Skip to content

GitLab

Unverified Commit a69754bb authored by Steven Liu's avatar Steven Liu Committed by GitHub
Browse files

[docs] Clean up pipeline apis (#3905) 

* start with stable diffusion

* fix

* finish stable diffusion pipelines

* fix path to pipeline output

* fix flax paths

* fix copies

* add up to score sde ve

* finish first pass of pipelines

* fix copies

* second review

* align doc titles

* more review fixes

* final review
parent bcc570b9
Hide whitespace changes
Inline Side-by-side
Showing with 720 additions and 808 deletions
src/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip.py
View file @ a69754bb
......@@ -54,15 +54,14 @@ class StableUnCLIPPipeline(DiffusionPipeline, TextualInversionLoaderMixin, LoraL
"""
Pipeline for text-to-image generation using stable unCLIP.
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.)
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.).
Args:
prior_tokenizer ([`CLIPTokenizer`]):
Tokenizer of class
[CLIPTokenizer](https://huggingface.co/docs/transformers/v4.21.0/en/model_doc/clip#transformers.CLIPTokenizer).
A [`CLIPTokenizer`].
prior_text_encoder ([`CLIPTextModelWithProjection`]):
Frozen text-encoder.
Frozen [`CLIPTextModelWithProjection`] text-encoder.
prior ([`PriorTransformer`]):
The canonincal unCLIP prior to approximate the image embedding from the text embedding.
prior_scheduler ([`KarrasDiffusionSchedulers`]):
......@@ -72,13 +71,13 @@ class StableUnCLIPPipeline(DiffusionPipeline, TextualInversionLoaderMixin, LoraL
embeddings after the noise has been applied.
image_noising_scheduler ([`KarrasDiffusionSchedulers`]):
Noise schedule for adding noise to the predicted image embeddings. The amount of noise to add is determined
by `noise_level` in `StableUnCLIPPipeline.__call__`.
tokenizer (`CLIPTokenizer`):
Tokenizer of class
[CLIPTokenizer](https://huggingface.co/docs/transformers/v4.21.0/en/model_doc/clip#transformers.CLIPTokenizer).
by the `noise_level`.
tokenizer ([`CLIPTokenizer`]):
A [`CLIPTokenizer`].
text_encoder ([`CLIPTextModel`]):
Frozen text-encoder.
unet ([`UNet2DConditionModel`]): Conditional U-Net architecture to denoise the encoded image latents.
Frozen [`CLIPTextModel`] text-encoder.
unet ([`UNet2DConditionModel`]):
A [`UNet2DConditionModel`] to denoise the encoded image latents.
scheduler ([`KarrasDiffusionSchedulers`]):
A scheduler to be used in combination with `unet` to denoise the encoded image latents.
vae ([`AutoencoderKL`]):
......@@ -145,27 +144,25 @@ class StableUnCLIPPipeline(DiffusionPipeline, TextualInversionLoaderMixin, LoraL
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_slicing
def enable_vae_slicing(self):
r"""
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.
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.
"""
self.vae.enable_slicing()
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.disable_vae_slicing
def disable_vae_slicing(self):
r"""
Disable sliced VAE decoding. If `enable_vae_slicing` was previously invoked, this method will go back to
Disable sliced VAE decoding. If `enable_vae_slicing` was previously enabled, this method will go back to
computing decoding in one step.
"""
self.vae.disable_slicing()
def enable_model_cpu_offload(self, gpu_id=0):
r"""
Offloads all models to CPU using accelerate, reducing memory usage with a low impact on performance. Compared
to `enable_sequential_cpu_offload`, this method moves one whole model at a time to the GPU when its `forward`
method is called, and the model remains in GPU until the next model runs. Memory savings are lower than with
`enable_sequential_cpu_offload`, but performance is much better due to the iterative execution of the `unet`.
Offload all models to CPU to reduce memory usage with a low impact on performance. 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 using `enable_sequential_cpu_offload`, but performance is much better due to the
iterative execution of the `unet`.
"""
if is_accelerate_available() and is_accelerate_version(">=", "0.17.0.dev0"):
from accelerate import cpu_offload_with_hook
......@@ -575,8 +572,8 @@ class StableUnCLIPPipeline(DiffusionPipeline, TextualInversionLoaderMixin, LoraL
Add noise to the image embeddings. The amount of noise is controlled by a `noise_level` input. A higher
`noise_level` increases the variance in the final un-noised images.
The noise is applied in two ways
1. A noise schedule is applied directly to the embeddings
The noise is applied in two ways:
1. A noise schedule is applied directly to the embeddings.
2. A vector of sinusoidal time embeddings are appended to the output.
In both cases, the amount of noise is controlled by the same `noise_level`.
......@@ -639,87 +636,76 @@ class StableUnCLIPPipeline(DiffusionPipeline, TextualInversionLoaderMixin, LoraL
prior_latents: Optional[torch.FloatTensor] = None,
):
"""
Function invoked when calling the pipeline for generation.
The call function to the pipeline for generation.
Args:
prompt (`str` or `List[str]`, *optional*):
The prompt or prompts to guide the image generation. If not defined, one has to pass `prompt_embeds`.
instead.
height (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):
The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`.
height (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
The height in pixels of the generated image.
width (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):
width (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
The width in pixels of the generated image.
num_inference_steps (`int`, *optional*, defaults to 20):
The number of denoising steps. More denoising steps usually lead to a higher quality image at the
expense of slower inference.
guidance_scale (`float`, *optional*, defaults to 10.0):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
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 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`).
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) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
[`schedulers.DDIMScheduler`], will be ignored for others.
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*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
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`.
tensor is generated by sampling using the supplied random `generator`.
prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
provided, text embeddings will be generated from `prompt` input argument.
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.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.
Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generate image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
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.
Whether or not to return a [`~pipelines.ImagePipelineOutput`] instead of a plain tuple.
callback (`Callable`, *optional*):
A function that will be called every `callback_steps` steps during inference. The function will be
called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
A function that calls every `callback_steps` steps during inference. The function is called with the
following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function will be called. If not specified, the callback will be
called at every step.
The frequency at which the `callback` function is called. If not specified, the callback is called at
every step.
cross_attention_kwargs (`dict`, *optional*):
A kwargs dictionary that if specified is passed along to the `AttentionProcessor` as defined under
`self.processor` in
[diffusers.cross_attention](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/cross_attention.py).
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/cross_attention.py).
noise_level (`int`, *optional*, defaults to `0`):
The amount of noise to add to the image embeddings. A higher `noise_level` increases the variance in
the final un-noised images. See `StableUnCLIPPipeline.noise_image_embeddings` for details.
the final un-noised images. See [`StableUnCLIPPipeline.noise_image_embeddings`] for more details.
prior_num_inference_steps (`int`, *optional*, defaults to 25):
The number of denoising steps in the prior denoising process. More denoising steps usually lead to a
higher quality image at the expense of slower inference.
prior_guidance_scale (`float`, *optional*, defaults to 4.0):
Guidance scale for the prior denoising process as defined in [Classifier-Free Diffusion
Guidance](https://arxiv.org/abs/2207.12598). `prior_guidance_scale` is defined as `w` of equation 2. of
[Imagen Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting
`guidance_scale > 1`. Higher guidance scale encourages to generate images that are closely linked to
the text `prompt`, usually at the expense of lower image quality.
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`.
prior_latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image
embedding generation in the prior denoising process. 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`.
different prompts. If not provided, a latents tensor is generated by sampling using the supplied random
`generator`.
Examples:
Returns:
[`~pipelines.ImagePipelineOutput`] or `tuple`: [`~ pipeline_utils.ImagePipelineOutput`] if `return_dict` is
True, otherwise a `tuple`. When returning a tuple, the first element is a list with the generated images.
[`~pipelines.ImagePipelineOutput`] or `tuple`:
[`~ pipeline_utils.ImagePipelineOutput`] if `return_dict` is True, otherwise a `tuple`. When returning
a tuple, the first element is a list with the generated images.
"""
# 0. Default height and width to unet
height = height or self.unet.config.sample_size * self.vae_scale_factor
......
src/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip_img2img.py
View file @ a69754bb
......@@ -65,10 +65,10 @@ EXAMPLE_DOC_STRING = """
class StableUnCLIPImg2ImgPipeline(DiffusionPipeline, TextualInversionLoaderMixin, LoraLoaderMixin):
"""
Pipeline for text-guided image to image generation using stable unCLIP.
Pipeline for text-guided image-to-image generation using stable unCLIP.
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.)
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.).
Args:
feature_extractor ([`CLIPImageProcessor`]):
......@@ -80,13 +80,13 @@ class StableUnCLIPImg2ImgPipeline(DiffusionPipeline, TextualInversionLoaderMixin
embeddings after the noise has been applied.
image_noising_scheduler ([`KarrasDiffusionSchedulers`]):
Noise schedule for adding noise to the predicted image embeddings. The amount of noise to add is determined
by `noise_level` in `StableUnCLIPPipeline.__call__`.
tokenizer (`CLIPTokenizer`):
Tokenizer of class
[CLIPTokenizer](https://huggingface.co/docs/transformers/v4.21.0/en/model_doc/clip#transformers.CLIPTokenizer).
text_encoder ([`CLIPTextModel`]):
Frozen text-encoder.
unet ([`UNet2DConditionModel`]): Conditional U-Net architecture to denoise the encoded image latents.
by the `noise_level`.
tokenizer (`~transformers.CLIPTokenizer`):
A [`~transformers.CLIPTokenizer`)].
text_encoder ([`~transformers.CLIPTextModel`]):
Frozen [`~transformers.CLIPTextModel`] text-encoder.
unet ([`UNet2DConditionModel`]):
A [`UNet2DConditionModel`] to denoise the encoded image latents.
scheduler ([`KarrasDiffusionSchedulers`]):
A scheduler to be used in combination with `unet` to denoise the encoded image latents.
vae ([`AutoencoderKL`]):
......@@ -147,27 +147,25 @@ class StableUnCLIPImg2ImgPipeline(DiffusionPipeline, TextualInversionLoaderMixin
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_slicing
def enable_vae_slicing(self):
r"""
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.
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.
"""
self.vae.enable_slicing()
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.disable_vae_slicing
def disable_vae_slicing(self):
r"""
Disable sliced VAE decoding. If `enable_vae_slicing` was previously invoked, this method will go back to
Disable sliced VAE decoding. If `enable_vae_slicing` was previously enabled, this method will go back to
computing decoding in one step.
"""
self.vae.disable_slicing()
def enable_model_cpu_offload(self, gpu_id=0):
r"""
Offloads all models to CPU using accelerate, reducing memory usage with a low impact on performance. Compared
to `enable_sequential_cpu_offload`, this method moves one whole model at a time to the GPU when its `forward`
method is called, and the model remains in GPU until the next model runs. Memory savings are lower than with
`enable_sequential_cpu_offload`, but performance is much better due to the iterative execution of the `unet`.
Offload all models to CPU to reduce memory usage with a low impact on performance. 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 using `enable_sequential_cpu_offload`, but performance is much better due to the
iterative execution of the `unet`.
"""
if is_accelerate_available() and is_accelerate_version(">=", "0.17.0.dev0"):
from accelerate import cpu_offload_with_hook
......@@ -542,8 +540,8 @@ class StableUnCLIPImg2ImgPipeline(DiffusionPipeline, TextualInversionLoaderMixin
Add noise to the image embeddings. The amount of noise is controlled by a `noise_level` input. A higher
`noise_level` increases the variance in the final un-noised images.
The noise is applied in two ways
1. A noise schedule is applied directly to the embeddings
The noise is applied in two ways:
1. A noise schedule is applied directly to the embeddings.
2. A vector of sinusoidal time embeddings are appended to the output.
In both cases, the amount of noise is controlled by the same `noise_level`.
......@@ -603,82 +601,73 @@ class StableUnCLIPImg2ImgPipeline(DiffusionPipeline, TextualInversionLoaderMixin
image_embeds: Optional[torch.FloatTensor] = None,
):
r"""
Function invoked when calling the pipeline for generation.
The call function to the pipeline for generation.
Args:
prompt (`str` or `List[str]`, *optional*):
The prompt or prompts to guide the image generation. If not defined, either `prompt_embeds` will be
used or prompt is initialized to `""`.
image (`torch.FloatTensor` or `PIL.Image.Image`):
`Image`, or tensor representing an image batch. The image will be encoded to its CLIP embedding which
the unet will be conditioned on. Note that the image is _not_ encoded by the vae and then used as the
latents in the denoising process such as in the standard stable diffusion text guided image variation
process.
height (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):
`Image` or tensor representing an image batch. The image is encoded to its CLIP embedding which the
`unet` is conditioned on. The image is _not_ encoded by the `vae` and then used as the latents in the
denoising process like it is in the standard Stable Diffusion text-guided image variation process.
height (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
The height in pixels of the generated image.
width (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):
width (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
The width in pixels of the generated image.
num_inference_steps (`int`, *optional*, defaults to 20):
The number of denoising steps. More denoising steps usually lead to a higher quality image at the
expense of slower inference.
guidance_scale (`float`, *optional*, defaults to 10.0):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
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 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`).
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) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
[`schedulers.DDIMScheduler`], will be ignored for others.
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*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
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`.
tensor is generated by sampling using the supplied random `generator`.
prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
provided, text embeddings will be generated from `prompt` input argument.
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.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.
Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generate image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
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.
Whether or not to return a [`~pipelines.ImagePipelineOutput`] instead of a plain tuple.
callback (`Callable`, *optional*):
A function that will be called every `callback_steps` steps during inference. The function will be
called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
A function that calls every `callback_steps` steps during inference. The function is called with the
following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function will be called. If not specified, the callback will be
called at every step.
The frequency at which the `callback` function is called. If not specified, the callback is called at
every step.
cross_attention_kwargs (`dict`, *optional*):
A kwargs dictionary that if specified is passed along to the `AttentionProcessor` as defined under
`self.processor` in
[diffusers.cross_attention](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/cross_attention.py).
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/cross_attention.py).
noise_level (`int`, *optional*, defaults to `0`):
The amount of noise to add to the image embeddings. A higher `noise_level` increases the variance in
the final un-noised images. See `StableUnCLIPPipeline.noise_image_embeddings` for details.
the final un-noised images. See [`StableUnCLIPPipeline.noise_image_embeddings`] for more details.
image_embeds (`torch.FloatTensor`, *optional*):
Pre-generated CLIP embeddings to condition the unet on. Note that these are not latents to be used in
the denoising process. If you want to provide pre-generated latents, pass them to `__call__` as
`latents`.
Pre-generated CLIP embeddings to condition the `unet` on. These latents are not used in the denoising
process. If you want to provide pre-generated latents, pass them to `__call__` as `latents`.
Examples:
Returns:
[`~pipelines.ImagePipelineOutput`] or `tuple`: [`~ pipeline_utils.ImagePipelineOutput`] if `return_dict` is
True, otherwise a `tuple`. When returning a tuple, the first element is a list with the generated images.
[`~pipelines.ImagePipelineOutput`] or `tuple`:
[`~ pipeline_utils.ImagePipelineOutput`] if `return_dict` is True, otherwise a `tuple`. When returning
a tuple, the first element is a list with the generated images.
"""
# 0. Default height and width to unet
height = height or self.unet.config.sample_size * self.vae_scale_factor
......
src/diffusers/pipelines/stable_diffusion_safe/pipeline_stable_diffusion_safe.py
View file @ a69754bb
......@@ -21,32 +21,29 @@ logger = logging.get_logger(__name__) # pylint: disable=invalid-name
class StableDiffusionPipelineSafe(DiffusionPipeline):
r"""
Pipeline for text-to-image generation using Safe Latent Diffusion.
Pipeline based on the [`StableDiffusionPipeline`] for text-to-image generation using Safe Latent Diffusion.
The implementation is based on the [`StableDiffusionPipeline`]
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.)
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.).
Args:
vae ([`AutoencoderKL`]):
Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations.
text_encoder ([`CLIPTextModel`]):
Frozen text-encoder. Stable Diffusion uses the text portion of
[CLIP](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTextModel), specifically
the [clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14) variant.
tokenizer (`CLIPTokenizer`):
Tokenizer of class
[CLIPTokenizer](https://huggingface.co/docs/transformers/v4.21.0/en/model_doc/clip#transformers.CLIPTokenizer).
unet ([`UNet2DConditionModel`]): Conditional U-Net architecture to denoise the encoded image latents.
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 details.
feature_extractor ([`CLIPImageProcessor`]):
Model that extracts features from generated images to be used as inputs for the `safety_checker`.
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`.
"""
_optional_components = ["safety_checker", "feature_extractor"]
......@@ -489,78 +486,82 @@ class StableDiffusionPipelineSafe(DiffusionPipeline):
sld_mom_beta: Optional[float] = 0.4,
):
r"""
Function invoked when calling the pipeline for generation.
The call function to the pipeline for generation.
Args:
prompt (`str` or `List[str]`):
The prompt or prompts to guide the image generation.
height (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):
The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`.
height (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
The height in pixels of the generated image.
width (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):
width (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
The width in pixels of the generated image.
num_inference_steps (`int`, *optional*, defaults to 50):
The number of denoising steps. More denoising steps usually lead to a higher quality image at the
expense of slower inference.
guidance_scale (`float`, *optional*, defaults to 7.5):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
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 not to guide the image generation. Ignored when not using guidance (i.e., ignored
if `guidance_scale` is less than `1`).
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) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
[`schedulers.DDIMScheduler`], will be ignored for others.
generator (`torch.Generator`, *optional*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
Corresponds to parameter eta (η) from the [DDIM](https://arxiv.org/abs/2010.02502) paper. Only applies
to the [`~schedulers.DDIMScheduler`], and is ignored in other schedulers.
generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
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`.
tensor is generated by sampling using the supplied random `generator`.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generate image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
The output format of the generated image. Choose between `PIL.Image` or `np.array`.
return_dict (`bool`, *optional*, defaults to `True`):
Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] instead of a
plain tuple.
callback (`Callable`, *optional*):
A function that 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)`.
A function that calls every `callback_steps` steps during inference. The function is called with the
following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function will be called. If not specified, the callback will be
called at every step.
The frequency at which the `callback` function is called. If not specified, the callback is called at
every step.
sld_guidance_scale (`float`, *optional*, defaults to 1000):
Safe latent guidance as defined in [Safe Latent Diffusion](https://arxiv.org/abs/2211.05105).
`sld_guidance_scale` is defined as sS of Eq. 6. If set to be less than 1, safety guidance will be
disabled.
If `sld_guidance_scale < 1`, safety guidance is disabled.
sld_warmup_steps (`int`, *optional*, defaults to 10):
Number of warmup steps for safety guidance. SLD will only be applied for diffusion steps greater than
`sld_warmup_steps`. `sld_warmup_steps` is defined as `delta` of [Safe Latent
Diffusion](https://arxiv.org/abs/2211.05105).
Number of warmup steps for safety guidance. SLD is only be applied for diffusion steps greater than
`sld_warmup_steps`.
sld_threshold (`float`, *optional*, defaults to 0.01):
Threshold that separates the hyperplane between appropriate and inappropriate images. `sld_threshold`
is defined as `lamda` of Eq. 5 in [Safe Latent Diffusion](https://arxiv.org/abs/2211.05105).
Threshold that separates the hyperplane between appropriate and inappropriate images.
sld_momentum_scale (`float`, *optional*, defaults to 0.3):
Scale of the SLD momentum to be added to the safety guidance at each diffusion step. If set to 0.0
momentum will be disabled. Momentum is already built up during warmup, i.e. for diffusion steps smaller
than `sld_warmup_steps`. `sld_momentum_scale` is defined as `sm` of Eq. 7 in [Safe Latent
Diffusion](https://arxiv.org/abs/2211.05105).
Scale of the SLD momentum to be added to the safety guidance at each diffusion step. If set to 0.0,
momentum is disabled. Momentum is built up during warmup for diffusion steps smaller than
`sld_warmup_steps`.
sld_mom_beta (`float`, *optional*, defaults to 0.4):
Defines how safety guidance momentum builds up. `sld_mom_beta` indicates how much of the previous
momentum will be kept. Momentum is already built up during warmup, i.e. for diffusion steps smaller
than `sld_warmup_steps`. `sld_mom_beta` is defined as `beta m` of Eq. 8 in [Safe Latent
Diffusion](https://arxiv.org/abs/2211.05105).
momentum is kept. Momentum is built up during warmup for diffusion steps smaller than
`sld_warmup_steps`.
Returns:
[`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] or `tuple`:
[`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] if `return_dict` is True, otherwise a `tuple.
When returning a tuple, the first element is a list with the generated images, and the second element is a
list of `bool`s denoting whether the corresponding generated image likely represents "not-safe-for-work"
(nsfw) content, according to the `safety_checker`.
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.
Examples:
```py
import torch
from diffusers import StableDiffusionPipelineSafe
pipeline = StableDiffusionPipelineSafe.from_pretrained(
"AIML-TUDA/stable-diffusion-safe", torch_dtype=torch.float16
)
prompt = "the four horsewomen of the apocalypse, painting by tom of finland, gaston bussiere, craig mullins, j. c. leyendecker"
image = pipeline(prompt=prompt, **SafetyConfig.MEDIUM).images[0]
```
"""
# 0. Default height and width to unet
height = height or self.unet.config.sample_size * self.vae_scale_factor
......
src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl.py
View file @ a69754bb
......@@ -146,17 +146,15 @@ class StableDiffusionXLPipeline(DiffusionPipeline, FromSingleFileMixin, LoraLoad
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_slicing
def enable_vae_slicing(self):
r"""
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.
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.
"""
self.vae.enable_slicing()
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.disable_vae_slicing
def disable_vae_slicing(self):
r"""
Disable sliced VAE decoding. If `enable_vae_slicing` was previously invoked, this method will go back to
Disable sliced VAE decoding. If `enable_vae_slicing` was previously enabled, this method will go back to
computing decoding in one step.
"""
self.vae.disable_slicing()
......@@ -164,17 +162,16 @@ class StableDiffusionXLPipeline(DiffusionPipeline, FromSingleFileMixin, LoraLoad
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_tiling
def enable_vae_tiling(self):
r"""
Enable tiled VAE decoding.
When this option is enabled, the VAE will split the input tensor into tiles to compute decoding and encoding in
several steps. This is useful to save a large amount of memory and to allow the processing of larger images.
Enable tiled VAE decoding. When this option is enabled, the VAE will split the input tensor into tiles to
compute decoding and encoding in several steps. This is useful for saving a large amount of memory and to allow
processing larger images.
"""
self.vae.enable_tiling()
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.disable_vae_tiling
def disable_vae_tiling(self):
r"""
Disable tiled VAE decoding. If `enable_vae_tiling` was previously invoked, this method will go back to
Disable tiled VAE decoding. If `enable_vae_tiling` was previously enabled, this method will go back to
computing decoding in one step.
"""
self.vae.disable_tiling()
......@@ -652,8 +649,8 @@ class StableDiffusionXLPipeline(DiffusionPipeline, FromSingleFileMixin, LoraLoad
The output format of the generate image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
return_dict (`bool`, *optional*, defaults to `True`):
Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionXLPipelineOutput`] instead of a
plain tuple.
Whether or not to return a [`~pipelines.stable_diffusion_xl.StableDiffusionXLPipelineOutput`] instead
of a plain tuple.
callback (`Callable`, *optional*):
A function that will be called every `callback_steps` steps during inference. The function will be
called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
......@@ -679,11 +676,9 @@ class StableDiffusionXLPipeline(DiffusionPipeline, FromSingleFileMixin, LoraLoad
Examples:
Returns:
[`~pipelines.stable_diffusion.StableDiffusionXLPipelineOutput`] or `tuple`:
[`~pipelines.stable_diffusion.StableDiffusionXLPipelineOutput`] if `return_dict` is True, otherwise a
`tuple. When returning a tuple, the first element is a list with the generated images, and the second
element is a list of `bool`s denoting whether the corresponding generated image likely represents
"not-safe-for-work" (nsfw) content, according to the `safety_checker`.
[`~pipelines.stable_diffusion_xl.StableDiffusionXLPipelineOutput`] or `tuple`:
[`~pipelines.stable_diffusion_xl.StableDiffusionXLPipelineOutput`] if `return_dict` is True, otherwise a
`tuple`. When returning a tuple, the first element is a list with the generated images.
"""
# 0. Default height and width to unet
height = height or self.default_sample_size * self.vae_scale_factor
......
src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_img2img.py
View file @ a69754bb
......@@ -153,17 +153,15 @@ class StableDiffusionXLImg2ImgPipeline(DiffusionPipeline, FromSingleFileMixin, L
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_slicing
def enable_vae_slicing(self):
r"""
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.
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.
"""
self.vae.enable_slicing()
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.disable_vae_slicing
def disable_vae_slicing(self):
r"""
Disable sliced VAE decoding. If `enable_vae_slicing` was previously invoked, this method will go back to
Disable sliced VAE decoding. If `enable_vae_slicing` was previously enabled, this method will go back to
computing decoding in one step.
"""
self.vae.disable_slicing()
......@@ -171,17 +169,16 @@ class StableDiffusionXLImg2ImgPipeline(DiffusionPipeline, FromSingleFileMixin, L
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_tiling
def enable_vae_tiling(self):
r"""
Enable tiled VAE decoding.
When this option is enabled, the VAE will split the input tensor into tiles to compute decoding and encoding in
several steps. This is useful to save a large amount of memory and to allow the processing of larger images.
Enable tiled VAE decoding. When this option is enabled, the VAE will split the input tensor into tiles to
compute decoding and encoding in several steps. This is useful for saving a large amount of memory and to allow
processing larger images.
"""
self.vae.enable_tiling()
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.disable_vae_tiling
def disable_vae_tiling(self):
r"""
Disable tiled VAE decoding. If `enable_vae_tiling` was previously invoked, this method will go back to
Disable tiled VAE decoding. If `enable_vae_tiling` was previously enabled, this method will go back to
computing decoding in one step.
"""
self.vae.disable_tiling()
......
src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_inpaint.py
View file @ a69754bb
......@@ -258,17 +258,15 @@ class StableDiffusionXLInpaintPipeline(
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_slicing
def enable_vae_slicing(self):
r"""
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.
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.
"""
self.vae.enable_slicing()
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.disable_vae_slicing
def disable_vae_slicing(self):
r"""
Disable sliced VAE decoding. If `enable_vae_slicing` was previously invoked, this method will go back to
Disable sliced VAE decoding. If `enable_vae_slicing` was previously enabled, this method will go back to
computing decoding in one step.
"""
self.vae.disable_slicing()
......@@ -276,17 +274,16 @@ class StableDiffusionXLInpaintPipeline(
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_tiling
def enable_vae_tiling(self):
r"""
Enable tiled VAE decoding.
When this option is enabled, the VAE will split the input tensor into tiles to compute decoding and encoding in
several steps. This is useful to save a large amount of memory and to allow the processing of larger images.
Enable tiled VAE decoding. When this option is enabled, the VAE will split the input tensor into tiles to
compute decoding and encoding in several steps. This is useful for saving a large amount of memory and to allow
processing larger images.
"""
self.vae.enable_tiling()
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.disable_vae_tiling
def disable_vae_tiling(self):
r"""
Disable tiled VAE decoding. If `enable_vae_tiling` was previously invoked, this method will go back to
Disable tiled VAE decoding. If `enable_vae_tiling` was previously enabled, this method will go back to
computing decoding in one step.
"""
self.vae.disable_tiling()
......
src/diffusers/pipelines/stochastic_karras_ve/pipeline_stochastic_karras_ve.py
View file @ a69754bb
......@@ -24,17 +24,13 @@ from ..pipeline_utils import DiffusionPipeline, ImagePipelineOutput
class KarrasVePipeline(DiffusionPipeline):
r"""
Stochastic sampling from Karras et al. [1] tailored to the Variance-Expanding (VE) models [2]. Use Algorithm 2 and
the VE column of Table 1 from [1] for reference.
[1] Karras, Tero, et al. "Elucidating the Design Space of Diffusion-Based Generative Models."
https://arxiv.org/abs/2206.00364 [2] Song, Yang, et al. "Score-based generative modeling through stochastic
differential equations." https://arxiv.org/abs/2011.13456
Pipeline for unconditional image generation.
Parameters:
unet ([`UNet2DModel`]): U-Net architecture to denoise the encoded image.
unet ([`UNet2DModel`]):
A `UNet2DModel` to denoise the encoded image.
scheduler ([`KarrasVeScheduler`]):
Scheduler for the diffusion process to be used in combination with `unet` to denoise the encoded image.
A scheduler to be used in combination with `unet` to denoise the encoded image.
"""
# add type hints for linting
......@@ -56,24 +52,28 @@ class KarrasVePipeline(DiffusionPipeline):
**kwargs,
) -> Union[Tuple, ImagePipelineOutput]:
r"""
The call function to the pipeline for generation.
Args:
batch_size (`int`, *optional*, defaults to 1):
The number of images to generate.
generator (`torch.Generator`, *optional*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
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.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generate image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
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.ImagePipelineOutput`] instead of a plain tuple.
Whether or not to return a [`ImagePipelineOutput`] instead of a plain tuple.
Example:
Returns:
[`~pipelines.ImagePipelineOutput`] or `tuple`: [`~pipelines.utils.ImagePipelineOutput`] if `return_dict` is
True, otherwise a `tuple. When returning a tuple, the first element is a list with the generated images.
[`~pipelines.ImagePipelineOutput`] or `tuple`:
If `return_dict` is `True`, [`~pipelines.ImagePipelineOutput`] is returned, otherwise a `tuple` is
returned where the first element is a list with the generated images.
"""
img_size = self.unet.config.sample_size
......
src/diffusers/pipelines/t2i_adapter/pipeline_stable_diffusion_adapter.py
View file @ a69754bb
......@@ -203,17 +203,15 @@ class StableDiffusionAdapterPipeline(DiffusionPipeline):
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_slicing
def enable_vae_slicing(self):
r"""
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.
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.
"""
self.vae.enable_slicing()
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.disable_vae_slicing
def disable_vae_slicing(self):
r"""
Disable sliced VAE decoding. If `enable_vae_slicing` was previously invoked, this method will go back to
Disable sliced VAE decoding. If `enable_vae_slicing` was previously enabled, this method will go back to
computing decoding in one step.
"""
self.vae.disable_slicing()
......
src/diffusers/pipelines/text_to_video_synthesis/__init__.py
View file @ a69754bb
......@@ -10,13 +10,12 @@ from ...utils import BaseOutput, OptionalDependencyNotAvailable, is_torch_availa
@dataclass
class TextToVideoSDPipelineOutput(BaseOutput):
"""
Output class for text to video pipelines.
Output class for text-to-video pipelines.
Args:
frames (`List[np.ndarray]` or `torch.FloatTensor`)
List of denoised frames (essentially images) as NumPy arrays of shape `(height, width, num_channels)` or as
a `torch` tensor. NumPy array present the denoised images of the diffusion pipeline. The length of the list
denotes the video length i.e., the number of frames.
a `torch` tensor. The length of the list denotes the video length (the number of frames).
"""
frames: Union[List[np.ndarray], torch.FloatTensor]
......
src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth.py
View file @ a69754bb
......@@ -77,18 +77,18 @@ class TextToVideoSDPipeline(DiffusionPipeline, TextualInversionLoaderMixin, Lora
r"""
Pipeline for text-to-video generation.
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.)
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.).
Args:
vae ([`AutoencoderKL`]):
Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations.
text_encoder ([`CLIPTextModel`]):
Frozen text-encoder. Same as Stable Diffusion 2.
Frozen text-encoder ([clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14)).
tokenizer (`CLIPTokenizer`):
Tokenizer of class
[CLIPTokenizer](https://huggingface.co/docs/transformers/v4.21.0/en/model_doc/clip#transformers.CLIPTokenizer).
unet ([`UNet3DConditionModel`]): Conditional U-Net architecture to denoise the encoded video latents.
A [`~transformers.CLIPTokenizer`] to tokenize text.
unet ([`UNet3DConditionModel`]):
A [`UNet3DConditionModel`] to denoise the encoded video 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`].
......@@ -116,17 +116,15 @@ class TextToVideoSDPipeline(DiffusionPipeline, TextualInversionLoaderMixin, Lora
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_slicing
def enable_vae_slicing(self):
r"""
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.
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.
"""
self.vae.enable_slicing()
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.disable_vae_slicing
def disable_vae_slicing(self):
r"""
Disable sliced VAE decoding. If `enable_vae_slicing` was previously invoked, this method will go back to
Disable sliced VAE decoding. If `enable_vae_slicing` was previously enabled, this method will go back to
computing decoding in one step.
"""
self.vae.disable_slicing()
......@@ -134,27 +132,26 @@ class TextToVideoSDPipeline(DiffusionPipeline, TextualInversionLoaderMixin, Lora
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_tiling
def enable_vae_tiling(self):
r"""
Enable tiled VAE decoding.
When this option is enabled, the VAE will split the input tensor into tiles to compute decoding and encoding in
several steps. This is useful to save a large amount of memory and to allow the processing of larger images.
Enable tiled VAE decoding. When this option is enabled, the VAE will split the input tensor into tiles to
compute decoding and encoding in several steps. This is useful for saving a large amount of memory and to allow
processing larger images.
"""
self.vae.enable_tiling()
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.disable_vae_tiling
def disable_vae_tiling(self):
r"""
Disable tiled VAE decoding. If `enable_vae_tiling` was previously invoked, this method will go back to
Disable tiled VAE decoding. If `enable_vae_tiling` was previously enabled, this method will go back to
computing decoding in one step.
"""
self.vae.disable_tiling()
def enable_model_cpu_offload(self, gpu_id=0):
r"""
Offloads all models to CPU using accelerate, reducing memory usage with a low impact on performance. Compared
to `enable_sequential_cpu_offload`, this method moves one whole model at a time to the GPU when its `forward`
method is called, and the model remains in GPU until the next model runs. Memory savings are lower than with
`enable_sequential_cpu_offload`, but performance is much better due to the iterative execution of the `unet`.
Offload all models to CPU to reduce memory usage with a low impact on performance. 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 using `enable_sequential_cpu_offload`, but performance is much better due to the
iterative execution of the `unet`.
"""
if is_accelerate_available() and is_accelerate_version(">=", "0.17.0.dev0"):
from accelerate import cpu_offload_with_hook
......@@ -466,15 +463,14 @@ class TextToVideoSDPipeline(DiffusionPipeline, TextualInversionLoaderMixin, Lora
cross_attention_kwargs: Optional[Dict[str, Any]] = None,
):
r"""
Function invoked when calling the pipeline for generation.
The call function to the pipeline for generation.
Args:
prompt (`str` or `List[str]`, *optional*):
The prompt or prompts to guide the video generation. If not defined, one has to pass `prompt_embeds`.
instead.
height (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):
The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`.
height (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
The height in pixels of the generated video.
width (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):
width (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
The width in pixels of the generated video.
num_frames (`int`, *optional*, defaults to 16):
The number of video frames that are generated. Defaults to 16 frames which at 8 frames per seconds
......@@ -483,55 +479,51 @@ class TextToVideoSDPipeline(DiffusionPipeline, TextualInversionLoaderMixin, Lora
The number of denoising steps. More denoising steps usually lead to a higher quality videos at the
expense of slower inference.
guidance_scale (`float`, *optional*, defaults to 7.5):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate videos that are closely linked to the text `prompt`,
usually at the expense of lower video quality.
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 not to guide the video 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`).
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) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
[`schedulers.DDIMScheduler`], will be ignored for others.
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*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for video
Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for video
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`. Latents should be of shape
tensor is generated by sampling using the supplied random `generator`. Latents should be of shape
`(batch_size, num_channel, num_frames, height, width)`.
prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
provided, text embeddings will be generated from `prompt` input argument.
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.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.
Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
output_type (`str`, *optional*, defaults to `"np"`):
The output format of the generate video. Choose between `torch.FloatTensor` or `np.array`.
The output format of the generated video. Choose between `torch.FloatTensor` or `np.array`.
return_dict (`bool`, *optional*, defaults to `True`):
Whether or not to return a [`~pipelines.stable_diffusion.TextToVideoSDPipelineOutput`] instead of a
plain tuple.
Whether or not to return a [`~pipelines.text_to_video_synthesis.TextToVideoSDPipelineOutput`] instead
of a plain tuple.
callback (`Callable`, *optional*):
A function that will be called every `callback_steps` steps during inference. The function will be
called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
A function that calls every `callback_steps` steps during inference. The function is called with the
following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function will be called. If not specified, the callback will be
called at every step.
The frequency at which the `callback` function is called. If not specified, the callback is called at
every step.
cross_attention_kwargs (`dict`, *optional*):
A kwargs dictionary that if specified is passed along to the `AttentionProcessor` as defined under
`self.processor` in
[diffusers.cross_attention](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/cross_attention.py).
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/cross_attention.py).
Examples:
Returns:
[`~pipelines.stable_diffusion.TextToVideoSDPipelineOutput`] or `tuple`:
[`~pipelines.stable_diffusion.TextToVideoSDPipelineOutput`] if `return_dict` is True, otherwise a `tuple.
When returning a tuple, the first element is a list with the generated frames.
[`~pipelines.text_to_video_synthesis.TextToVideoSDPipelineOutput`] or `tuple`:
If `return_dict` is `True`, [`~pipelines.text_to_video_synthesis.TextToVideoSDPipelineOutput`] is
returned, otherwise a `tuple` is returned where the first element is a list with the generated frames.
"""
# 0. Default height and width to unet
height = height or self.unet.config.sample_size * self.vae_scale_factor
......
src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth_img2img.py
View file @ a69754bb
......@@ -137,20 +137,20 @@ def preprocess_video(video):
class VideoToVideoSDPipeline(DiffusionPipeline, TextualInversionLoaderMixin, LoraLoaderMixin):
r"""
Pipeline for text-to-video generation.
Pipeline for text-guided video-to-video generation.
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.)
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.).
Args:
vae ([`AutoencoderKL`]):
Variational Auto-Encoder (VAE) Model to encode and decode videos to and from latent representations.
text_encoder ([`CLIPTextModel`]):
Frozen text-encoder. Same as Stable Diffusion 2.
Frozen text-encoder ([clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14)).
tokenizer (`CLIPTokenizer`):
Tokenizer of class
[CLIPTokenizer](https://huggingface.co/docs/transformers/v4.21.0/en/model_doc/clip#transformers.CLIPTokenizer).
unet ([`UNet3DConditionModel`]): Conditional U-Net architecture to denoise the encoded video latents.
A [`~transformers.CLIPTokenizer`] to tokenize text.
unet ([`UNet3DConditionModel`]):
A [`UNet3DConditionModel`] to denoise the encoded video 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`].
......@@ -178,17 +178,15 @@ class VideoToVideoSDPipeline(DiffusionPipeline, TextualInversionLoaderMixin, Lor
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_slicing
def enable_vae_slicing(self):
r"""
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.
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.
"""
self.vae.enable_slicing()
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.disable_vae_slicing
def disable_vae_slicing(self):
r"""
Disable sliced VAE decoding. If `enable_vae_slicing` was previously invoked, this method will go back to
Disable sliced VAE decoding. If `enable_vae_slicing` was previously enabled, this method will go back to
computing decoding in one step.
"""
self.vae.disable_slicing()
......@@ -196,27 +194,26 @@ class VideoToVideoSDPipeline(DiffusionPipeline, TextualInversionLoaderMixin, Lor
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_tiling
def enable_vae_tiling(self):
r"""
Enable tiled VAE decoding.
When this option is enabled, the VAE will split the input tensor into tiles to compute decoding and encoding in
several steps. This is useful to save a large amount of memory and to allow the processing of larger images.
Enable tiled VAE decoding. When this option is enabled, the VAE will split the input tensor into tiles to
compute decoding and encoding in several steps. This is useful for saving a large amount of memory and to allow
processing larger images.
"""
self.vae.enable_tiling()
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.disable_vae_tiling
def disable_vae_tiling(self):
r"""
Disable tiled VAE decoding. If `enable_vae_tiling` was previously invoked, this method will go back to
Disable tiled VAE decoding. If `enable_vae_tiling` was previously enabled, this method will go back to
computing decoding in one step.
"""
self.vae.disable_tiling()
def enable_model_cpu_offload(self, gpu_id=0):
r"""
Offloads all models to CPU using accelerate, reducing memory usage with a low impact on performance. Compared
to `enable_sequential_cpu_offload`, this method moves one whole model at a time to the GPU when its `forward`
method is called, and the model remains in GPU until the next model runs. Memory savings are lower than with
`enable_sequential_cpu_offload`, but performance is much better due to the iterative execution of the `unet`.
Offload all models to CPU to reduce memory usage with a low impact on performance. 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 using `enable_sequential_cpu_offload`, but performance is much better due to the
iterative execution of the `unet`.
"""
if is_accelerate_available() and is_accelerate_version(">=", "0.17.0.dev0"):
from accelerate import cpu_offload_with_hook
......@@ -550,75 +547,67 @@ class VideoToVideoSDPipeline(DiffusionPipeline, TextualInversionLoaderMixin, Lor
cross_attention_kwargs: Optional[Dict[str, Any]] = None,
):
r"""
Function invoked when calling the pipeline for generation.
The call function to the pipeline for generation.
Args:
prompt (`str` or `List[str]`, *optional*):
The prompt or prompts to guide the video generation. If not defined, one has to pass `prompt_embeds`.
instead.
video: (`List[np.ndarray]` or `torch.FloatTensor`):
`video` frames or tensor representing a video batch, that will be used as the starting point for the
process. Can also accpet video latents as `image`, if passing latents directly, it will not be encoded
again.
The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`.
video (`List[np.ndarray]` or `torch.FloatTensor`):
`video` frames or tensor representing a video batch to be used as the starting point for the process.
Can also accpet video latents as `image`, if passing latents directly, it will not be encoded again.
strength (`float`, *optional*, defaults to 0.8):
Conceptually, indicates how much to transform the reference `image`. Must be between 0 and 1. `image`
will be used as a starting point, adding more noise to it the larger the `strength`. The number of
denoising steps depends on the amount of noise initially added. When `strength` is 1, added noise will
be maximum and the denoising process will run for the full number of iterations specified in
`num_inference_steps`. A value of 1, therefore, essentially ignores `image`.
Indicates extent to transform the reference `video`. Must be between 0 and 1. `video` is used as a
starting point, adding more noise to it the larger 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 `video`.
num_inference_steps (`int`, *optional*, defaults to 50):
The number of denoising steps. More denoising steps usually lead to a higher quality videos at the
expense of slower inference.
guidance_scale (`float`, *optional*, defaults to 7.5):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate videos that are closely linked to the text `prompt`,
usually at the expense of lower video quality.
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 not to guide the video 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`).
The prompt or prompts to guide what to not include in video generation. If not defined, you need to
pass `negative_prompt_embeds` instead. Ignored when not using guidance (`guidance_scale < 1`).
eta (`float`, *optional*, defaults to 0.0):
Corresponds to parameter eta (η) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
[`schedulers.DDIMScheduler`], will be ignored for others.
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*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for video
Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for video
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`. Latents should be of shape
tensor is generated by sampling using the supplied random `generator`. Latents should be of shape
`(batch_size, num_channel, num_frames, height, width)`.
prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
provided, text embeddings will be generated from `prompt` input argument.
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.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.
Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
output_type (`str`, *optional*, defaults to `"np"`):
The output format of the generate video. Choose between `torch.FloatTensor` or `np.array`.
The output format of the generated video. Choose between `torch.FloatTensor` or `np.array`.
return_dict (`bool`, *optional*, defaults to `True`):
Whether or not to return a [`~pipelines.stable_diffusion.TextToVideoSDPipelineOutput`] instead of a
plain tuple.
Whether or not to return a [`~pipelines.text_to_video_synthesis.TextToVideoSDPipelineOutput`] instead
of a plain tuple.
callback (`Callable`, *optional*):
A function that will be called every `callback_steps` steps during inference. The function will be
called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
A function that calls every `callback_steps` steps during inference. The function is called with the
following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function will be called. If not specified, the callback will be
called at every step.
The frequency at which the `callback` function is called. If not specified, the callback is called at
every step.
cross_attention_kwargs (`dict`, *optional*):
A kwargs dictionary that if specified is passed along to the `AttentionProcessor` as defined under
`self.processor` in
[diffusers.cross_attention](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/cross_attention.py).
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/cross_attention.py).
Examples:
Returns:
[`~pipelines.stable_diffusion.TextToVideoSDPipelineOutput`] or `tuple`:
[`~pipelines.stable_diffusion.TextToVideoSDPipelineOutput`] if `return_dict` is True, otherwise a `tuple.
When returning a tuple, the first element is a list with the generated frames.
[`~pipelines.text_to_video_synthesis.TextToVideoSDPipelineOutput`] or `tuple`:
If `return_dict` is `True`, [`~pipelines.text_to_video_synthesis.TextToVideoSDPipelineOutput`] is
returned, otherwise a `tuple` is returned where the first element is a list with the generated frames.
"""
# 0. Default height and width to unet
num_images_per_prompt = 1
......
src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_zero.py
View file @ a69754bb
......@@ -172,6 +172,17 @@ class CrossFrameAttnProcessor2_0:
@dataclass
class TextToVideoPipelineOutput(BaseOutput):
r"""
Output class for zero-shot text-to-video pipeline.
Args:
images (`[List[PIL.Image.Image]`, `np.ndarray`]):
List of denoised PIL images of length `batch_size` or NumPy array of shape `(batch_size, height, width,
num_channels)`.
nsfw_content_detected (`[List[bool]]`):
List indicating whether the corresponding generated image contains "not-safe-for-work" (nsfw) content or
`None` if safety checking could not be performed.
"""
images: Union[List[PIL.Image.Image], np.ndarray]
nsfw_content_detected: Optional[List[bool]]
......@@ -264,28 +275,27 @@ class TextToVideoZeroPipeline(StableDiffusionPipeline):
r"""
Pipeline for zero-shot text-to-video generation using Stable Diffusion.
This model inherits from [`StableDiffusionPipeline`]. 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.)
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.).
Args:
vae ([`AutoencoderKL`]):
Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations.
text_encoder ([`CLIPTextModel`]):
Frozen text-encoder. Stable Diffusion uses the text portion of
[CLIP](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTextModel), specifically
the [clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14) variant.
Frozen text-encoder ([clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14)).
tokenizer (`CLIPTokenizer`):
Tokenizer of class
[CLIPTokenizer](https://huggingface.co/docs/transformers/v4.21.0/en/model_doc/clip#transformers.CLIPTokenizer).
unet ([`UNet2DConditionModel`]): Conditional U-Net architecture to denoise the encoded image latents.
A [`~transformers.CLIPTokenizer`] to tokenize text.
unet ([`UNet2DConditionModel`]):
A [`UNet3DConditionModel`] to denoise the encoded video latents.
scheduler ([`SchedulerMixin`]):
A scheduler to be used in combination with `unet` to denoise the encoded image latents. Can be one of
[`DDIMScheduler`], [`LMSDiscreteScheduler`], or [`PNDMScheduler`].
safety_checker ([`StableDiffusionSafetyChecker`]):
Classification module that estimates whether generated images could be considered offensive or harmful.
Please, refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for details.
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 ([`CLIPImageProcessor`]):
Model that extracts features from generated images to be used as inputs for the `safety_checker`.
A [`CLIPImageProcessor`] to extract features from generated images; used as inputs to the `safety_checker`.
"""
def __init__(
......@@ -311,16 +321,22 @@ class TextToVideoZeroPipeline(StableDiffusionPipeline):
def forward_loop(self, x_t0, t0, t1, generator):
"""
Perform ddpm forward process from time t0 to t1. This is the same as adding noise with corresponding variance.
Perform DDPM forward process from time t0 to t1. This is the same as adding noise with corresponding variance.
Args:
x_t0: latent code at time t0
t0: t0
t1: t1
generator: torch.Generator object
x_t0:
Latent code at time t0.
t0:
Timestep at t0.
t1:
Timestamp at t1.
generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
Returns:
x_t1: forward process applied to x_t0 from time t0 to t1.
x_t1:
Forward process applied to x_t0 from time t0 to t1.
"""
eps = torch.randn(x_t0.size(), generator=generator, dtype=x_t0.dtype, device=x_t0.device)
alpha_vec = torch.prod(self.scheduler.alphas[t0:t1])
......@@ -340,30 +356,35 @@ class TextToVideoZeroPipeline(StableDiffusionPipeline):
cross_attention_kwargs=None,
):
"""
Perform backward process given list of time steps
Perform backward process given list of time steps.
Args:
latents: Latents at time timesteps[0].
timesteps: time steps, along which to perform backward process.
prompt_embeds: Pre-generated text embeddings
latents:
Latents at time timesteps[0].
timesteps:
Time steps along which to perform backward process.
prompt_embeds:
Pre-generated text embeddings.
guidance_scale:
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
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`.
callback (`Callable`, *optional*):
A function that will be called every `callback_steps` steps during inference. The function will be
called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
A function that calls every `callback_steps` steps during inference. The function is called with the
following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function will be called. If not specified, the callback will be
called at every step.
extra_step_kwargs: extra_step_kwargs.
cross_attention_kwargs: cross_attention_kwargs.
num_warmup_steps: number of warmup steps.
The frequency at which the `callback` function is called. If not specified, the callback is called at
every step.
extra_step_kwargs:
Extra_step_kwargs.
cross_attention_kwargs:
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/cross_attention.py).
num_warmup_steps:
number of warmup steps.
Returns:
latents: latents of backward process output at time timesteps[-1]
latents:
Latents of backward process output at time timesteps[-1].
"""
do_classifier_free_guidance = guidance_scale > 1.0
num_steps = (len(timesteps) - num_warmup_steps) // self.scheduler.order
......@@ -421,53 +442,50 @@ class TextToVideoZeroPipeline(StableDiffusionPipeline):
frame_ids: Optional[List[int]] = None,
):
"""
Function invoked when calling the pipeline for generation.
The call function to the pipeline for generation.
Args:
prompt (`str` or `List[str]`, *optional*):
The prompt or prompts to guide the image generation. If not defined, one has to pass `prompt_embeds`.
instead.
video_length (`int`, *optional*, defaults to 8): The number of generated video frames
height (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):
The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`.
video_length (`int`, *optional*, defaults to 8):
The number of generated video frames.
height (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
The height in pixels of the generated image.
width (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):
width (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
The width in pixels of the generated image.
num_inference_steps (`int`, *optional*, defaults to 50):
The number of denoising steps. More denoising steps usually lead to a higher quality image at the
expense of slower inference.
guidance_scale (`float`, *optional*, defaults to 7.5):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
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 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`).
The prompt or prompts to guide what to not include in video generation. If not defined, you need to
pass `negative_prompt_embeds` instead. Ignored when not using guidance (`guidance_scale < 1`).
num_videos_per_prompt (`int`, *optional*, defaults to 1):
The number of videos to generate per prompt.
eta (`float`, *optional*, defaults to 0.0):
Corresponds to parameter eta (η) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
[`schedulers.DDIMScheduler`], will be ignored for others.
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*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for video
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`.
tensor is generated by sampling using the supplied random `generator`.
output_type (`str`, *optional*, defaults to `"numpy"`):
The output format of the generated image. Choose between `"latent"` and `"numpy"`.
The output format of the generated video. Choose between `"latent"` and `"numpy"`.
return_dict (`bool`, *optional*, defaults to `True`):
Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] instead of a
plain tuple.
Whether or not to return a
[`~pipelines.text_to_video_synthesis.pipeline_text_to_video_zero.TextToVideoPipelineOutput`] instead of
a plain tuple.
callback (`Callable`, *optional*):
A function that will be called every `callback_steps` steps during inference. The function will be
called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
A function that calls every `callback_steps` steps during inference. The function is called with the
following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function will be called. If not specified, the callback will be
called at every step.
The frequency at which the `callback` function is called. If not specified, the callback is called at
every step.
motion_field_strength_x (`float`, *optional*, defaults to 12):
Strength of motion in generated video along x-axis. See the [paper](https://arxiv.org/abs/2303.13439),
Sect. 3.3.1.
......@@ -485,10 +503,10 @@ class TextToVideoZeroPipeline(StableDiffusionPipeline):
chunk-by-chunk.
Returns:
[`~pipelines.text_to_video_synthesis.TextToVideoPipelineOutput`]:
The output contains a ndarray of the generated images, when output_type != 'latent', otherwise a latent
codes of generated image, and a list of `bool`s denoting whether the corresponding generated image
likely represents "not-safe-for-work" (nsfw) content, according to the `safety_checker`.
[`~pipelines.text_to_video_synthesis.pipeline_text_to_video_zero.TextToVideoPipelineOutput`]:
The output contains a `ndarray` of the generated video, when `output_type` != `"latent"`, otherwise a
latent code of generated videos and a list of `bool`s indicating whether the corresponding generated
video contains "not-safe-for-work" (nsfw) content..
"""
assert video_length > 0
if frame_ids is None:
......
src/diffusers/pipelines/unclip/pipeline_unclip.py
View file @ a69754bb
......@@ -33,33 +33,32 @@ logger = logging.get_logger(__name__) # pylint: disable=invalid-name
class UnCLIPPipeline(DiffusionPipeline):
"""
Pipeline for text-to-image generation using unCLIP
Pipeline for text-to-image generation using unCLIP.
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.)
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.).
Args:
text_encoder ([`CLIPTextModelWithProjection`]):
text_encoder ([`~transformers.CLIPTextModelWithProjection`]):
Frozen text-encoder.
tokenizer (`CLIPTokenizer`):
Tokenizer of class
[CLIPTokenizer](https://huggingface.co/docs/transformers/v4.21.0/en/model_doc/clip#transformers.CLIPTokenizer).
tokenizer ([`~transformers.CLIPTokenizer`]):
A `CLIPTokenizer` to tokenize text.
prior ([`PriorTransformer`]):
The canonincal unCLIP prior to approximate the image embedding from the text embedding.
The canonical unCLIP prior to approximate the image embedding from the text embedding.
text_proj ([`UnCLIPTextProjModel`]):
Utility class to prepare and combine the embeddings before they are passed to the decoder.
decoder ([`UNet2DConditionModel`]):
The decoder to invert the image embedding into an image.
super_res_first ([`UNet2DModel`]):
Super resolution unet. Used in all but the last step of the super resolution diffusion process.
Super resolution UNet. Used in all but the last step of the super resolution diffusion process.
super_res_last ([`UNet2DModel`]):
Super resolution unet. Used in the last step of the super resolution diffusion process.
Super resolution UNet. Used in the last step of the super resolution diffusion process.
prior_scheduler ([`UnCLIPScheduler`]):
Scheduler used in the prior denoising process. Just a modified DDPMScheduler.
Scheduler used in the prior denoising process (a modified [`DDPMScheduler`]).
decoder_scheduler ([`UnCLIPScheduler`]):
Scheduler used in the decoder denoising process. Just a modified DDPMScheduler.
Scheduler used in the decoder denoising process (a modified [`DDPMScheduler`]).
super_res_scheduler ([`UnCLIPScheduler`]):
Scheduler used in the super resolution denoising process. Just a modified DDPMScheduler.
Scheduler used in the super resolution denoising process (a modified [`DDPMScheduler`]).
"""
......@@ -227,12 +226,12 @@ class UnCLIPPipeline(DiffusionPipeline):
return_dict: bool = True,
):
"""
Function invoked when calling the pipeline for generation.
The call function to the pipeline for generation.
Args:
prompt (`str` or `List[str]`):
The prompt or prompts to guide the image generation. This can only be left undefined if
`text_model_output` and `text_attention_mask` is passed.
The prompt or prompts to guide image generation. This can only be left undefined if `text_model_output`
and `text_attention_mask` is passed.
num_images_per_prompt (`int`, *optional*, defaults to 1):
The number of images to generate per prompt.
prior_num_inference_steps (`int`, *optional*, defaults to 25):
......@@ -245,8 +244,8 @@ class UnCLIPPipeline(DiffusionPipeline):
The number of denoising steps for super resolution. More denoising steps usually lead to a higher
quality image at the expense of slower inference.
generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
prior_latents (`torch.FloatTensor` of shape (batch size, embeddings dimension), *optional*):
Pre-generated noisy latents to be used as inputs for the prior.
decoder_latents (`torch.FloatTensor` of shape (batch size, channels, height, width), *optional*):
......@@ -254,29 +253,27 @@ class UnCLIPPipeline(DiffusionPipeline):
super_res_latents (`torch.FloatTensor` of shape (batch size, channels, super res height, super res width), *optional*):
Pre-generated noisy latents to be used as inputs for the decoder.
prior_guidance_scale (`float`, *optional*, defaults to 4.0):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
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`.
decoder_guidance_scale (`float`, *optional*, defaults to 4.0):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
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`.
text_model_output (`CLIPTextModelOutput`, *optional*):
Pre-defined CLIPTextModel outputs that can be derived from the text encoder. Pre-defined text outputs
can be passed for tasks like text embedding interpolations. Make sure to also pass
`text_attention_mask` in this case. `prompt` can the be left to `None`.
Pre-defined [`CLIPTextModel`] outputs that can be derived from the text encoder. Pre-defined text
outputs can be passed for tasks like text embedding interpolations. Make sure to also pass
`text_attention_mask` in this case. `prompt` can the be left `None`.
text_attention_mask (`torch.Tensor`, *optional*):
Pre-defined CLIP text attention mask that can be derived from the tokenizer. Pre-defined text attention
masks are necessary when passing `text_model_output`.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generated image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
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.ImagePipelineOutput`] instead of a plain tuple.
Returns:
[`~pipelines.ImagePipelineOutput`] or `tuple`:
If `return_dict` is `True`, [`~pipelines.ImagePipelineOutput`] is returned, otherwise a `tuple` is
returned where the first element is a list with the generated images.
"""
if prompt is not None:
if isinstance(prompt, str):
......
src/diffusers/pipelines/unclip/pipeline_unclip_image_variation.py
View file @ a69754bb
......@@ -37,36 +37,32 @@ logger = logging.get_logger(__name__) # pylint: disable=invalid-name
class UnCLIPImageVariationPipeline(DiffusionPipeline):
"""
Pipeline to generate variations from an input image using unCLIP
Pipeline to generate image variations from an input image using UnCLIP.
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.)
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.).
Args:
text_encoder ([`CLIPTextModelWithProjection`]):
text_encoder ([`~transformers.CLIPTextModelWithProjection`]):
Frozen text-encoder.
tokenizer (`CLIPTokenizer`):
Tokenizer of class
[CLIPTokenizer](https://huggingface.co/docs/transformers/v4.21.0/en/model_doc/clip#transformers.CLIPTokenizer).
feature_extractor ([`CLIPImageProcessor`]):
tokenizer ([`~transformers.CLIPTokenizer`]):
A `CLIPTokenizer` to tokenize text.
feature_extractor ([`~transformers.CLIPImageProcessor`]):
Model that extracts features from generated images to be used as inputs for the `image_encoder`.
image_encoder ([`CLIPVisionModelWithProjection`]):
Frozen CLIP image-encoder. unCLIP Image Variation uses the vision portion of
[CLIP](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPVisionModelWithProjection),
specifically the [clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14) variant.
image_encoder ([`~transformers.CLIPVisionModelWithProjection`]):
Frozen CLIP image-encoder ([clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14)).
text_proj ([`UnCLIPTextProjModel`]):
Utility class to prepare and combine the embeddings before they are passed to the decoder.
decoder ([`UNet2DConditionModel`]):
The decoder to invert the image embedding into an image.
super_res_first ([`UNet2DModel`]):
Super resolution unet. Used in all but the last step of the super resolution diffusion process.
Super resolution UNet. Used in all but the last step of the super resolution diffusion process.
super_res_last ([`UNet2DModel`]):
Super resolution unet. Used in the last step of the super resolution diffusion process.
Super resolution UNet. Used in the last step of the super resolution diffusion process.
decoder_scheduler ([`UnCLIPScheduler`]):
Scheduler used in the decoder denoising process. Just a modified DDPMScheduler.
Scheduler used in the decoder denoising process (a modified [`DDPMScheduler`]).
super_res_scheduler ([`UnCLIPScheduler`]):
Scheduler used in the super resolution denoising process. Just a modified DDPMScheduler.
Scheduler used in the super resolution denoising process (a modified [`DDPMScheduler`]).
"""
decoder: UNet2DConditionModel
......@@ -214,14 +210,14 @@ class UnCLIPImageVariationPipeline(DiffusionPipeline):
return_dict: bool = True,
):
"""
Function invoked when calling the pipeline for generation.
The call function to the pipeline for generation.
Args:
image (`PIL.Image.Image` or `List[PIL.Image.Image]` or `torch.FloatTensor`):
The image or images to guide the image generation. If you provide a tensor, it needs to comply with the
configuration of
[this](https://huggingface.co/fusing/karlo-image-variations-diffusers/blob/main/feature_extractor/preprocessor_config.json)
`CLIPImageProcessor`. Can be left to `None` only when `image_embeddings` are passed.
`Image` or tensor representing an image batch to be used as the starting point. If you provide a
tensor, it needs to be compatible with the [`CLIPImageProcessor`]
[configuration](https://huggingface.co/fusing/karlo-image-variations-diffusers/blob/main/feature_extractor/preprocessor_config.json).
Can be left as `None` only when `image_embeddings` are passed.
num_images_per_prompt (`int`, *optional*, defaults to 1):
The number of images to generate per prompt.
decoder_num_inference_steps (`int`, *optional*, defaults to 25):
......@@ -231,26 +227,27 @@ class UnCLIPImageVariationPipeline(DiffusionPipeline):
The number of denoising steps for super resolution. More denoising steps usually lead to a higher
quality image at the expense of slower inference.
generator (`torch.Generator`, *optional*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
decoder_latents (`torch.FloatTensor` of shape (batch size, channels, height, width), *optional*):
Pre-generated noisy latents to be used as inputs for the decoder.
super_res_latents (`torch.FloatTensor` of shape (batch size, channels, super res height, super res width), *optional*):
Pre-generated noisy latents to be used as inputs for the decoder.
decoder_guidance_scale (`float`, *optional*, defaults to 4.0):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
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`.
image_embeddings (`torch.Tensor`, *optional*):
Pre-defined image embeddings that can be derived from the image encoder. Pre-defined image embeddings
can be passed for tasks like image interpolations. `image` can the be left to `None`.
can be passed for tasks like image interpolations. `image` can be left as `None`.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generated image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
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.ImagePipelineOutput`] instead of a plain tuple.
Returns:
[`~pipelines.ImagePipelineOutput`] or `tuple`:
If `return_dict` is `True`, [`~pipelines.ImagePipelineOutput`] is returned, otherwise a `tuple` is
returned where the first element is a list with the generated images.
"""
if image is not None:
if isinstance(image, PIL.Image.Image):
......
src/diffusers/pipelines/unidiffuser/pipeline_unidiffuser.py
View file @ a69754bb
......@@ -81,43 +81,33 @@ class ImageTextPipelineOutput(BaseOutput):
class UniDiffuserPipeline(DiffusionPipeline):
r"""
Pipeline for a bimodal image-text [UniDiffuser](https://arxiv.org/pdf/2303.06555.pdf) model, which supports
unconditional text and image generation, text-conditioned image generation, image-conditioned text generation, and
joint image-text generation.
Pipeline for a bimodal image-text model which supports unconditional text and image generation, text-conditioned
image generation, image-conditioned text generation, and joint image-text generation.
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.)
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.).
Args:
vae ([`AutoencoderKL`]):
Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations. This
is part of the UniDiffuser image representation, along with the CLIP vision encoding.
Variational Auto-Encoder (VAE) model to encode and decode images to and from latent representations. This
is part of the UniDiffuser image representation along with the CLIP vision encoding.
text_encoder ([`CLIPTextModel`]):
Frozen text-encoder. Similar to Stable Diffusion, UniDiffuser uses the text portion of
[CLIP](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTextModel) to encode text
prompts.
Frozen text-encoder ([clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14)).
image_encoder ([`CLIPVisionModel`]):
UniDiffuser uses the vision portion of
[CLIP](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPVisionModel) to encode
images as part of its image representation, along with the VAE latent representation.
A [`~transformers.CLIPVisionModel`] to encode images as part of its image representation along with the VAE
latent representation.
image_processor ([`CLIPImageProcessor`]):
CLIP image processor of class
[CLIPImageProcessor](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPImageProcessor),
used to preprocess the image before CLIP encoding it with `image_encoder`.
[`~transformers.CLIPImageProcessor`] to preprocess an image before CLIP encoding it with `image_encoder`.
clip_tokenizer ([`CLIPTokenizer`]):
Tokenizer of class
[CLIPTokenizer](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTokenizer) which
is used to tokenizer a prompt before encoding it with `text_encoder`.
A [`~transformers.CLIPTokenizer`] to tokenize the prompt before encoding it with `text_encoder`.
text_decoder ([`UniDiffuserTextDecoder`]):
Frozen text decoder. This is a GPT-style model which is used to generate text from the UniDiffuser
embedding.
text_tokenizer ([`GPT2Tokenizer`]):
Tokenizer of class
[GPT2Tokenizer](https://huggingface.co/docs/transformers/model_doc/gpt2#transformers.GPT2Tokenizer) which
is used along with the `text_decoder` to decode text for text generation.
A [`~transformers.GPT2Tokenizer`] to decode text for text generation; used along with the `text_decoder`.
unet ([`UniDiffuserModel`]):
UniDiffuser uses a [U-ViT](https://github.com/baofff/U-ViT) model architecture, which is similar to a
[`Transformer2DModel`] with U-Net-style skip connections between transformer layers.
A [U-ViT](https://github.com/baofff/U-ViT) model with UNNet-style skip connections between transformer
layers to denoise the encoded image latents.
scheduler ([`SchedulerMixin`]):
A scheduler to be used in combination with `unet` to denoise the encoded image and/or text latents. The
original UniDiffuser paper uses the [`DPMSolverMultistepScheduler`] scheduler.
......@@ -1062,14 +1052,14 @@ class UniDiffuserPipeline(DiffusionPipeline):
callback_steps: int = 1,
):
r"""
Function invoked when calling the pipeline for generation.
The call function to the pipeline for generation.
Args:
prompt (`str` or `List[str]`, *optional*):
The prompt or prompts to guide the image generation. If not defined, one has to pass `prompt_embeds`
instead. Required for text-conditioned image generation (`text2img`) mode.
The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`.
Required for text-conditioned image generation (`text2img`) mode.
image (`torch.FloatTensor` or `PIL.Image.Image`, *optional*):
`Image`, or tensor representing an image batch. Required for image-conditioned text generation
`Image` or tensor representing an image batch. Required for image-conditioned text generation
(`img2text`) mode.
height (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
The height in pixels of the generated image.
......@@ -1077,78 +1067,74 @@ class UniDiffuserPipeline(DiffusionPipeline):
The width in pixels of the generated image.
data_type (`int`, *optional*, defaults to 1):
The data type (either 0 or 1). Only used if you are loading a checkpoint which supports a data type
embedding; this is added for compatibility with the UniDiffuser-v1 checkpoint.
embedding; this is added for compatibility with the
[UniDiffuser-v1](https://huggingface.co/thu-ml/unidiffuser-v1) checkpoint.
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.
guidance_scale (`float`, *optional*, defaults to 8.0):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality. Note that the original [UniDiffuser
paper](https://arxiv.org/pdf/2303.06555.pdf) uses a different definition of the guidance scale `w'`,
which satisfies `w = w' + 1`.
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 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`). Used in text-conditioned image generation (`text2img`) mode.
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`). Used in
text-conditioned image generation (`text2img`) mode.
num_images_per_prompt (`int`, *optional*, defaults to 1):
The number of images to generate per prompt. Used in `text2img` (text-conditioned image generation) and
`img` mode. If the mode is joint and both `num_images_per_prompt` and `num_prompts_per_image` are
supplied, `min(num_images_per_prompt, num_prompts_per_image)` samples will be generated.
supplied, `min(num_images_per_prompt, num_prompts_per_image)` samples are generated.
num_prompts_per_image (`int`, *optional*, defaults to 1):
The number of prompts to generate per image. Used in `img2text` (image-conditioned text generation) and
`text` mode. If the mode is joint and both `num_images_per_prompt` and `num_prompts_per_image` are
supplied, `min(num_images_per_prompt, num_prompts_per_image)` samples will be generated.
supplied, `min(num_images_per_prompt, num_prompts_per_image)` samples are generated.
eta (`float`, *optional*, defaults to 0.0):
Corresponds to parameter eta (η) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
[`schedulers.DDIMScheduler`], will be ignored for others.
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*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for joint
Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for joint
image-text generation. Can be used to tweak the same generation with different prompts. If not
provided, a latents tensor will be generated by sampling using the supplied random `generator`. Note
that this is assumed to be a full set of VAE, CLIP, and text latents, if supplied, this will override
the value of `prompt_latents`, `vae_latents`, and `clip_latents`.
provided, a latents tensor is generated by sampling using the supplied random `generator`. This assumes
a full set of VAE, CLIP, and text latents, if supplied, overrides the value of `prompt_latents`,
`vae_latents`, and `clip_latents`.
prompt_latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for text
Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for text
generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
tensor will be generated by sampling using the supplied random `generator`.
tensor is generated by sampling using the supplied random `generator`.
vae_latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
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 be generated by sampling using the supplied random `generator`.
tensor is generated by sampling using the supplied random `generator`.
clip_latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
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 be generated by sampling using the supplied random `generator`.
tensor is generated by sampling using the supplied random `generator`.
prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
provided, text embeddings will be generated from `prompt` input argument. Used in text-conditioned
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. Used in text-conditioned
image generation (`text2img`) mode.
negative_prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
argument. Used in text-conditioned image generation (`text2img`) mode.
Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
not provided, `negative_prompt_embeds` are be generated from the `negative_prompt` input argument. Used
in text-conditioned image generation (`text2img`) mode.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generate image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
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.unidiffuser.ImageTextPipelineOutput`] instead of a plain tuple.
Whether or not to return a [`~pipelines.ImageTextPipelineOutput`] instead of a plain tuple.
callback (`Callable`, *optional*):
A function that will be called every `callback_steps` steps during inference. The function will be
called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
A function that calls every `callback_steps` steps during inference. The function is called with the
following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function will be called. If not specified, the callback will be
called at every step.
The frequency at which the `callback` function is called. If not specified, the callback is called at
every step.
Returns:
[`~pipelines.unidiffuser.ImageTextPipelineOutput`] or `tuple`:
[`pipelines.unidiffuser.ImageTextPipelineOutput`] 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 generated texts.
If `return_dict` is `True`, [`~pipelines.unidiffuser.ImageTextPipelineOutput`] 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 generated texts.
"""
# 0. Default height and width to unet
......
src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion.py
View file @ a69754bb
......@@ -21,28 +21,27 @@ class VersatileDiffusionPipeline(DiffusionPipeline):
r"""
Pipeline for text-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.)
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.).
Args:
vae ([`AutoencoderKL`]):
Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations.
text_encoder ([`CLIPTextModel`]):
Frozen text-encoder. Stable Diffusion uses the text portion of
[CLIP](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTextModel), specifically
the [clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14) variant.
tokenizer (`CLIPTokenizer`):
Tokenizer of class
[CLIPTokenizer](https://huggingface.co/docs/transformers/v4.21.0/en/model_doc/clip#transformers.CLIPTokenizer).
unet ([`UNet2DConditionModel`]): Conditional U-Net architecture to denoise the encoded image latents.
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 ([`StableDiffusionMegaSafetyChecker`]):
safety_checker ([`StableDiffusionSafetyChecker`]):
Classification module that estimates whether generated images could be considered offensive or harmful.
Please, refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for details.
feature_extractor ([`CLIPImageProcessor`]):
Model that extracts features from generated images to be used as inputs for the `safety_checker`.
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`.
"""
tokenizer: CLIPTokenizer
......@@ -98,51 +97,47 @@ class VersatileDiffusionPipeline(DiffusionPipeline):
callback_steps: int = 1,
):
r"""
Function invoked when calling the pipeline for generation.
The call function to the pipeline for generation.
Args:
image (`PIL.Image.Image`, `List[PIL.Image.Image]` or `torch.Tensor`):
The image prompt or prompts to guide the image generation.
height (`int`, *optional*, defaults to self.image_unet.config.sample_size * self.vae_scale_factor):
height (`int`, *optional*, defaults to `self.image_unet.config.sample_size * self.vae_scale_factor`):
The height in pixels of the generated image.
width (`int`, *optional*, defaults to self.image_unet.config.sample_size * self.vae_scale_factor):
width (`int`, *optional*, defaults to `self.image_unet.config.sample_size * self.vae_scale_factor`):
The width in pixels of the generated image.
num_inference_steps (`int`, *optional*, defaults to 50):
The number of denoising steps. More denoising steps usually lead to a higher quality image at the
expense of slower inference.
guidance_scale (`float`, *optional*, defaults to 7.5):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
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 not to guide the image generation. Ignored when not using guidance (i.e., ignored
if `guidance_scale` is less than `1`).
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) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
[`schedulers.DDIMScheduler`], will be ignored for others.
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`, *optional*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
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`.
tensor is generated by sampling using the supplied random `generator`.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generate image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
The output format of the generated image. Choose between `PIL.Image` or `np.array`.
return_dict (`bool`, *optional*, defaults to `True`):
Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] instead of a
plain tuple.
callback (`Callable`, *optional*):
A function that 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)`.
A function that calls every `callback_steps` steps during inference. The function is called with the
following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function will be called. If not specified, the callback will be
called at every step.
The frequency at which the `callback` function is called. If not specified, the callback is called at
every step.
Examples:
......@@ -171,10 +166,10 @@ class VersatileDiffusionPipeline(DiffusionPipeline):
Returns:
[`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] or `tuple`:
[`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] if `return_dict` is True, otherwise a `tuple.
When returning a tuple, the first element is a list with the generated images, and the second element is a
list of `bool`s denoting whether the corresponding generated image likely represents "not-safe-for-work"
(nsfw) content, according to the `safety_checker`.
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.
"""
expected_components = inspect.signature(VersatileDiffusionImageVariationPipeline.__init__).parameters.keys()
components = {name: component for name, component in self.components.items() if name in expected_components}
......@@ -214,51 +209,47 @@ class VersatileDiffusionPipeline(DiffusionPipeline):
callback_steps: int = 1,
):
r"""
Function invoked when calling the pipeline for generation.
The call function to the pipeline for generation.
Args:
prompt (`str` or `List[str]`):
The prompt or prompts to guide the image generation.
height (`int`, *optional*, defaults to self.image_unet.config.sample_size * self.vae_scale_factor):
The prompt or prompts to guide image generation.
height (`int`, *optional*, defaults to `self.image_unet.config.sample_size * self.vae_scale_factor`):
The height in pixels of the generated image.
width (`int`, *optional*, defaults to self.image_unet.config.sample_size * self.vae_scale_factor):
width (`int`, *optional*, defaults to `self.image_unet.config.sample_size * self.vae_scale_factor`):
The width in pixels of the generated image.
num_inference_steps (`int`, *optional*, defaults to 50):
The number of denoising steps. More denoising steps usually lead to a higher quality image at the
expense of slower inference.
guidance_scale (`float`, *optional*, defaults to 7.5):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
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 not to guide the image generation. Ignored when not using guidance (i.e., ignored
if `guidance_scale` is less than `1`).
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) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
[`schedulers.DDIMScheduler`], will be ignored for others.
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`, *optional*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
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`.
tensor is generated by sampling using the supplied random `generator`.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generate image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
The output format of the generated image. Choose between `PIL.Image` or `np.array`.
return_dict (`bool`, *optional*, defaults to `True`):
Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] instead of a
plain tuple.
callback (`Callable`, *optional*):
A function that 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)`.
A function that calls every `callback_steps` steps during inference. The function is called with the
following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function will be called. If not specified, the callback will be
called at every step.
The frequency at which the `callback` function is called. If not specified, the callback is called at
every step.
Examples:
......@@ -278,10 +269,10 @@ class VersatileDiffusionPipeline(DiffusionPipeline):
Returns:
[`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] or `tuple`:
[`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] if `return_dict` is True, otherwise a `tuple.
When returning a tuple, the first element is a list with the generated images, and the second element is a
list of `bool`s denoting whether the corresponding generated image likely represents "not-safe-for-work"
(nsfw) content, according to the `safety_checker`.
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.
"""
expected_components = inspect.signature(VersatileDiffusionTextToImagePipeline.__init__).parameters.keys()
components = {name: component for name, component in self.components.items() if name in expected_components}
......@@ -327,51 +318,47 @@ class VersatileDiffusionPipeline(DiffusionPipeline):
callback_steps: int = 1,
):
r"""
Function invoked when calling the pipeline for generation.
The call function to the pipeline for generation.
Args:
prompt (`str` or `List[str]`):
The prompt or prompts to guide the image generation.
height (`int`, *optional*, defaults to self.image_unet.config.sample_size * self.vae_scale_factor):
The prompt or prompts to guide image generation.
height (`int`, *optional*, defaults to `self.image_unet.config.sample_size * self.vae_scale_factor`):
The height in pixels of the generated image.
width (`int`, *optional*, defaults to self.image_unet.config.sample_size * self.vae_scale_factor):
width (`int`, *optional*, defaults to `self.image_unet.config.sample_size * self.vae_scale_factor`):
The width in pixels of the generated image.
num_inference_steps (`int`, *optional*, defaults to 50):
The number of denoising steps. More denoising steps usually lead to a higher quality image at the
expense of slower inference.
guidance_scale (`float`, *optional*, defaults to 7.5):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
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 not to guide the image generation. Ignored when not using guidance (i.e., ignored
if `guidance_scale` is less than `1`).
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) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
[`schedulers.DDIMScheduler`], will be ignored for others.
generator (`torch.Generator`, *optional*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
Corresponds to parameter eta (η) from the [DDIM](https://arxiv.org/abs/2010.02502) paper. Only applies
to the [`~schedulers.DDIMScheduler`], and is ignored in other schedulers.
generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
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`.
tensor is generated by sampling using the supplied random `generator`.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generate image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
The output format of the generated image. Choose between `PIL.Image` or `np.array`.
return_dict (`bool`, *optional*, defaults to `True`):
Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] instead of a
plain tuple.
callback (`Callable`, *optional*):
A function that 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)`.
A function that calls every `callback_steps` steps during inference. The function is called with the
following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function will be called. If not specified, the callback will be
called at every step.
The frequency at which the `callback` function is called. If not specified, the callback is called at
every step.
Examples:
......@@ -404,9 +391,9 @@ class VersatileDiffusionPipeline(DiffusionPipeline):
```
Returns:
[`~pipelines.stable_diffusion.ImagePipelineOutput`] or `tuple`:
[`~pipelines.stable_diffusion.ImagePipelineOutput`] if `return_dict` is True, otherwise a `tuple. When
returning a tuple, the first element is a list with the generated images.
[`~pipelines.ImagePipelineOutput`] or `tuple`:
If `return_dict` is `True`, [`~pipelines.ImagePipelineOutput`] is returned, otherwise a `tuple` is
returned where the first element is a list with the generated images.
"""
expected_components = inspect.signature(VersatileDiffusionDualGuidedPipeline.__init__).parameters.keys()
......
src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion_dual_guided.py
View file @ a69754bb
......@@ -40,18 +40,20 @@ logger = logging.get_logger(__name__) # pylint: disable=invalid-name
class VersatileDiffusionDualGuidedPipeline(DiffusionPipeline):
r"""
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.)
Pipeline for image-text dual-guided generation using Versatile 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.).
Parameters:
vqvae ([`VQModel`]):
Vector-quantized (VQ) Model to encode and decode images to and from latent representations.
Vector-quantized (VQ) model to encode and decode images to and from latent representations.
bert ([`LDMBertModel`]):
Text-encoder model based on [BERT](https://huggingface.co/docs/transformers/model_doc/bert) architecture.
tokenizer (`transformers.BertTokenizer`):
Tokenizer of class
[BertTokenizer](https://huggingface.co/docs/transformers/model_doc/bert#transformers.BertTokenizer).
unet ([`UNet2DConditionModel`]): Conditional U-Net architecture to denoise the encoded image latents.
Text-encoder model based on [`~transformers.BERT`].
tokenizer ([`~transformers.BertTokenizer`]):
A `BertTokenizer` 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`].
......@@ -395,51 +397,46 @@ class VersatileDiffusionDualGuidedPipeline(DiffusionPipeline):
**kwargs,
):
r"""
Function invoked when calling the pipeline for generation.
The call function to the pipeline for generation.
Args:
prompt (`str` or `List[str]`):
The prompt or prompts to guide the image generation.
height (`int`, *optional*, defaults to self.image_unet.config.sample_size * self.vae_scale_factor):
The prompt or prompts to guide image generation.
height (`int`, *optional*, defaults to `self.image_unet.config.sample_size * self.vae_scale_factor`):
The height in pixels of the generated image.
width (`int`, *optional*, defaults to self.image_unet.config.sample_size * self.vae_scale_factor):
width (`int`, *optional*, defaults to `self.image_unet.config.sample_size * self.vae_scale_factor`):
The width in pixels of the generated image.
num_inference_steps (`int`, *optional*, defaults to 50):
The number of denoising steps. More denoising steps usually lead to a higher quality image at the
expense of slower inference.
guidance_scale (`float`, *optional*, defaults to 7.5):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
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 not to guide the image generation. Ignored when not using guidance (i.e., ignored
if `guidance_scale` is less than `1`).
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) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
[`schedulers.DDIMScheduler`], will be ignored for others.
generator (`torch.Generator`, *optional*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
Corresponds to parameter eta (η) from the [DDIM](https://arxiv.org/abs/2010.02502) paper. Only applies
to the [`~schedulers.DDIMScheduler`], and is ignored in other schedulers.
generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
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`.
tensor is generated by sampling using the supplied random `generator`.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generate image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
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.
Whether or not to return a [`~pipelines.ImagePipelineOutput`] instead of a plain tuple.
callback (`Callable`, *optional*):
A function that will be called every `callback_steps` steps during inference. The function will be
called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
A function that calls every `callback_steps` steps during inference. The function is called with the
following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function will be called. If not specified, the callback will be
called at every step.
The frequency at which the `callback` function is called. If not specified, the callback is called at
every step.
Examples:
......@@ -473,9 +470,9 @@ class VersatileDiffusionDualGuidedPipeline(DiffusionPipeline):
```
Returns:
[`~pipelines.stable_diffusion.ImagePipelineOutput`] or `tuple`:
[`~pipelines.stable_diffusion.ImagePipelineOutput`] if `return_dict` is True, otherwise a `tuple. When
returning a tuple, the first element is a list with the generated images.
[`~pipelines.ImagePipelineOutput`] or `tuple`:
If `return_dict` is `True`, [`~pipelines.ImagePipelineOutput`] is returned, otherwise a `tuple` is
returned where the first element is a list with the generated images.
"""
# 0. Default height and width to unet
height = height or self.image_unet.config.sample_size * self.vae_scale_factor
......
src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion_image_variation.py
View file @ a69754bb
......@@ -34,18 +34,20 @@ logger = logging.get_logger(__name__) # pylint: disable=invalid-name
class VersatileDiffusionImageVariationPipeline(DiffusionPipeline):
r"""
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.)
Pipeline for image variation using Versatile 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.).
Parameters:
vqvae ([`VQModel`]):
Vector-quantized (VQ) Model to encode and decode images to and from latent representations.
Vector-quantized (VQ) model to encode and decode images to and from latent representations.
bert ([`LDMBertModel`]):
Text-encoder model based on [BERT](https://huggingface.co/docs/transformers/model_doc/bert) architecture.
tokenizer (`transformers.BertTokenizer`):
Tokenizer of class
[BertTokenizer](https://huggingface.co/docs/transformers/model_doc/bert#transformers.BertTokenizer).
unet ([`UNet2DConditionModel`]): Conditional U-Net architecture to denoise the encoded image latents.
Text-encoder model based on [`~transformers.BERT`].
tokenizer ([`~transformers.BertTokenizer`]):
A `BertTokenizer` 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`].
......@@ -247,51 +249,47 @@ class VersatileDiffusionImageVariationPipeline(DiffusionPipeline):
**kwargs,
):
r"""
Function invoked when calling the pipeline for generation.
The call function to the pipeline for generation.
Args:
image (`PIL.Image.Image`, `List[PIL.Image.Image]` or `torch.Tensor`):
The image prompt or prompts to guide the image generation.
height (`int`, *optional*, defaults to self.image_unet.config.sample_size * self.vae_scale_factor):
height (`int`, *optional*, defaults to `self.image_unet.config.sample_size * self.vae_scale_factor`):
The height in pixels of the generated image.
width (`int`, *optional*, defaults to self.image_unet.config.sample_size * self.vae_scale_factor):
width (`int`, *optional*, defaults to `self.image_unet.config.sample_size * self.vae_scale_factor`):
The width in pixels of the generated image.
num_inference_steps (`int`, *optional*, defaults to 50):
The number of denoising steps. More denoising steps usually lead to a higher quality image at the
expense of slower inference.
guidance_scale (`float`, *optional*, defaults to 7.5):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
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 not to guide the image generation. Ignored when not using guidance (i.e., ignored
if `guidance_scale` is less than `1`).
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) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
[`schedulers.DDIMScheduler`], will be ignored for others.
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`, *optional*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
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`.
tensor is generated by sampling using the supplied random `generator`.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generate image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
The output format of the generated image. Choose between `PIL.Image` or `np.array`.
return_dict (`bool`, *optional*, defaults to `True`):
Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] instead of a
plain tuple.
callback (`Callable`, *optional*):
A function that 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)`.
A function that calls every `callback_steps` steps during inference. The function is called with the
following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function will be called. If not specified, the callback will be
called at every step.
The frequency at which the `callback` function is called. If not specified, the callback is called at
every step.
Examples:
......@@ -320,10 +318,8 @@ class VersatileDiffusionImageVariationPipeline(DiffusionPipeline):
Returns:
[`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] or `tuple`:
[`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] if `return_dict` is True, otherwise a `tuple.
When returning a tuple, the first element is a list with the generated images, and the second element is a
list of `bool`s denoting whether the corresponding generated image likely represents "not-safe-for-work"
(nsfw) content, according to the `safety_checker`.
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.
"""
# 0. Default height and width to unet
height = height or self.image_unet.config.sample_size * self.vae_scale_factor
......
src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion_text_to_image.py
View file @ a69754bb
......@@ -33,18 +33,20 @@ logger = logging.get_logger(__name__) # pylint: disable=invalid-name
class VersatileDiffusionTextToImagePipeline(DiffusionPipeline):
r"""
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.)
Pipeline for text-to-image generation using Versatile 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.).
Parameters:
vqvae ([`VQModel`]):
Vector-quantized (VQ) Model to encode and decode images to and from latent representations.
Vector-quantized (VQ) model to encode and decode images to and from latent representations.
bert ([`LDMBertModel`]):
Text-encoder model based on [BERT](https://huggingface.co/docs/transformers/model_doc/bert) architecture.
tokenizer (`transformers.BertTokenizer`):
Tokenizer of class
[BertTokenizer](https://huggingface.co/docs/transformers/model_doc/bert#transformers.BertTokenizer).
unet ([`UNet2DConditionModel`]): Conditional U-Net architecture to denoise the encoded image latents.
Text-encoder model based on [`~transformers.BERT`].
tokenizer ([`~transformers.BertTokenizer`]):
A `BertTokenizer` 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`].
......@@ -329,51 +331,47 @@ class VersatileDiffusionTextToImagePipeline(DiffusionPipeline):
**kwargs,
):
r"""
Function invoked when calling the pipeline for generation.
The call function to the pipeline for generation.
Args:
prompt (`str` or `List[str]`):
The prompt or prompts to guide the image generation.
height (`int`, *optional*, defaults to self.image_unet.config.sample_size * self.vae_scale_factor):
The prompt or prompts to guide image generation.
height (`int`, *optional*, defaults to `self.image_unet.config.sample_size * self.vae_scale_factor`):
The height in pixels of the generated image.
width (`int`, *optional*, defaults to self.image_unet.config.sample_size * self.vae_scale_factor):
width (`int`, *optional*, defaults to `self.image_unet.config.sample_size * self.vae_scale_factor`):
The width in pixels of the generated image.
num_inference_steps (`int`, *optional*, defaults to 50):
The number of denoising steps. More denoising steps usually lead to a higher quality image at the
expense of slower inference.
guidance_scale (`float`, *optional*, defaults to 7.5):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
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 not to guide the image generation. Ignored when not using guidance (i.e., ignored
if `guidance_scale` is less than `1`).
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) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
[`schedulers.DDIMScheduler`], will be ignored for others.
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`, *optional*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
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`.
tensor is generated by sampling using the supplied random `generator`.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generate image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
The output format of the generated image. Choose between `PIL.Image` or `np.array`.
return_dict (`bool`, *optional*, defaults to `True`):
Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] instead of a
plain tuple.
callback (`Callable`, *optional*):
A function that 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)`.
A function that calls every `callback_steps` steps during inference. The function is called with the
following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function will be called. If not specified, the callback will be
called at every step.
The frequency at which the `callback` function is called. If not specified, the callback is called at
every step.
Examples:
......@@ -394,10 +392,8 @@ class VersatileDiffusionTextToImagePipeline(DiffusionPipeline):
Returns:
[`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] or `tuple`:
[`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] if `return_dict` is True, otherwise a `tuple.
When returning a tuple, the first element is a list with the generated images, and the second element is a
list of `bool`s denoting whether the corresponding generated image likely represents "not-safe-for-work"
(nsfw) content, according to the `safety_checker`.
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.
"""
# 0. Default height and width to unet
height = height or self.image_unet.config.sample_size * self.vae_scale_factor
......
src/diffusers/pipelines/vq_diffusion/pipeline_vq_diffusion.py
View file @ a69754bb
......@@ -51,24 +51,21 @@ class LearnedClassifierFreeSamplingEmbeddings(ModelMixin, ConfigMixin):
class VQDiffusionPipeline(DiffusionPipeline):
r"""
Pipeline for text-to-image generation using VQ Diffusion
Pipeline for text-to-image generation using VQ 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.)
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.).
Args:
vqvae ([`VQModel`]):
Vector Quantized Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent
Vector Quantized Variational Auto-Encoder (VAE) model to encode and decode images to and from latent
representations.
text_encoder ([`CLIPTextModel`]):
Frozen text-encoder. VQ Diffusion uses the text portion of
[CLIP](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTextModel), specifically
the [clip-vit-base-patch32](https://huggingface.co/openai/clip-vit-base-patch32) variant.
tokenizer (`CLIPTokenizer`):
Tokenizer of class
[CLIPTokenizer](https://huggingface.co/docs/transformers/v4.21.0/en/model_doc/clip#transformers.CLIPTokenizer).
text_encoder ([`~transformers.CLIPTextModel`]):
Frozen text-encoder ([clip-vit-base-patch32](https://huggingface.co/openai/clip-vit-base-patch32)).
tokenizer ([`~transformers.CLIPTokenizer`]):
A `CLIPTokenizer` to tokenize text.
transformer ([`Transformer2DModel`]):
Conditional transformer to denoise the encoded image latents.
A conditional `Transformer2DModel` to denoise the encoded image latents.
scheduler ([`VQDiffusionScheduler`]):
A scheduler to be used in combination with `transformer` to denoise the encoded image latents.
"""
......@@ -179,20 +176,17 @@ class VQDiffusionPipeline(DiffusionPipeline):
callback_steps: int = 1,
) -> Union[ImagePipelineOutput, Tuple]:
"""
Function invoked when calling the pipeline for generation.
The call function to the pipeline for generation.
Args:
prompt (`str` or `List[str]`):
The prompt or prompts to guide the image generation.
The prompt or prompts to guide image generation.
num_inference_steps (`int`, *optional*, defaults to 100):
The number of denoising steps. More denoising steps usually lead to a higher quality image at the
expense of slower inference.
guidance_scale (`float`, *optional*, defaults to 7.5):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
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`.
truncation_rate (`float`, *optional*, defaults to 1.0 (equivalent to no truncation)):
Used to "truncate" the predicted classes for x_0 such that the cumulative probability for a pixel is at
most `truncation_rate`. The lowest probabilities that would increase the cumulative probability above
......@@ -200,27 +194,27 @@ class VQDiffusionPipeline(DiffusionPipeline):
num_images_per_prompt (`int`, *optional*, defaults to 1):
The number of images to generate per prompt.
generator (`torch.Generator`, *optional*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
latents (`torch.FloatTensor` of shape (batch), *optional*):
Pre-generated noisy latents to be used as inputs for image generation. Must be valid embedding indices.
Can be used to tweak the same generation with different prompts. If not provided, a latents tensor will
be generated of completely masked latent pixels.
Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image
generation. Must be valid embedding indices.If not provided, a latents tensor will be generated of
completely masked latent pixels.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generated image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
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.ImagePipelineOutput`] instead of a plain tuple.
callback (`Callable`, *optional*):
A function that will be called every `callback_steps` steps during inference. The function will be
called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
A function that calls every `callback_steps` steps during inference. The function is called with the
following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function will be called. If not specified, the callback will be
called at every step.
The frequency at which the `callback` function is called. If not specified, the callback is called at
every step.
Returns:
[`~pipelines.ImagePipelineOutput`] or `tuple`: [`~ pipeline_utils.ImagePipelineOutput `] if `return_dict`
is True, otherwise a `tuple. When returning a tuple, the first element is a list with the generated images.
[`~pipelines.ImagePipelineOutput`] or `tuple`:
If `return_dict` is `True`, [`~pipelines.ImagePipelineOutput`] is returned, otherwise a `tuple` is
returned where the first element is a list with the generated images.
"""
if isinstance(prompt, str):
batch_size = 1
......@@ -309,8 +303,9 @@ class VQDiffusionPipeline(DiffusionPipeline):
def truncate(self, log_p_x_0: torch.FloatTensor, truncation_rate: float) -> torch.FloatTensor:
"""
Truncates log_p_x_0 such that for each column vector, the total cumulative probability is `truncation_rate` The
lowest probabilities that would increase the cumulative probability above `truncation_rate` are set to zero.
Truncates `log_p_x_0` such that for each column vector, the total cumulative probability is `truncation_rate`
The lowest probabilities that would increase the cumulative probability above `truncation_rate` are set to
zero.
"""
sorted_log_p_x_0, indices = torch.sort(log_p_x_0, 1, descending=True)
sorted_p_x_0 = torch.exp(sorted_log_p_x_0)
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment