diff --git a/CHANGELOG.md b/CHANGELOG.md index 0f6afa646..1f92845ab 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -18,12 +18,12 @@ can be used with manually created masks or with auto-generated depth-maps uses general denoising strength value simply enable from *img2img -> scripts -> differential diffusion* - *note*: supports sd-xl models only + *note*: supports sd15 and sdxl models - [Regional prompting](https://github.com/huggingface/diffusers/blob/main/examples/community/README.md#regional-prompting-pipeline) as a built-in solution usage is same as original implementation from @hako-mikan click on title to open docs and see examples of full syntax on how to use it simply enable from *scripts -> regional prompting* - *note*: supports sd-15 models only + *note*: supports sd15 models only - [DeepCache](https://github.com/horseee/DeepCache) model acceleration it can produce massive speedups (2x-5x) with no overhead, but with some loss of quality *settings -> compute -> model compile -> deep-cache* and *settings -> compute -> model compile -> cache interval* diff --git a/modules/prompt_parser_diffusers.py b/modules/prompt_parser_diffusers.py index 8349c8d5f..680ca05cd 100644 --- a/modules/prompt_parser_diffusers.py +++ b/modules/prompt_parser_diffusers.py @@ -163,8 +163,10 @@ def prepare_embedding_providers(pipe, clip_skip): def pad_to_same_length(pipe, embeds): + if not hasattr(pipe, 'encode_prompt'): + return embeds device = pipe.device if str(pipe.device) != 'meta' else devices.device - try: # SDXL + try: # SDXL empty_embed = pipe.encode_prompt("") except TypeError: # SD1.5 empty_embed = pipe.encode_prompt("", device, 1, False) diff --git a/scripts/differential_diffusion.py b/scripts/differential_diffusion.py index ee9d02964..e876f05ed 100644 --- a/scripts/differential_diffusion.py +++ b/scripts/differential_diffusion.py @@ -13,7 +13,8 @@ from typing import Any, Callable, Dict, List, Optional, Tuple, Union import numpy as np import PIL.Image import torch -from transformers import CLIPTextModel, CLIPTextModelWithProjection, CLIPTokenizer +from packaging import version +from transformers import CLIPFeatureExtractor, CLIPTextModel, CLIPTextModelWithProjection, CLIPTokenizer import torchvision from diffusers.image_processor import VaeImageProcessor @@ -25,16 +26,21 @@ from diffusers.models.attention_processor import ( LoRAXFormersAttnProcessor, XFormersAttnProcessor, ) +from diffusers.configuration_utils import FrozenDict from diffusers.schedulers import KarrasDiffusionSchedulers from diffusers.utils import ( + PIL_INTERPOLATION, + logging, + deprecate, is_accelerate_available, is_accelerate_version, - logging, replace_example_docstring, ) from diffusers.utils.torch_utils import randn_tensor from diffusers.pipelines.pipeline_utils import DiffusionPipeline from diffusers.pipelines.stable_diffusion_xl.pipeline_output import StableDiffusionXLPipelineOutput +from diffusers.pipelines.stable_diffusion.pipeline_output import StableDiffusionPipelineOutput +from diffusers.pipelines.stable_diffusion.safety_checker import StableDiffusionSafetyChecker logger = logging.get_logger(__name__) # pylint: disable=invalid-name @@ -1031,6 +1037,823 @@ class StableDiffusionXLDiffImg2ImgPipeline(DiffusionPipeline, FromSingleFileMixi return StableDiffusionXLPipelineOutput(images=image) + +class StableDiffusionDiffImg2ImgPipeline(DiffusionPipeline): + r""" + Pipeline for text-guided image to image generation using Stable Diffusion. + + 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/runwayml/stable-diffusion-v1-5) for details. + feature_extractor ([`CLIPFeatureExtractor`]): + Model that extracts features from generated images to be used as inputs for the `safety_checker`. + """ + _optional_components = ["safety_checker", "feature_extractor"] + + # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.__init__ + def __init__( + self, + vae: AutoencoderKL, + text_encoder: CLIPTextModel, + tokenizer: CLIPTokenizer, + unet: UNet2DConditionModel, + scheduler: KarrasDiffusionSchedulers, + safety_checker: StableDiffusionSafetyChecker, + feature_extractor: CLIPFeatureExtractor, + 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.register_to_config(requires_safety_checker=requires_safety_checker) + + # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_sequential_cpu_offload + def enable_sequential_cpu_offload(self, gpu_id=0): # pylint: disable=arguments-differ + r""" + Offloads all models to CPU using accelerate, significantly reducing memory usage. When called, unet, + text_encoder, vae and safety checker have their state dicts saved to CPU and then are moved to a + `torch.device('meta') and loaded to GPU only when their specific submodule has its `forward` method called. + Note that offloading happens on a submodule basis. Memory savings are higher than with + `enable_model_cpu_offload`, but performance is lower. + """ + if is_accelerate_available() and is_accelerate_version(">=", "0.14.0"): + from accelerate import cpu_offload + else: + raise ImportError("`enable_sequential_cpu_offload` requires `accelerate v0.14.0` or higher") + + device = torch.device(f"cuda:{gpu_id}") + + if self.device.type != "cpu": + self.to("cpu", silence_dtype_warnings=True) + torch.cuda.empty_cache() # otherwise we don't see the memory savings (but they probably exist) + + for cpu_offloaded_model in [self.unet, self.text_encoder, self.vae]: + cpu_offload(cpu_offloaded_model, device) + + if self.safety_checker is not None: + cpu_offload(self.safety_checker, execution_device=device, offload_buffers=True) + + # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_model_cpu_offload + def enable_model_cpu_offload(self, gpu_id=0): # pylint: disable=arguments-differ + r""" + Offloads all models to CPU using accelerate, reducing memory usage with a low impact on performance. Compared + to `enable_sequential_cpu_offload`, this method moves one whole model at a time to the GPU when its `forward` + method is called, and the model remains in GPU until the next model runs. Memory savings are lower than with + `enable_sequential_cpu_offload`, but performance is much better due to the iterative execution of the `unet`. + """ + if is_accelerate_available() and is_accelerate_version(">=", "0.17.0.dev0"): + from accelerate import cpu_offload_with_hook + else: + raise ImportError("`enable_model_offload` requires `accelerate v0.17.0` or higher.") + + device = torch.device(f"cuda:{gpu_id}") + + if self.device.type != "cpu": + self.to("cpu", silence_dtype_warnings=True) + torch.cuda.empty_cache() # otherwise we don't see the memory savings (but they probably exist) + + hook = None + for cpu_offloaded_model in [self.text_encoder, self.unet, self.vae]: + _, hook = cpu_offload_with_hook(cpu_offloaded_model, device, prev_module_hook=hook) + + if self.safety_checker is not None: + _, hook = cpu_offload_with_hook(self.safety_checker, device, prev_module_hook=hook) + + # We'll offload the last model manually. + self.final_offload_hook = hook # pylint: disable=attribute-defined-outside-init + + @property + # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline._execution_device + def _execution_device(self): + r""" + Returns the device on which the pipeline's models will be executed. After calling + `pipeline.enable_sequential_cpu_offload()` the execution device can only be inferred from Accelerate's module + hooks. + """ + if not hasattr(self.unet, "_hf_hook"): + return self.device + for module in self.unet.modules(): + if ( + hasattr(module, "_hf_hook") + and hasattr(module._hf_hook, "execution_device") # pylint: disable=protected-access + and module._hf_hook.execution_device is not None # pylint: disable=protected-access + ): + return torch.device(module._hf_hook.execution_device) # pylint: disable=protected-access + return self.device + + # 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.FloatTensor] = None, + negative_prompt_embeds: Optional[torch.FloatTensor] = None, + ): + r""" + Encodes the prompt into text encoder hidden states. + + Args: + prompt (`str` or `List[str]`, *optional*): + prompt to be encoded + device: (`torch.device`): + torch device + num_images_per_prompt (`int`): + number of images that should be generated per prompt + do_classifier_free_guidance (`bool`): + whether to use classifier free guidance or not + negative_prompt (`str` or `List[str]`, *optional*): + The prompt or prompts not to guide the image generation. If not defined, one has to pass + `negative_prompt_embeds`. instead. If not defined, one has to pass `negative_prompt_embeds`. instead. + Ignored when not using guidance (i.e., ignored if `guidance_scale` is less than `1`). + prompt_embeds (`torch.FloatTensor`, *optional*): + Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not + provided, text embeddings will be generated from `prompt` input argument. + negative_prompt_embeds (`torch.FloatTensor`, *optional*): + Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt + weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input + argument. + """ + if prompt is not None and isinstance(prompt, str): + batch_size = 1 + elif prompt is not None and isinstance(prompt, list): + batch_size = len(prompt) + else: + batch_size = prompt_embeds.shape[0] + + if prompt_embeds is None: + text_inputs = self.tokenizer( + prompt, + padding="max_length", + max_length=self.tokenizer.model_max_length, + truncation=True, + return_tensors="pt", + ) + text_input_ids = text_inputs.input_ids + untruncated_ids = self.tokenizer(prompt, padding="longest", return_tensors="pt").input_ids + + if untruncated_ids.shape[-1] >= text_input_ids.shape[-1] and not torch.equal( + text_input_ids, untruncated_ids + ): + removed_text = self.tokenizer.batch_decode( + untruncated_ids[:, self.tokenizer.model_max_length - 1 : -1] + ) + logger.warning( + "The following part of your input was truncated because CLIP can only handle sequences up to" + f" {self.tokenizer.model_max_length} tokens: {removed_text}" + ) + + if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask: + attention_mask = text_inputs.attention_mask.to(device) + else: + attention_mask = None + + prompt_embeds = self.text_encoder( + text_input_ids.to(device), + attention_mask=attention_mask, + ) + prompt_embeds = prompt_embeds[0] + + prompt_embeds = prompt_embeds.to(dtype=self.text_encoder.dtype, device=device) + + bs_embed, seq_len, _ = prompt_embeds.shape + # duplicate text embeddings for each generation per prompt, using mps friendly method + prompt_embeds = prompt_embeds.repeat(1, num_images_per_prompt, 1) + prompt_embeds = prompt_embeds.view(bs_embed * num_images_per_prompt, seq_len, -1) + + # get unconditional embeddings for classifier free guidance + if do_classifier_free_guidance and negative_prompt_embeds is None: + uncond_tokens: List[str] + if negative_prompt is None: + uncond_tokens = [""] * batch_size + elif type(prompt) is not type(negative_prompt): + raise TypeError( + f"`negative_prompt` should be the same type to `prompt`, but got {type(negative_prompt)} !=" + f" {type(prompt)}." + ) + elif isinstance(negative_prompt, str): + uncond_tokens = [negative_prompt] + elif batch_size != len(negative_prompt): + raise ValueError( + f"`negative_prompt`: {negative_prompt} has batch size {len(negative_prompt)}, but `prompt`:" + f" {prompt} has batch size {batch_size}. Please make sure that passed `negative_prompt` matches" + " the batch size of `prompt`." + ) + else: + uncond_tokens = negative_prompt + + max_length = prompt_embeds.shape[1] + uncond_input = self.tokenizer( + uncond_tokens, + padding="max_length", + max_length=max_length, + truncation=True, + return_tensors="pt", + ) + + if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask: + attention_mask = uncond_input.attention_mask.to(device) + else: + attention_mask = None + + negative_prompt_embeds = self.text_encoder( + uncond_input.input_ids.to(device), + attention_mask=attention_mask, + ) + negative_prompt_embeds = negative_prompt_embeds[0] + + if do_classifier_free_guidance: + # duplicate unconditional embeddings for each generation per prompt, using mps friendly method + seq_len = negative_prompt_embeds.shape[1] + + negative_prompt_embeds = negative_prompt_embeds.to(dtype=self.text_encoder.dtype, device=device) + + negative_prompt_embeds = negative_prompt_embeds.repeat(1, num_images_per_prompt, 1) + negative_prompt_embeds = negative_prompt_embeds.view(batch_size * num_images_per_prompt, seq_len, -1) + + # For classifier free guidance, we need to do two forward passes. + # Here we concatenate the unconditional and text embeddings into a single batch + # to avoid doing two forward passes + prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds]) + + return 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 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 + + # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.decode_latents + 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 + + # 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, strength, callback_steps, negative_prompt=None, prompt_embeds=None, negative_prompt_embeds=None + ): + 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): # pylint: disable=unused-argument + # 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:] + + return timesteps, num_inference_steps - t_start + + def prepare_latents(self, image, timestep, batch_size, num_images_per_prompt, dtype, device, generator=None): + if not isinstance(image, (torch.Tensor, PIL.Image.Image, list)): + raise ValueError( + f"`image` has to be of type `torch.Tensor`, `PIL.Image.Image` or list but is {type(image)}" + ) + + image = image.to(device=device, dtype=dtype) + + batch_size = batch_size * num_images_per_prompt + if isinstance(generator, list) and len(generator) != batch_size: + raise ValueError( + f"You have passed a list of generators of length {len(generator)}, but requested an effective batch" + f" size of {batch_size}. Make sure the batch size matches the length of the generators." + ) + + if isinstance(generator, list): + init_latents = [ + self.vae.encode(image[i : i + 1]).latent_dist.sample(generator[i]) for i in range(batch_size) + ] + init_latents = torch.cat(init_latents, dim=0) + else: + init_latents = self.vae.encode(image).latent_dist.sample(generator) + + init_latents = self.vae.config.scaling_factor * init_latents + + if batch_size > init_latents.shape[0] and batch_size % init_latents.shape[0] == 0: + # expand init_latents for batch_size + deprecation_message = ( + f"You have passed {batch_size} text prompts (`prompt`), but only {init_latents.shape[0]} initial" + " images (`image`). Initial images are now duplicating to match the number of text prompts. Note" + " that this behavior is deprecated and will be removed in a version 1.0.0. Please make sure to update" + " your script to pass as many initial images as text prompts to suppress this warning." + ) + deprecate("len(prompt) != len(image)", "1.0.0", deprecation_message, standard_warn=False) + additional_image_per_prompt = batch_size // init_latents.shape[0] + init_latents = torch.cat([init_latents] * additional_image_per_prompt, dim=0) + elif batch_size > init_latents.shape[0] and batch_size % init_latents.shape[0] != 0: + raise ValueError( + f"Cannot duplicate `image` of batch size {init_latents.shape[0]} to {batch_size} text prompts." + ) + else: + init_latents = torch.cat([init_latents], dim=0) + + shape = init_latents.shape + noise = randn_tensor(shape, generator=generator, device=device, dtype=dtype) + + # get latents + init_latents = self.scheduler.add_noise(init_latents, noise, timestep) + latents = init_latents + + return latents + + def encode_prompt( + self, + prompt, + device, + num_images_per_prompt, + do_classifier_free_guidance, + negative_prompt=None, + prompt_embeds: Optional[torch.FloatTensor] = None, + negative_prompt_embeds: Optional[torch.FloatTensor] = None, + lora_scale: Optional[float] = None, + clip_skip: Optional[int] = None, + ): + if prompt is not None and isinstance(prompt, str): + batch_size = 1 + elif prompt is not None and isinstance(prompt, list): + batch_size = len(prompt) + else: + batch_size = prompt_embeds.shape[0] + + if prompt_embeds is None: + # textual inversion: procecss multi-vector tokens if necessary + if isinstance(self, TextualInversionLoaderMixin): + prompt = self.maybe_convert_prompt(prompt, self.tokenizer) + + text_inputs = self.tokenizer( + prompt, + padding="max_length", + max_length=self.tokenizer.model_max_length, + truncation=True, + return_tensors="pt", + ) + text_input_ids = text_inputs.input_ids + untruncated_ids = self.tokenizer(prompt, padding="longest", return_tensors="pt").input_ids + + if untruncated_ids.shape[-1] >= text_input_ids.shape[-1] and not torch.equal( + text_input_ids, untruncated_ids + ): + removed_text = self.tokenizer.batch_decode( + untruncated_ids[:, self.tokenizer.model_max_length - 1 : -1] + ) + logger.warning( + "The following part of your input was truncated because CLIP can only handle sequences up to" + f" {self.tokenizer.model_max_length} tokens: {removed_text}" + ) + + if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask: + attention_mask = text_inputs.attention_mask.to(device) + else: + attention_mask = None + + 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: procecss multi-vector tokens if necessary + if isinstance(self, TextualInversionLoaderMixin): + uncond_tokens = self.maybe_convert_prompt(uncond_tokens, self.tokenizer) + + max_length = prompt_embeds.shape[1] + uncond_input = self.tokenizer( + uncond_tokens, + padding="max_length", + max_length=max_length, + truncation=True, + return_tensors="pt", + ) + + if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask: + attention_mask = uncond_input.attention_mask.to(device) + else: + attention_mask = None + + negative_prompt_embeds = self.text_encoder( + uncond_input.input_ids.to(device), + attention_mask=attention_mask, + ) + negative_prompt_embeds = negative_prompt_embeds[0] + + if do_classifier_free_guidance: + # duplicate unconditional embeddings for each generation per prompt, using mps friendly method + seq_len = negative_prompt_embeds.shape[1] + + negative_prompt_embeds = negative_prompt_embeds.to(dtype=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) + + return prompt_embeds, negative_prompt_embeds + + + def preprocess(self, image): + if isinstance(image, torch.Tensor): + return image + elif isinstance(image, PIL.Image.Image): + image = [image] + + if isinstance(image[0], PIL.Image.Image): + w, h = image[0].size + w, h = map(lambda x: x - x % 8, (w, h)) # resize to integer multiple of 8 # noqa: C417 + + image = [np.array(i.resize((w, h), resample=PIL_INTERPOLATION["lanczos"]))[None, :] for i in image] + image = np.concatenate(image, axis=0) + image = np.array(image).astype(np.float32) / 255.0 + image = image.transpose(0, 3, 1, 2) + image = 2.0 * image - 1.0 + image = torch.from_numpy(image) + elif isinstance(image[0], torch.Tensor): + image = torch.cat(image, dim=0) + return image + + @torch.no_grad() + @replace_example_docstring(EXAMPLE_DOC_STRING) + def __call__( + self, + prompt: Union[str, List[str]] = None, + image: Union[torch.FloatTensor, PIL.Image.Image] = None, + strength: float = 1, + num_inference_steps: Optional[int] = 50, + guidance_scale: Optional[float] = 7.5, + negative_prompt: Optional[Union[str, List[str]]] = None, + 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.FloatTensor] = None, + negative_prompt_embeds: Optional[torch.FloatTensor] = None, + output_type: Optional[str] = "pil", + return_dict: bool = True, + callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None, + callback_steps: int = 1, + map:torch.FloatTensor = None, # pylint: disable=redefined-builtin + ): + r""" + Function invoked when calling the pipeline for generation. + + Args: + prompt (`str` or `List[str]`, *optional*): + The prompt or prompts to guide the image generation. If not defined, one has to pass `prompt_embeds`. + instead. + image (`torch.FloatTensor` or `PIL.Image.Image`): + `Image`, or tensor representing an image batch, that will be used as the starting point for the + process. + strength (`float`, *optional*, defaults to 1): + Repealed in favor of the map. + 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. + 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`). + 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`, *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.FloatTensor`, *optional*): + Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not + provided, text embeddings will be generated from `prompt` input argument. + negative_prompt_embeds (`torch.FloatTensor`, *optional*): + Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt + weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input + argument. + 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.FloatTensor)`. + 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. + Examples: + + Returns: + [`~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`. + """ + # 1. Check inputs. Raise error if not correct + self.check_inputs(prompt, 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 + + # 3. Encode input prompt + prompt_embeds = self._encode_prompt( + prompt, + device, + num_images_per_prompt, + do_classifier_free_guidance, + negative_prompt, + prompt_embeds=prompt_embeds, + negative_prompt_embeds=negative_prompt_embeds, + ) + + # 4. Preprocess image + # image = self.preprocess(image) + + # 5. set timesteps + self.scheduler.set_timesteps(num_inference_steps, device=device) + timesteps, num_inference_steps = self.get_timesteps(num_inference_steps, strength, device) + + + # 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) + map = torchvision.transforms.Resize(tuple(s // self.vae_scale_factor for s in image.shape[2:]),antialias=None)(map) + + # 8. Denoising loop + num_warmup_steps = len(timesteps) - num_inference_steps * self.scheduler.order + + # prepartions + original_with_noise = self.prepare_latents( + image, timesteps, batch_size, num_images_per_prompt, prompt_embeds.dtype, device, generator + ) + thresholds = torch.arange(len(timesteps), dtype=map.dtype) / len(timesteps) + thresholds = thresholds.unsqueeze(1).unsqueeze(1).to(device) + masks = map > thresholds + # end diff diff preparations + + with self.progress_bar(total=num_inference_steps) as progress_bar: + + for i, t in enumerate(timesteps): + # diff diff + if i == 0: + latents = original_with_noise[:1] + else: + mask = masks[i].unsqueeze(0) + # cast mask to the same type as latents etc + mask = mask.to(latents.dtype) + mask = mask.unsqueeze(1) # fit shape + latents = original_with_noise[i] * mask + latents * (1 - mask) + # end diff diff + # 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).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 + + # call the callback, if provided + if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0): + progress_bar.update() + if callback is not None and i % callback_steps == 0: + callback(i, t, latents) + + # 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) + #has_nsfw_concept = False + + # 11. Convert to PIL + # if output_type == "pil": + # image = self.numpy_to_pil(image) + + # 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=latents, nsfw_content_detected=False) + ### pipeline end ### script start @@ -1069,7 +1892,7 @@ class Script(scripts.Script): image = gr.Image(label="Image map", show_label=False, type="pil", source="upload", interactive=True, tool="editor", visible=True, image_mode='RGB') return enabled, strength, invert, model, image - def depthmap(self, image_init: Image.Image, image_map: Image.Image, model: str, strength: float, invert: bool): + def depthmap(self, image_init: Image.Image, image_map: Image.Image, model: str, strength: float, invert: bool, output_type="tensor"): global detector # pylint: disable=global-statement from modules.control.proc.dpt import DPTDetector if image_init is None: @@ -1092,22 +1915,27 @@ class Script(scripts.Script): if shared.opts.save_init_img: init_img_hash = hashlib.sha256(image_map.tobytes()).hexdigest()[0:8] # pylint: disable=attribute-defined-outside-init images.save_image(image_map, path=shared.opts.outdir_init_images, basename=None, forced_filename=init_img_hash, suffix="-init-image") + else: + return None, None + if output_type == "tensor": image_map = transforms.ToTensor()(image_map) - image_init = 2 * transforms.ToTensor()(image_init) - 1 - image_init = image_init.unsqueeze(0) - return image_init.to(devices.device), image_map.to(devices.device) + image_map = image_map.to(devices.device) + image_init = 2 * transforms.ToTensor()(image_init) - 1 + image_init = image_init.unsqueeze(0) + image_init = image_init.to(devices.device) + return image_init, image_map def run(self, p: processing.StableDiffusionProcessingImg2Img, enabled, strength, invert, model, image): # pylint: disable=arguments-differ if not enabled: return - if shared.sd_model_type != 'sdxl': + if shared.sd_model_type != 'sdxl' and shared.sd_model_type != 'sd': shared.log.error(f'Differential-diffusion: incorrect base model: {shared.sd_model.__class__.__name__}') return if not hasattr(p, 'init_images') or len(p.init_images) == 0: shared.log.error('Differential-diffusion: no input images') return - image_init, image_map = self.depthmap(p.init_images[0], image, model, strength, invert) + image_init, image_map = self.depthmap(p.init_images[0], image, model, strength, invert, output_type="tensor") if image_map is None: shared.log.error('Differential-diffusion: no image map') return @@ -1116,21 +1944,35 @@ class Script(scripts.Script): pipe = None try: diffusers.pipelines.auto_pipeline.AUTO_IMAGE2IMAGE_PIPELINES_MAPPING["StableDiffusionXLDiffImg2ImgPipeline"] = StableDiffusionXLDiffImg2ImgPipeline - pipe = StableDiffusionXLDiffImg2ImgPipeline( - vae=shared.sd_model.vae, - text_encoder=shared.sd_model.text_encoder, - text_encoder_2=shared.sd_model.text_encoder_2, - tokenizer=shared.sd_model.tokenizer, - tokenizer_2=shared.sd_model.tokenizer_2, - unet=shared.sd_model.unet, - scheduler=shared.sd_model.scheduler, - ) + diffusers.pipelines.auto_pipeline.AUTO_IMAGE2IMAGE_PIPELINES_MAPPING["StableDiffusionDiffImg2ImgPipeline"] = StableDiffusionDiffImg2ImgPipeline + if shared.sd_model_type == 'sdxl': + pipe = StableDiffusionXLDiffImg2ImgPipeline( + text_encoder=shared.sd_model.text_encoder, + text_encoder_2=shared.sd_model.text_encoder_2, + tokenizer=shared.sd_model.tokenizer, + tokenizer_2=shared.sd_model.tokenizer_2, + unet=shared.sd_model.unet, + vae=shared.sd_model.vae, + scheduler=shared.sd_model.scheduler, + ) + elif shared.sd_model_type == 'sd': + pipe = StableDiffusionDiffImg2ImgPipeline( + text_encoder=shared.sd_model.text_encoder, + tokenizer=shared.sd_model.tokenizer, + unet=shared.sd_model.unet, + vae=shared.sd_model.vae, + scheduler=shared.sd_model.scheduler, + feature_extractor=getattr(shared.sd_model, 'feature_extractor', None), + safety_checker=None, + requires_safety_checker=False, + ) sd_models.copy_diffuser_options(pipe, shared.sd_model) sd_models.set_diffuser_options(pipe) p.task_args['image'] = image_init - p.task_args['original_image'] = image_init p.task_args['map'] = image_map - shared.log.debug(f'Differential-diffusion: pipeline={pipe.__class__.__name__} strength={strength} model={model} image={image_init.shape} map={image_map.shape}') + if shared.sd_model_type == 'sdxl': + p.task_args['original_image'] = image_init + shared.log.debug(f'Differential-diffusion: pipeline={pipe.__class__.__name__} strength={strength} model={model} auto={image is None}') shared.sd_model = pipe except Exception as e: shared.log.error(f'Differential-diffusion: pipeline creation failed: {e}')