractor,180        )181        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)182        self.register_to_config(requires_safety_checker=requires_safety_checker)183 184    def get_input_example(self, prompt, height=None, width=None, guidance_scale=7.5, num_images_per_prompt=1):185        prompt_embeds = None186        negative_prompt_embeds = None187        negative_prompt = None188        callback_steps = 1189        generator = None190        latents = None191 192        # 0. Default height and width to unet193        height = height or self.unet.config.sample_size * self.vae_scale_factor194        width = width or self.unet.config.sample_size * self.vae_scale_factor195 196        # 1. Check inputs. Raise error if not correct197        self.check_inputs(198            prompt, height, width, callback_steps, negative_prompt, prompt_embeds, negative_prompt_embeds199        )200 201        # 2. Define call parameters202        if prompt is not None and isinstance(prompt, str):203            batch_size = 1204        elif prompt is not None and isinstance(prompt, list):205            batch_size = len(prompt)206 207        device = "cpu"208        # here `guidance_scale` is defined analog to the guidance weight `w` of equation (2)209        # of the Imagen paper: https://arxiv.org/pdf/2205.11487.pdf . `guidance_scale = 1`210        # corresponds to doing no classifier free guidance.211        do_classifier_free_guidance = guidance_scale > 1.0212 213        # 3. Encode input prompt214        prompt_embeds = self._encode_prompt(215            prompt,216            device,217            num_images_per_prompt,218            do_classifier_free_guidance,219            negative_prompt,220            prompt_embeds=prompt_embeds,221            negative_prompt_embeds=negative_prompt_embeds,222        )223 224        # 5. Prepare latent variables225        latents = self.prepare_latents(226            batch_size * num_images_per_prompt,227            self.unet.config.in_channels,228            height,229            width,230            prompt_embeds.dtype,231            device,232            generator,233            latents,234        )235        dummy = torch.ones(1, dtype=torch.int32)236        latent_model_input = torch.cat([latents] * 2) if do_classifier_free_guidance else latents237        latent_model_input = self.scheduler.scale_model_input(latent_model_input, dummy)238 239        unet_input_example = (latent_model_input, dummy, prompt_embeds)240        vae_decoder_input_example = latents241 242        return unet_input_example, vae_decoder_input_example243 244    def prepare_for_ipex(self, promt, dtype=torch.float32, height=None, width=None, guidance_scale=7.5):245        self.unet = self.unet.to(memory_format=torch.channels_last)246        self.vae.decoder = self.vae.decoder.to(memory_format=torch.channels_last)247        self.text_encoder = self.text_encoder.to(memory_format=torch.channels_last)248        if self.safety_checker is not None:249            self.safety_checker = self.safety_checker.to(memory_format=torch.channels_last)250 251        unet_input_example, vae_decoder_input_example = self.get_input_example(promt, height, width, guidance_scale)252 253        # optimize with ipex254        if dtype == torch.bfloat16:255            self.unet = ipex.optimize(self.unet.eval(), dtype=torch.bfloat16, inplace=True)256            self.vae.decoder = ipex.optimize(self.vae.decoder.eval(), dtype=torch.bfloat16, inplace=True)257            self.text_encoder = ipex.optimize(self.text_encoder.eval(), dtype=torch.bfloat16, inplace=True)258            if self.safety_checker is not None:259                self.safety_checker = ipex.optimize(self.safety_checker.eval(), dtype=torch.bfloat16, inplace=True)260        elif dtype == torch.float32:261            self.unet = ipex.optimize(262                self.unet.eval(),263                dtype=torch.float32,264                inplace=True,265                weights_prepack=True,266                auto_kernel_selection=False,267            )268            self.vae.decoder = ipex.optimize(269                self.vae.decoder.eval(),270                dtype=torch.float32,271                inplace=True,272                weights_prepack=True,273                auto_kernel_selection=False,274            )275            self.text_encoder = ipex.optimize(276                self.text_encoder.eval(),277                dtype=torch.float32,278                inplace=True,279                weights_prepack=True,280                auto_kernel_selection=False,281            )282            if self.safety_checker is not None:283                self.safety_checker = ipex.optimize(284                    self.safety_checker.eval(),285                    dtype=torch.float32,286                    inplace=True,287                    weights_prepack=True,288                    auto_kernel_selection=False,289                )290        else:291            raise ValueError(" The value of 'dtype' should be 'torch.bfloat16' or 'torch.float32' !")292 293        # trace unet model to get better performance on IPEX294        with torch.cpu.amp.autocast(enabled=dtype == torch.bfloat16), torch.no_grad():295            unet_trace_model = torch.jit.trace(self.unet, unet_input_example, check_trace=False, strict=False)296            unet_trace_model = torch.jit.freeze(unet_trace_model)297        self.unet.forward = unet_trace_model.forward298 299        # trace vae.decoder model to get better performance on IPEX300        with torch.cpu.amp.autocast(enabled=dtype == torch.bfloat16), torch.no_grad():301            ave_decoder_trace_model = torch.jit.trace(302                self.vae.decoder, vae_decoder_input_example, check_trace=False, strict=False303            )304            ave_decoder_trace_model = torch.jit.freeze(ave_decoder_trace_model)305        self.vae.decoder.forward = ave_decoder_trace_model.forward306 307    def _encode_prompt(308        self,309        prompt,310        device,311        num_images_per_prompt,312        do_classifier_free_guidance,313        negative_prompt=None,314        prompt_embeds: Optional[torch.Tensor] = None,315        negative_prompt_embeds: Optional[torch.Tensor] = None,316    ):317        r"""318        Encodes the prompt into text encoder hidden states.319 320        Args:321             prompt (`str` or `List[str]`, *optional*):322                prompt to be encoded323            device: (`torch.device`):324                torch device325            num_images_per_prompt (`int`):326                number of images that should be generated per prompt327            do_classifier_free_guidance (`bool`):328                whether to use classifier free guidance or not329            negative_prompt (`str` or `List[str]`, *optional*):330                The prompt or prompts not to guide the image generation. If not defined, one has to pass331                `negative_prompt_embeds`. instead. If not defined, one has to pass `negative_prompt_embeds`. instead.332                Ignored when not using guidance (i.e., ignored if `guidance_scale` is less than `1`).333            prompt_embeds (`torch.Tensor`, *optional*):334                Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not335                provided, text embeddings will be generated from `prompt` input argument.336            negative_prompt_embeds (`torch.Tensor`, *optional*):337                Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt338                weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input339                argument.340        """341        if prompt is not None and isinstance(prompt, str):342            batch_size = 1343        elif prompt is not None and isinstance(prompt, list):344            batch_size = len(prompt)345        else:346            batch_size = prompt_embeds.shape[0]347 348        if prompt_embeds is None:349            # textual inversion: process multi-vector tokens if necessary350            if isinstance(self, TextualInversionLoaderMixin):351                prompt = self.maybe_convert_prompt(prompt, self.tokenizer)352 353            text_inputs = self.tokenizer(354                prompt,355                padding="max_length",356                max_length=self.tokenizer.model_max_length,357                truncation=True,358                return_tensors="pt",359            )360            text_input_ids = text_inputs.input_ids361            untruncated_ids = self.tokenizer(prompt, padding="longest", return_tensors="pt").input_ids362 363            if untruncated_ids.shape[-1] >= text_input_ids.shape[-1] and not torch.equal(364                text_input_ids, untruncated_ids365            ):366                removed_text = self.tokenizer.batch_decode(367                    untruncated_ids[:, self.tokenizer.model_max_length - 1 : -1]368                )369                logger.warning(370                    "The following part of your input was truncated because CLIP can only handle sequences up to"371                    f" {self.tokenizer.model_max_length} tokens: {removed_text}"372                )373 374            if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:375                attention_mask = text_inputs.attention_mask.to(device)376            else:377                attention_mask = None378 379            prompt_embeds = self.text_encoder(380                text_input_ids.to(device),381                attention_mask=attention_mask,382            )383            prompt_embeds = prompt_embeds[0]384 385        prompt_embeds = prompt_embeds.to(dtype=self.text_encoder.dtype, device=device)386 387        bs_embed, seq_len, _ = prompt_embeds.shape388        # duplicate text embeddings for each generation per prompt, using mps friendly method389        prompt_embeds = prompt_embeds.repeat(1, num_images_per_prompt, 1)390        prompt_embeds = prompt_embeds.view(bs_embed * num_images_per_prompt, seq_len, -1)391 392        # get unconditional embeddings for classifier free guidance393        if do_classifier_free_guidance and negative_prompt_embeds is None:394            uncond_tokens: List[str]395            if negative_prompt is None:396                uncond_tokens = [""] * batch_size397            elif type(prompt) is not type(negative_prompt):398                raise TypeError(399                    f"`negative_prompt` should be the same type to `prompt`, but got {type(negative_prompt)} !="400                    f" {type(prompt)}."401                )402            elif isinstance(negative_prompt, str):403                uncond_tokens = [negative_prompt]404            elif batch_size != len(negative_prompt):405                raise ValueError(406                    f"`negative_prompt`: {negative_prompt} has batch size {len(negative_prompt)}, but `prompt`:"407                    f" {prompt} has batch size {batch_size}. Please make sure that passed `negative_prompt` matches"408                    " the batch size of `prompt`."409                )410            else:411                uncond_tokens = negative_prompt412 413            # textual inversion: process multi-vector tokens if necessary414            if isinstance(self, TextualInversionLoaderMixin):415                uncond_tokens = self.maybe_convert_prompt(uncond_tokens, self.tokenizer)416 417            max_length = prompt_embeds.shape[1]418            uncond_input = self.tokenizer(419                uncond_tokens,420                padding="max_length",421                max_length=max_length,422                truncation=True,423                return_tensors="pt",424            )425 426            if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:427                attention_mask = uncond_input.attention_mask.to(device)428            else:429                attention_mask = None430 431            negative_prompt_embeds = self.text_encoder(432                uncond_input.input_ids.to(device),433                attention_mask=attention_mask,434            )435            negative_prompt_embeds = negative_prompt_embeds[0]436 437        if do_classifier_free_guidance:438            # duplicate unconditional embeddings for each generation per prompt, using mps friendly method439            seq_len = negative_prompt_embeds.shape[1]440 441            negative_prompt_embeds = negative_prompt_embeds.to(dtype=self.text_encoder.dtype, device=device)442 443            negative_prompt_embeds = negative_prompt_embeds.repeat(1, num_images_per_prompt, 1)444            negative_prompt_embeds = negative_prompt_embeds.view(batch_size * num_images_per_prompt, seq_len, -1)445 446            # For classifier free guidance, we need to do two forward passes.447            # Here we concatenate the unconditional and text embeddings into a single batch448            # to avoid doing two forward passes449            prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds])450 451        return prompt_embeds452 453    def run_safety_checker(self, image, device, dtype):454        if self.safety_checker is not None:455            safety_checker_input = self.feature_extractor(self.numpy_to_pil(image), return_tensors="pt").to(device)456            image, has_nsfw_concept = self.safety_checker(457                images=image, clip_input=safety_checker_input.pixel_values.to(dtype)458            )459        else:460            has_nsfw_concept = None461        return image, has_nsfw_concept462 463    def decode_latents(self, latents):464        latents = 1 / self.vae.config.scaling_factor * latents465        image = self.vae.decode(latents).sample466        image = (image / 2 + 0.5).clamp(0, 1)467        # we always cast to float32 as this does not cause significant overhead and is compatible with bfloat16468        image = image.cpu().permute(0, 2, 3, 1).float().numpy()469        return image470 471    def prepare_extra_step_kwargs(self, generator, eta):472        # prepare extra kwargs for the scheduler step, since not all schedulers have the same signature473        # eta (η) is only used with the DDIMScheduler, it will be ignored for other schedulers.474        # eta corresponds to η in DDIM paper: https://arxiv.org/abs/2010.02502475        # and should be between [0, 1]476 477        accepts_eta = "eta" in set(inspect.signature(self.scheduler.step).parameters.keys())478        extra_step_kwargs = {}479        if accepts_eta:480            extra_step_kwargs["eta"] = eta481 482        # check if the scheduler accepts generator483        accepts_generator = "generator" in set(inspect.signature(self.scheduler.step).parameters.keys())484        if accepts_generator:485            extra_step_kwargs["generator"] = generator486        return extra_step_kwargs487 488    def check_inputs(489        self,490        prompt,491        height,492        width,493        callback_steps,494        negative_prompt=None,495        prompt_embeds=None,496        negative_prompt_embeds=None,497    ):498        if height % 8 != 0 or width % 8 != 0:499            raise ValueError(f"`height` and `width` have to be divisible by 8 but are {height} and {width}.")500 501        if (callback_steps is None) or (502            callback_steps is not None and (not isinstance(callback_steps, int) or callback_steps <= 0)503        ):504            raise ValueError(505                f"`callback_steps` has to be a positive integer but is {callback_steps} of type"506                f" {type(callback_steps)}."507            )508 509        if prompt is not None and prompt_embeds is not None:510            raise ValueError(511                f"Cannot forward both `prompt`: {prompt} and `prompt_embeds`: {prompt_embeds}. Please make sure to"512                " only forward one of the two."513            )514        elif prompt is None and prompt_embeds is None:515            raise ValueError(516                "Provide either `prompt` or `prompt_embeds`. Cannot leave both `prompt` and `prompt_embeds` undefined."517            )518        elif prompt is not None and (not isinstance(prompt, str) and not isinstance(prompt, list)):519            raise ValueError(f"`prompt` has to be of type `str` or `list` but is {type(prompt)}")520 521        if negative_prompt is not None and negative_prompt_embeds is not None:522            raise ValueError(523                f"Cannot forward both `negative_prompt`: {negative_prompt} and `negative_prompt_embeds`:"524                f" {negative_prompt_embeds}. Please make sure to only forward one of the two."525            )526 527        if prompt_embeds is not None and negative_prompt_embeds is not None:528            if prompt_embeds.shape != negative_prompt_embeds.shape:529                raise ValueError(530                    "`prompt_embeds` and `negative_prompt_embeds` must have the same shape when passed directly, but"531                    f" got: `prompt_embeds` {prompt_embeds.shape} != `negative_prompt_embeds`"532                    f" {negative_prompt_embeds.shape}."533                )534 535    def prepare_latents(self, batch_size, num_channels_latents, height, width, dtype, device, generator, latents=None):536        shape = (537            batch_size,538            num_channels_latents,539            int(height) // self.vae_scale_factor,540            int(width) // self.vae_scale_factor,541        )542        if isinstance(generator, list) and len(generator) != batch_size:543            raise ValueError(544                f"You have passed a list of generators of length {len(generator)}, but requested an effective batch"545                f" size of {batch_size}. Make sure the batch size matches the length of the generators."546            )547 548        if latents is None:549            latents = randn_tensor(shape, generator=generator, device=device, dtype=dtype)550        else:551            latents = latents.to(device)552 553        # scale the initial noise by the standard deviation required by the scheduler554        latents = latents * self.scheduler.init_noise_sigma555        return latents556 557    @torch.no_grad()558    @replace_example_docstring(EXAMPLE_DOC_STRING)559    def __call__(560        self,561        prompt: Union[str, List[str]] = None,562        height: Optional[int] = None,563        width: Optional[int] = None,564        num_inference_steps: int = 50,565        guidance_scale: float = 7.5,566        negative_prompt: Optional[Union[str, List[str]]] = None,567        num_images_per_prompt: Optional[int] = 1,568        eta: float = 0.0,569        generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,570        latents: Optional[torch.Tensor] = None,571        prompt_embeds: Optional[torch.Tensor] = None,572        negative_prompt_embeds: Optional[torch.Tensor] = None,573        output_type: Optional[str] = "pil",574        return_dict: bool = True,575        callback: Optional[Callable[[int, int, torch.Tensor], None]] = None,576        callback_steps: int = 1,577        cross_attention_kwargs: Optional[Dict[str, Any]] = None,578    ):579        r"""580        Function invoked when calling the pipeline for generation.581 582        Args:583            prompt (`str` or `List[str]`, *optional*):584                The prompt or prompts to guide the image generation. If not defined, one has to pass `prompt_embeds`.585                instead.586            height (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):587                The height in pixels of the generated image.588            width (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):589                The width in pixels of the generated image.590            num_inference_steps (`int`, *optional*, defaults to 50):591                The number of denoising steps. More denoising steps usually lead to a higher quality image at the592                expense of slower inference.593            guidance_scale (`float`, *optional*, defaults to 7.5):594                Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).595                `guidance_scale` is defined as `w` of equation 2. of [Imagen596                Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >597                1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,598                usually at the expense of lower image quality.599            negative_prompt (`str` or `List[str]`, *optional*):600                The prompt or prompts not to guide the image generation. If not defined, one has to pass601                `negative_prompt_embeds`. instead. If not defined, one has to pass `negative_prompt_embeds`. instead.602                Ignored when not using guidance (i.e., ignored if `guidance_scale` is less than `1`).603            num_images_per_prompt (`int`, *optional*, defaults to 1):604                The number of images to generate per prompt.605            eta (`float`, *optional*, defaults to 0.0):606                Corresponds to parameter eta (η) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to607                [`schedulers.DDIMScheduler`], will be ignored for others.608            generator (`torch.Generator` or `List[torch.Generator]`, *optional*):609                One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)610                to make generation deterministic.611            latents (`torch.Tensor`, *optional*):612                Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image613                generation. Can be used to tweak the same generation with different prompts. If not provided, a latents614                tensor will ge generated by sampling using the supplied random `generator`.615            prompt_embeds (`torch.Tensor`, *optional*):616                Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not617                provided, text embeddings will be generated from `prompt` input argument.618            negative_prompt_embeds (`torch.Tensor`, *optional*):619                Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt620                weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input621                argument.622            output_type (`str`, *optional*, defaults to `"pil"`):623                The output format of the generate image. Choose between624                [PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.625            return_dict (`bool`, *optional*, defaults to `True`):626                Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] instead of a627                plain tuple.628            callback (`Callable`, *optional*):629                A function that will be called every `callback_steps` steps during inference. The function will be630                called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`.631            callback_steps (`int`, *optional*, defaults to 1):632                The frequency at which the `callback` function will be called. If not specified, the callback will be633                called at every step.634            cross_attention_kwargs (`dict`, *optional*):635                A kwargs dictionary that if specified is passed along to the `AttnProcessor` as defined under636                `self.processor` in637                [diffusers.models.attention_processor](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).638 639        Examples:640 641        Returns:642            [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] or `tuple`:643            [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] if `return_dict` is True, otherwise a `tuple.644            When returning a tuple, the first element is a list with the generated images, and the second element is a645            list of `bool`s denoting whether the corresponding generated image likely represents "not-safe-for-work"646            (nsfw) content, according to the `safety_checker`.647        """648        # 0. Default height and width to unet649        height = height or self.unet.config.sample_size * self.vae_scale_factor650        width = width or self.unet.config.sample_size * self.vae_scale_factor651 652        # 1. Check inputs. Raise error if not correct653        self.check_inputs(654            prompt, height, width, callback_steps, negative_prompt, prompt_embeds, negative_prompt_embeds655        )656 657        # 2. Define call parameters658        if prompt is not None and isinstance(prompt, str):659            batch_size = 1660        elif prompt is not None and isinstance(prompt, list):661            batch_size = len(prompt)662        else:663            batch_size = prompt_embeds.shape[0]664 665        device = self._execution_device666        # here `guidance_scale` is defined analog to the guidance weight `w` of equation (2)667        # of the Imagen paper: https://arxiv.org/pdf/2205.11487.pdf . `guidance_scale = 1`668        # corresponds to doing no classifier free guidance.669        do_classifier_free_guidance = guidance_scale > 1.0670 671        # 3. Encode input prompt672        prompt_embeds = self._encode_prompt(673            prompt,674            device,675            num_images_per_prompt,676            do_classifier_free_guidance,677            negative_prompt,678            prompt_embeds=prompt_embeds,679            negative_prompt_embeds=negative_prompt_embeds,680        )681 682        # 4. Prepare timesteps683        self.scheduler.set_timesteps(num_inference_steps, device=device)684        timesteps = self.scheduler.timesteps685 686        # 5. Prepare latent variables687        num_channels_latents = self.unet.config.in_channels688        latents = self.prepare_latents(689            batch_size * num_images_per_prompt,690            num_channels_latents,691            height,692            width,693            prompt_embeds.dtype,694            device,695            generator,696            latents,697        )698 699        # 6. Prepare extra step kwargs. TODO: Logic should ideally just be moved out of the pipeline700        extra_step_kwargs = self.prepare_extra_step_kwargs(generator, eta)701 702        # 7. Denoising loop703        num_warmup_steps = len(timesteps) - num_inference_steps * self.scheduler.order704        with self.progress_bar(total=num_inference_steps) as progress_bar:705            for i, t in enumerate(timesteps):706                # expand the latents if we are doing classifier free guidance707                latent_model_input = torch.cat([latents] * 2) if do_classifier_free_guidance else latents708                latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)709 710                # predict the noise residual711                noise_pred = self.unet(latent_model_input, t, encoder_hidden_states=prompt_embeds)["sample"]712 713                # perform guidance714                if do_classifier_free_guidance:715                    noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)716                    noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)717 718                # compute the previous noisy sample x_t -> x_t-1719                latents = self.scheduler.step(noise_pred, t, latents, **extra_step_kwargs).prev_sample720 721                # call the callback, if provided722                if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0):723                    progress_bar.update()724                    if callback is not None and i % callback_steps == 0:725                        step_idx = i // getattr(self.scheduler, "order", 1)726                        callback(step_idx, t, latents)727 728        if output_type == "latent":729            image = latents730            has_nsfw_concept = None731        elif output_type == "pil":732            # 8. Post-processing733            image = self.decode_latents(latents)734 735            # 9. Run safety checker736            image, has_nsfw_concept = self.run_safety_checker(image, device, prompt_embeds.dtype)737 738            # 10. Convert to PIL739            image = self.numpy_to_pil(image)740        else:741            # 8. Post-processing742            image = self.decode_latents(latents)743 744            # 9. Run safety checker745            image, has_nsfw_concept = self.run_safety_checker(image, device, prompt_embeds.dtype)746 747        # Offload last model to CPU748        if hasattr(self, "final_offload_hook") and self.final_offload_hook is not None:749            self.final_offload_hook.offload()750 751        if not return_dict:752            return (image, has_nsfw_concept)753 754        return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=has_nsfw_concept)755