# Copyright 2024 ParaDiGMS authors and The HuggingFace Team. All rights reserved. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. import inspect from typing import Any, Callable, Dict, List, Optional, Union import torch from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer from ....image_processor import VaeImageProcessor from ....loaders import FromSingleFileMixin, LoraLoaderMixin, TextualInversionLoaderMixin from ....models import AutoencoderKL, UNet2DConditionModel from ....models.lora import adjust_lora_scale_text_encoder from ....schedulers import KarrasDiffusionSchedulers from ....utils import ( USE_PEFT_BACKEND, deprecate, logging, replace_example_docstring, scale_lora_layers, unscale_lora_layers, ) from ....utils.torch_utils import randn_tensor from ...pipeline_utils import DiffusionPipeline, StableDiffusionMixin from ...stable_diffusion.pipeline_output import StableDiffusionPipelineOutput from ...stable_diffusion.safety_checker import StableDiffusionSafetyChecker logger = logging.get_logger(__name__) # pylint: disable=invalid-name EXAMPLE_DOC_STRING = """ Examples: ```py >>> import torch >>> from diffusers import DDPMParallelScheduler >>> from diffusers import StableDiffusionParadigmsPipeline >>> scheduler = DDPMParallelScheduler.from_pretrained("runwayml/stable-diffusion-v1-5", subfolder="scheduler") >>> pipe = StableDiffusionParadigmsPipeline.from_pretrained( ... "runwayml/stable-diffusion-v1-5", scheduler=scheduler, torch_dtype=torch.float16 ... ) >>> pipe = pipe.to("cuda") >>> ngpu, batch_per_device = torch.cuda.device_count(), 5 >>> pipe.wrapped_unet = torch.nn.DataParallel(pipe.unet, device_ids=[d for d in range(ngpu)]) >>> prompt = "a photo of an astronaut riding a horse on mars" >>> image = pipe(prompt, parallel=ngpu * batch_per_device, num_inference_steps=1000).images[0] ``` """ class StableDiffusionParadigmsPipeline( DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin, LoraLoaderMixin, FromSingleFileMixin ): r""" Pipeline for text-to-image generation using a parallelized version of Stable Diffusion. This model inherits from [`DiffusionPipeline`]. Check the superclass documentation for the generic methods implemented for all pipelines (downloading, saving, running on a particular device, etc.). The pipeline also inherits the following loading methods: - [`~loaders.TextualInversionLoaderMixin.load_textual_inversion`] for loading textual inversion embeddings - [`~loaders.LoraLoaderMixin.load_lora_weights`] for loading LoRA weights - [`~loaders.LoraLoaderMixin.save_lora_weights`] for saving LoRA weights - [`~loaders.FromSingleFileMixin.from_single_file`] for loading `.ckpt` files Args: vae ([`AutoencoderKL`]): Variational Auto-Encoder (VAE) model to encode and decode images to and from latent representations. text_encoder ([`~transformers.CLIPTextModel`]): Frozen text-encoder ([clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14)). tokenizer ([`~transformers.CLIPTokenizer`]): A `CLIPTokenizer` to tokenize text. unet ([`UNet2DConditionModel`]): A `UNet2DConditionModel` to denoise the encoded image latents. scheduler ([`SchedulerMixin`]): A scheduler to be used in combination with `unet` to denoise the encoded image latents. Can be one of [`DDIMScheduler`], [`LMSDiscreteScheduler`], or [`PNDMScheduler`]. safety_checker ([`StableDiffusionSafetyChecker`]): Classification module that estimates whether generated images could be considered offensive or harmful. Please refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for more details about a model's potential harms. feature_extractor ([`~transformers.CLIPImageProcessor`]): A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`. """ model_cpu_offload_seq = "text_encoder->unet->vae" _optional_components = ["safety_checker", "feature_extractor"] _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 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." ) 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) # attribute to wrap the unet with torch.nn.DataParallel when running multiple denoising steps on multiple GPUs self.wrapped_unet = self.unet # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline._encode_prompt def _encode_prompt( self, prompt, device, num_images_per_prompt, do_classifier_free_guidance, negative_prompt=None, prompt_embeds: Optional[torch.Tensor] = None, negative_prompt_embeds: Optional[torch.Tensor] = None, lora_scale: Optional[float] = None, **kwargs, ): deprecation_message = "`_encode_prompt()` is deprecated and it will be removed in a future version. Use `encode_prompt()` instead. Also, be aware that the output format changed from a concatenated tensor to a tuple." deprecate("_encode_prompt()", "1.0.0", deprecation_message, standard_warn=False) prompt_embeds_tuple = self.encode_prompt( prompt=prompt, device=device, num_images_per_prompt=num_images_per_prompt, do_classifier_free_guidance=do_classifier_free_guidance, negative_prompt=negative_prompt, prompt_embeds=prompt_embeds, negative_prompt_embeds=negative_prompt_embeds, lora_scale=lora_scale, **kwargs, ) # concatenate for backwards comp prompt_embeds = torch.cat([prompt_embeds_tuple[1], prompt_embeds_tuple[0]]) return prompt_embeds # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.encode_prompt def encode_prompt( self, prompt, device, num_images_per_prompt, do_classifier_free_guidance, negative_prompt=None, prompt_embeds: Optional[torch.Tensor] = None, negative_prompt_embeds: Optional[torch.Tensor] = None, lora_scale: Optional[float] = None, clip_skip: Optional[int] = None, ): r""" Encodes the prompt into text encoder hidden states. Args: prompt (`str` or `List[str]`, *optional*): prompt to be encoded device: (`torch.device`): torch device num_images_per_prompt (`int`): number of images that should be generated per prompt do_classifier_free_guidance (`bool`): whether to use classifier free guidance or not negative_prompt (`str` or `List[str]`, *optional*): The prompt or prompts not to guide the image generation. If not defined, one has to pass `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is less than `1`). prompt_embeds (`torch.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. lora_scale (`float`, *optional*): A LoRA scale that will be applied to all LoRA layers of the text encoder if LoRA layers are loaded. clip_skip (`int`, *optional*): Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that the output of the pre-final layer will be used for computing the prompt embeddings. """ # set lora scale so that monkey patched LoRA # function of text encoder can correctly access it if lora_scale is not None and isinstance(self, LoraLoaderMixin): self._lora_scale = lora_scale # dynamically adjust the LoRA scale if not USE_PEFT_BACKEND: adjust_lora_scale_text_encoder(self.text_encoder, lora_scale) else: scale_lora_layers(self.text_encoder, lora_scale) if prompt is not None and isinstance(prompt, str): batch_size = 1 elif prompt is not None and isinstance(prompt, list): batch_size = len(prompt) else: batch_size = prompt_embeds.shape[0] if prompt_embeds is None: # textual inversion: process multi-vector tokens if necessary if isinstance(self, TextualInversionLoaderMixin): prompt = self.maybe_convert_prompt(prompt, self.tokenizer) text_inputs = self.tokenizer( prompt, padding="max_length", max_length=self.tokenizer.model_max_length, truncation=True, return_tensors="pt", ) text_input_ids = text_inputs.input_ids untruncated_ids = self.tokenizer(prompt, padding="longest", return_tensors="pt").input_ids if untruncated_ids.shape[-1] >= text_input_ids.shape[-1] and not torch.equal( text_input_ids, untruncated_ids ): removed_text = self.tokenizer.batch_decode( untruncated_ids[:, self.tokenizer.model_max_length - 1 : -1] ) logger.warning( "The following part of your input was truncated because CLIP can only handle sequences up to" f" {self.tokenizer.model_max_length} tokens: {removed_text}" ) if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask: attention_mask = text_inputs.attention_mask.to(device) else: attention_mask = None if clip_skip is None: prompt_embeds = self.text_encoder(text_input_ids.to(device), attention_mask=attention_mask) prompt_embeds = prompt_embeds[0] else: prompt_embeds = self.text_encoder( text_input_ids.to(device), attention_mask=attention_mask, output_hidden_states=True ) # Access the `hidden_states` first, that contains a tuple of # all the hidden states from the encoder layers. Then index into # the tuple to access the hidden states from the desired layer. prompt_embeds = prompt_embeds[-1][-(clip_skip + 1)] # We also need to apply the final LayerNorm here to not mess with the # representations. The `last_hidden_states` that we typically use for # obtaining the final prompt representations passes through the LayerNorm # layer. prompt_embeds = self.text_encoder.text_model.final_layer_norm(prompt_embeds) if self.text_encoder is not None: prompt_embeds_dtype = self.text_encoder.dtype elif self.unet is not None: prompt_embeds_dtype = self.unet.dtype else: prompt_embeds_dtype = prompt_embeds.dtype prompt_embeds = prompt_embeds.to(dtype=prompt_embeds_dtype, device=device) bs_embed, seq_len, _ = prompt_embeds.shape # duplicate text embeddings for each generation per prompt, using mps friendly method prompt_embeds = prompt_embeds.repeat(1, num_images_per_prompt, 1) prompt_embeds = prompt_embeds.view(bs_embed * num_images_per_prompt, seq_len, -1) # get unconditional embeddings for classifier free guidance if do_classifier_free_guidance and negative_prompt_embeds is None: uncond_tokens: List[str] if negative_prompt is None: uncond_tokens = [""] * batch_size elif prompt is not None and type(prompt) is not type(negative_prompt): raise TypeError( f"`negative_prompt` should be the same type to `prompt`, but got {type(negative_prompt)} !=" f" {type(prompt)}." ) elif isinstance(negative_prompt, str): uncond_tokens = [negative_prompt] elif batch_size != len(negative_prompt): raise ValueError( f"`negative_prompt`: {negative_prompt} has batch size {len(negative_prompt)}, but `prompt`:" f" {prompt} has batch size {batch_size}. Please make sure that passed `negative_prompt` matches" " the batch size of `prompt`." ) else: uncond_tokens = negative_prompt # textual inversion: process multi-vector tokens if necessary if isinstance(self, TextualInversionLoaderMixin): uncond_tokens = self.maybe_convert_prompt(uncond_tokens, self.tokenizer) max_length = prompt_embeds.shape[1] uncond_input = self.tokenizer( uncond_tokens, padding="max_length", max_length=max_length, truncation=True, return_tensors="pt", ) if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask: attention_mask = uncond_input.attention_mask.to(device) else: attention_mask = None negative_prompt_embeds = self.text_encoder( uncond_input.input_ids.to(device), attention_mask=attention_mask, ) negative_prompt_embeds = negative_prompt_embeds[0] if do_classifier_free_guidance: # duplicate unconditional embeddings for each generation per prompt, using mps friendly method seq_len = negative_prompt_embeds.shape[1] negative_prompt_embeds = negative_prompt_embeds.to(dtype=prompt_embeds_dtype, device=device) negative_prompt_embeds = negative_prompt_embeds.repeat(1, num_images_per_prompt, 1) negative_prompt_embeds = negative_prompt_embeds.view(batch_size * num_images_per_prompt, seq_len, -1) if self.text_encoder is not None: if isinstance(self, LoraLoaderMixin) and USE_PEFT_BACKEND: # Retrieve the original scale by scaling back the LoRA layers unscale_lora_layers(self.text_encoder, lora_scale) return prompt_embeds, negative_prompt_embeds # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.run_safety_checker def run_safety_checker(self, image, device, dtype): if self.safety_checker is None: has_nsfw_concept = None else: if torch.is_tensor(image): feature_extractor_input = self.image_processor.postprocess(image, output_type="pil") else: feature_extractor_input = self.image_processor.numpy_to_pil(image) safety_checker_input = self.feature_extractor(feature_extractor_input, return_tensors="pt").to(device) image, has_nsfw_concept = self.safety_checker( images=image, clip_input=safety_checker_input.pixel_values.to(dtype) ) return image, has_nsfw_concept # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_extra_step_kwargs def prepare_extra_step_kwargs(self, generator, eta): # prepare extra kwargs for the scheduler step, since not all schedulers have the same signature # eta (η) is only used with the DDIMScheduler, it will be ignored for other schedulers. # eta corresponds to η in DDIM paper: https://arxiv.org/abs/2010.02502 # and should be between [0, 1] accepts_eta = "eta" in set(inspect.signature(self.scheduler.step).parameters.keys()) extra_step_kwargs = {} if accepts_eta: extra_step_kwargs["eta"] = eta # check if the scheduler accepts generator accepts_generator = "generator" in set(inspect.signature(self.scheduler.step).parameters.keys()) if accepts_generator: extra_step_kwargs["generator"] = generator return extra_step_kwargs def check_inputs( self, prompt, height, width, callback_steps, negative_prompt=None, prompt_embeds=None, negative_prompt_embeds=None, callback_on_step_end_tensor_inputs=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 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 callback_on_step_end_tensor_inputs is not None and not all( k in self._callback_tensor_inputs for k in callback_on_step_end_tensor_inputs ): raise ValueError( f"`callback_on_step_end_tensor_inputs` has to be in {self._callback_tensor_inputs}, but found {[k for k in callback_on_step_end_tensor_inputs if k not in self._callback_tensor_inputs]}" ) 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}." ) # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_latents def prepare_latents(self, batch_size, num_channels_latents, height, width, dtype, device, generator, latents=None): shape = ( batch_size, num_channels_latents, int(height) // self.vae_scale_factor, int(width) // self.vae_scale_factor, ) if isinstance(generator, list) and len(generator) != batch_size: raise ValueError( f"You have passed a list of generators of length {len(generator)}, but requested an effective batch" f" size of {batch_size}. Make sure the batch size matches the length of the generators." ) if latents is None: latents = randn_tensor(shape, generator=generator, device=device, dtype=dtype) else: latents = latents.to(device) # scale the initial noise by the standard deviation required by the scheduler latents = latents * self.scheduler.init_noise_sigma return latents def _cumsum(self, input, dim, debug=False): if debug: # cumsum_cuda_kernel does not have a deterministic implementation # so perform cumsum on cpu for debugging purposes return torch.cumsum(input.cpu().float(), dim=dim).to(input.device) else: return torch.cumsum(input, dim=dim) @torch.no_grad() @replace_example_docstring(EXAMPLE_DOC_STRING) def __call__( self, prompt: Union[str, List[str]] = None, height: Optional[int] = None, width: Optional[int] = None, num_inference_steps: int = 50, parallel: int = 10, tolerance: float = 0.1, guidance_scale: float = 7.5, negative_prompt: Optional[Union[str, List[str]]] = None, num_images_per_prompt: Optional[int] = 1, eta: float = 0.0, generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None, latents: Optional[torch.Tensor] = None, prompt_embeds: Optional[torch.Tensor] = None, negative_prompt_embeds: Optional[torch.Tensor] = None, output_type: Optional[str] = "pil", return_dict: bool = True, callback: Optional[Callable[[int, int, torch.Tensor], None]] = None, callback_steps: int = 1, cross_attention_kwargs: Optional[Dict[str, Any]] = None, debug: bool = False, clip_skip: int = None, ): r""" The call function to the pipeline for generation. Args: prompt (`str` or `List[str]`, *optional*): The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`. height (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`): The height in pixels of the generated image. width (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`): 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. parallel (`int`, *optional*, defaults to 10): The batch size to use when doing parallel sampling. More parallelism may lead to faster inference but requires higher memory usage and can also require more total FLOPs. tolerance (`float`, *optional*, defaults to 0.1): The error tolerance for determining when to slide the batch window forward for parallel sampling. Lower tolerance usually leads to less or no degradation. Higher tolerance is faster but can risk degradation of sample quality. The tolerance is specified as a ratio of the scheduler's noise magnitude. guidance_scale (`float`, *optional*, defaults to 7.5): A higher guidance scale value encourages the model to generate images closely linked to the text `prompt` at the expense of lower image quality. Guidance scale is enabled when `guidance_scale > 1`. negative_prompt (`str` or `List[str]`, *optional*): The prompt or prompts to guide what to not include in image generation. If not defined, you need to pass `negative_prompt_embeds` instead. Ignored when not using guidance (`guidance_scale < 1`). num_images_per_prompt (`int`, *optional*, defaults to 1): The number of images to generate per prompt. eta (`float`, *optional*, defaults to 0.0): Corresponds to parameter eta (η) from the [DDIM](https://arxiv.org/abs/2010.02502) paper. Only applies to the [`~schedulers.DDIMScheduler`], and is ignored in other schedulers. generator (`torch.Generator` or `List[torch.Generator]`, *optional*): A [`torch.Generator`](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 is 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 (prompt weighting). If not provided, text embeddings are generated from the `prompt` input argument. negative_prompt_embeds (`torch.Tensor`, *optional*): Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument. output_type (`str`, *optional*, defaults to `"pil"`): The output format of the generated image. Choose between `PIL.Image` or `np.array`. return_dict (`bool`, *optional*, defaults to `True`): Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] instead of a plain tuple. callback (`Callable`, *optional*): A function that calls every `callback_steps` steps during inference. The function is called with the following arguments: `callback(step: int, timestep: int, latents: torch.Tensor)`. callback_steps (`int`, *optional*, defaults to 1): The frequency at which the `callback` function is called. If not specified, the callback is called at every step. cross_attention_kwargs (`dict`, *optional*): A kwargs dictionary that if specified is passed along to the [`AttentionProcessor`] as defined in [`self.processor`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py). debug (`bool`, *optional*, defaults to `False`): Whether or not to run in debug mode. In debug mode, `torch.cumsum` is evaluated using the CPU. clip_skip (`int`, *optional*): Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that the output of the pre-final layer will be used for computing the prompt embeddings. Examples: Returns: [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] or `tuple`: If `return_dict` is `True`, [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] is returned, otherwise a `tuple` is returned where the first element is a list with the generated images and the second element is a list of `bool`s indicating whether the corresponding generated image contains "not-safe-for-work" (nsfw) content. """ # 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, 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 # 3. Encode input prompt prompt_embeds, negative_prompt_embeds = self.encode_prompt( prompt, device, num_images_per_prompt, do_classifier_free_guidance, negative_prompt, prompt_embeds=prompt_embeds, negative_prompt_embeds=negative_prompt_embeds, clip_skip=clip_skip, ) # For classifier free guidance, we need to do two forward passes. # Here we concatenate the unconditional and text embeddings into a single batch # to avoid doing two forward passes if do_classifier_free_guidance: prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds]) # 4. Prepare timesteps self.scheduler.set_timesteps(num_inference_steps, device=device) # 5. Prepare latent variables num_channels_latents = self.unet.config.in_channels latents = self.prepare_latents( batch_size * num_images_per_prompt, num_channels_latents, height, width, prompt_embeds.dtype, device, generator, latents, ) # 6. 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) extra_step_kwargs.pop("generator", None) # # 7. Denoising loop scheduler = self.scheduler parallel = min(parallel, len(scheduler.timesteps)) begin_idx = 0 end_idx = parallel latents_time_evolution_buffer = torch.stack([latents] * (len(scheduler.timesteps) + 1)) # We must make sure the noise of stochastic schedulers such as DDPM is sampled only once per timestep. # Sampling inside the parallel denoising loop will mess this up, so we pre-sample the noise vectors outside the denoising loop. noise_array = torch.zeros_like(latents_time_evolution_buffer) for j in range(len(scheduler.timesteps)): base_noise = randn_tensor( shape=latents.shape, generator=generator, device=latents.device, dtype=prompt_embeds.dtype ) noise = (self.scheduler._get_variance(scheduler.timesteps[j]) ** 0.5) * base_noise noise_array[j] = noise.clone() # We specify the error tolerance as a ratio of the scheduler's noise magnitude. We similarly compute the error tolerance # outside of the denoising loop to avoid recomputing it at every step. # We will be dividing the norm of the noise, so we store its inverse here to avoid a division at every step. inverse_variance_norm = 1.0 / torch.tensor( [scheduler._get_variance(scheduler.timesteps[j]) for j in range(len(scheduler.timesteps))] + [0] ).to(noise_array.device) latent_dim = noise_array[0, 0].numel() inverse_variance_norm = inverse_variance_norm[:, None] / latent_dim scaled_tolerance = tolerance**2 with self.progress_bar(total=num_inference_steps) as progress_bar: steps = 0 while begin_idx < len(scheduler.timesteps): # these have shape (parallel_dim, 2*batch_size, ...) # parallel_len is at most parallel, but could be less if we are at the end of the timesteps # we are processing batch window of timesteps spanning [begin_idx, end_idx) parallel_len = end_idx - begin_idx block_prompt_embeds = torch.stack([prompt_embeds] * parallel_len) block_latents = latents_time_evolution_buffer[begin_idx:end_idx] block_t = scheduler.timesteps[begin_idx:end_idx, None].repeat(1, batch_size * num_images_per_prompt) t_vec = block_t if do_classifier_free_guidance: t_vec = t_vec.repeat(1, 2) # expand the latents if we are doing classifier free guidance latent_model_input = ( torch.cat([block_latents] * 2, dim=1) if do_classifier_free_guidance else block_latents ) latent_model_input = self.scheduler.scale_model_input(latent_model_input, t_vec) # if parallel_len is small, no need to use multiple GPUs net = self.wrapped_unet if parallel_len > 3 else self.unet # predict the noise residual, shape is now [parallel_len * 2 * batch_size * num_images_per_prompt, ...] model_output = net( latent_model_input.flatten(0, 1), t_vec.flatten(0, 1), encoder_hidden_states=block_prompt_embeds.flatten(0, 1), cross_attention_kwargs=cross_attention_kwargs, return_dict=False, )[0] per_latent_shape = model_output.shape[1:] if do_classifier_free_guidance: model_output = model_output.reshape( parallel_len, 2, batch_size * num_images_per_prompt, *per_latent_shape ) noise_pred_uncond, noise_pred_text = model_output[:, 0], model_output[:, 1] model_output = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond) model_output = model_output.reshape( parallel_len * batch_size * num_images_per_prompt, *per_latent_shape ) block_latents_denoise = scheduler.batch_step_no_noise( model_output=model_output, timesteps=block_t.flatten(0, 1), sample=block_latents.flatten(0, 1), **extra_step_kwargs, ).reshape(block_latents.shape) # back to shape (parallel_dim, batch_size, ...) # now we want to add the pre-sampled noise # parallel sampling algorithm requires computing the cumulative drift from the beginning # of the window, so we need to compute cumulative sum of the deltas and the pre-sampled noises. delta = block_latents_denoise - block_latents cumulative_delta = self._cumsum(delta, dim=0, debug=debug) cumulative_noise = self._cumsum(noise_array[begin_idx:end_idx], dim=0, debug=debug) # if we are using an ODE-like scheduler (like DDIM), we don't want to add noise if scheduler._is_ode_scheduler: cumulative_noise = 0 block_latents_new = ( latents_time_evolution_buffer[begin_idx][None,] + cumulative_delta + cumulative_noise ) cur_error = torch.linalg.norm( (block_latents_new - latents_time_evolution_buffer[begin_idx + 1 : end_idx + 1]).reshape( parallel_len, batch_size * num_images_per_prompt, -1 ), dim=-1, ).pow(2) error_ratio = cur_error * inverse_variance_norm[begin_idx + 1 : end_idx + 1] # find the first index of the vector error_ratio that is greater than error tolerance # we can shift the window for the next iteration up to this index error_ratio = torch.nn.functional.pad( error_ratio, (0, 0, 0, 1), value=1e9 ) # handle the case when everything is below ratio, by padding the end of parallel_len dimension any_error_at_time = torch.max(error_ratio > scaled_tolerance, dim=1).values.int() ind = torch.argmax(any_error_at_time).item() # compute the new begin and end idxs for the window new_begin_idx = begin_idx + min(1 + ind, parallel) new_end_idx = min(new_begin_idx + parallel, len(scheduler.timesteps)) # store the computed latents for the current window in the global buffer latents_time_evolution_buffer[begin_idx + 1 : end_idx + 1] = block_latents_new # initialize the new sliding window latents with the end of the current window, # should be better than random initialization latents_time_evolution_buffer[end_idx : new_end_idx + 1] = latents_time_evolution_buffer[end_idx][ None, ] steps += 1 progress_bar.update(new_begin_idx - begin_idx) if callback is not None and steps % callback_steps == 0: callback(begin_idx, block_t[begin_idx], latents_time_evolution_buffer[begin_idx]) begin_idx = new_begin_idx end_idx = new_end_idx latents = latents_time_evolution_buffer[-1] if not output_type == "latent": image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False)[0] image, has_nsfw_concept = self.run_safety_checker(image, device, prompt_embeds.dtype) else: image = latents has_nsfw_concept = None if has_nsfw_concept is None: do_denormalize = [True] * image.shape[0] else: do_denormalize = [not has_nsfw for has_nsfw in has_nsfw_concept] image = self.image_processor.postprocess(image, output_type=output_type, do_denormalize=do_denormalize) # Offload all models self.maybe_free_model_hooks() if not return_dict: return (image, has_nsfw_concept) return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=has_nsfw_concept)