-
Notifications
You must be signed in to change notification settings - Fork 2
/
train_fusion.py
168 lines (131 loc) · 5.36 KB
/
train_fusion.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
166
167
168
from utils.pointnet import PointNet
from utils.dualnet import DualNet
from utils.datasets import MonoDataset, DuoDataset
import argparse
import numpy as np
import datetime
import os
import shutil
current_time = datetime.datetime.now()
prefix = current_time.strftime("%m-%d:%H:%M") + "fusion/"
path = "test_results/" + prefix
if not os.path.exists(path):
os.makedirs(path)
else:
shutil.rmtree(path)
os.makedirs(path)
# initialize final variables for return
test_acc = []
test_conf = []
test_energy = []
train_paths = ["train/zero/", "train/one/", "train/two/", "train/three/", "train/four/", "train/five/", "train/thumbup/", "train/ell/", "train/frame/", "train/bird/"]
test_paths = ["test/zero/", "test/one/", "test/two/", "test/three/", "test/four/", "test/five/", "test/thumbup/", "test/ell/", "test/frame/", "test/bird/"]
# initialize running variables for collection
trial_acc = []
trial_conf = np.zeros((len(test_paths),len(test_paths)))
trial_energy = np.zeros(len(test_paths)*201)
# IMPORTANT
num_trials = 25
epochs = 100
bs = 67
num_points = 320
# Single test LL
dataset = MonoDataset(left=True, right=False, num_points=num_points, file_paths=train_paths)
test_dataset = MonoDataset(left=True, right=False, num_points=num_points, file_paths=test_paths)
for j in range(num_trials):
print("LL: ", j)
pnt = PointNet(num_points=320, num_classes=len(test_paths), num_epoch=epochs, batchsize=bs, ptype='small', alpha=0.002, beta=0.01)
res = pnt.train(dataset, test_dataset)
trial_acc.append(res[0])
trial_conf += res[1]
trial_energy += res[2]
test_acc.append(trial_acc)
test_conf.append(trial_conf)
test_energy.append(trial_energy)
# reinitialize running variables for collection
trial_acc = []
trial_conf = np.zeros((len(test_paths),len(test_paths)))
trial_energy = np.zeros(len(test_paths)*201)
np.save(path + "acc", np.array(test_acc))
np.save(path + "conf", np.array(test_conf))
np.save(path + "energy", np.array(test_energy))
# Single test RR
dataset = MonoDataset(left=False, right=True, num_points=num_points, file_paths=train_paths)
test_dataset = MonoDataset(left=False, right=True, num_points=num_points, file_paths=test_paths)
for j in range(num_trials):
print("RR: ", j)
pnt = PointNet(num_points=320, num_classes=len(test_paths), num_epoch=epochs, batchsize=bs, ptype='small', alpha=0.002, beta=0.01)
res = pnt.train(dataset, test_dataset)
trial_acc.append(res[0])
trial_conf += res[1]
trial_energy += res[2]
test_acc.append(trial_acc)
test_conf.append(trial_conf)
test_energy.append(trial_energy)
# reinitialize running variables for collection
trial_acc = []
trial_conf = np.zeros((len(test_paths),len(test_paths)))
trial_energy = np.zeros(len(test_paths)*201)
np.save(path + "acc", np.array(test_acc))
np.save(path + "conf", np.array(test_conf))
np.save(path + "energy", np.array(test_energy))
# Single test F
dataset = MonoDataset(left=True, right=True, num_points=num_points, file_paths=train_paths)
test_dataset = MonoDataset(left=True, right=True, num_points=num_points, file_paths=test_paths)
for j in range(num_trials):
print("F: ", j)
pnt = PointNet(num_points=320, num_classes=len(test_paths), num_epoch=epochs, batchsize=bs, ptype='small', alpha=0.002, beta=0.01)
res = pnt.train(dataset, test_dataset)
trial_acc.append(res[0])
trial_conf += res[1]
trial_energy += res[2]
test_acc.append(trial_acc)
test_conf.append(trial_conf)
test_energy.append(trial_energy)
# reinitialize running variables for collection
trial_acc = []
trial_conf = np.zeros((len(test_paths),len(test_paths)))
trial_energy = np.zeros(len(test_paths)*201)
np.save(path + "acc", np.array(test_acc))
np.save(path + "conf", np.array(test_conf))
np.save(path + "energy", np.array(test_energy))
# Dual test
dataset = DuoDataset(num_points=320, file_paths=train_paths)
test_dataset = DuoDataset(num_points=320, file_paths=test_paths)
for j in range(num_trials):
print("Dual: ", j)
pnt = DualNet(num_points=320, num_classes=len(test_paths), num_epoch=epochs, batchsize=bs, ptype='', alpha=0.002, beta=0.01)
res = pnt.train(dataset, test_dataset)
trial_acc.append(res[0])
trial_conf += res[1]
trial_energy += res[2]
test_acc.append(trial_acc)
test_conf.append(trial_conf)
test_energy.append(trial_energy)
# reinitialize running variables for collection
trial_acc = []
trial_conf = np.zeros((len(test_paths),len(test_paths)))
trial_energy = np.zeros(len(test_paths)*201)
np.save(path + "acc", np.array(test_acc))
np.save(path + "conf", np.array(test_conf))
np.save(path + "energy", np.array(test_energy))
# Dual test modified
dataset = DuoDataset(num_points=320, file_paths=train_paths)
test_dataset = DuoDataset(num_points=320, file_paths=test_paths)
for j in range(num_trials):
print("Dual Mod: ", j)
pnt = DualNet(num_points=320, num_classes=len(test_paths), num_epoch=epochs, batchsize=bs, ptype='modified', alpha=0.002, beta=0.01)
res = pnt.train(dataset, test_dataset)
trial_acc.append(res[0])
trial_conf += res[1]
trial_energy += res[2]
test_acc.append(trial_acc)
test_conf.append(trial_conf)
test_energy.append(trial_energy)
# reinitialize running variables for collection
trial_acc = []
trial_conf = np.zeros((len(test_paths),len(test_paths)))
trial_energy = np.zeros(len(test_paths)*201)
np.save(path + "acc", np.array(test_acc))
np.save(path + "conf", np.array(test_conf))
np.save(path + "energy", np.array(test_energy))