o
pi% @ s d dl mZ d dlmZmZmZ d dlZd dlZddl m
Z
m Z ddl m
Z
ddlmZ dd lmZ eG d
d d e
ZG d d
d
ee
ZdS ) ) dataclass)OptionalTupleUnionN ) ConfigMixinregister_to_config)
BaseOutput) randn_tensor )SchedulerMixinc @ s8 e Zd ZU dZejed< ejed< dZeej ed< dS )KarrasVeOutputa
Output class for the scheduler's step function output.
Args:
prev_sample (`torch.Tensor` of shape `(batch_size, num_channels, height, width)` for images):
Computed sample (x_{t-1}) of previous timestep. `prev_sample` should be used as next model input in the
denoising loop.
derivative (`torch.Tensor` of shape `(batch_size, num_channels, height, width)` for images):
Derivative of predicted original image sample (x_0).
pred_original_sample (`torch.Tensor` of shape `(batch_size, num_channels, height, width)` for images):
The predicted denoised sample (x_{0}) based on the model output from the current timestep.
`pred_original_sample` can be used to preview progress or for guidance.
prev_sample
derivativeNpred_original_sample) __name__
__module__ __qualname____doc__torchTensor__annotations__r r r r r/home/ubuntu/SoloSpeech/.venv/lib/python3.10/site-packages/diffusers/schedulers/deprecated/scheduling_karras_ve.pyr
s
r
c @ s. e Zd ZdZdZe d-d ed
ed ed ed
edef ddZd.dej de
e dej fddZ d.de de
eejf fddZ d.dej ded e
ej deej ef fd d Z d/d ej d!ed"ed#ej d$ede
eef f d%d&Z d/d ej d!ed"ed#ej d'ej d(ej d$ede
eef fd)d*Zd+d, ZdS )0KarrasVeScheduleraE
A stochastic scheduler tailored to variance-expanding models.
This model inherits from [`SchedulerMixin`] and [`ConfigMixin`]. Check the superclass documentation for the generic
methods the library implements for all schedulers such as loading and saving.
For more details on the parameters, see [Appendix E](https://arxiv.org/abs/2206.00364). The grid search values used
to find the optimal `{s_noise, s_churn, s_min, s_max}` for a specific model are described in Table 5 of the paper.
Args:
sigma_min (`float`, defaults to 0.02):
The minimum noise magnitude.
sigma_max (`float`, defaults to 100):
The maximum noise magnitude.
s_noise (`float`, defaults to 1.007):
The amount of additional noise to counteract loss of detail during sampling. A reasonable range is [1.000,
1.011].
s_churn (`float`, defaults to 80):
The parameter controlling the overall amount of stochasticity. A reasonable range is [0, 100].
s_min (`float`, defaults to 0.05):
The start value of the sigma range to add noise (enable stochasticity). A reasonable range is [0, 10].
s_max (`float`, defaults to 50):
The end value of the sigma range to add noise. A reasonable range is [0.2, 80].
r {Gz?d &1 ?P 皙?2 sigma_min sigma_maxs_noises_churns_mins_maxc C s || _ d | _d | _d | _d S N)init_noise_sigmanum_inference_steps timestepsschedule)selfr! r" r# r$ r% r& r r r __init__Q s
zKarrasVeScheduler.__init__Nsampletimestepreturnc C s |S )a
Ensures interchangeability with schedulers that need to scale the denoising model input depending on the
current timestep.
Args:
sample (`torch.Tensor`):
The input sample.
timestep (`int`, *optional*):
The current timestep in the diffusion chain.
Returns:
`torch.Tensor`:
A scaled input sample.
r )r, r. r/ r r r scale_model_inputc s z#KarrasVeScheduler.scale_model_inputr) devicec sb _ tdj ddd }t||_ fddjD }tj|tj |d_
dS )a
Sets the discrete timesteps used for the diffusion chain (to be run before inference).
Args:
num_inference_steps (`int`):
The number of diffusion steps used when generating samples with a pre-trained model.
device (`str` or `torch.device`, *optional*):
The device to which the timesteps should be moved to. If `None`, the timesteps are not moved.
r Nc s< g | ]}j jd j jd j jd | d qS )r )configr" r! ).0ir) r, r r
s
"z3KarrasVeScheduler.set_timesteps..)dtyper2 ) r) nparangecopyr
from_numpytor* tensorfloat32r+ )r, r) r2 r* r+ r r8 r
set_timestepst s
z KarrasVeScheduler.set_timestepssigma generatorc C s | j j| kr| j jkrn n t| j j| j d}nd}| j jt|j|d |j
}||| }||d |d d | }||fS )u
Explicit Langevin-like "churn" step of adding noise to the sample according to a `gamma_i ≥ 0` to reach a
higher noise level `sigma_hat = sigma_i + gamma_i*sigma_i`.
Args:
sample (`torch.Tensor`):
The input sample.
sigma (`float`):
generator (`torch.Generator`, *optional*):
A random number generator.
g4y?r )rD r ?) r5 r% r& minr$ r) r# r
shaper? r2 )r, r. rC rD gammaeps sigma_hat
sample_hatr r r add_noise_to_input s z$KarrasVeScheduler.add_noise_to_inputT model_outputrJ
sigma_prevrK return_dictc C sB ||| }|| | }||| | }|s||fS t |||dS )a
Predict the sample from the previous timestep by reversing the SDE. This function propagates the diffusion
process from the learned model outputs (most often the predicted noise).
Args:
model_output (`torch.Tensor`):
The direct output from learned diffusion model.
sigma_hat (`float`):
sigma_prev (`float`):
sample_hat (`torch.Tensor`):
return_dict (`bool`, *optional*, defaults to `True`):
Whether or not to return a [`~schedulers.scheduling_karras_ve.KarrasVESchedulerOutput`] or `tuple`.
Returns:
[`~schedulers.scheduling_karras_ve.KarrasVESchedulerOutput`] or `tuple`:
If return_dict is `True`, [`~schedulers.scheduling_karras_ve.KarrasVESchedulerOutput`] is returned,
otherwise a tuple is returned where the first element is the sample tensor.
r r r r
) r, rM rJ rN rK rO r r sample_prevr r r step s zKarrasVeScheduler.steprR r c
C sN ||| }|| | } ||| d| d| }|s ||fS t |||dS )a
Corrects the predicted sample based on the `model_output` of the network.
Args:
model_output (`torch.Tensor`):
The direct output from learned diffusion model.
sigma_hat (`float`): TODO
sigma_prev (`float`): TODO
sample_hat (`torch.Tensor`): TODO
sample_prev (`torch.Tensor`): TODO
derivative (`torch.Tensor`): TODO
return_dict (`bool`, *optional*, defaults to `True`):
Whether or not to return a [`~schedulers.scheduling_ddpm.DDPMSchedulerOutput`] or `tuple`.
Returns:
prev_sample (TODO): updated sample in the diffusion chain. derivative (TODO): TODO
rE rP rQ )
r, rM rJ rN rK rR r rO r derivative_corrr r r step_correct s z KarrasVeScheduler.step_correctc C s t r' )NotImplementedError)r, original_samplesnoiser* r r r add_noise s zKarrasVeScheduler.add_noise)r r r r r r r' )T)r r r r orderr floatr- r r r intr1 r strr2 rB Generatorr rL boolr
rS rU rY r r r r r 1 s
/
'r ) dataclassesr typingr r r numpyr; r configuration_utilsr r utilsr utils.torch_utilsr
scheduling_utilsr r
r r r r r s