From c072fd79aa71707a5e8c05732d26f8a83d34ea1d Mon Sep 17 00:00:00 2001 From: SahilCarterr <110806554+SahilCarterr@users.noreply.github.com> Date: Wed, 18 Sep 2024 19:04:24 +0530 Subject: [PATCH 01/17] added pag to sd img2img pipeline --- src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py | 0 1 file changed, 0 insertions(+), 0 deletions(-) create mode 100644 src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py diff --git a/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py b/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py new file mode 100644 index 000000000000..e69de29bb2d1 From c622bd60b77b97fcd8c88d8c6788ec57ad61fb9e Mon Sep 17 00:00:00 2001 From: SahilCarterr <110806554+SahilCarterr@users.noreply.github.com> Date: Wed, 18 Sep 2024 19:07:15 +0530 Subject: [PATCH 02/17] Update pipeline_pag_sd_img2img.py --- .../pipelines/pag/pipeline_pag_sd_img2img.py | 1120 +++++++++++++++++ 1 file changed, 1120 insertions(+) diff --git a/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py b/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py index e69de29bb2d1..5b34f7932224 100644 --- a/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py +++ b/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py @@ -0,0 +1,1120 @@ +# Copyright 2024 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 numpy as np +import PIL.Image +import torch +from packaging import version +from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer, CLIPVisionModelWithProjection + +from ...callbacks import MultiPipelineCallbacks, PipelineCallback +from ...configuration_utils import FrozenDict +from ...image_processor import PipelineImageInput, VaeImageProcessor +from ...loaders import FromSingleFileMixin, IPAdapterMixin, StableDiffusionLoraLoaderMixin, TextualInversionLoaderMixin +from ...models import AutoencoderKL, ImageProjection, UNet2DConditionModel +from ...models.lora import adjust_lora_scale_text_encoder +from ...schedulers import KarrasDiffusionSchedulers +from ...utils import ( + PIL_INTERPOLATION, + 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 +from .pag_utils import PAGMixin + + +logger = logging.get_logger(__name__) # pylint: disable=invalid-name + +EXAMPLE_DOC_STRING = """ + Examples: + ```py + >>> import torch + >>> from diffusers import AutoPipelineForImage2Image + >>> from diffusers.utils import load_image + + >>> pipe = AutoPipelineForImage2Image.from_pretrained( + ... "runwayml/stable-diffusion-v1-5", + ... torch_dtype=torch.float16, + ... enable_pag=True, + ... ) + >>> pipe = pipe.to("cuda") + >>> url = "https://huggingface.co/datasets/patrickvonplaten/images/resolve/main/aa_xl/000000009.png" + + >>> init_image = load_image(url).convert("RGB") + >>> prompt = "a photo of an astronaut riding a horse on mars" + >>> image = pipe(prompt, image=init_image, pag_scale=0.3).images[0] + ``` +""" + +# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.retrieve_timesteps +def retrieve_latents( + encoder_output: torch.Tensor, generator: Optional[torch.Generator] = None, sample_mode: str = "sample" +): + if hasattr(encoder_output, "latent_dist") and sample_mode == "sample": + return encoder_output.latent_dist.sample(generator) + elif hasattr(encoder_output, "latent_dist") and sample_mode == "argmax": + return encoder_output.latent_dist.mode() + elif hasattr(encoder_output, "latents"): + return encoder_output.latents + else: + raise AttributeError("Could not access latents of provided encoder_output") + + +# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.retrieve_timesteps +def retrieve_timesteps( + scheduler, + num_inference_steps: Optional[int] = None, + device: Optional[Union[str, torch.device]] = None, + timesteps: Optional[List[int]] = None, + sigmas: Optional[List[float]] = None, + **kwargs, +): + """ + Calls the scheduler's `set_timesteps` method and retrieves timesteps from the scheduler after the call. Handles + custom timesteps. Any kwargs will be supplied to `scheduler.set_timesteps`. + + Args: + scheduler (`SchedulerMixin`): + The scheduler to get timesteps from. + num_inference_steps (`int`): + The number of diffusion steps used when generating samples with a pre-trained model. If used, `timesteps` + must be `None`. + device (`str` or `torch.device`, *optional*): + The device to which the timesteps should be moved to. If `None`, the timesteps are not moved. + timesteps (`List[int]`, *optional*): + Custom timesteps used to override the timestep spacing strategy of the scheduler. If `timesteps` is passed, + `num_inference_steps` and `sigmas` must be `None`. + sigmas (`List[float]`, *optional*): + Custom sigmas used to override the timestep spacing strategy of the scheduler. If `sigmas` is passed, + `num_inference_steps` and `timesteps` must be `None`. + + Returns: + `Tuple[torch.Tensor, int]`: A tuple where the first element is the timestep schedule from the scheduler and the + second element is the number of inference steps. + """ + if timesteps is not None and sigmas is not None: + raise ValueError("Only one of `timesteps` or `sigmas` can be passed. Please choose one to set custom values") + if timesteps is not None: + accepts_timesteps = "timesteps" in set(inspect.signature(scheduler.set_timesteps).parameters.keys()) + if not accepts_timesteps: + raise ValueError( + f"The current scheduler class {scheduler.__class__}'s `set_timesteps` does not support custom" + f" timestep schedules. Please check whether you are using the correct scheduler." + ) + scheduler.set_timesteps(timesteps=timesteps, device=device, **kwargs) + timesteps = scheduler.timesteps + num_inference_steps = len(timesteps) + elif sigmas is not None: + accept_sigmas = "sigmas" in set(inspect.signature(scheduler.set_timesteps).parameters.keys()) + if not accept_sigmas: + raise ValueError( + f"The current scheduler class {scheduler.__class__}'s `set_timesteps` does not support custom" + f" sigmas schedules. Please check whether you are using the correct scheduler." + ) + scheduler.set_timesteps(sigmas=sigmas, device=device, **kwargs) + timesteps = scheduler.timesteps + num_inference_steps = len(timesteps) + else: + scheduler.set_timesteps(num_inference_steps, device=device, **kwargs) + timesteps = scheduler.timesteps + return timesteps, num_inference_steps + + +class StableDiffusionPAGImg2ImgPipeline( + DiffusionPipeline, + StableDiffusionMixin, + TextualInversionLoaderMixin, + IPAdapterMixin, + StableDiffusionLoraLoaderMixin, + FromSingleFileMixin, + PAGMixin, +): + 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 + 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.StableDiffusionLoraLoaderMixin.load_lora_weights`] for loading LoRA weights + - [`~loaders.StableDiffusionLoraLoaderMixin.save_lora_weights`] for saving LoRA weights + - [`~loaders.FromSingleFileMixin.from_single_file`] for loading `.ckpt` files + - [`~loaders.IPAdapterMixin.load_ip_adapter`] for loading IP Adapters + + 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->image_encoder->unet->vae" + _optional_components = ["safety_checker", "feature_extractor", "image_encoder"] + _exclude_from_cpu_offload = ["safety_checker"] + _callback_tensor_inputs = ["latents", "prompt_embeds", "negative_prompt_embeds"] + + def __init__( + self, + vae: AutoencoderKL, + text_encoder: CLIPTextModel, + tokenizer: CLIPTokenizer, + unet: UNet2DConditionModel, + scheduler: KarrasDiffusionSchedulers, + safety_checker: StableDiffusionSafetyChecker, + feature_extractor: CLIPImageProcessor, + image_encoder: CLIPVisionModelWithProjection = None, + requires_safety_checker: bool = True, + pag_applied_layers: Union[str, List[str]] = "mid", # ["mid"], ["down.block_1", "up.block_0.attentions_0"] + ): + 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, + image_encoder=image_encoder, + ) + 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) + + self.set_pag_applied_layers(pag_applied_layers) + + # 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, StableDiffusionLoraLoaderMixin): + 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, StableDiffusionLoraLoaderMixin) 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.encode_image + def encode_image(self, image, device, num_images_per_prompt, output_hidden_states=None): + dtype = next(self.image_encoder.parameters()).dtype + + if not isinstance(image, torch.Tensor): + image = self.feature_extractor(image, return_tensors="pt").pixel_values + + image = image.to(device=device, dtype=dtype) + if output_hidden_states: + image_enc_hidden_states = self.image_encoder(image, output_hidden_states=True).hidden_states[-2] + image_enc_hidden_states = image_enc_hidden_states.repeat_interleave(num_images_per_prompt, dim=0) + uncond_image_enc_hidden_states = self.image_encoder( + torch.zeros_like(image), output_hidden_states=True + ).hidden_states[-2] + uncond_image_enc_hidden_states = uncond_image_enc_hidden_states.repeat_interleave( + num_images_per_prompt, dim=0 + ) + return image_enc_hidden_states, uncond_image_enc_hidden_states + else: + image_embeds = self.image_encoder(image).image_embeds + image_embeds = image_embeds.repeat_interleave(num_images_per_prompt, dim=0) + uncond_image_embeds = torch.zeros_like(image_embeds) + + return image_embeds, uncond_image_embeds + + # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_ip_adapter_image_embeds + def prepare_ip_adapter_image_embeds( + self, ip_adapter_image, ip_adapter_image_embeds, device, num_images_per_prompt, do_classifier_free_guidance + ): + image_embeds = [] + if do_classifier_free_guidance: + negative_image_embeds = [] + if ip_adapter_image_embeds is None: + if not isinstance(ip_adapter_image, list): + ip_adapter_image = [ip_adapter_image] + + if len(ip_adapter_image) != len(self.unet.encoder_hid_proj.image_projection_layers): + raise ValueError( + f"`ip_adapter_image` must have same length as the number of IP Adapters. Got {len(ip_adapter_image)} images and {len(self.unet.encoder_hid_proj.image_projection_layers)} IP Adapters." + ) + + for single_ip_adapter_image, image_proj_layer in zip( + ip_adapter_image, self.unet.encoder_hid_proj.image_projection_layers + ): + output_hidden_state = not isinstance(image_proj_layer, ImageProjection) + single_image_embeds, single_negative_image_embeds = self.encode_image( + single_ip_adapter_image, device, 1, output_hidden_state + ) + + image_embeds.append(single_image_embeds[None, :]) + if do_classifier_free_guidance: + negative_image_embeds.append(single_negative_image_embeds[None, :]) + else: + for single_image_embeds in ip_adapter_image_embeds: + if do_classifier_free_guidance: + single_negative_image_embeds, single_image_embeds = single_image_embeds.chunk(2) + negative_image_embeds.append(single_negative_image_embeds) + image_embeds.append(single_image_embeds) + + ip_adapter_image_embeds = [] + for i, single_image_embeds in enumerate(image_embeds): + single_image_embeds = torch.cat([single_image_embeds] * num_images_per_prompt, dim=0) + if do_classifier_free_guidance: + single_negative_image_embeds = torch.cat([negative_image_embeds[i]] * num_images_per_prompt, dim=0) + single_image_embeds = torch.cat([single_negative_image_embeds, single_image_embeds], dim=0) + + single_image_embeds = single_image_embeds.to(device=device) + ip_adapter_image_embeds.append(single_image_embeds) + + return ip_adapter_image_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, + strength, + callback_steps, + negative_prompt=None, + prompt_embeds=None, + negative_prompt_embeds=None, + ip_adapter_image=None, + ip_adapter_image_embeds=None, + callback_on_step_end_tensor_inputs=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 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}." + ) + + if ip_adapter_image is not None and ip_adapter_image_embeds is not None: + raise ValueError( + "Provide either `ip_adapter_image` or `ip_adapter_image_embeds`. Cannot leave both `ip_adapter_image` and `ip_adapter_image_embeds` defined." + ) + + if ip_adapter_image_embeds is not None: + if not isinstance(ip_adapter_image_embeds, list): + raise ValueError( + f"`ip_adapter_image_embeds` has to be of type `list` but is {type(ip_adapter_image_embeds)}" + ) + elif ip_adapter_image_embeds[0].ndim not in [3, 4]: + raise ValueError( + f"`ip_adapter_image_embeds` has to be a list of 3D or 4D tensors but is {ip_adapter_image_embeds[0].ndim}D" + ) + + def get_timesteps(self, num_inference_steps, strength, device): + # get the original timestep using init_timestep + init_timestep = min(int(num_inference_steps * strength), num_inference_steps) + + t_start = max(num_inference_steps - init_timestep, 0) + timesteps = self.scheduler.timesteps[t_start * self.scheduler.order :] + if hasattr(self.scheduler, "set_begin_index"): + self.scheduler.set_begin_index(t_start * self.scheduler.order) + + 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 image.shape[1] == 4: + init_latents = image + + else: + if isinstance(generator, list) and len(generator) != batch_size: + raise ValueError( + f"You have passed a list of generators of length {len(generator)}, but requested an effective batch" + f" size of {batch_size}. Make sure the batch size matches the length of the generators." + ) + + elif isinstance(generator, list): + if image.shape[0] < batch_size and batch_size % image.shape[0] == 0: + image = torch.cat([image] * (batch_size // image.shape[0]), dim=0) + elif image.shape[0] < batch_size and batch_size % image.shape[0] != 0: + raise ValueError( + f"Cannot duplicate `image` of batch size {image.shape[0]} to effective batch_size {batch_size} " + ) + + init_latents = [ + retrieve_latents(self.vae.encode(image[i : i + 1]), generator=generator[i]) + for i in range(batch_size) + ] + init_latents = torch.cat(init_latents, dim=0) + else: + init_latents = retrieve_latents(self.vae.encode(image), generator=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 + + # Copied from diffusers.pipelines.latent_consistency_models.pipeline_latent_consistency_text2img.LatentConsistencyModelPipeline.get_guidance_scale_embedding + def get_guidance_scale_embedding( + self, w: torch.Tensor, embedding_dim: int = 512, dtype: torch.dtype = torch.float32 + ) -> torch.Tensor: + """ + See https://github.com/google-research/vdm/blob/dc27b98a554f65cdc654b800da5aa1846545d41b/model_vdm.py#L298 + + Args: + w (`torch.Tensor`): + Generate embedding vectors with a specified guidance scale to subsequently enrich timestep embeddings. + embedding_dim (`int`, *optional*, defaults to 512): + Dimension of the embeddings to generate. + dtype (`torch.dtype`, *optional*, defaults to `torch.float32`): + Data type of the generated embeddings. + + Returns: + `torch.Tensor`: Embedding vectors with shape `(len(w), embedding_dim)`. + """ + assert len(w.shape) == 1 + w = w * 1000.0 + + half_dim = embedding_dim // 2 + emb = torch.log(torch.tensor(10000.0)) / (half_dim - 1) + emb = torch.exp(torch.arange(half_dim, dtype=dtype) * -emb) + emb = w.to(dtype)[:, None] * emb[None, :] + emb = torch.cat([torch.sin(emb), torch.cos(emb)], dim=1) + if embedding_dim % 2 == 1: # zero pad + emb = torch.nn.functional.pad(emb, (0, 1)) + assert emb.shape == (w.shape[0], embedding_dim) + return emb + + @property + def guidance_scale(self): + return self._guidance_scale + + @property + def clip_skip(self): + return self._clip_skip + + # 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. + @property + def do_classifier_free_guidance(self): + return self._guidance_scale > 1 and self.unet.config.time_cond_proj_dim is None + + @property + def cross_attention_kwargs(self): + return self._cross_attention_kwargs + + @property + def num_timesteps(self): + return self._num_timesteps + + @property + def interrupt(self): + return self._interrupt + + @torch.no_grad() + @replace_example_docstring(EXAMPLE_DOC_STRING) + def __call__( + self, + prompt: Union[str, List[str]] = None, + image: PipelineImageInput = None, + strength: float = 0.8, + num_inference_steps: Optional[int] = 50, + timesteps: List[int] = None, + sigmas: List[float] = None, + 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.Tensor] = None, + negative_prompt_embeds: Optional[torch.Tensor] = None, + ip_adapter_image: Optional[PipelineImageInput] = None, + ip_adapter_image_embeds: Optional[List[torch.Tensor]] = None, + output_type: Optional[str] = "pil", + return_dict: bool = True, + cross_attention_kwargs: Optional[Dict[str, Any]] = None, + clip_skip: int = None, + callback_on_step_end: Optional[ + Union[Callable[[int, int, Dict], None], PipelineCallback, MultiPipelineCallbacks] + ] = None, + callback_on_step_end_tensor_inputs: List[str] = ["latents"], + pag_scale: float = 3.0, + pag_adaptive_scale: float = 0.0, + **kwargs, + ): + r""" + The call function to the pipeline for generation. + + Args: + prompt (`str` or `List[str]`, *optional*): + The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`. + image (`torch.Tensor`, `PIL.Image.Image`, `np.ndarray`, `List[torch.Tensor]`, `List[PIL.Image.Image]`, or `List[np.ndarray]`): + `Image`, numpy array or tensor representing an image batch to be used as the starting point. For both + numpy array and pytorch tensor, the expected value range is between `[0, 1]` If it's a tensor or a list + or tensors, the expected shape should be `(B, C, H, W)` or `(C, H, W)`. If it is a numpy array or a + list of arrays, the expected shape should be `(B, H, W, C)` or `(H, W, C)` It can also accept image + latents as `image`, but if passing latents directly it is not encoded again. + strength (`float`, *optional*, defaults to 0.8): + Indicates extent to transform the reference `image`. Must be between 0 and 1. `image` is used as a + starting point and more noise is added the higher the `strength`. The number of denoising steps depends + on the amount of noise initially added. When `strength` is 1, added noise is maximum and the denoising + process runs for the full number of iterations specified in `num_inference_steps`. A value of 1 + essentially ignores `image`. + num_inference_steps (`int`, *optional*, defaults to 50): + The number of denoising steps. More denoising steps usually lead to a higher quality image at the + expense of slower inference. This parameter is modulated by `strength`. + timesteps (`List[int]`, *optional*): + Custom timesteps to use for the denoising process with schedulers which support a `timesteps` argument + in their `set_timesteps` method. If not defined, the default behavior when `num_inference_steps` is + passed will be used. Must be in descending order. + sigmas (`List[float]`, *optional*): + Custom sigmas to use for the denoising process with schedulers which support a `sigmas` argument in + their `set_timesteps` method. If not defined, the default behavior when `num_inference_steps` is passed + will be used. + 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. + 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. + ip_adapter_image: (`PipelineImageInput`, *optional*): Optional image input to work with IP Adapters. + ip_adapter_image_embeds (`List[torch.Tensor]`, *optional*): + Pre-generated image embeddings for IP-Adapter. It should be a list of length same as number of + IP-adapters. Each element should be a tensor of shape `(batch_size, num_images, emb_dim)`. It should + contain the negative image embedding if `do_classifier_free_guidance` is set to `True`. If not + provided, embeddings are computed from the `ip_adapter_image` 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. + 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). + clip_skip (`int`, *optional*): + Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that + the output of the pre-final layer will be used for computing the prompt embeddings. + callback_on_step_end (`Callable`, `PipelineCallback`, `MultiPipelineCallbacks`, *optional*): + A function or a subclass of `PipelineCallback` or `MultiPipelineCallbacks` that is called at the end of + each denoising step during the inference. with the following arguments: `callback_on_step_end(self: + DiffusionPipeline, step: int, timestep: int, callback_kwargs: Dict)`. `callback_kwargs` will include a + list of all tensors as specified by `callback_on_step_end_tensor_inputs`. + callback_on_step_end_tensor_inputs (`List`, *optional*): + The list of tensor inputs for the `callback_on_step_end` function. The tensors specified in the list + will be passed as `callback_kwargs` argument. You will only be able to include variables listed in the + `._callback_tensor_inputs` attribute of your pipeline class. + pag_scale (`float`, *optional*, defaults to 3.0): + The scale factor for the perturbed attention guidance. If it is set to 0.0, the perturbed attention + guidance will not be used. + pag_adaptive_scale (`float`, *optional*, defaults to 0.0): + The adaptive scale factor for the perturbed attention guidance. If it is set to 0.0, `pag_scale` is + used. + + 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. + """ + + callback = kwargs.pop("callback", None) + callback_steps = kwargs.pop("callback_steps", None) + + if callback is not None: + deprecate( + "callback", + "1.0.0", + "Passing `callback` as an input argument to `__call__` is deprecated, consider use `callback_on_step_end`", + ) + if callback_steps is not None: + deprecate( + "callback_steps", + "1.0.0", + "Passing `callback_steps` as an input argument to `__call__` is deprecated, consider use `callback_on_step_end`", + ) + + if isinstance(callback_on_step_end, (PipelineCallback, MultiPipelineCallbacks)): + callback_on_step_end_tensor_inputs = callback_on_step_end.tensor_inputs + + # 1. Check inputs. Raise error if not correct + self.check_inputs( + prompt, + strength, + callback_steps, + negative_prompt, + prompt_embeds, + negative_prompt_embeds, + ip_adapter_image, + ip_adapter_image_embeds, + callback_on_step_end_tensor_inputs, + ) + + self._guidance_scale = guidance_scale + self._clip_skip = clip_skip + self._cross_attention_kwargs = cross_attention_kwargs + self._interrupt = False + + self._pag_scale = pag_scale + self._pag_adaptive_scale = pag_adaptive_scale + + # 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 + + # 3. Encode input prompt + text_encoder_lora_scale = ( + self.cross_attention_kwargs.get("scale", None) if self.cross_attention_kwargs is not None else None + ) + prompt_embeds, negative_prompt_embeds = self.encode_prompt( + prompt, + device, + num_images_per_prompt, + self.do_classifier_free_guidance, + negative_prompt, + prompt_embeds=prompt_embeds, + negative_prompt_embeds=negative_prompt_embeds, + lora_scale=text_encoder_lora_scale, + clip_skip=self.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 self.do_perturbed_attention_guidance: + prompt_embeds = self._prepare_perturbed_attention_guidance( + prompt_embeds, negative_prompt_embeds, self.do_classifier_free_guidance + ) + elif self.do_classifier_free_guidance: + prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds]) + + if ip_adapter_image is not None or ip_adapter_image_embeds is not None: + ip_adapter_image_embeds = self.prepare_ip_adapter_image_embeds( + ip_adapter_image, + ip_adapter_image_embeds, + device, + batch_size * num_images_per_prompt, + self.do_classifier_free_guidance, + ) + + for i, image_embeds in enumerate(ip_adapter_image_embeds): + negative_image_embeds = None + if self.do_classifier_free_guidance: + negative_image_embeds, image_embeds = image_embeds.chunk(2) + if self.do_perturbed_attention_guidance: + image_embeds = self._prepare_perturbed_attention_guidance( + image_embeds, negative_image_embeds, self.do_classifier_free_guidance + ) + + elif self.do_classifier_free_guidance: + image_embeds = torch.cat([negative_image_embeds, image_embeds], dim=0) + image_embeds = image_embeds.to(device) + ip_adapter_image_embeds[i] = image_embeds + + # 4. Preprocess image + image = self.image_processor.preprocess(image) + + # 5. set timesteps + timesteps, num_inference_steps = retrieve_timesteps( + self.scheduler, num_inference_steps, device, timesteps, sigmas + ) + timesteps, num_inference_steps = self.get_timesteps(num_inference_steps, strength, device) + latent_timestep = timesteps[:1].repeat(batch_size * num_images_per_prompt) + + # 6. Prepare latent variables + latents = self.prepare_latents( + image, + latent_timestep, + batch_size, + num_images_per_prompt, + prompt_embeds.dtype, + device, + generator, + ) + + # 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) + + # 7.1 Add image embeds for IP-Adapter + added_cond_kwargs = ( + {"image_embeds": image_embeds} + if ip_adapter_image is not None or ip_adapter_image_embeds is not None + else None + ) + + # 7.2 Optionally get Guidance Scale Embedding + timestep_cond = None + if self.unet.config.time_cond_proj_dim is not None: + guidance_scale_tensor = torch.tensor(self.guidance_scale - 1).repeat(batch_size * num_images_per_prompt) + timestep_cond = self.get_guidance_scale_embedding( + guidance_scale_tensor, embedding_dim=self.unet.config.time_cond_proj_dim + ).to(device=device, dtype=latents.dtype) + + # 8. Denoising loop + num_warmup_steps = len(timesteps) - num_inference_steps * self.scheduler.order + if self.do_perturbed_attention_guidance: + original_attn_proc = self.unet.attn_processors + self._set_pag_attn_processor( + pag_applied_layers=self.pag_applied_layers, + do_classifier_free_guidance=self.do_classifier_free_guidance, + ) + self._num_timesteps = len(timesteps) + + num_warmup_steps = max(len(timesteps) - num_inference_steps * self.scheduler.order, 0) + with self.progress_bar(total=num_inference_steps) as progress_bar: + for i, t in enumerate(timesteps): + if self.interrupt: + continue + + # expand the latents if we are doing classifier free guidance + latent_model_input = torch.cat([latents] * (prompt_embeds.shape[0] // latents.shape[0])) + 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, + timestep_cond=timestep_cond, + cross_attention_kwargs=self.cross_attention_kwargs, + added_cond_kwargs=added_cond_kwargs, + return_dict=False, + )[0] + + # perform guidance + if self.do_perturbed_attention_guidance: + noise_pred = self._apply_perturbed_attention_guidance( + noise_pred, self.do_classifier_free_guidance, self.guidance_scale, t + ) + elif self.do_classifier_free_guidance: + noise_pred_uncond, noise_pred_text = noise_pred.chunk(2) + noise_pred = noise_pred_uncond + self.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, return_dict=False)[0] + + if callback_on_step_end is not None: + callback_kwargs = {} + for k in callback_on_step_end_tensor_inputs: + callback_kwargs[k] = locals()[k] + callback_outputs = callback_on_step_end(self, i, t, callback_kwargs) + + latents = callback_outputs.pop("latents", latents) + prompt_embeds = callback_outputs.pop("prompt_embeds", prompt_embeds) + negative_prompt_embeds = callback_outputs.pop("negative_prompt_embeds", negative_prompt_embeds) + + # 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: + step_idx = i // getattr(self.scheduler, "order", 1) + callback(step_idx, t, latents) + + if not output_type == "latent": + image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False, generator=generator)[ + 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 self.do_perturbed_attention_guidance: + self.unet.set_attn_processor(original_attn_proc) + + if not return_dict: + return (image, has_nsfw_concept) + + return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=has_nsfw_concept) \ No newline at end of file From 555215de6980d9b6d4567f16cf94bee8e31208cb Mon Sep 17 00:00:00 2001 From: SahilCarterr <110806554+SahilCarterr@users.noreply.github.com> Date: Wed, 18 Sep 2024 19:25:47 +0530 Subject: [PATCH 03/17] auto pipeline added --- src/diffusers/pipelines/__init__.py | 2 ++ src/diffusers/pipelines/auto_pipeline.py | 2 ++ src/diffusers/pipelines/pag/__init__.py | 2 ++ 3 files changed, 6 insertions(+) diff --git a/src/diffusers/pipelines/__init__.py b/src/diffusers/pipelines/__init__.py index ddab5122d870..cf1b8f06579a 100644 --- a/src/diffusers/pipelines/__init__.py +++ b/src/diffusers/pipelines/__init__.py @@ -163,6 +163,7 @@ "HunyuanDiTPAGPipeline", "StableDiffusion3PAGPipeline", "StableDiffusionPAGPipeline", + "StableDiffusionPAGImg2ImgPipeline" "StableDiffusionControlNetPAGPipeline", "StableDiffusionXLPAGPipeline", "StableDiffusionXLPAGInpaintPipeline", @@ -568,6 +569,7 @@ StableDiffusion3PAGPipeline, StableDiffusionControlNetPAGPipeline, StableDiffusionPAGPipeline, + StableDiffusionPAGImg2ImgPipeline, StableDiffusionXLControlNetPAGImg2ImgPipeline, StableDiffusionXLControlNetPAGPipeline, StableDiffusionXLPAGImg2ImgPipeline, diff --git a/src/diffusers/pipelines/auto_pipeline.py b/src/diffusers/pipelines/auto_pipeline.py index 39ceadb5acef..96111c7da35d 100644 --- a/src/diffusers/pipelines/auto_pipeline.py +++ b/src/diffusers/pipelines/auto_pipeline.py @@ -56,6 +56,7 @@ StableDiffusion3PAGPipeline, StableDiffusionControlNetPAGPipeline, StableDiffusionPAGPipeline, + StableDiffusionPAGImg2ImgPipeline, StableDiffusionXLControlNetPAGImg2ImgPipeline, StableDiffusionXLControlNetPAGPipeline, StableDiffusionXLPAGImg2ImgPipeline, @@ -123,6 +124,7 @@ ("kandinsky22", KandinskyV22Img2ImgCombinedPipeline), ("kandinsky3", Kandinsky3Img2ImgPipeline), ("stable-diffusion-controlnet", StableDiffusionControlNetImg2ImgPipeline), + ("stable-diffusion-pag", StableDiffusionPAGImg2ImgPipeline), ("stable-diffusion-xl-controlnet", StableDiffusionXLControlNetImg2ImgPipeline), ("stable-diffusion-xl-pag", StableDiffusionXLPAGImg2ImgPipeline), ("stable-diffusion-xl-controlnet-pag", StableDiffusionXLControlNetPAGImg2ImgPipeline), diff --git a/src/diffusers/pipelines/pag/__init__.py b/src/diffusers/pipelines/pag/__init__.py index d8842ce91175..a60eea732acd 100644 --- a/src/diffusers/pipelines/pag/__init__.py +++ b/src/diffusers/pipelines/pag/__init__.py @@ -29,6 +29,7 @@ _import_structure["pipeline_pag_kolors"] = ["KolorsPAGPipeline"] _import_structure["pipeline_pag_pixart_sigma"] = ["PixArtSigmaPAGPipeline"] _import_structure["pipeline_pag_sd"] = ["StableDiffusionPAGPipeline"] + _import_structure["pipeline_pag_sd_img2img"] = ["StableDiffusionPAGImg2ImgPipeline"] _import_structure["pipeline_pag_sd_3"] = ["StableDiffusion3PAGPipeline"] _import_structure["pipeline_pag_sd_animatediff"] = ["AnimateDiffPAGPipeline"] _import_structure["pipeline_pag_sd_xl"] = ["StableDiffusionXLPAGPipeline"] @@ -50,6 +51,7 @@ from .pipeline_pag_kolors import KolorsPAGPipeline from .pipeline_pag_pixart_sigma import PixArtSigmaPAGPipeline from .pipeline_pag_sd import StableDiffusionPAGPipeline + from .pipeline_pag_sd_img2img import StableDiffusionPAGImg2ImgPipeline from .pipeline_pag_sd_3 import StableDiffusion3PAGPipeline from .pipeline_pag_sd_animatediff import AnimateDiffPAGPipeline from .pipeline_pag_sd_xl import StableDiffusionXLPAGPipeline From b6150da732a3dc18bb8e177022ef8edf10aba2df Mon Sep 17 00:00:00 2001 From: SahilCarterr <110806554+SahilCarterr@users.noreply.github.com> Date: Wed, 18 Sep 2024 21:38:20 +0530 Subject: [PATCH 04/17] docs update --- docs/source/en/api/pipelines/pag.md | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/docs/source/en/api/pipelines/pag.md b/docs/source/en/api/pipelines/pag.md index aa69598ae290..ecc57468d14e 100644 --- a/docs/source/en/api/pipelines/pag.md +++ b/docs/source/en/api/pipelines/pag.md @@ -53,6 +53,11 @@ Since RegEx is supported as a way for matching layer identifiers, it is crucial - all - __call__ +## StableDiffusionPAGImg2ImgPipeline +[[autodoc]] StableDiffusionPAGImg2ImgPipeline + - all + - __call__ + ## StableDiffusionControlNetPAGPipeline [[autodoc]] StableDiffusionControlNetPAGPipeline - all From ce8b22d5c5d5f0116ac1db861bf487481c15412a Mon Sep 17 00:00:00 2001 From: SahilCarterr <110806554+SahilCarterr@users.noreply.github.com> Date: Wed, 18 Sep 2024 22:54:29 +0530 Subject: [PATCH 05/17] add test --- tests/pipelines/pag/test_pag_sd_img2img.py | 422 +++++++++++++++++++++ 1 file changed, 422 insertions(+) create mode 100644 tests/pipelines/pag/test_pag_sd_img2img.py diff --git a/tests/pipelines/pag/test_pag_sd_img2img.py b/tests/pipelines/pag/test_pag_sd_img2img.py new file mode 100644 index 000000000000..1cfe42193e7e --- /dev/null +++ b/tests/pipelines/pag/test_pag_sd_img2img.py @@ -0,0 +1,422 @@ +# coding=utf-8 +# Copyright 2024 HuggingFace Inc. +# +# 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 gc +import random +import traceback +import unittest + +import numpy as np +import torch +from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer + +from diffusers import ( + AutoencoderKL, + AutoencoderTiny, + DDIMScheduler, + DPMSolverMultistepScheduler, + HeunDiscreteScheduler, + LCMScheduler, + LMSDiscreteScheduler, + PNDMScheduler, + StableDiffusionImg2ImgPipeline, + UNet2DConditionModel, +) +from diffusers.utils.testing_utils import ( + enable_full_determinism, + floats_tensor, + is_torch_compile, + load_image, + load_numpy, + nightly, + require_torch_2, + require_torch_gpu, + run_test_in_subprocess, + skip_mps, + slow, + torch_device, +) + +from ..pipeline_params import ( + IMAGE_TO_IMAGE_IMAGE_PARAMS, + TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS, + TEXT_GUIDED_IMAGE_VARIATION_PARAMS, + TEXT_TO_IMAGE_CALLBACK_CFG_PARAMS, +) +from ..test_pipelines_common import ( + IPAdapterTesterMixin, + PipelineKarrasSchedulerTesterMixin, + PipelineLatentTesterMixin, + PipelineTesterMixin, +) + + +enable_full_determinism() + + + + +class StableDiffusionPAGImg2ImgPipelineFastTests( + IPAdapterTesterMixin, + PipelineLatentTesterMixin, + PipelineKarrasSchedulerTesterMixin, + PipelineTesterMixin, + unittest.TestCase, +): + pipeline_class = StableDiffusionPAGImg2ImgPipeline + params = TEXT_GUIDED_IMAGE_VARIATION_PARAMS - {"height", "width"} + required_optional_params = PipelineTesterMixin.required_optional_params - {"latents"} + batch_params = TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS + image_params = IMAGE_TO_IMAGE_IMAGE_PARAMS + image_latents_params = IMAGE_TO_IMAGE_IMAGE_PARAMS + callback_cfg_params = TEXT_TO_IMAGE_CALLBACK_CFG_PARAMS + + def get_dummy_components(self, time_cond_proj_dim=None): + torch.manual_seed(0) + unet = UNet2DConditionModel( + block_out_channels=(32, 64), + layers_per_block=2, + time_cond_proj_dim=time_cond_proj_dim, + sample_size=32, + in_channels=4, + out_channels=4, + down_block_types=("DownBlock2D", "CrossAttnDownBlock2D"), + up_block_types=("CrossAttnUpBlock2D", "UpBlock2D"), + cross_attention_dim=32, + ) + scheduler = PNDMScheduler(skip_prk_steps=True) + torch.manual_seed(0) + vae = AutoencoderKL( + block_out_channels=[32, 64], + in_channels=3, + out_channels=3, + down_block_types=["DownEncoderBlock2D", "DownEncoderBlock2D"], + up_block_types=["UpDecoderBlock2D", "UpDecoderBlock2D"], + latent_channels=4, + ) + torch.manual_seed(0) + text_encoder_config = CLIPTextConfig( + bos_token_id=0, + eos_token_id=2, + hidden_size=32, + intermediate_size=37, + layer_norm_eps=1e-05, + num_attention_heads=4, + num_hidden_layers=5, + pad_token_id=1, + vocab_size=1000, + ) + text_encoder = CLIPTextModel(text_encoder_config) + tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip") + + components = { + "unet": unet, + "scheduler": scheduler, + "vae": vae, + "text_encoder": text_encoder, + "tokenizer": tokenizer, + "safety_checker": None, + "feature_extractor": None, + "image_encoder": None, + } + return components + + def get_dummy_tiny_autoencoder(self): + return AutoencoderTiny(in_channels=3, out_channels=3, latent_channels=4) + + def get_dummy_inputs(self, device, seed=0): + image = floats_tensor((1, 3, 32, 32), rng=random.Random(seed)).to(device) + image = image / 2 + 0.5 + if str(device).startswith("mps"): + generator = torch.manual_seed(seed) + else: + generator = torch.Generator(device=device).manual_seed(seed) + inputs = { + "prompt": "A painting of a squirrel eating a burger", + "image": image, + "generator": generator, + "num_inference_steps": 2, + "guidance_scale": 6.0, + "output_type": "np", + } + return inputs + + def test_pag_disable_enable(self): + device = "cpu" # ensure determinism for the device-dependent torch.Generator + components = self.get_dummy_components() + + # base pipeline (expect same output when pag is disabled) + pipe_sd = StableDiffusionPipeline(**components) + pipe_sd = pipe_sd.to(device) + pipe_sd.set_progress_bar_config(disable=None) + + inputs = self.get_dummy_inputs(device) + del inputs["pag_scale"] + assert ( + "pag_scale" not in inspect.signature(pipe_sd.__call__).parameters + ), f"`pag_scale` should not be a call parameter of the base pipeline {pipe_sd.__class__.__name__}." + out = pipe_sd(**inputs).images[0, -3:, -3:, -1] + + # pag disabled with pag_scale=0.0 + pipe_pag = self.pipeline_class(**components) + pipe_pag = pipe_pag.to(device) + pipe_pag.set_progress_bar_config(disable=None) + + inputs = self.get_dummy_inputs(device) + inputs["pag_scale"] = 0.0 + out_pag_disabled = pipe_pag(**inputs).images[0, -3:, -3:, -1] + + # pag enabled + pipe_pag = self.pipeline_class(**components, pag_applied_layers=["mid", "up", "down"]) + pipe_pag = pipe_pag.to(device) + pipe_pag.set_progress_bar_config(disable=None) + + inputs = self.get_dummy_inputs(device) + out_pag_enabled = pipe_pag(**inputs).images[0, -3:, -3:, -1] + + assert np.abs(out.flatten() - out_pag_disabled.flatten()).max() < 1e-3 + assert np.abs(out.flatten() - out_pag_enabled.flatten()).max() > 1e-3 + + def test_pag_applied_layers(self): + device = "cpu" # ensure determinism for the device-dependent torch.Generator + components = self.get_dummy_components() + + # base pipeline + pipe = self.pipeline_class(**components) + pipe = pipe.to(device) + pipe.set_progress_bar_config(disable=None) + + # pag_applied_layers = ["mid","up","down"] should apply to all self-attention layers + all_self_attn_layers = [k for k in pipe.unet.attn_processors.keys() if "attn1" in k] + original_attn_procs = pipe.unet.attn_processors + pag_layers = [ + "down", + "mid", + "up", + ] + pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) + assert set(pipe.pag_attn_processors) == set(all_self_attn_layers) + + # pag_applied_layers = ["mid"], or ["mid.block_0"] or ["mid.block_0.attentions_0"] should apply to all self-attention layers in mid_block, i.e. + # mid_block.attentions.0.transformer_blocks.0.attn1.processor + # mid_block.attentions.0.transformer_blocks.1.attn1.processor + all_self_attn_mid_layers = [ + "mid_block.attentions.0.transformer_blocks.0.attn1.processor", + # "mid_block.attentions.0.transformer_blocks.1.attn1.processor", + ] + pipe.unet.set_attn_processor(original_attn_procs.copy()) + pag_layers = ["mid"] + pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) + assert set(pipe.pag_attn_processors) == set(all_self_attn_mid_layers) + + pipe.unet.set_attn_processor(original_attn_procs.copy()) + pag_layers = ["mid_block"] + pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) + assert set(pipe.pag_attn_processors) == set(all_self_attn_mid_layers) + + pipe.unet.set_attn_processor(original_attn_procs.copy()) + pag_layers = ["mid_block.attentions.0"] + pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) + assert set(pipe.pag_attn_processors) == set(all_self_attn_mid_layers) + + # pag_applied_layers = ["mid.block_0.attentions_1"] does not exist in the model + pipe.unet.set_attn_processor(original_attn_procs.copy()) + pag_layers = ["mid_block.attentions.1"] + with self.assertRaises(ValueError): + pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) + + # pag_applied_layers = "down" should apply to all self-attention layers in down_blocks + # down_blocks.1.attentions.0.transformer_blocks.0.attn1.processor + # down_blocks.1.attentions.0.transformer_blocks.1.attn1.processor + # down_blocks.1.attentions.0.transformer_blocks.0.attn1.processor + + pipe.unet.set_attn_processor(original_attn_procs.copy()) + pag_layers = ["down"] + pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) + assert len(pipe.pag_attn_processors) == 2 + + pipe.unet.set_attn_processor(original_attn_procs.copy()) + pag_layers = ["down_blocks.0"] + with self.assertRaises(ValueError): + pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) + + pipe.unet.set_attn_processor(original_attn_procs.copy()) + pag_layers = ["down_blocks.1"] + pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) + assert len(pipe.pag_attn_processors) == 2 + + pipe.unet.set_attn_processor(original_attn_procs.copy()) + pag_layers = ["down_blocks.1.attentions.1"] + pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) + assert len(pipe.pag_attn_processors) == 1 + + def test_pag_inference(self): + device = "cpu" # ensure determinism for the device-dependent torch.Generator + components = self.get_dummy_components() + + pipe_pag = self.pipeline_class(**components, pag_applied_layers=["mid", "up", "down"]) + pipe_pag = pipe_pag.to(device) + pipe_pag.set_progress_bar_config(disable=None) + + inputs = self.get_dummy_inputs(device) + image = pipe_pag(**inputs).images + image_slice = image[0, -3:, -3:, -1] + + assert image.shape == ( + 1, + 64, + 64, + 3, + ), f"the shape of the output image should be (1, 64, 64, 3) but got {image.shape}" + + expected_slice = np.array( + [0.22802538, 0.44626093, 0.48905736, 0.29633686, 0.36400637, 0.4724258, 0.4678891, 0.32260418, 0.41611585] + ) + max_diff = np.abs(image_slice.flatten() - expected_slice).max() + self.assertLessEqual(max_diff, 1e-3) + +@slow +@require_torch_gpu +class StableDiffusionPAGImg2ImgPipelineSlowTests(unittest.TestCase): + def setUp(self): + super().setUp() + gc.collect() + torch.cuda.empty_cache() + + def tearDown(self): + super().tearDown() + gc.collect() + torch.cuda.empty_cache() + + def get_inputs(self, device, generator_device="cpu", dtype=torch.float32, seed=0): + generator = torch.Generator(device=generator_device).manual_seed(seed) + init_image = load_image( + "https://huggingface.co/datasets/diffusers/test-arrays/resolve/main" + "/stable_diffusion_img2img/sketch-mountains-input.png" + ) + inputs = { + "prompt": "a fantasy landscape, concept art, high resolution", + "image": init_image, + "generator": generator, + "num_inference_steps": 3, + "strength": 0.75, + "guidance_scale": 7.5, + "output_type": "np", + } + return inputs + + def test_pag_cfg(self): + pipeline = AutoPipelineForText2Image.from_pretrained(self.repo_id, enable_pag=True, torch_dtype=torch.float16) + pipeline.enable_model_cpu_offload() + pipeline.set_progress_bar_config(disable=None) + + inputs = self.get_inputs(torch_device) + image = pipeline(**inputs).images + + image_slice = image[0, -3:, -3:, -1].flatten() + assert image.shape == (1, 512, 512, 3) + print(image_slice.flatten()) + expected_slice = np.array( + [0.58251953, 0.5722656, 0.5683594, 0.55029297, 0.52001953, 0.52001953, 0.49951172, 0.45410156, 0.50146484] + ) + assert ( + np.abs(image_slice.flatten() - expected_slice).max() < 1e-3 + ), f"output is different from expected, {image_slice.flatten()}" + + def test_pag_uncond(self): + pipeline = AutoPipelineForText2Image.from_pretrained(self.repo_id, enable_pag=True, torch_dtype=torch.float16) + pipeline.enable_model_cpu_offload() + pipeline.set_progress_bar_config(disable=None) + + inputs = self.get_inputs(torch_device, guidance_scale=0.0) + image = pipeline(**inputs).images + + image_slice = image[0, -3:, -3:, -1].flatten() + assert image.shape == (1, 512, 512, 3) + expected_slice = np.array( + [0.5986328, 0.52441406, 0.3972168, 0.4741211, 0.34985352, 0.22705078, 0.4128418, 0.2866211, 0.31713867] + ) + print(image_slice.flatten()) + assert ( + np.abs(image_slice.flatten() - expected_slice).max() < 1e-3 + ), f"output is different from expected, {image_slice.flatten()}" + +@nightly +@require_torch_gpu +class StableDiffusionImg2ImgPipelineNightlyTests(unittest.TestCase): + pipeline_class = StableDiffusionPAGImage2ImagePipeline + repo_id = "Jiali/stable-diffusion-1.5" + def setUp(self): + super().setUp() + gc.collect() + torch.cuda.empty_cache() + + def tearDown(self): + super().tearDown() + gc.collect() + torch.cuda.empty_cache() + + def get_inputs(self, device, generator_device="cpu", dtype=torch.float32, seed=0): + generator = torch.Generator(device=generator_device).manual_seed(seed) + init_image = load_image( + "https://huggingface.co/datasets/diffusers/test-arrays/resolve/main" + "/stable_diffusion_img2img/sketch-mountains-input.png" + ) + inputs = { + "prompt": "a fantasy landscape, concept art, high resolution", + "image": init_image, + "generator": generator, + "num_inference_steps": 50, + "strength": 0.75, + "guidance_scale": 7.5, + "output_type": "np", + } + return inputs + + def test_pag_cfg(self): + pipeline = AutoPipelineForImage2Image.from_pretrained(self.repo_id, enable_pag=True, torch_dtype=torch.float16) + pipeline.enable_model_cpu_offload() + pipeline.set_progress_bar_config(disable=None) + + inputs = self.get_inputs(torch_device) + image = pipeline(**inputs).images + + image_slice = image[0, -3:, -3:, -1].flatten() + assert image.shape == (1, 512, 512, 3) + print(image_slice.flatten()) + expected_slice = np.array( + [0.58251953, 0.5722656, 0.5683594, 0.55029297, 0.52001953, 0.52001953, 0.49951172, 0.45410156, 0.50146484] + ) + assert ( + np.abs(image_slice.flatten() - expected_slice).max() < 1e-3 + ), f"output is different from expected, {image_slice.flatten()}" + + def test_pag_uncond(self): + pipeline = AutoPipelineForImage2Image.from_pretrained(self.repo_id, enable_pag=True, torch_dtype=torch.float16) + pipeline.enable_model_cpu_offload() + pipeline.set_progress_bar_config(disable=None) + + inputs = self.get_inputs(torch_device, guidance_scale=0.0) + image = pipeline(**inputs).images + + image_slice = image[0, -3:, -3:, -1].flatten() + assert image.shape == (1, 512, 512, 3) + expected_slice = np.array( + [0.5986328, 0.52441406, 0.3972168, 0.4741211, 0.34985352, 0.22705078, 0.4128418, 0.2866211, 0.31713867] + ) + print(image_slice.flatten()) + assert ( + np.abs(image_slice.flatten() - expected_slice).max() < 1e-3 + ), f"output is different from expected, {image_slice.flatten()}" From 00dfea69e6053c2a9699caac1c1515370577052c Mon Sep 17 00:00:00 2001 From: SahilCarterr <110806554+SahilCarterr@users.noreply.github.com> Date: Thu, 19 Sep 2024 19:07:14 +0530 Subject: [PATCH 06/17] fix #copied from and tests --- src/diffusers/pipelines/__init__.py | 2 +- src/diffusers/pipelines/auto_pipeline.py | 2 +- src/diffusers/pipelines/pag/__init__.py | 2 +- .../pipelines/pag/pipeline_pag_sd_img2img.py | 31 +----- tests/pipelines/pag/test_pag_sd_img2img.py | 97 ++----------------- 5 files changed, 14 insertions(+), 120 deletions(-) diff --git a/src/diffusers/pipelines/__init__.py b/src/diffusers/pipelines/__init__.py index cf1b8f06579a..2486f2255ad1 100644 --- a/src/diffusers/pipelines/__init__.py +++ b/src/diffusers/pipelines/__init__.py @@ -568,8 +568,8 @@ PixArtSigmaPAGPipeline, StableDiffusion3PAGPipeline, StableDiffusionControlNetPAGPipeline, - StableDiffusionPAGPipeline, StableDiffusionPAGImg2ImgPipeline, + StableDiffusionPAGPipeline, StableDiffusionXLControlNetPAGImg2ImgPipeline, StableDiffusionXLControlNetPAGPipeline, StableDiffusionXLPAGImg2ImgPipeline, diff --git a/src/diffusers/pipelines/auto_pipeline.py b/src/diffusers/pipelines/auto_pipeline.py index 96111c7da35d..c9eb1b8fdfb0 100644 --- a/src/diffusers/pipelines/auto_pipeline.py +++ b/src/diffusers/pipelines/auto_pipeline.py @@ -55,8 +55,8 @@ PixArtSigmaPAGPipeline, StableDiffusion3PAGPipeline, StableDiffusionControlNetPAGPipeline, - StableDiffusionPAGPipeline, StableDiffusionPAGImg2ImgPipeline, + StableDiffusionPAGPipeline, StableDiffusionXLControlNetPAGImg2ImgPipeline, StableDiffusionXLControlNetPAGPipeline, StableDiffusionXLPAGImg2ImgPipeline, diff --git a/src/diffusers/pipelines/pag/__init__.py b/src/diffusers/pipelines/pag/__init__.py index a60eea732acd..60329ad903b9 100644 --- a/src/diffusers/pipelines/pag/__init__.py +++ b/src/diffusers/pipelines/pag/__init__.py @@ -51,9 +51,9 @@ from .pipeline_pag_kolors import KolorsPAGPipeline from .pipeline_pag_pixart_sigma import PixArtSigmaPAGPipeline from .pipeline_pag_sd import StableDiffusionPAGPipeline - from .pipeline_pag_sd_img2img import StableDiffusionPAGImg2ImgPipeline from .pipeline_pag_sd_3 import StableDiffusion3PAGPipeline from .pipeline_pag_sd_animatediff import AnimateDiffPAGPipeline + from .pipeline_pag_sd_img2img import StableDiffusionPAGImg2ImgPipeline from .pipeline_pag_sd_xl import StableDiffusionXLPAGPipeline from .pipeline_pag_sd_xl_img2img import StableDiffusionXLPAGImg2ImgPipeline from .pipeline_pag_sd_xl_inpaint import StableDiffusionXLPAGInpaintPipeline diff --git a/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py b/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py index 5b34f7932224..bd41d6e1c3ab 100644 --- a/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py +++ b/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py @@ -15,7 +15,6 @@ import inspect from typing import Any, Callable, Dict, List, Optional, Union -import numpy as np import PIL.Image import torch from packaging import version @@ -29,7 +28,6 @@ from ...models.lora import adjust_lora_scale_text_encoder from ...schedulers import KarrasDiffusionSchedulers from ...utils import ( - PIL_INTERPOLATION, USE_PEFT_BACKEND, deprecate, logging, @@ -570,6 +568,7 @@ def prepare_extra_step_kwargs(self, generator, eta): extra_step_kwargs["generator"] = generator return extra_step_kwargs + # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion_img2img.check_inputs def check_inputs( self, prompt, @@ -638,6 +637,7 @@ def check_inputs( f"`ip_adapter_image_embeds` has to be a list of 3D or 4D tensors but is {ip_adapter_image_embeds[0].ndim}D" ) + # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion_img2img.get_timesteps def get_timesteps(self, num_inference_steps, strength, device): # get the original timestep using init_timestep init_timestep = min(int(num_inference_steps * strength), num_inference_steps) @@ -649,6 +649,7 @@ def get_timesteps(self, num_inference_steps, strength, device): return timesteps, num_inference_steps - t_start + # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion_img2img.prepare_latents 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( @@ -801,7 +802,6 @@ def __call__( callback_on_step_end_tensor_inputs: List[str] = ["latents"], pag_scale: float = 3.0, pag_adaptive_scale: float = 0.0, - **kwargs, ): r""" The call function to the pipeline for generation. @@ -895,22 +895,6 @@ def __call__( "not-safe-for-work" (nsfw) content. """ - callback = kwargs.pop("callback", None) - callback_steps = kwargs.pop("callback_steps", None) - - if callback is not None: - deprecate( - "callback", - "1.0.0", - "Passing `callback` as an input argument to `__call__` is deprecated, consider use `callback_on_step_end`", - ) - if callback_steps is not None: - deprecate( - "callback_steps", - "1.0.0", - "Passing `callback_steps` as an input argument to `__call__` is deprecated, consider use `callback_on_step_end`", - ) - if isinstance(callback_on_step_end, (PipelineCallback, MultiPipelineCallbacks)): callback_on_step_end_tensor_inputs = callback_on_step_end.tensor_inputs @@ -1085,13 +1069,6 @@ def __call__( prompt_embeds = callback_outputs.pop("prompt_embeds", prompt_embeds) negative_prompt_embeds = callback_outputs.pop("negative_prompt_embeds", negative_prompt_embeds) - # 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: - step_idx = i // getattr(self.scheduler, "order", 1) - callback(step_idx, t, latents) - if not output_type == "latent": image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False, generator=generator)[ 0 @@ -1117,4 +1094,4 @@ def __call__( if not return_dict: return (image, has_nsfw_concept) - return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=has_nsfw_concept) \ No newline at end of file + return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=has_nsfw_concept) diff --git a/tests/pipelines/pag/test_pag_sd_img2img.py b/tests/pipelines/pag/test_pag_sd_img2img.py index 1cfe42193e7e..23210c8f7814 100644 --- a/tests/pipelines/pag/test_pag_sd_img2img.py +++ b/tests/pipelines/pag/test_pag_sd_img2img.py @@ -15,7 +15,6 @@ import gc import random -import traceback import unittest import numpy as np @@ -25,26 +24,16 @@ from diffusers import ( AutoencoderKL, AutoencoderTiny, - DDIMScheduler, - DPMSolverMultistepScheduler, - HeunDiscreteScheduler, - LCMScheduler, - LMSDiscreteScheduler, + AutoPipelineForImage2Image, PNDMScheduler, - StableDiffusionImg2ImgPipeline, + StableDiffusionPAGImage2ImagePipeline, UNet2DConditionModel, ) from diffusers.utils.testing_utils import ( enable_full_determinism, floats_tensor, - is_torch_compile, load_image, - load_numpy, - nightly, - require_torch_2, require_torch_gpu, - run_test_in_subprocess, - skip_mps, slow, torch_device, ) @@ -76,7 +65,7 @@ class StableDiffusionPAGImg2ImgPipelineFastTests( unittest.TestCase, ): pipeline_class = StableDiffusionPAGImg2ImgPipeline - params = TEXT_GUIDED_IMAGE_VARIATION_PARAMS - {"height", "width"} + params = TEXT_GUIDED_IMAGE_VARIATION_PARAMS.union({"pag_scale", "pag_adaptive_scale"}) required_optional_params = PipelineTesterMixin.required_optional_params - {"latents"} batch_params = TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS image_params = IMAGE_TO_IMAGE_IMAGE_PARAMS @@ -152,7 +141,7 @@ def get_dummy_inputs(self, device, seed=0): "output_type": "np", } return inputs - + def test_pag_disable_enable(self): device = "cpu" # ensure determinism for the device-dependent torch.Generator components = self.get_dummy_components() @@ -189,78 +178,6 @@ def test_pag_disable_enable(self): assert np.abs(out.flatten() - out_pag_disabled.flatten()).max() < 1e-3 assert np.abs(out.flatten() - out_pag_enabled.flatten()).max() > 1e-3 - def test_pag_applied_layers(self): - device = "cpu" # ensure determinism for the device-dependent torch.Generator - components = self.get_dummy_components() - - # base pipeline - pipe = self.pipeline_class(**components) - pipe = pipe.to(device) - pipe.set_progress_bar_config(disable=None) - - # pag_applied_layers = ["mid","up","down"] should apply to all self-attention layers - all_self_attn_layers = [k for k in pipe.unet.attn_processors.keys() if "attn1" in k] - original_attn_procs = pipe.unet.attn_processors - pag_layers = [ - "down", - "mid", - "up", - ] - pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) - assert set(pipe.pag_attn_processors) == set(all_self_attn_layers) - - # pag_applied_layers = ["mid"], or ["mid.block_0"] or ["mid.block_0.attentions_0"] should apply to all self-attention layers in mid_block, i.e. - # mid_block.attentions.0.transformer_blocks.0.attn1.processor - # mid_block.attentions.0.transformer_blocks.1.attn1.processor - all_self_attn_mid_layers = [ - "mid_block.attentions.0.transformer_blocks.0.attn1.processor", - # "mid_block.attentions.0.transformer_blocks.1.attn1.processor", - ] - pipe.unet.set_attn_processor(original_attn_procs.copy()) - pag_layers = ["mid"] - pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) - assert set(pipe.pag_attn_processors) == set(all_self_attn_mid_layers) - - pipe.unet.set_attn_processor(original_attn_procs.copy()) - pag_layers = ["mid_block"] - pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) - assert set(pipe.pag_attn_processors) == set(all_self_attn_mid_layers) - - pipe.unet.set_attn_processor(original_attn_procs.copy()) - pag_layers = ["mid_block.attentions.0"] - pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) - assert set(pipe.pag_attn_processors) == set(all_self_attn_mid_layers) - - # pag_applied_layers = ["mid.block_0.attentions_1"] does not exist in the model - pipe.unet.set_attn_processor(original_attn_procs.copy()) - pag_layers = ["mid_block.attentions.1"] - with self.assertRaises(ValueError): - pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) - - # pag_applied_layers = "down" should apply to all self-attention layers in down_blocks - # down_blocks.1.attentions.0.transformer_blocks.0.attn1.processor - # down_blocks.1.attentions.0.transformer_blocks.1.attn1.processor - # down_blocks.1.attentions.0.transformer_blocks.0.attn1.processor - - pipe.unet.set_attn_processor(original_attn_procs.copy()) - pag_layers = ["down"] - pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) - assert len(pipe.pag_attn_processors) == 2 - - pipe.unet.set_attn_processor(original_attn_procs.copy()) - pag_layers = ["down_blocks.0"] - with self.assertRaises(ValueError): - pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) - - pipe.unet.set_attn_processor(original_attn_procs.copy()) - pag_layers = ["down_blocks.1"] - pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) - assert len(pipe.pag_attn_processors) == 2 - - pipe.unet.set_attn_processor(original_attn_procs.copy()) - pag_layers = ["down_blocks.1.attentions.1"] - pipe._set_pag_attn_processor(pag_applied_layers=pag_layers, do_classifier_free_guidance=False) - assert len(pipe.pag_attn_processors) == 1 def test_pag_inference(self): device = "cpu" # ensure determinism for the device-dependent torch.Generator @@ -289,7 +206,7 @@ def test_pag_inference(self): @slow @require_torch_gpu -class StableDiffusionPAGImg2ImgPipelineSlowTests(unittest.TestCase): +class StableDiffusionPAGImg2ImgPipelineIntegrationTests(unittest.TestCase): def setUp(self): super().setUp() gc.collect() @@ -353,7 +270,7 @@ def test_pag_uncond(self): np.abs(image_slice.flatten() - expected_slice).max() < 1e-3 ), f"output is different from expected, {image_slice.flatten()}" -@nightly +@slow @require_torch_gpu class StableDiffusionImg2ImgPipelineNightlyTests(unittest.TestCase): pipeline_class = StableDiffusionPAGImage2ImagePipeline @@ -384,7 +301,7 @@ def get_inputs(self, device, generator_device="cpu", dtype=torch.float32, seed=0 "output_type": "np", } return inputs - + def test_pag_cfg(self): pipeline = AutoPipelineForImage2Image.from_pretrained(self.repo_id, enable_pag=True, torch_dtype=torch.float16) pipeline.enable_model_cpu_offload() From bd002acbab38bd0990147480b2702a1ddeecf2ed Mon Sep 17 00:00:00 2001 From: SahilCarterr <110806554+SahilCarterr@users.noreply.github.com> Date: Thu, 19 Sep 2024 19:42:29 +0530 Subject: [PATCH 07/17] fix typo --- src/diffusers/__init__.py | 2 ++ src/diffusers/pipelines/__init__.py | 2 +- 2 files changed, 3 insertions(+), 1 deletion(-) diff --git a/src/diffusers/__init__.py b/src/diffusers/__init__.py index dedb6f5c7f14..0beddc6db222 100644 --- a/src/diffusers/__init__.py +++ b/src/diffusers/__init__.py @@ -344,6 +344,7 @@ "StableDiffusionLDM3DPipeline", "StableDiffusionModelEditingPipeline", "StableDiffusionPAGPipeline", + "StableDiffusionPAGImg2ImgPipeline", "StableDiffusionPanoramaPipeline", "StableDiffusionParadigmsPipeline", "StableDiffusionPipeline", @@ -793,6 +794,7 @@ StableDiffusionLatentUpscalePipeline, StableDiffusionLDM3DPipeline, StableDiffusionModelEditingPipeline, + StableDiffusionPAGImg2ImgPipeline, StableDiffusionPAGPipeline, StableDiffusionPanoramaPipeline, StableDiffusionParadigmsPipeline, diff --git a/src/diffusers/pipelines/__init__.py b/src/diffusers/pipelines/__init__.py index 2486f2255ad1..86420a8570ce 100644 --- a/src/diffusers/pipelines/__init__.py +++ b/src/diffusers/pipelines/__init__.py @@ -163,7 +163,7 @@ "HunyuanDiTPAGPipeline", "StableDiffusion3PAGPipeline", "StableDiffusionPAGPipeline", - "StableDiffusionPAGImg2ImgPipeline" + "StableDiffusionPAGImg2ImgPipeline", "StableDiffusionControlNetPAGPipeline", "StableDiffusionXLPAGPipeline", "StableDiffusionXLPAGInpaintPipeline", From ec0f21ab31b4ded1e19a65ed5e495ef4292974f1 Mon Sep 17 00:00:00 2001 From: SahilCarterr <110806554+SahilCarterr@users.noreply.github.com> Date: Fri, 20 Sep 2024 07:33:09 +0530 Subject: [PATCH 08/17] remove test --- .../pipelines/pag/pipeline_pag_sd_img2img.py | 9 +-- tests/pipelines/pag/test_pag_sd_img2img.py | 77 ++----------------- 2 files changed, 8 insertions(+), 78 deletions(-) diff --git a/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py b/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py index bd41d6e1c3ab..3762f0006b81 100644 --- a/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py +++ b/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py @@ -573,7 +573,6 @@ def check_inputs( self, prompt, strength, - callback_steps, negative_prompt=None, prompt_embeds=None, negative_prompt_embeds=None, @@ -584,12 +583,6 @@ def check_inputs( 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 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 ): @@ -884,7 +877,7 @@ def __call__( pag_adaptive_scale (`float`, *optional*, defaults to 0.0): The adaptive scale factor for the perturbed attention guidance. If it is set to 0.0, `pag_scale` is used. - + Examples: Returns: diff --git a/tests/pipelines/pag/test_pag_sd_img2img.py b/tests/pipelines/pag/test_pag_sd_img2img.py index 23210c8f7814..7f4bc5ddc318 100644 --- a/tests/pipelines/pag/test_pag_sd_img2img.py +++ b/tests/pipelines/pag/test_pag_sd_img2img.py @@ -14,6 +14,7 @@ # limitations under the License. import gc +import inspect import random import unittest @@ -26,7 +27,8 @@ AutoencoderTiny, AutoPipelineForImage2Image, PNDMScheduler, - StableDiffusionPAGImage2ImagePipeline, + StableDiffusionImg2ImgPipeline, + StableDiffusionPAGImg2ImgPipeline, UNet2DConditionModel, ) from diffusers.utils.testing_utils import ( @@ -147,7 +149,7 @@ def test_pag_disable_enable(self): components = self.get_dummy_components() # base pipeline (expect same output when pag is disabled) - pipe_sd = StableDiffusionPipeline(**components) + pipe_sd = StableDiffusionImg2ImgPipeline(**components) pipe_sd = pipe_sd.to(device) pipe_sd.set_progress_bar_config(disable=None) @@ -207,74 +209,9 @@ def test_pag_inference(self): @slow @require_torch_gpu class StableDiffusionPAGImg2ImgPipelineIntegrationTests(unittest.TestCase): - def setUp(self): - super().setUp() - gc.collect() - torch.cuda.empty_cache() - - def tearDown(self): - super().tearDown() - gc.collect() - torch.cuda.empty_cache() - - def get_inputs(self, device, generator_device="cpu", dtype=torch.float32, seed=0): - generator = torch.Generator(device=generator_device).manual_seed(seed) - init_image = load_image( - "https://huggingface.co/datasets/diffusers/test-arrays/resolve/main" - "/stable_diffusion_img2img/sketch-mountains-input.png" - ) - inputs = { - "prompt": "a fantasy landscape, concept art, high resolution", - "image": init_image, - "generator": generator, - "num_inference_steps": 3, - "strength": 0.75, - "guidance_scale": 7.5, - "output_type": "np", - } - return inputs - - def test_pag_cfg(self): - pipeline = AutoPipelineForText2Image.from_pretrained(self.repo_id, enable_pag=True, torch_dtype=torch.float16) - pipeline.enable_model_cpu_offload() - pipeline.set_progress_bar_config(disable=None) - - inputs = self.get_inputs(torch_device) - image = pipeline(**inputs).images - - image_slice = image[0, -3:, -3:, -1].flatten() - assert image.shape == (1, 512, 512, 3) - print(image_slice.flatten()) - expected_slice = np.array( - [0.58251953, 0.5722656, 0.5683594, 0.55029297, 0.52001953, 0.52001953, 0.49951172, 0.45410156, 0.50146484] - ) - assert ( - np.abs(image_slice.flatten() - expected_slice).max() < 1e-3 - ), f"output is different from expected, {image_slice.flatten()}" - - def test_pag_uncond(self): - pipeline = AutoPipelineForText2Image.from_pretrained(self.repo_id, enable_pag=True, torch_dtype=torch.float16) - pipeline.enable_model_cpu_offload() - pipeline.set_progress_bar_config(disable=None) - - inputs = self.get_inputs(torch_device, guidance_scale=0.0) - image = pipeline(**inputs).images - - image_slice = image[0, -3:, -3:, -1].flatten() - assert image.shape == (1, 512, 512, 3) - expected_slice = np.array( - [0.5986328, 0.52441406, 0.3972168, 0.4741211, 0.34985352, 0.22705078, 0.4128418, 0.2866211, 0.31713867] - ) - print(image_slice.flatten()) - assert ( - np.abs(image_slice.flatten() - expected_slice).max() < 1e-3 - ), f"output is different from expected, {image_slice.flatten()}" - -@slow -@require_torch_gpu -class StableDiffusionImg2ImgPipelineNightlyTests(unittest.TestCase): - pipeline_class = StableDiffusionPAGImage2ImagePipeline + pipeline_class = StableDiffusionPAGImg2ImgPipeline repo_id = "Jiali/stable-diffusion-1.5" + def setUp(self): super().setUp() gc.collect() @@ -295,7 +232,7 @@ def get_inputs(self, device, generator_device="cpu", dtype=torch.float32, seed=0 "prompt": "a fantasy landscape, concept art, high resolution", "image": init_image, "generator": generator, - "num_inference_steps": 50, + "num_inference_steps": 3, "strength": 0.75, "guidance_scale": 7.5, "output_type": "np", From bc12960fbf05739f5cd16ca2ec0bc6bb0c853965 Mon Sep 17 00:00:00 2001 From: YiYi Xu Date: Fri, 20 Sep 2024 12:16:55 -1000 Subject: [PATCH 09/17] Update src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py --- src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py | 1 - 1 file changed, 1 deletion(-) diff --git a/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py b/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py index 3762f0006b81..22890aef1f73 100644 --- a/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py +++ b/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py @@ -895,7 +895,6 @@ def __call__( self.check_inputs( prompt, strength, - callback_steps, negative_prompt, prompt_embeds, negative_prompt_embeds, From f4f615357807155eab9121d49235160f71c85864 Mon Sep 17 00:00:00 2001 From: yiyixuxu Date: Sat, 21 Sep 2024 00:28:26 +0200 Subject: [PATCH 10/17] style + copy --- src/diffusers/__init__.py | 2 +- src/diffusers/pipelines/pag/__init__.py | 2 +- .../pipelines/pag/pipeline_pag_sd_img2img.py | 8 ++++---- .../utils/dummy_torch_and_transformers_objects.py | 15 +++++++++++++++ tests/pipelines/pag/test_pag_sd_img2img.py | 4 +--- 5 files changed, 22 insertions(+), 9 deletions(-) diff --git a/src/diffusers/__init__.py b/src/diffusers/__init__.py index 0beddc6db222..21b5dbcc72c3 100644 --- a/src/diffusers/__init__.py +++ b/src/diffusers/__init__.py @@ -343,8 +343,8 @@ "StableDiffusionLatentUpscalePipeline", "StableDiffusionLDM3DPipeline", "StableDiffusionModelEditingPipeline", - "StableDiffusionPAGPipeline", "StableDiffusionPAGImg2ImgPipeline", + "StableDiffusionPAGPipeline", "StableDiffusionPanoramaPipeline", "StableDiffusionParadigmsPipeline", "StableDiffusionPipeline", diff --git a/src/diffusers/pipelines/pag/__init__.py b/src/diffusers/pipelines/pag/__init__.py index 60329ad903b9..e3fb03cee5d8 100644 --- a/src/diffusers/pipelines/pag/__init__.py +++ b/src/diffusers/pipelines/pag/__init__.py @@ -29,9 +29,9 @@ _import_structure["pipeline_pag_kolors"] = ["KolorsPAGPipeline"] _import_structure["pipeline_pag_pixart_sigma"] = ["PixArtSigmaPAGPipeline"] _import_structure["pipeline_pag_sd"] = ["StableDiffusionPAGPipeline"] - _import_structure["pipeline_pag_sd_img2img"] = ["StableDiffusionPAGImg2ImgPipeline"] _import_structure["pipeline_pag_sd_3"] = ["StableDiffusion3PAGPipeline"] _import_structure["pipeline_pag_sd_animatediff"] = ["AnimateDiffPAGPipeline"] + _import_structure["pipeline_pag_sd_img2img"] = ["StableDiffusionPAGImg2ImgPipeline"] _import_structure["pipeline_pag_sd_xl"] = ["StableDiffusionXLPAGPipeline"] _import_structure["pipeline_pag_sd_xl_img2img"] = ["StableDiffusionXLPAGImg2ImgPipeline"] _import_structure["pipeline_pag_sd_xl_inpaint"] = ["StableDiffusionXLPAGInpaintPipeline"] diff --git a/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py b/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py index 22890aef1f73..3181b8fd5719 100644 --- a/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py +++ b/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py @@ -65,7 +65,8 @@ ``` """ -# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.retrieve_timesteps + +# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion_img2img.retrieve_latents def retrieve_latents( encoder_output: torch.Tensor, generator: Optional[torch.Generator] = None, sample_mode: str = "sample" ): @@ -568,7 +569,6 @@ def prepare_extra_step_kwargs(self, generator, eta): extra_step_kwargs["generator"] = generator return extra_step_kwargs - # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion_img2img.check_inputs def check_inputs( self, prompt, @@ -630,7 +630,7 @@ def check_inputs( f"`ip_adapter_image_embeds` has to be a list of 3D or 4D tensors but is {ip_adapter_image_embeds[0].ndim}D" ) - # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion_img2img.get_timesteps + # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion_img2img.StableDiffusionImg2ImgPipeline.get_timesteps def get_timesteps(self, num_inference_steps, strength, device): # get the original timestep using init_timestep init_timestep = min(int(num_inference_steps * strength), num_inference_steps) @@ -642,7 +642,7 @@ def get_timesteps(self, num_inference_steps, strength, device): return timesteps, num_inference_steps - t_start - # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion_img2img.prepare_latents + # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion_img2img.StableDiffusionImg2ImgPipeline.prepare_latents 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( diff --git a/src/diffusers/utils/dummy_torch_and_transformers_objects.py b/src/diffusers/utils/dummy_torch_and_transformers_objects.py index 4f22501ce7ec..d49e419b3838 100644 --- a/src/diffusers/utils/dummy_torch_and_transformers_objects.py +++ b/src/diffusers/utils/dummy_torch_and_transformers_objects.py @@ -1577,6 +1577,21 @@ def from_pretrained(cls, *args, **kwargs): requires_backends(cls, ["torch", "transformers"]) +class StableDiffusionPAGImg2ImgPipeline(metaclass=DummyObject): + _backends = ["torch", "transformers"] + + def __init__(self, *args, **kwargs): + requires_backends(self, ["torch", "transformers"]) + + @classmethod + def from_config(cls, *args, **kwargs): + requires_backends(cls, ["torch", "transformers"]) + + @classmethod + def from_pretrained(cls, *args, **kwargs): + requires_backends(cls, ["torch", "transformers"]) + + class StableDiffusionPAGPipeline(metaclass=DummyObject): _backends = ["torch", "transformers"] diff --git a/tests/pipelines/pag/test_pag_sd_img2img.py b/tests/pipelines/pag/test_pag_sd_img2img.py index 7f4bc5ddc318..ba678c307ba2 100644 --- a/tests/pipelines/pag/test_pag_sd_img2img.py +++ b/tests/pipelines/pag/test_pag_sd_img2img.py @@ -57,8 +57,6 @@ enable_full_determinism() - - class StableDiffusionPAGImg2ImgPipelineFastTests( IPAdapterTesterMixin, PipelineLatentTesterMixin, @@ -180,7 +178,6 @@ def test_pag_disable_enable(self): assert np.abs(out.flatten() - out_pag_disabled.flatten()).max() < 1e-3 assert np.abs(out.flatten() - out_pag_enabled.flatten()).max() > 1e-3 - def test_pag_inference(self): device = "cpu" # ensure determinism for the device-dependent torch.Generator components = self.get_dummy_components() @@ -206,6 +203,7 @@ def test_pag_inference(self): max_diff = np.abs(image_slice.flatten() - expected_slice).max() self.assertLessEqual(max_diff, 1e-3) + @slow @require_torch_gpu class StableDiffusionPAGImg2ImgPipelineIntegrationTests(unittest.TestCase): From d030f3979cce739cf9b5899ddc28655520c5e984 Mon Sep 17 00:00:00 2001 From: SahilCarterr <110806554+SahilCarterr@users.noreply.github.com> Date: Wed, 25 Sep 2024 22:44:52 +0530 Subject: [PATCH 11/17] fix test --- tests/pipelines/pag/test_pag_sd_img2img.py | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/tests/pipelines/pag/test_pag_sd_img2img.py b/tests/pipelines/pag/test_pag_sd_img2img.py index ba678c307ba2..db1b4fa2e369 100644 --- a/tests/pipelines/pag/test_pag_sd_img2img.py +++ b/tests/pipelines/pag/test_pag_sd_img2img.py @@ -65,7 +65,7 @@ class StableDiffusionPAGImg2ImgPipelineFastTests( unittest.TestCase, ): pipeline_class = StableDiffusionPAGImg2ImgPipeline - params = TEXT_GUIDED_IMAGE_VARIATION_PARAMS.union({"pag_scale", "pag_adaptive_scale"}) + params = TEXT_GUIDED_IMAGE_VARIATION_PARAMS.union({"pag_scale", "pag_adaptive_scale"}) - {"height", "width"} required_optional_params = PipelineTesterMixin.required_optional_params - {"latents"} batch_params = TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS image_params = IMAGE_TO_IMAGE_IMAGE_PARAMS @@ -138,6 +138,7 @@ def get_dummy_inputs(self, device, seed=0): "generator": generator, "num_inference_steps": 2, "guidance_scale": 6.0, + "pag_scale": 0.9, "output_type": "np", } return inputs @@ -233,6 +234,7 @@ def get_inputs(self, device, generator_device="cpu", dtype=torch.float32, seed=0 "num_inference_steps": 3, "strength": 0.75, "guidance_scale": 7.5, + "pag_scale": 3.0, "output_type": "np", } return inputs From ec32e90a7f4201f3a042faa663095d42077d5353 Mon Sep 17 00:00:00 2001 From: SahilCarterr <110806554+SahilCarterr@users.noreply.github.com> Date: Sun, 6 Oct 2024 11:11:49 +0530 Subject: [PATCH 12/17] fix inference test --- tests/pipelines/pag/test_pag_sd_img2img.py | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/tests/pipelines/pag/test_pag_sd_img2img.py b/tests/pipelines/pag/test_pag_sd_img2img.py index db1b4fa2e369..c6d79c33151e 100644 --- a/tests/pipelines/pag/test_pag_sd_img2img.py +++ b/tests/pipelines/pag/test_pag_sd_img2img.py @@ -193,13 +193,13 @@ def test_pag_inference(self): assert image.shape == ( 1, - 64, - 64, + 32, + 32, 3, - ), f"the shape of the output image should be (1, 64, 64, 3) but got {image.shape}" + ), f"the shape of the output image should be (1, 32, 32, 3) but got {image.shape}" expected_slice = np.array( - [0.22802538, 0.44626093, 0.48905736, 0.29633686, 0.36400637, 0.4724258, 0.4678891, 0.32260418, 0.41611585] + [0.46088067, 0.32335982, 0.40687063, 0.45717427, 0.4627515, 0.41006276, 0.41850817 ,0.46726882 ,0.45808175] ) max_diff = np.abs(image_slice.flatten() - expected_slice).max() self.assertLessEqual(max_diff, 1e-3) From c8d0a49bd0742d4562f86c490d8f8f055ff69ac3 Mon Sep 17 00:00:00 2001 From: SahilCarterr <110806554+SahilCarterr@users.noreply.github.com> Date: Sun, 6 Oct 2024 11:41:01 +0530 Subject: [PATCH 13/17] fix ip_adapter error --- src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py | 2 ++ 1 file changed, 2 insertions(+) diff --git a/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py b/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py index 3181b8fd5719..e4f26494d5c3 100644 --- a/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py +++ b/src/diffusers/pipelines/pag/pipeline_pag_sd_img2img.py @@ -1029,6 +1029,8 @@ def __call__( latent_model_input = self.scheduler.scale_model_input(latent_model_input, t) # predict the noise residual + if ip_adapter_image_embeds is not None: + added_cond_kwargs["image_embeds"] = ip_adapter_image_embeds noise_pred = self.unet( latent_model_input, t, From 2e8295bf531a2e2ad2915af9b584b00246c588c1 Mon Sep 17 00:00:00 2001 From: SahilCarterr <110806554+SahilCarterr@users.noreply.github.com> Date: Sun, 6 Oct 2024 13:03:27 +0530 Subject: [PATCH 14/17] fix test errors --- tests/pipelines/pag/test_pag_sd_img2img.py | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/tests/pipelines/pag/test_pag_sd_img2img.py b/tests/pipelines/pag/test_pag_sd_img2img.py index c6d79c33151e..63461d40041e 100644 --- a/tests/pipelines/pag/test_pag_sd_img2img.py +++ b/tests/pipelines/pag/test_pag_sd_img2img.py @@ -26,7 +26,7 @@ AutoencoderKL, AutoencoderTiny, AutoPipelineForImage2Image, - PNDMScheduler, + EulerDiscreteScheduler, StableDiffusionImg2ImgPipeline, StableDiffusionPAGImg2ImgPipeline, UNet2DConditionModel, @@ -85,7 +85,13 @@ def get_dummy_components(self, time_cond_proj_dim=None): up_block_types=("CrossAttnUpBlock2D", "UpBlock2D"), cross_attention_dim=32, ) - scheduler = PNDMScheduler(skip_prk_steps=True) + scheduler = EulerDiscreteScheduler( + beta_start=0.00085, + beta_end=0.012, + steps_offset=1, + beta_schedule="scaled_linear", + timestep_spacing="leading", + ) torch.manual_seed(0) vae = AutoencoderKL( block_out_channels=[32, 64], @@ -94,6 +100,7 @@ def get_dummy_components(self, time_cond_proj_dim=None): down_block_types=["DownEncoderBlock2D", "DownEncoderBlock2D"], up_block_types=["UpDecoderBlock2D", "UpDecoderBlock2D"], latent_channels=4, + sample_size=128, ) torch.manual_seed(0) text_encoder_config = CLIPTextConfig( From 56517cccbeaad560291b5c7c130d8b12fbcde6eb Mon Sep 17 00:00:00 2001 From: SahilCarterr <110806554+SahilCarterr@users.noreply.github.com> Date: Sun, 6 Oct 2024 13:08:15 +0530 Subject: [PATCH 15/17] fixed arrestion --- tests/pipelines/pag/test_pag_sd_img2img.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/pipelines/pag/test_pag_sd_img2img.py b/tests/pipelines/pag/test_pag_sd_img2img.py index 63461d40041e..869af1629630 100644 --- a/tests/pipelines/pag/test_pag_sd_img2img.py +++ b/tests/pipelines/pag/test_pag_sd_img2img.py @@ -206,7 +206,7 @@ def test_pag_inference(self): ), f"the shape of the output image should be (1, 32, 32, 3) but got {image.shape}" expected_slice = np.array( - [0.46088067, 0.32335982, 0.40687063, 0.45717427, 0.4627515, 0.41006276, 0.41850817 ,0.46726882 ,0.45808175] + [0.44203848, 0.49598145, 0.42248967, 0.6707724, 0.5683791, 0.43603387, 0.58316565, 0.60077155, 0.5174199] ) max_diff = np.abs(image_slice.flatten() - expected_slice).max() self.assertLessEqual(max_diff, 1e-3) From 93b561ff23ac6dfa6b5aaf62c457f01db7bfc1bc Mon Sep 17 00:00:00 2001 From: SahilCarterr <110806554+SahilCarterr@users.noreply.github.com> Date: Sun, 6 Oct 2024 13:22:30 +0530 Subject: [PATCH 16/17] fix tests --- tests/pipelines/pag/test_pag_sd_img2img.py | 1 - 1 file changed, 1 deletion(-) diff --git a/tests/pipelines/pag/test_pag_sd_img2img.py b/tests/pipelines/pag/test_pag_sd_img2img.py index 869af1629630..3c9050badf6d 100644 --- a/tests/pipelines/pag/test_pag_sd_img2img.py +++ b/tests/pipelines/pag/test_pag_sd_img2img.py @@ -102,7 +102,6 @@ def get_dummy_components(self, time_cond_proj_dim=None): latent_channels=4, sample_size=128, ) - torch.manual_seed(0) text_encoder_config = CLIPTextConfig( bos_token_id=0, eos_token_id=2, From abcb2abbe8c6f061fa139267ccde0b864fdd632a Mon Sep 17 00:00:00 2001 From: SahilCarterr <110806554+SahilCarterr@users.noreply.github.com> Date: Wed, 9 Oct 2024 08:02:16 +0530 Subject: [PATCH 17/17] fix make error --- tests/pipelines/pag/test_pag_sd_img2img.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/pipelines/pag/test_pag_sd_img2img.py b/tests/pipelines/pag/test_pag_sd_img2img.py index 3c9050badf6d..ec8cde23c31d 100644 --- a/tests/pipelines/pag/test_pag_sd_img2img.py +++ b/tests/pipelines/pag/test_pag_sd_img2img.py @@ -205,7 +205,7 @@ def test_pag_inference(self): ), f"the shape of the output image should be (1, 32, 32, 3) but got {image.shape}" expected_slice = np.array( - [0.44203848, 0.49598145, 0.42248967, 0.6707724, 0.5683791, 0.43603387, 0.58316565, 0.60077155, 0.5174199] + [0.44203848, 0.49598145, 0.42248967, 0.6707724, 0.5683791, 0.43603387, 0.58316565, 0.60077155, 0.5174199] ) max_diff = np.abs(image_slice.flatten() - expected_slice).max() self.assertLessEqual(max_diff, 1e-3)