-
Notifications
You must be signed in to change notification settings - Fork 15
/
Copy pathexp.sh
126 lines (105 loc) · 16 KB
/
exp.sh
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
# README: Since our script generating the random images for training/validation on the fly using memory,
# sometime deadlocks may happen in PyTorch dataloader when num_workers > 0.
# The solution is to use "OMP_NUM_THREADS=1 MKL_NUM_THREADS=1" as prefix to each of those scripts.
# Demosaicing Algorithm: Malvar2004
# 512+64=576 image size, while random cropping 512 within 16px boundary
# Achieve 99% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 576 --crop_size 512 --image_type both --crop random_crop_inside_boundary
# Stuck at ~50% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 576 --crop_size 512 --image_type jpeg --crop random_crop_inside_boundary
# lr tuning, and none achieve a better accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 576 --crop_size 512 --image_type jpeg --crop random_crop_inside_boundary --learning_rate 0.1
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 576 --crop_size 512 --image_type jpeg --crop random_crop_inside_boundary --learning_rate 0.01
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 576 --crop_size 512 --image_type jpeg --crop random_crop_inside_boundary --learning_rate 0.0001
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 576 --crop_size 512 --image_type jpeg --crop random_crop_inside_boundary --learning_rate 0.00001
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 576 --crop_size 512 --image_type jpeg --crop random_crop_inside_boundary --learning_rate 0.000001
# Stuck at ~50% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 576 --crop_size 512 --image_type demosaic --crop random_crop_inside_boundary
# lr tuning, and none achieve a better accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 576 --crop_size 512 --image_type demosaic --crop random_crop_inside_boundary --learning_rate 0.1
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 576 --crop_size 512 --image_type demosaic --crop random_crop_inside_boundary --learning_rate 0.01
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 576 --crop_size 512 --image_type demosaic --crop random_crop_inside_boundary --learning_rate 0.0001
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 576 --crop_size 512 --image_type demosaic --crop random_crop_inside_boundary --learning_rate 0.00001
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 576 --crop_size 512 --image_type demosaic --crop random_crop_inside_boundary --learning_rate 0.000001
# 99 No crop
# Achieve 99% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 99 --image_type both --crop none
# Achieve 99% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 99 --image_type jpeg --crop none
# Stuck at ~50% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 99 --image_type demosaic --crop none
# lr tuning, and none achieve a better accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 99 --image_type demosaic --crop none --learning_rate 0.1
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 99 --image_type demosaic --crop none --learning_rate 0.01
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 99 --image_type demosaic --crop none --learning_rate 0.0001
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 99 --image_type demosaic --crop none --learning_rate 0.00001
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 99 --image_type demosaic --crop none --learning_rate 0.000001
# 100 No crop
# All below 3 achieve 99% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 100 --image_type both --crop none
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 100 --image_type jpeg --crop none
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 100 --image_type demosaic --crop none
# 112 No crop
# Achieve 99% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 112 --image_type both --crop none
# Stuck at ~50% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 112 --image_type jpeg --crop none
# lr tuning, and none achieve a better accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 112 --image_type jpeg --crop none --learning_rate 0.1
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 112 --image_type jpeg --crop none --learning_rate 0.01
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 112 --image_type jpeg --crop none --learning_rate 0.0001
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 112 --image_type jpeg --crop none --learning_rate 0.00001
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 112 --image_type jpeg --crop none --learning_rate 0.000001
# Achieve 99% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo Malvar2004 --image_size 112 --image_type demosaic --crop none
# Demosaicing Algorithm: bilinear
# 512+64=576 image size, while random cropping 512 within 16px boundary
# Achieve 99% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 576 --crop_size 512 --image_type both --crop random_crop_inside_boundary
# Stuck at ~50% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 576 --crop_size 512 --image_type jpeg --crop random_crop_inside_boundary
# lr tuning, and none achieve a better accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 576 --crop_size 512 --image_type jpeg --crop random_crop_inside_boundary --learning_rate 0.1
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 576 --crop_size 512 --image_type jpeg --crop random_crop_inside_boundary --learning_rate 0.01
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 576 --crop_size 512 --image_type jpeg --crop random_crop_inside_boundary --learning_rate 0.0001
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 576 --crop_size 512 --image_type jpeg --crop random_crop_inside_boundary --learning_rate 0.00001
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 576 --crop_size 512 --image_type jpeg --crop random_crop_inside_boundary --learning_rate 0.000001
# Stuck at ~50% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 576 --crop_size 512 --image_type demosaic --crop random_crop_inside_boundary
# lr tuning, and none achieve a better accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 576 --crop_size 512 --image_type demosaic --crop random_crop_inside_boundary --learning_rate 0.1
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 576 --crop_size 512 --image_type demosaic --crop random_crop_inside_boundary --learning_rate 0.01
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 576 --crop_size 512 --image_type demosaic --crop random_crop_inside_boundary --learning_rate 0.0001
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 576 --crop_size 512 --image_type demosaic --crop random_crop_inside_boundary --learning_rate 0.00001
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 576 --crop_size 512 --image_type demosaic --crop random_crop_inside_boundary --learning_rate 0.000001
# 99 No crop
# Achieve 99% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 99 --image_type both --crop none
# Achieve 99% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 99 --image_type jpeg --crop none
# Stuck at ~50% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 99 --image_type demosaic --crop none
# lr tuning, and none achieve a better accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 99 --image_type demosaic --crop none --learning_rate 0.1
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 99 --image_type demosaic --crop none --learning_rate 0.01
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 99 --image_type demosaic --crop none --learning_rate 0.0001
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 99 --image_type demosaic --crop none --learning_rate 0.00001
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 99 --image_type demosaic --crop none --learning_rate 0.000001
# 100 No crop
# All below 3 achieve 99% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 100 --image_type both --crop none
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 100 --image_type jpeg --crop none
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 100 --image_type demosaic --crop none
# 112 No crop
# Achieve 99% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 112 --image_type both --crop none
# Stuck at ~50% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 112 --image_type jpeg --crop none
# lr tuning, and none achieve a better accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 112 --image_type jpeg --crop none --learning_rate 0.1
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 112 --image_type jpeg --crop none --learning_rate 0.01
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 112 --image_type jpeg --crop none --learning_rate 0.0001
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 112 --image_type jpeg --crop none --learning_rate 0.00001
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 112 --image_type jpeg --crop none --learning_rate 0.000001
# Achieve 99% accuracy
OMP_NUM_THREADS=1 MKL_NUM_THREADS=1 python train.py --train_size 100000 --val_size 5000 --image_pattern gaussian_rgb --demosaic_algo bilinear --image_size 112 --image_type demosaic --crop none