Spaces:
Running
Running
File size: 4,116 Bytes
c9311b7 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 | from abc import ABC, abstractmethod
from typing import List
import torch
from torch import nn
from sampling.sampleable import Sampleable, IsotropicGaussian
from sampling.noise_scheduling import Alpha, Beta
class ConditionalProbabilityPath(nn.Module, ABC):
"""
Abstract class for conditional probability paths.
"""
def __init__(self, p_simple: Sampleable, p_data: Sampleable):
"""
:param p_simple: A simple probability distribution,
:param p_data: Probability distribution of the data
"""
super().__init__()
self.p_simple = p_simple
self.p_data = p_data
def sample_marginal_path(self, t: torch.Tensor) -> torch.Tensor:
"""
Samples from the marginal distribution p_t(x) = p_t(x|z) p(z).
:param t: time, shape (num_samples, 1, 1, 1)
:return: samples from p_t(x), shape (num_samples, c, h, w)
"""
num_samples = t.shape[0]
# Sample conditioning variable z ~ p(z)
z, _ = self.sample_conditioning_variable(num_samples) # (num_samples, c, h, w)
# Sample conditional probability path x ~ p_t(x|z)
x = self.sample_conditional_path(z, t) # (num_samples, c, h, w)
return x
@abstractmethod
def sample_conditioning_variable(self, num_samples: int) -> torch.Tensor:
"""
Samples the conditioning variable z.
:param num_samples: number of samples
:return: z, shape (num_samples, c, h, w)
"""
pass
@abstractmethod
def sample_conditional_path(self, z: torch.Tensor, t: torch.Tensor) -> torch.Tensor:
"""
Samples from the conditional distribution p_t(x|z).
:param z: conditioning variable, shape (num_samples, c, h, w)
:param t: time, shape (num_samples, 1, 1, 1)
:return: x: samples from p_t(x|z) ,shape (num_samples, c, h, w)
"""
pass
@abstractmethod
def conditional_vector_field(self, x: torch.Tensor, z: torch.Tensor, t: torch.Tensor) -> torch.Tensor:
"""
Evaluates the conditional vector field u_t(x|z).
:param x: position variable, shape (num_samples, c, h, w)
:param z: conditioning variable, shape (num_samples, c, h, w)
:param t: time, shape (num_samples, 1, 1, 1)
:return: conditional vector field, shape (num_samples, c, h, w)
"""
pass
@abstractmethod
def conditional_score(self, x: torch.Tensor, z: torch.Tensor, t: torch.Tensor) -> torch.Tensor:
"""
Evaluates the conditional score of p_t(x|z).
:param x: position variable, shape (num_samples, c, h, w)
:param z: conditioning variable, shape (num_samples, c, h, w)
:param t: time, shape (num_samples, 1, 1, 1)
:return: conditional score, shape (num_samples, c, h, w)
"""
pass
class GaussianConditionalProbabilityPath(ConditionalProbabilityPath):
def __init__(self, p_data: Sampleable, p_simple_shape: List[int], alpha: Alpha, beta: Beta):
p_simple = IsotropicGaussian(shape=p_simple_shape, std=1.0)
super().__init__(p_simple, p_data)
self.alpha = alpha
self.beta = beta
def sample_conditioning_variable(self, num_samples: int) -> torch.Tensor:
return self.p_data.sample(num_samples)
def sample_conditional_path(self, z: torch.Tensor, t: torch.Tensor) -> torch.Tensor:
return self.alpha(t) * z + self.beta(t) * torch.randn_like(z)
def conditional_vector_field(self, x: torch.Tensor, z: torch.Tensor, t: torch.Tensor) -> torch.Tensor:
alpha_t = self.alpha(t) # (num_samples, 1, 1, 1)
beta_t = self.beta(t) # (num_samples, 1, 1, 1)
dt_alpha_t = self.alpha.dt(t) # (num_samples, 1, 1, 1)
dt_beta_t = self.beta.dt(t) # (num_samples, 1, 1, 1)
return (dt_alpha_t - dt_beta_t / beta_t * alpha_t) * z + dt_beta_t / beta_t * x
def conditional_score(self, x: torch.Tensor, z: torch.Tensor, t: torch.Tensor) -> torch.Tensor:
alpha_t = self.alpha(t)
beta_t = self.beta(t)
return (z * alpha_t - x) / beta_t ** 2 |