Cycle Diffusion
Overview
Cycle Diffusion is a Text-Guided Image-to-Image Generation model proposed in Unifying Diffusion Models’ Latent Space, with Applications to CycleDiffusion and Guidance by Chen Henry Wu, Fernando De la Torre.
The abstract of the paper is the following:
Diffusion models have achieved unprecedented performance in generative modeling. The commonly-adopted formulation of the latent code of diffusion models is a sequence of gradually denoised samples, as opposed to the simpler (e.g., Gaussian) latent space of GANs, VAEs, and normalizing flows. This paper provides an alternative, Gaussian formulation of the latent space of various diffusion models, as well as an invertible DPM-Encoder that maps images into the latent space. While our formulation is purely based on the definition of diffusion models, we demonstrate several intriguing consequences. (1) Empirically, we observe that a common latent space emerges from two diffusion models trained independently on related domains. In light of this finding, we propose CycleDiffusion, which uses DPM-Encoder for unpaired image-to-image translation. Furthermore, applying CycleDiffusion to text-to-image diffusion models, we show that large-scale text-to-image diffusion models can be used as zero-shot image-to-image editors. (2) One can guide pre-trained diffusion models and GANs by controlling the latent codes in a unified, plug-and-play formulation based on energy-based models. Using the CLIP model and a face recognition model as guidance, we demonstrate that diffusion models have better coverage of low-density sub-populations and individuals than GANs.
Tips:
- The Cycle Diffusion pipeline is fully compatible with any Stable Diffusion checkpoints
- Currently Cycle Diffusion only works with the DDIMScheduler.
Example:
In the following we should how to best use the CycleDiffusionPipeline
import requests
import torch
from PIL import Image
from io import BytesIO
from diffusers import CycleDiffusionPipeline, DDIMScheduler
# load the pipeline
# make sure you're logged in with `huggingface-cli login`
model_id_or_path = "CompVis/stable-diffusion-v1-4"
scheduler = DDIMScheduler.from_pretrained(model_id_or_path, subfolder="scheduler")
pipe = CycleDiffusionPipeline.from_pretrained(model_id_or_path, scheduler=scheduler).to("cuda")
# let's download an initial image
url = "https://raw.githubusercontent.com/ChenWu98/cycle-diffusion/main/data/dalle2/An%20astronaut%20riding%20a%20horse.png"
response = requests.get(url)
init_image = Image.open(BytesIO(response.content)).convert("RGB")
init_image = init_image.resize((512, 512))
init_image.save("horse.png")
# let's specify a prompt
source_prompt = "An astronaut riding a horse"
prompt = "An astronaut riding an elephant"
# call the pipeline
image = pipe(
prompt=prompt,
source_prompt=source_prompt,
image=init_image,
num_inference_steps=100,
eta=0.1,
strength=0.8,
guidance_scale=2,
source_guidance_scale=1,
).images[0]
image.save("horse_to_elephant.png")
# let's try another example
# See more samples at the original repo: https://github.com/ChenWu98/cycle-diffusion
url = "https://raw.githubusercontent.com/ChenWu98/cycle-diffusion/main/data/dalle2/A%20black%20colored%20car.png"
response = requests.get(url)
init_image = Image.open(BytesIO(response.content)).convert("RGB")
init_image = init_image.resize((512, 512))
init_image.save("black.png")
source_prompt = "A black colored car"
prompt = "A blue colored car"
# call the pipeline
torch.manual_seed(0)
image = pipe(
prompt=prompt,
source_prompt=source_prompt,
image=init_image,
num_inference_steps=100,
eta=0.1,
strength=0.85,
guidance_scale=3,
source_guidance_scale=1,
).images[0]
image.save("black_to_blue.png")CycleDiffusionPipeline
class diffusers.CycleDiffusionPipeline
< source >( vae: AutoencoderKL text_encoder: CLIPTextModel tokenizer: CLIPTokenizer unet: UNet2DConditionModel scheduler: DDIMScheduler safety_checker: StableDiffusionSafetyChecker feature_extractor: CLIPFeatureExtractor requires_safety_checker: bool = True )
Parameters
- vae (AutoencoderKL) — Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations.
-
text_encoder (
CLIPTextModel) — Frozen text-encoder. Stable Diffusion uses the text portion of CLIP, specifically the clip-vit-large-patch14 variant. -
tokenizer (
CLIPTokenizer) — Tokenizer of class CLIPTokenizer. - unet (UNet2DConditionModel) — Conditional U-Net architecture to denoise the encoded image latents.
-
scheduler (SchedulerMixin) —
A scheduler to be used in combination with
unetto 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 for details. -
feature_extractor (
CLIPFeatureExtractor) — Model that extracts features from generated images to be used as inputs for thesafety_checker.
Pipeline for text-guided image to image generation using Stable Diffusion.
This model inherits from DiffusionPipeline. Check the superclass documentation for the generic methods the library implements for all the pipelines (such as downloading or saving, running on a particular device, etc.)
__call__
< source >(
prompt: typing.Union[str, typing.List[str]]
source_prompt: typing.Union[str, typing.List[str]]
image: typing.Union[torch.FloatTensor, PIL.Image.Image] = None
strength: float = 0.8
num_inference_steps: typing.Optional[int] = 50
guidance_scale: typing.Optional[float] = 7.5
source_guidance_scale: typing.Optional[float] = 1
num_images_per_prompt: typing.Optional[int] = 1
eta: typing.Optional[float] = 0.1
generator: typing.Union[torch._C.Generator, typing.List[torch._C.Generator], NoneType] = None
prompt_embeds: typing.Optional[torch.FloatTensor] = None
output_type: typing.Optional[str] = 'pil'
return_dict: bool = True
callback: typing.Union[typing.Callable[[int, int, torch.FloatTensor], NoneType], NoneType] = None
callback_steps: int = 1
**kwargs
)
→
StableDiffusionPipelineOutput or tuple
Parameters
-
prompt (
strorList[str]) — The prompt or prompts to guide the image generation. -
image (
torch.FloatTensororPIL.Image.Image) —Image, or tensor representing an image batch, that will be used as the starting point for the process. -
strength (
float, optional, defaults to 0.8) — Conceptually, indicates how much to transform the referenceimage. Must be between 0 and 1.imagewill be used as a starting point, adding more noise to it the larger thestrength. The number of denoising steps depends on the amount of noise initially added. Whenstrengthis 1, added noise will be maximum and the denoising process will run for the full number of iterations specified innum_inference_steps. A value of 1, therefore, essentially ignoresimage. -
num_inference_steps (
int, optional, defaults to 50) — The number of denoising steps. More denoising steps usually lead to a higher quality image at the expense of slower inference. This parameter will be modulated bystrength. -
guidance_scale (
float, optional, defaults to 7.5) — Guidance scale as defined in Classifier-Free Diffusion Guidance.guidance_scaleis defined aswof equation 2. of Imagen Paper. Guidance scale is enabled by settingguidance_scale > 1. Higher guidance scale encourages to generate images that are closely linked to the textprompt, usually at the expense of lower image quality. -
source_guidance_scale (
float, optional, defaults to 1) — Guidance scale for the source prompt. This is useful to control the amount of influence the source prompt for encoding. -
num_images_per_prompt (
int, optional, defaults to 1) — The number of images to generate per prompt. -
eta (
float, optional, defaults to 0.1) — Corresponds to parameter eta (η) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to schedulers.DDIMScheduler, will be ignored for others. -
generator (
torch.Generator, optional) — One or a list of torch generator(s) to make generation deterministic. -
prompt_embeds (
torch.FloatTensor, optional) — Pre-generated text embeddings. Can be used to easily tweak text inputs, e.g. prompt weighting. If not provided, text embeddings will be generated frompromptinput argument. -
negative_prompt_embeds (
torch.FloatTensor, optional) — Pre-generated negative text embeddings. Can be used to easily tweak text inputs, e.g. prompt weighting. If not provided, negative_prompt_embeds will be generated fromnegative_promptinput argument. -
output_type (
str, optional, defaults to"pil") — The output format of the generate image. Choose between PIL:PIL.Image.Imageornp.array. -
return_dict (
bool, optional, defaults toTrue) — Whether or not to return a StableDiffusionPipelineOutput instead of a plain tuple. -
callback (
Callable, optional) — A function that will be called everycallback_stepssteps during inference. The function will be called with the following arguments:callback(step: int, timestep: int, latents: torch.FloatTensor). -
callback_steps (
int, optional, defaults to 1) — The frequency at which thecallbackfunction will be called. If not specified, the callback will be called at every step.
Returns
StableDiffusionPipelineOutput or tuple
StableDiffusionPipelineOutput if return_dict is True, otherwise a tuple. When returning a tuple, the first element is a list with the generated images, and the second element is a list of bools denoting whether the corresponding generated image likely represents "not-safe-for-work" (nsfw) content, according to the safety_checker`.
Function invoked when calling the pipeline for generation.
Offloads all models to CPU using accelerate, reducing memory usage with a low impact on performance. Compared
to enable_sequential_cpu_offload, this method moves one whole model at a time to the GPU when its forward
method is called, and the model remains in GPU until the next model runs. Memory savings are lower than with
enable_sequential_cpu_offload, but performance is much better due to the iterative execution of the unet.
Offloads all models to CPU using accelerate, significantly reducing memory usage. When called, unet,
text_encoder, vae and safety checker have their state dicts saved to CPU and then are moved to a
torch.device('meta') and loaded to GPU only when their specific submodule has its forwardmethod called. Note that offloading happens on a submodule basis. Memory savings are higher than withenable_model_cpu_offload`, but performance is lower.