-
Notifications
You must be signed in to change notification settings - Fork 20
/
gradient_descent.py
137 lines (124 loc) · 3.53 KB
/
gradient_descent.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
from diffusercam_gd import grad_descent, GradientDescentUpdate
from lensless.recon.gd import FISTA
from lensless.recon.apgd import APGD, APGDPriors
from lensless.utils.io import load_data
import time
import pathlib as plib
psf_fp = "data/psf/tape_rgb.png"
data_fp = "data/raw_data/thumbs_up_rgb.png"
n_iter = 300
downsample = 4
gray = True
dtype = "float32"
n_trials = 3
psf, data = load_data(
psf_fp=psf_fp,
data_fp=data_fp,
downsample=downsample,
plot=False,
gray=gray,
dtype=dtype,
)
save = "profile_gradient_descent"
save = plib.Path(__file__).parent / save
save.mkdir(exist_ok=True)
""" LenslessPiCam """
recon = FISTA(psf, dtype=dtype)
recon.set_data(data)
# res = recon.apply(n_iter=n_iter, save=save, disp_iter=None)
# recon.reset()
total_time = 0
for _ in range(n_trials):
start_time = time.time()
res = recon.apply(n_iter=n_iter, disp_iter=None)
total_time += time.time() - start_time
recon.reset()
print(f"LenslessPiCam (avg) : {total_time / n_trials} s")
# -- using Pycsou, complex conv
recon = APGD(
psf,
max_iter=n_iter,
acceleration="BT",
diff_penalty=None,
prox_penalty=APGDPriors.NONNEG,
realconv=False,
)
recon.set_data(data)
res = recon.apply(n_iter=n_iter, save=save, disp_iter=None)
recon.reset()
total_time = 0
for _ in range(n_trials):
start_time = time.time()
res = recon.apply(n_iter=n_iter, disp_iter=None)
total_time += time.time() - start_time
recon.reset()
print(f"LenslessPiCam (Pycsou, complex) : {total_time / n_trials} s")
# -- using Pycsou, real conv
recon = APGD(
psf,
max_iter=n_iter,
acceleration="BT",
diff_penalty=None,
prox_penalty=APGDPriors.NONNEG,
realconv=True,
)
recon.set_data(data)
res = recon.apply(n_iter=n_iter, save=save, disp_iter=None)
recon.reset()
total_time = 0
for _ in range(n_trials):
start_time = time.time()
res = recon.apply(n_iter=n_iter, disp_iter=None)
total_time += time.time() - start_time
recon.reset()
print(f"LenslessPiCam (Pycsou, real) : {total_time / n_trials} s")
""" DiffuserCam"""
method = GradientDescentUpdate.FISTA
grad_descent(psf, data, n_iter=n_iter, update_method=method, disp_iter=None, save=save)
start_time = time.time()
for _ in range(n_trials):
grad_descent(psf, data, n_iter=n_iter, update_method=method, disp_iter=None)
print(f"DiffuserCam : {(time.time() - start_time) / n_trials} s")
""" PyTorch CPU """
psf, data = load_data(
psf_fp=psf_fp,
data_fp=data_fp,
downsample=downsample,
plot=False,
gray=gray,
dtype=dtype,
use_torch=True,
)
recon = FISTA(psf, dtype=dtype)
recon.set_data(data)
res = recon.apply(n_iter=n_iter, save=save, disp_iter=None, plot=False)
recon.reset()
total_time = 0
for _ in range(n_trials):
start_time = time.time()
res = recon.apply(n_iter=n_iter, disp_iter=None, plot=False)
total_time += time.time() - start_time
recon.reset()
print(f"LenslessPiCam, PyTorch CPU (avg) : {total_time / n_trials} s")
""" PyTorch GPU """
psf, data = load_data(
psf_fp=psf_fp,
data_fp=data_fp,
downsample=downsample,
plot=False,
gray=gray,
dtype=dtype,
use_torch=True,
torch_device="cuda",
)
recon = FISTA(psf, dtype=dtype)
recon.set_data(data)
res = recon.apply(n_iter=n_iter, save=save, disp_iter=None, plot=False)
recon.reset()
total_time = 0
for _ in range(n_trials):
start_time = time.time()
res = recon.apply(n_iter=n_iter, disp_iter=None, plot=False)
total_time += time.time() - start_time
recon.reset()
print(f"LenslessPiCam, PyTorch GPU (avg) : {total_time / n_trials} s")