forked from WongKinYiu/yolov7
-
Notifications
You must be signed in to change notification settings - Fork 0
/
DatasetSplitter.py
95 lines (62 loc) · 3.29 KB
/
DatasetSplitter.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
import pandas as pd
import argparse
import os
import shutil
from sklearn.model_selection import train_test_split
parser = argparse.ArgumentParser(description='test program')
parser.add_argument('-init', '--settings', help='print the supplied argument.', nargs='*')
args = parser.parse_args()
"""
"""
def main():
if len(args.settings) < 3:
print('too few arguments')
print('example: python3 DatasetSplitter.py -init datasets/sodascapes/SODA_Labels datasets/sodascapes/InfraredImagesPNG datasets/sodascapes/')
return 0
else:
print(
'example: python3 DatasetSplitter.py -init datasets/sodascapes/SODA_Labels datasets/sodascapes/InfraredImagesPNG datasets/sodascapes/')
labels_dir = args.settings[0]
frames_dir = args.settings[1]
results_dir = args.settings[2]
print(frames_dir, labels_dir, results_dir)
# sorting the frames and labels and saving list of names and directories
frames = sorted(os.listdir(frames_dir))
labels = sorted(os.listdir(labels_dir))
# Let's say we want to split the data in 80:10:10 for train:valid:test dataset
# In the first step we will split the data in training and remaining dataset
X_train, X_rem, y_train, y_rem = train_test_split(frames, labels, train_size=0.7)
# Now since we want the valid and test size to be equal (10% each of overall data).
# we have to define valid_size=0.5 (that is 50% of remaining data)
test_size = 0.5
X_valid, X_test, y_valid, y_test = train_test_split(X_rem, y_rem, test_size=0.5)
print(len(X_train)), print(len(y_train))
print(len(X_test)), print(len(y_test))
print(len(X_valid)), print(len(y_valid))
# print(X_train)
# Saving the splits intro their respective directors
sourcefolder = frames_dir.split('/')[len(frames_dir.split('/'))-1]
print(sourcefolder)
for i in range(len(X_train)):
folder_name = [os.path.join(labels_dir, y_train[i]), os.path.join(results_dir, 'labels', 'train',sourcefolder)]
os.makedirs(folder_name[1], exist_ok=True)
shutil.copy(folder_name[0],folder_name[1])
folder_name = [os.path.join(frames_dir, X_train[i]), os.path.join(results_dir, 'images', 'train',sourcefolder)]
os.makedirs(folder_name[1], exist_ok=True)
shutil.copy(folder_name[0],folder_name[1])
for i in range(len(X_test)):
folder_name = [os.path.join(labels_dir, y_test[i]), os.path.join(results_dir, 'labels', 'test',sourcefolder)]
os.makedirs(folder_name[1], exist_ok=True)
shutil.copy(folder_name[0],folder_name[1])
folder_name = [os.path.join(frames_dir, X_test[i]), os.path.join(results_dir, 'images', 'test',sourcefolder)]
os.makedirs(folder_name[1], exist_ok=True)
shutil.copy(folder_name[0],folder_name[1])
for i in range(len(X_valid)):
folder_name = [os.path.join(labels_dir, y_valid[i]), os.path.join(results_dir, 'labels', 'val',sourcefolder)]
os.makedirs(folder_name[1], exist_ok=True)
shutil.copy(folder_name[0],folder_name[1])
folder_name = [os.path.join(frames_dir, X_valid[i]), os.path.join(results_dir, 'images', 'val',sourcefolder)]
os.makedirs(folder_name[1], exist_ok=True)
shutil.copy(folder_name[0],folder_name[1])
if __name__ == '__main__':
main()