-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathsdedit.py
186 lines (155 loc) · 6.07 KB
/
sdedit.py
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
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
# References:
# https://nn.labml.ai/diffusion/ddpm/index.html
# https://github.com/davidADSP/Generative_Deep_Learning_2nd_Edition/blob/main/notebooks/08_diffusion/01_ddm/ddm.ipynb
# https://huggingface.co/blog/annotated-diffusion
import torch
from torch import nn
import numpy as np
from tqdm import tqdm
import numpy as np
from celeba import CelebADS
from utils import image_to_grid
from stroke import StrokeSimulator
class DDPM(nn.Module):
def linearly_schedule_beta(self):
self.beta = torch.linspace(
self.init_beta,
self.fin_beta,
self.n_diffusion_steps,
device=self.device,
)
self.alpha = 1 - self.beta
self.alpha_bar = torch.cumprod(self.alpha, dim=0)
self.signal_rate = self.alpha_bar ** 0.5 # "$\alpha(t)$"
self.noise_rate = (1 - self.alpha ** 2) ** 0.5 # "$\sigma(t)$"
def __init__(
self,
model,
img_size,
device,
image_channels=3,
n_diffusion_steps=1000,
init_beta=0.0001,
fin_beta=0.02,
):
super().__init__()
self.img_size = img_size
self.device = device
self.image_channels = image_channels
self.n_diffusion_steps = n_diffusion_steps
self.init_beta = init_beta
self.fin_beta = fin_beta
self.model = model.to(device)
self.linearly_schedule_beta()
@staticmethod
def index(x, diffusion_step):
return x[diffusion_step][:, None, None, None]
def sample_noise(self, batch_size):
return torch.randn(
size=(
batch_size, self.image_channels, self.img_size, self.img_size,
),
device=self.device,
)
def sample_diffusion_step(self, batch_size):
return torch.randint(
0, self.n_diffusion_steps, size=(batch_size,), device=self.device,
)
def batchify_diffusion_steps(self, step, batch_size):
return torch.full(
size=(batch_size,),
fill_value=step,
dtype=torch.long,
device=self.device,
)
def perform_diffusion_process(
self, ori_image, diffusion_step, rand_noise=None,
):
"""
$\mathbf{x}(t)
= \alpha(t)\mathbf{x}(0) + \sigma(t)\mathbf{z},
\mathbf{z} \sim \mathcal{N}(\mathbf{0}, \mathbf{I})$
"""
signal_rate_t = self.index(
self.signal_rate, diffusion_step=diffusion_step,
)
noise_rate_t = self.index(self.noise_rate, diffusion_step=diffusion_step)
if rand_noise is None:
rand_noise = self.sample_noise(batch_size=ori_image.size(0))
noisy_image = signal_rate_t * ori_image + noise_rate_t * rand_noise
return noisy_image
def forward(self, noisy_image, diffusion_step):
return self.model(
noisy_image=noisy_image.to(self.device), diffusion_step=diffusion_step,
)
@torch.inference_mode()
def take_denoising_step(self, noisy_image, step):
diffusion_step = self.batchify_diffusion_steps(
step=step, batch_size=noisy_image.size(0),
)
alpha_t = self.index(self.alpha, diffusion_step=diffusion_step)
beta_t = self.index(self.beta, diffusion_step=diffusion_step)
alpha_bar_t = self.index(self.alpha_bar, diffusion_step=diffusion_step)
pred_noise = self(
noisy_image=noisy_image.detach(), diffusion_step=diffusion_step,
)
model_mean = (1 / (alpha_t ** 0.5)) * (
noisy_image - ((beta_t / ((1 - alpha_bar_t) ** 0.5)) * pred_noise)
)
model_var = beta_t
if step > 0:
rand_noise = self.sample_noise(batch_size=noisy_image.size(0))
else:
rand_noise = torch.zeros(
size=(noisy_image.size(0), self.image_channels, self.img_size, self.img_size),
device=self.device,
)
return model_mean + (model_var ** 0.5) * rand_noise
@staticmethod
def _get_frame(x):
grid = image_to_grid(x, n_cols=int(x.size(0) ** 0.5))
frame = np.array(grid)
return frame
def perform_denoising_process(self, noisy_image, start_step, n_frames=None):
if n_frames is not None:
frames = list()
x = noisy_image
pbar = tqdm(range(start_step, -1, -1), leave=False)
for step in pbar:
pbar.set_description("Denoising...")
x = self.take_denoising_step(x, step=step)
if n_frames is not None and (
step % (self.n_diffusion_steps // n_frames) == 0
):
frames.append(self._get_frame(x))
return frames if n_frames is not None else x
class SDEdit(DDPM):
def __init__(self, model, data_dir, img_size, device, kernel_size=3):
super().__init__(model=model, img_size=img_size, device=device)
self.stroke_sim = StrokeSimulator(kernel_size=kernel_size)
self.ds = CelebADS(
data_dir=data_dir, split="test", img_size=self.img_size, hflip=False,
)
def select_and_batchify_ref(self, ref_idx, batch_size):
# if dataset == "celeba":
return self.ds[ref_idx][None, ...].to(self.device).repeat(batch_size, 1, 1, 1)
def time_to_step(self, time):
return int(time * self.n_diffusion_steps)
def sample_from_stroke(self, ref_idx, interm_time, n_colors, batch_size):
ref = self.select_and_batchify_ref(
ref_idx=ref_idx, batch_size=batch_size - 2,
)
stroke = self.stroke_sim(ref, n_colors=n_colors).to(self.device)
interm_step = self.time_to_step(interm_time)
diffusion_step = self.batchify_diffusion_steps(
step=interm_step - 1, batch_size=batch_size - 2,
)
noisy_stroke = self.perform_diffusion_process(
ori_image=stroke,
diffusion_step=diffusion_step,
)
denoised_stroke = self.perform_denoising_process(
noisy_image=noisy_stroke,
start_step=interm_step - 1,
)
return torch.cat([ref[: 1, ...], stroke[: 1, ...], denoised_stroke], dim=0)