-
Notifications
You must be signed in to change notification settings - Fork 5
/
sample_condition.py
165 lines (128 loc) · 7.04 KB
/
sample_condition.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
import os
from functools import partial
import argparse
import yaml
import torch
import torchvision.transforms as transforms
import matplotlib.pyplot as plt
import numpy as np
from piq import psnr, ssim
from piq.perceptual import LPIPS
from guided_diffusion.condition_methods import get_conditioning_method
from guided_diffusion.measurements import get_noise, get_operator
from guided_diffusion.unet import create_model
from guided_diffusion.gaussian_diffusion import create_sampler
from guided_diffusion.svd_replacement import Deblurring, Deblurring2D
from data.dataloader import get_dataset, get_dataloader
from util.img_utils import clear_color, mask_generator, _transform, Blurkernel
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():
parser = argparse.ArgumentParser()
parser.add_argument('--model_config', type=str)
parser.add_argument('--diffusion_config', type=str)
parser.add_argument('--task_config', type=str)
parser.add_argument('--gpu', type=int, default=0)
parser.add_argument('--save_dir', type=str, default='./results')
parser.add_argument('--c_rate', type=float, default=0.95)
parser.add_argument('--particle_size', type=int, default=5)
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
model_config = load_yaml(args.model_config)
diffusion_config = load_yaml(args.diffusion_config)
task_config = load_yaml(args.task_config)
#assert model_config['learn_sigma'] == diffusion_config['learn_sigma'], \
#"learn_sigma must be the same for model and diffusion configuartion."
# Load model
model = create_model(**model_config)
model = model.to(device)
model.eval()
# 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'])
measurement_cond_fn = cond_method.conditioning
logger.info(f"Conditioning method : {task_config['conditioning']['method']}")
# Load diffusion sampler
sampler = create_sampler(**diffusion_config, c_rate=args.c_rate, particle_size=args.particle_size)
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'])
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']
batch_size = 1 # Do not change this value. Larger batch size is not available for particle size > 1.
transform = transforms.Compose([transforms.ToTensor(),
transforms.CenterCrop((256, 256)),
transforms.Resize((256, 256)),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))]) # Preprocessing shared by FFHQ and ImageNet.
dataset = get_dataset(**data_config, transforms=transform)
loader = get_dataloader(dataset, batch_size=batch_size, num_workers=0, train=False)
# (Exception) In case of inpainting, we need to generate a mask
if measure_config['operator']['name'] == 'inpainting':
mask_gen = mask_generator(
**measure_config['mask_opt']
)
# Do inference
for i, ref_img in enumerate(loader):
logger.info(f"Inference for image {i}")
fnames = [str(j).zfill(5) + '.png' for j in range(i * batch_size, (i+1) * batch_size)]
ref_img = ref_img.to(device)
if measure_config['operator'] ['name'] == 'inpainting':
# Masks only exist in the inpainting tasks.
mask = mask_gen(ref_img)
mask = mask[0, 0, :, :].unsqueeze(dim=0).unsqueeze(dim=0)
measurement_cond_fn = partial(cond_method.conditioning, mask=mask)
sample_fn = partial(sample_fn, measurement_cond_fn=measurement_cond_fn, operator = operator, op = 'inpainting', mask = mask)
y = operator.forward(ref_img, mask=mask)
y_n = noiser(y)
elif measure_config['operator'] ['name'] == 'gaussian_blur':
sample_fn = partial(sample_fn, operator = operator, op = 'gaussian_blur', mask = None)
kernel = operator.get_kernel().type(torch.float64).reshape(61,61)
kernel = kernel[30,:] / torch.sqrt(kernel[30,30])
task_Svd = Deblurring(kernel=kernel, channels=3, img_dim=256, device=device)
y = task_Svd.forward(ref_img)
y_n = noiser(y)
elif measure_config['operator'] ['name'] == 'motion_blur':
sample_fn = partial(sample_fn, operator = operator, op = 'motion_blur', mask = None)
kernel = operator.get_kernel().type(torch.float64).reshape(61,61)
kernel1 = kernel[30,:] / torch.sum(kernel[30,:])
kernel1 = torch.tensor([0.0] * 30 + [1.0] + [0.0] * 30)
conv2 = Blurkernel(blur_type='gaussian',
kernel_size=61,
std=0.5,
device=device).to(device)
kernel2 = conv2.get_kernel().view(1,1,61,61).type(torch.float64).reshape(61,61)
kernel2 = kernel2[30,:] / torch.sum(kernel2[30,:])
task_Svd = Deblurring2D(kernel1=kernel1, kernel2=kernel2, channels=3, img_dim=256, device=device)
y = task_Svd.forward(ref_img)
y_n = noiser(y)
else:
sample_fn = partial(sample_fn, operator = operator, op = 'super_resolution', mask = None)
y = operator.forward(ref_img)
y_n = noiser(y)
# Sampling
# If you wish to record the intermediate steps, turn record = True below.
x_start = torch.randn(ref_img.shape, device=device).requires_grad_()
sample = sample_fn(x_start=x_start, measurement=y_n, record=False, save_root=out_path).requires_grad_()
for _ in range(batch_size):
plt.imsave(os.path.join(out_path, 'input', fnames[_]), clear_color(y_n[_,:,:,:].unsqueeze(dim=0)))
plt.imsave(os.path.join(out_path, 'label', fnames[_]), clear_color(ref_img[_,:,:,:].unsqueeze(dim=0)))
plt.imsave(os.path.join(out_path, 'recon', fnames[_]), clear_color(sample[_,:,:,:].unsqueeze(dim=0)))
if __name__ == '__main__':
main()