-
Notifications
You must be signed in to change notification settings - Fork 13
/
Copy pathturbulence_demo.py
154 lines (126 loc) · 6.39 KB
/
turbulence_demo.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
import os
import argparse
import yaml
import numpy as np
import torch
import torchvision.transforms as transforms
import matplotlib.pyplot as plt
from functools import partial
from guided_diffusion.blind_condition_methods import get_conditioning_method
from guided_diffusion.measurements import get_operator, get_noise
from guided_diffusion.unet import create_model
from guided_diffusion.gaussian_diffusion import create_sampler
from data.dataloader import get_dataset, get_dataloader
from util.img_utils import Blurkernel, clear_color, generate_tilt_map
from util.logger import get_logger
def load_yaml(file_path: str) -> dict:
with open(file_path) as f:
config = yaml.load(f, Loader=yaml.FullLoader)
return config
def main():
# Configurations
parser = argparse.ArgumentParser()
parser.add_argument('--img_model_config', type=str, default='configs/model_config.yaml')
parser.add_argument('--kernel_model_config', type=str, default='configs/kernel_model_config.yaml')
parser.add_argument('--tilt_model_config', type=str, default='configs/tilt_model_config.yaml')
parser.add_argument('--diffusion_config', type=str, default='configs/diffusion_config.yaml')
parser.add_argument('--task_config', type=str, default='configs/turbulence_config.yaml')
# Training
parser.add_argument('--gpu', type=int, default=0)
parser.add_argument('--save_dir', type=str, default='./results')
# Regularization
parser.add_argument('--reg_scale', type=float, default=0.0)
parser.add_argument('--reg_ord', type=int, default=0, choices=[0, 1])
args = parser.parse_args()
# logger
logger = get_logger()
# Device setting
device_str = f"cuda:{args.gpu}" if torch.cuda.is_available() else 'cpu'
logger.info(f"Device set to {device_str}.")
device = torch.device(device_str)
# Load configurations
img_model_config = load_yaml(args.img_model_config)
kernel_model_config = load_yaml(args.kernel_model_config)
tilt_model_config = load_yaml(args.tilt_model_config)
diffusion_config = load_yaml(args.diffusion_config)
task_config = load_yaml(args.task_config)
# configs to namespace save space
args.kernel = task_config["kernel"]
args.kernel_size = task_config["kernel_size"]
args.intensity = task_config["intensity"]
# Load model
img_model = create_model(**img_model_config)
img_model = img_model.to(device)
img_model.eval()
kernel_model = create_model(**kernel_model_config)
kernel_model = kernel_model.to(device)
kernel_model.eval()
tilt_model = create_model(**tilt_model_config)
tilt_model = tilt_model.to(device)
tilt_model.eval()
model = {'img': img_model, 'kernel': kernel_model, 'tilt': tilt_model}
# Prepare Operator and noise
measure_config = task_config['measurement']
operator = get_operator(device=device, **measure_config['operator'])
noiser = get_noise(**measure_config['noise'])
logger.info(f"Operation: {measure_config['operator']['name']} / Noise: {measure_config['noise']['name']}")
# Prepare conditioning method
cond_config = task_config['conditioning']
cond_method = get_conditioning_method(cond_config['method'], operator, noiser, **cond_config['params'])
logger.info(f"Conditioning method : {task_config['conditioning']['method']}")
measurement_cond_fn = cond_method.conditioning
# We will not use regularization. So skip the part.
# Load diffusion sampler
sampler = create_sampler(**diffusion_config)
sample_fn = partial(sampler.p_sample_loop, model=model, measurement_cond_fn=measurement_cond_fn)
# Working directory
out_path = os.path.join(args.save_dir, measure_config['operator']['name'])
logger.info(f"work directory is created as {out_path}")
os.makedirs(out_path, exist_ok=True)
for img_dir in ['input', 'recon', 'progress', 'label']:
os.makedirs(os.path.join(out_path, img_dir), exist_ok=True)
# Prepare dataloader
data_config = task_config['data']
transform = transforms.Compose([transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])
dataset = get_dataset(**data_config, transforms=transform)
loader = get_dataloader(dataset, batch_size=1, num_workers=0, train=False)
# set seed for reproduce
np.random.seed(123)
# Do Inference
for i, ref_img in enumerate(loader):
logger.info(f"Inference for image {i}")
fname = str(i).zfill(5) + '.png'
ref_img = ref_img.to(device)
# blur
conv = Blurkernel('gaussian', kernel_size=args.kernel_size, device=device)
kernel = conv.get_kernel().type(torch.float32)
kernel = kernel.to(device).view(1, 1, args.kernel_size, args.kernel_size)
# tilt
img_size = img_model_config["image_size"]
# tile_map requires loop for generation that could be slow.
tilt = generate_tilt_map(img_h=img_size, img_w=img_size, kernel_size=7, device=device)
tilt = torch.clip(tilt, -2.5, 2.5)
# Forward measurement model (Ax + n)
y = operator.forward(ref_img, kernel, tilt)
y_n = noiser(y)
# Set initial sample
# !All values will be given to operator.forward(). Please be aware it.
x_start = {'img': torch.randn(ref_img.shape, device=device).requires_grad_(),
'kernel': torch.randn(kernel.shape, device=device).requires_grad_(),
'tilt': torch.randn(tilt.shape, device=device).requires_grad_()}
# !prior check: keys of model (line 74) must be the same as those of x_start to use diffusion prior.
for k in x_start:
if k in model.keys():
logger.info(f"{k} will use diffusion prior")
else:
logger.info(f"{k} will use uniform prior.")
# sample
sample = sample_fn(x_start=x_start, measurement=y_n, record=False, save_root=out_path)
plt.imsave(os.path.join(out_path, 'input', fname), clear_color(y_n))
plt.imsave(os.path.join(out_path, 'label', 'ker_'+fname), clear_color(kernel))
plt.imsave(os.path.join(out_path, 'label', 'img_'+fname), clear_color(ref_img))
plt.imsave(os.path.join(out_path, 'recon', 'img_'+fname), clear_color(sample['img']))
plt.imsave(os.path.join(out_path, 'recon', 'ker_'+fname), clear_color(sample['kernel']))
if __name__ == '__main__':
main()