Attend and Excite for Stable Diffusion was proposed in Attend-and-Excite: Attention-Based Semantic Guidance for Text-to-Image Diffusion Models and provides textual attention control over the image generation.
The abstract of the paper is the following:
Text-to-image diffusion models have recently received a lot of interest for their astonishing ability to produce high-fidelity images from text only. However, achieving one-shot generation that aligns with the user’s intent is nearly impossible, yet small changes to the input prompt often result in very different images. This leaves the user with little semantic control. To put the user in control, we show how to interact with the diffusion process to flexibly steer it along semantic directions. This semantic guidance (SEGA) allows for subtle and extensive edits, changes in composition and style, as well as optimizing the overall artistic conception. We demonstrate SEGA’s effectiveness on a variety of tasks and provide evidence for its versatility and flexibility.
Resources
Pipeline | Tasks | Colab | Demo |
---|---|---|---|
pipeline_semantic_stable_diffusion_attend_and_excite.py | Text-to-Image Generation | - | https://huggingface.co/spaces/AttendAndExcite/Attend-and-Excite |
import torch
from diffusers import StableDiffusionAttendAndExcitePipeline
model_id = "CompVis/stable-diffusion-v1-4"
pipe = StableDiffusionAttendAndExcitePipeline.from_pretrained(model_id, torch_dtype=torch.float16).to("cuda")
pipe = pipe.to("cuda")
prompt = "a cat and a frog"
# use get_indices function to find out indices of the tokens you want to alter
pipe.get_indices(prompt)
token_indices = [2, 5]
seed = 6141
generator = torch.Generator("cuda").manual_seed(seed)
images = pipe(
prompt=prompt,
token_indices=token_indices,
guidance_scale=7.5,
generator=generator,
num_inference_steps=50,
max_iter_to_alter=25,
).images
image = images[0]
image.save(f"../images/{prompt}_{seed}.png")
( vae: AutoencoderKL text_encoder: CLIPTextModel tokenizer: CLIPTokenizer unet: UNet2DConditionModel scheduler: KarrasDiffusionSchedulers safety_checker: StableDiffusionSafetyChecker feature_extractor: CLIPImageProcessor requires_safety_checker: bool = True )
Parameters
CLIPTextModel
) —
Frozen text-encoder. Stable Diffusion uses the text portion of
CLIP, specifically
the clip-vit-large-patch14 variant.
CLIPTokenizer
) —
Tokenizer of class
CLIPTokenizer.
unet
to denoise the encoded image latents. Can be one of
DDIMScheduler, LMSDiscreteScheduler, or PNDMScheduler.
StableDiffusionSafetyChecker
) —
Classification module that estimates whether generated images could be considered offensive or harmful.
Please, refer to the model card for details.
CLIPImageProcessor
) —
Model that extracts features from generated images to be used as inputs for the safety_checker
.
Pipeline for text-to-image generation using Stable Diffusion and Attend and Excite.
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.)
(
prompt: typing.Union[str, typing.List[str]]
token_indices: typing.Union[typing.List[int], typing.List[typing.List[int]]]
height: typing.Optional[int] = None
width: typing.Optional[int] = None
num_inference_steps: int = 50
guidance_scale: float = 7.5
negative_prompt: typing.Union[str, typing.List[str], NoneType] = None
num_images_per_prompt: int = 1
eta: float = 0.0
generator: typing.Union[torch._C.Generator, typing.List[torch._C.Generator], NoneType] = None
latents: typing.Optional[torch.FloatTensor] = None
prompt_embeds: typing.Optional[torch.FloatTensor] = None
negative_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
cross_attention_kwargs: typing.Union[typing.Dict[str, typing.Any], NoneType] = None
max_iter_to_alter: int = 25
thresholds: dict = {0: 0.05, 10: 0.5, 20: 0.8}
scale_factor: int = 20
attn_res: int = 16
)
→
StableDiffusionPipelineOutput or tuple
Parameters
str
or List[str]
, optional) —
The prompt or prompts to guide the image generation. If not defined, one has to pass prompt_embeds
.
instead.
List[int]
) —
The token indices to alter with attend-and-excite.
int
, optional, defaults to self.unet.config.sample_size * self.vae_scale_factor) —
The height in pixels of the generated image.
int
, optional, defaults to self.unet.config.sample_size * self.vae_scale_factor) —
The width in pixels of the generated image.
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.
float
, optional, defaults to 7.5) —
Guidance scale as defined in Classifier-Free Diffusion Guidance.
guidance_scale
is defined as w
of equation 2. of Imagen
Paper. Guidance scale is enabled by setting guidance_scale > 1
. Higher guidance scale encourages to generate images that are closely linked to the text prompt
,
usually at the expense of lower image quality.
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
).
int
, optional, defaults to 1) —
The number of images to generate per prompt.
float
, optional, defaults to 0.0) —
Corresponds to parameter eta (η) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
schedulers.DDIMScheduler, will be ignored for others.
torch.Generator
or List[torch.Generator]
, optional) —
One or a list of torch generator(s)
to make generation deterministic.
torch.FloatTensor
, optional) —
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
tensor will ge generated by sampling using the supplied random generator
.
torch.FloatTensor
, optional) —
Pre-generated text embeddings. Can be used to easily tweak text inputs, e.g. prompt weighting. If not
provided, text embeddings will be generated from prompt
input argument.
torch.FloatTensor
, optional) —
Pre-generated negative text embeddings. Can be used to easily tweak text inputs, e.g. prompt
weighting. If not provided, negative_prompt_embeds will be generated from negative_prompt
input
argument.
str
, optional, defaults to "pil"
) —
The output format of the generate image. Choose between
PIL: PIL.Image.Image
or np.array
.
bool
, optional, defaults to True
) —
Whether or not to return a StableDiffusionPipelineOutput instead of a
plain tuple.
Callable
, optional) —
A function that will be called every callback_steps
steps during inference. The function will be
called with the following arguments: callback(step: int, timestep: int, latents: torch.FloatTensor)
.
int
, optional, defaults to 1) —
The frequency at which the callback
function will be called. If not specified, the callback will be
called at every step.
dict
, optional) —
A kwargs dictionary that if specified is passed along to the AttentionProcessor
as defined under
self.processor
in
diffusers.cross_attention.
int
, optional, defaults to 25
) —
Number of denoising steps to apply attend-and-excite. The first max_iter_to_alter
is 25 and there are a total of 30
denoising steps, the first 25 denoising steps will apply attend-and-excite and the last 5 will not
apply attend-and-excite.dict
, optional, defaults to {0 -- 0.05, 10: 0.5, 20: 0.8}
):
Dictionary defining the iterations and desired thresholds to apply iterative latent refinement in.
int
, optional, default to 20) —
Scale factor that controls the step size of each Attend and Excite update.
int
, optional, default to 16) —
The resolution of most semantic attention map.
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`. :type attention_store: object
Function invoked when calling the pipeline for generation.
Examples:
>>> import torch
>>> from diffusers import StableDiffusionAttendAndExcitePipeline
>>> pipe = StableDiffusionAttendAndExcitePipeline.from_pretrained(
... "CompVis/stable-diffusion-v1-4", torch_dtype=torch.float16
... ).to("cuda")
>>> prompt = "a cat and a frog"
>>> # use get_indices function to find out indices of the tokens you want to alter
>>> pipe.get_indices(prompt)
{0: '<|startoftext|>', 1: 'a</w>', 2: 'cat</w>', 3: 'and</w>', 4: 'a</w>', 5: 'frog</w>', 6: '<|endoftext|>'}
>>> token_indices = [2, 5]
>>> seed = 6141
>>> generator = torch.Generator("cuda").manual_seed(seed)
>>> images = pipe(
... prompt=prompt,
... token_indices=token_indices,
... guidance_scale=7.5,
... generator=generator,
... num_inference_steps=50,
... max_iter_to_alter=25,
... ).images
>>> image = images[0]
>>> image.save(f"../images/{prompt}_{seed}.png")
Disable sliced VAE decoding. If enable_vae_slicing
was previously invoked, this method will go back to
computing decoding in one step.
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 with
enable_model_cpu_offload`, but performance is lower.
Enable sliced VAE decoding.
When this option is enabled, the VAE will split the input tensor in slices to compute decoding in several steps. This is useful to save some memory and allow larger batch sizes.
Utility function to list the indices of the tokens you wish to alte