|
import enum |
|
import math |
|
|
|
import numpy as np |
|
import torch as th |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
class GaussianDiffusion_SEQDIFF: |
|
""" |
|
T = number of timesteps to set up diffuser with |
|
|
|
schedule = type of noise schedule to use linear, cosine, gaussian |
|
|
|
noise = type of ditribution to sample from; DEFAULT - normal_gaussian |
|
|
|
""" |
|
|
|
def __init__(self, |
|
T=1000, |
|
schedule='sqrt', |
|
sample_distribution='normal', |
|
sample_distribution_gmm_means=[-1.0, 1.0], |
|
sample_distribution_gmm_variances=[1.0, 1.0], |
|
F=1, |
|
): |
|
|
|
|
|
betas = np.array(get_named_beta_schedule(schedule, T), dtype=np.float64) |
|
self.betas = betas |
|
assert len(betas.shape) == 1, "betas must be 1-D" |
|
assert (betas > 0).all() and (betas <= 1).all() |
|
|
|
self.num_timesteps = int(betas.shape[0]) |
|
self.F = F |
|
|
|
alphas = 1.0 - betas |
|
self.alphas_cumprod = np.cumprod(alphas, axis=0) |
|
self.alphas_cumprod_prev = np.append(1.0, self.alphas_cumprod[:-1]) |
|
self.alphas_cumprod_next = np.append(self.alphas_cumprod[1:], 0.0) |
|
assert self.alphas_cumprod_prev.shape == (self.num_timesteps,) |
|
|
|
|
|
self.posterior_variance = (betas * (1.0 - self.alphas_cumprod_prev) / (1.0 - self.alphas_cumprod)) |
|
|
|
|
|
self.posterior_log_variance_clipped = np.log(np.append(self.posterior_variance[1], self.posterior_variance[1:])) |
|
self.posterior_mean_coef1 = (betas * np.sqrt(self.alphas_cumprod_prev) / (1.0 - self.alphas_cumprod)) |
|
self.posterior_mean_coef2 = ((1.0 - self.alphas_cumprod_prev) * np.sqrt(alphas) / (1.0 - self.alphas_cumprod)) |
|
|
|
|
|
self.sqrt_alphas_cumprod = np.sqrt(self.alphas_cumprod) |
|
self.sqrt_one_minus_alphas_cumprod = np.sqrt(1.0 - self.alphas_cumprod) |
|
self.log_one_minus_alphas_cumprod = np.log(1.0 - self.alphas_cumprod) |
|
self.sqrt_recip_alphas_cumprod = np.sqrt(1.0 / self.alphas_cumprod) |
|
|
|
|
|
self.sample_distribution = sample_distribution |
|
self.sample_distribution_gmm_means = [float(mean) for mean in sample_distribution_gmm_means] |
|
self.sample_distribution_gmm_variances = [float(variance) for variance in sample_distribution_gmm_variances] |
|
|
|
if self.sample_distribution == 'normal': |
|
self.noise_function = th.randn_like |
|
else: |
|
self.noise_function = self.randnmixture_like |
|
|
|
|
|
def q_mean_variance(self, x_start, t): |
|
""" |
|
Get the distribution q(x_t | x_0). |
|
:param x_start: the [N x C x ...] tensor of noiseless inputs. |
|
:param t: the number of diffusion steps (minus 1). Here, 0 means one step. |
|
:return: A tuple (mean, variance, log_variance), all of x_start's shape. |
|
""" |
|
mean = ( |
|
_extract(self.sqrt_alphas_cumprod, t, x_start.shape) * x_start |
|
) |
|
variance = _extract(1.0 - self.alphas_cumprod, t, x_start.shape) |
|
log_variance = _extract( |
|
self.log_one_minus_alphas_cumprod, t, x_start.shape |
|
) |
|
return mean, variance, log_variance |
|
|
|
def q_sample(self, x_start, t, mask=None, DEVICE=None): |
|
""" |
|
Diffuse the data for a given number of diffusion steps. |
|
In other words, sample from q(x_t | x_0). |
|
:param x_start: the initial data batch. |
|
:param t: the number of diffusion steps (minus 1). Here, 0 means one step. |
|
:param noise: if specified, the split-out normal noise. |
|
:return: A noisy version of x_start. |
|
""" |
|
|
|
|
|
noise = self.noise_function(x_start)*(self.F**2) |
|
if DEVICE != None: |
|
noise = noise.to(DEVICE) |
|
|
|
assert noise.shape == x_start.shape |
|
x_sample = ( |
|
_extract(self.sqrt_alphas_cumprod, t, x_start.shape) * x_start |
|
+ _extract(self.sqrt_one_minus_alphas_cumprod, t, x_start.shape) |
|
* noise) |
|
|
|
if mask is not None: |
|
x_sample[mask]=x_start[mask] |
|
|
|
return x_sample |
|
|
|
|
|
def q_posterior_mean_variance(self, x_start, x_t, t): |
|
""" |
|
Compute the mean and variance of the diffusion posterior: |
|
q(x_{t-1} | x_t, x_0) |
|
""" |
|
assert x_start.shape == x_t.shape |
|
|
|
posterior_mean = (_extract(self.posterior_mean_coef1, t, x_t.shape) * x_start |
|
+ _extract(self.posterior_mean_coef2, t, x_t.shape) * x_t) |
|
|
|
posterior_variance = _extract(self.posterior_variance, t, x_t.shape) |
|
|
|
posterior_log_variance_clipped = _extract(self.posterior_log_variance_clipped, t, x_t.shape) |
|
|
|
assert ( |
|
posterior_mean.shape[0] |
|
== posterior_variance.shape[0] |
|
== posterior_log_variance_clipped.shape[0] |
|
== x_start.shape[0] |
|
) |
|
return posterior_mean, posterior_variance, posterior_log_variance_clipped |
|
|
|
|
|
def randnmixture_like(self, tensor_like, number_normal=3, weights_normal=None): |
|
|
|
if self.sample_distribution_gmm_means and self.sample_distribution_gmm_variances: |
|
assert len(self.sample_distribution_gmm_means) == len(self.sample_distribution_gmm_variances) |
|
|
|
if not weights_normal: |
|
mix = th.distributions.Categorical(th.ones(len(self.sample_distribution_gmm_means))) |
|
else: |
|
assert len(weights_normal) == number_normal |
|
mix = th.distributions.Categorical(weights_normal) |
|
|
|
comp = th.distributions.Normal(th.tensor(self.sample_distribution_gmm_means), th.tensor(self.sample_distribution_gmm_variances)) |
|
|
|
|
|
|
|
gmm = th.distributions.mixture_same_family.MixtureSameFamily(mix, comp) |
|
return th.tensor([gmm.sample() for _ in range(np.prod(tensor_like.shape))]).reshape(tensor_like.shape) |
|
|
|
|
|
|
|
def get_named_beta_schedule(schedule_name, num_diffusion_timesteps): |
|
""" |
|
Get a pre-defined beta schedule for the given name. |
|
The beta schedule library consists of beta schedules which remain similar |
|
in the limit of num_diffusion_timesteps. |
|
Beta schedules may be added, but should not be removed or changed once |
|
they are committed to maintain backwards compatibility. |
|
""" |
|
if schedule_name == "linear": |
|
|
|
|
|
scale = 1000 / num_diffusion_timesteps |
|
beta_start = scale * 0.0001 |
|
beta_end = scale * 0.02 |
|
return np.linspace(beta_start, beta_end, num_diffusion_timesteps, dtype=np.float64) |
|
|
|
elif schedule_name == "cosine": |
|
return betas_for_alpha_bar(num_diffusion_timesteps, lambda t: math.cos((t + 0.008) / 1.008 * math.pi / 2) ** 2,) |
|
|
|
elif schedule_name == 'sqrt': |
|
return betas_for_alpha_bar(num_diffusion_timesteps, lambda t: 1-np.sqrt(t + 0.0001),) |
|
|
|
else: |
|
raise NotImplementedError(f"unknown beta schedule: {schedule_name}") |
|
|
|
def betas_for_alpha_bar(num_diffusion_timesteps, alpha_bar, max_beta=0.999): |
|
""" |
|
Create a beta schedule that discretizes the given alpha_t_bar function, |
|
which defines the cumulative product of (1-beta) over time from t = [0,1]. |
|
:param num_diffusion_timesteps: the number of betas to produce. |
|
:param alpha_bar: a lambda that takes an argument t from 0 to 1 and |
|
produces the cumulative product of (1-beta) up to that |
|
part of the diffusion process. |
|
:param max_beta: the maximum beta to use; use values lower than 1 to |
|
prevent singularities. |
|
""" |
|
betas = [] |
|
for i in range(num_diffusion_timesteps): |
|
t1 = i / num_diffusion_timesteps |
|
t2 = (i + 1) / num_diffusion_timesteps |
|
betas.append(min(1 - alpha_bar(t2) / alpha_bar(t1), max_beta)) |
|
return np.array(betas) |
|
|
|
|
|
def _extract(arr, timesteps, broadcast_shape): |
|
""" |
|
Extract values from a 1-D numpy array for a batch of indices. |
|
:param arr: the 1-D numpy array. |
|
:param timesteps: a tensor of indices into the array to extract. |
|
:param broadcast_shape: a larger shape of K dimensions with the batch |
|
dimension equal to the length of timesteps. |
|
:return: a tensor of shape [batch_size, 1, ...] where the shape has K dims. |
|
""" |
|
res = th.from_numpy(arr).to(device=timesteps.device)[timesteps].float() |
|
while len(res.shape) < len(broadcast_shape): |
|
res = res[..., None] |
|
return res.expand(broadcast_shape) |
|
|