-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathevaluate_regression_models.py
164 lines (132 loc) · 5.64 KB
/
evaluate_regression_models.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
import matplotlib.pyplot as plt
import numpy as np
import matplotlib.pyplot as plt
import torch
from data import CrowdDataSet
from data import default_train_transforms, default_val_transforms
import argparse
from utils import get_density_map_gaussian
from sklearn.metrics import mean_squared_error, r2_score
import matplotlib.gridspec as gridspes
def run_argparse():
parser = argparse.ArgumentParser()
parser.add_argument("--model", type=str)
return parser.parse_args()
def main(args):
loaders = {
"train": CrowdDataSet(
'part_A/train_data', default_train_transforms()
),
"val": CrowdDataSet(
'part_A/test_data', default_val_transforms()
),
"test_unbalanced": CrowdDataSet(
'part_B/test_data', default_val_transforms()
)
}
model = torch.load('saved_models/' + args.model)
model.eval()
train_vgg16_predictions = []
train_vgg16_actual = []
val_vgg16_predictions = []
val_vgg16_actual = []
test_b_vgg16_predictions = []
test_b_vgg16_actual = []
test_ub_vgg16_predictions = []
test_ub_vgg16_actual = []
print('Evaluating Training...')
for i, data in enumerate(loaders['train'], 0):
dt = data
image = dt['image'].to()
gt = dt['gt']
model.eval()
predictions = model(image[None, ...].float())
predictions = predictions.squeeze().data.cpu().numpy()
count = np.sum(predictions) / 100
train_vgg16_predictions.append(count)
train_vgg16_actual.append(len(gt))
limit = int(len(loaders['val']) / 2)
print('Evaluating Validation...')
for i, data in enumerate(loaders['val'], 0):
if i >= limit: break
dt = data
image = dt['image'].to()
gt = dt['gt']
model.eval()
predictions = model(image[None, ...].float())
predictions = predictions.squeeze().data.cpu().numpy()
count = np.sum(predictions) / 100
val_vgg16_predictions.append(count)
val_vgg16_actual.append(len(gt))
print('Evaluating Testing (Balanced)...')
for i, data in enumerate(loaders['val'], limit):
dt = data
image = dt['image'].to()
gt = dt['gt']
model.eval()
predictions = model(image[None, ...].float())
predictions = predictions.squeeze().data.cpu().numpy()
count = np.sum(predictions) / 100
test_b_vgg16_predictions.append(count)
test_b_vgg16_actual.append(len(gt))
print('Evaluating Testing (Unbalanced)...')
for i, data in enumerate(loaders['test_unbalanced'], 0):
dt = data
image = dt['image'].to()
gt = dt['gt']
model.eval()
predictions = model(image[None, ...].float())
predictions = predictions.squeeze().data.cpu().numpy()
count = np.sum(predictions) / 100
test_ub_vgg16_predictions.append(count)
test_ub_vgg16_actual.append(len(gt))
train_r2 = r2_score(train_vgg16_actual, train_vgg16_predictions)
val_r2 = r2_score(val_vgg16_actual, val_vgg16_predictions)
test_b_r2 = r2_score(test_b_vgg16_actual, test_b_vgg16_predictions)
test_ub_r2 = r2_score(test_ub_vgg16_actual, test_ub_vgg16_predictions)
train_mse = mean_squared_error(train_vgg16_actual, train_vgg16_predictions)
val_mse = mean_squared_error(val_vgg16_actual, val_vgg16_predictions)
test_b_mse = mean_squared_error(test_b_vgg16_actual, test_b_vgg16_predictions)
test_ub_mse = mean_squared_error(test_ub_vgg16_actual, test_ub_vgg16_predictions)
print("{}".format(args.model))
print('================================')
print('Training r2: {}'.format(train_r2))
print('Validation r2: {}'.format(val_r2))
print('Testing (BalanceD) r2: {}'.format(test_b_r2))
print('Testing (Unbalanced) r2: {}'.format(test_ub_r2))
print('================================')
print('Training MSE: {}'.format(train_mse))
print('Validation MSE: {}'.format(val_mse))
print('Testing (Balanced) MSE: {}'.format(test_b_mse))
print('Testing (Unbalanced) MSE: {}'.format(test_ub_mse))
fg, (p1, p2, p3, p4) = plt.subplots(1, 4, figsize=(15, 4))
x = np.linspace(0, max(train_vgg16_actual), 1000)
y = x
p1.plot(x, y, '-r', label='Ground Truths')
p1.scatter(train_vgg16_actual, train_vgg16_predictions, label='Training Data')
p1.legend()
p1.set_title('Training\nMSE: {:.2f}\n r2: {:.2f}'.format(train_mse, train_r2))
x = np.linspace(0, max(val_vgg16_actual), 1000)
y = x
p2.plot(x, y, '-r', label='Ground Truths')
p2.scatter(val_vgg16_actual, val_vgg16_predictions, label='Validation Data')
p2.legend()
p2.set_title('Validation\nMSE: {:.2f}\n r2: {:.2f}'.format(val_mse, val_r2))
x = np.linspace(0, max(test_b_vgg16_actual), 1000)
y = x
p3.plot(x, y, '-r', label='Ground Truths')
p3.scatter(test_b_vgg16_actual, test_b_vgg16_predictions, label='Testing (Balanced) Data')
p3.legend()
p3.set_title('Testing (Balanced)\nMSE: {:.2f}\n r2: {:.2f}'.format(test_b_mse, test_b_r2))
x = np.linspace(0, max(test_ub_vgg16_actual), 1000)
y = x
p4.plot(x, y, '-r', label='Ground Truths')
p4.scatter(test_ub_vgg16_actual, test_ub_vgg16_predictions, label='Testing (Unbalanced) Data')
p4.legend()
p4.set_title('Testing (Unbalanced)\nMSE: {:.2f}\n r2: {:.2f}'.format(test_ub_mse, test_ub_r2))
fg.text(0.04, 0.5, 'Predicted', va='center', rotation='vertical')
fg.text(0.5, 0.04, 'Actual', ha='center')
fg.savefig('results/{}_results'.format(args.model))
if __name__=='__main__':
args = run_argparse()
main(args)