-
Notifications
You must be signed in to change notification settings - Fork 0
/
model_test_onnx.py
248 lines (191 loc) · 9.66 KB
/
model_test_onnx.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
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
import argparse
import json
import time
import os
import cv2
import numpy as np
import onnxruntime
import random
random.seed(100)
colormap = []
for i in range(1000):
colormap.append((random.randint(0, 255), random.randint(0, 255), random.randint(0, 255)))
def test(config:dict):
if config['run_cpu']: #cpu
ep_list = ['CPUExecutionProvider']
else: #gpu
ep_list = ['CUDAExecutionProvider', 'CPUExecutionProvider']
ort_session = onnxruntime.InferenceSession(f"{config['model_dir']}/model.onnx", providers=ep_list)
print(f"ep_list : {ep_list}")
print(f"onnx session : {ort_session.get_providers()}")
w_input = config['network']['input_size'][1]
h_input = config['network']['input_size'][0]
w_output = 600
h_output = 600
line_thickness = 1
text_thickness = 1
text_scale = 0.5
# output
vc_out = None
if config['save_video'] is not None:
vc_out = cv2.VideoWriter(f"{config['model_dir']}/{config['save_video']}", cv2.VideoWriter_fourcc(*'DIVX'), 30, (w_output, h_output))
# cam
if config['cam']:
vc = cv2.VideoCapture(0)
assert vc.isOpened(), '동영상 파일을 열 수 없습니다.'
while True:
ret, img = vc.read()
if not ret:
continue
start = time.time()
img = cv2.resize(img, (w_input, h_input))
img_out = cv2.resize(img, (w_output, h_output)).astype(np.uint8)
img = np.expand_dims(img.transpose(2, 0, 1).astype(np.float32), axis=0)
ort_inputs = {ort_session.get_inputs()[0].name: img}
detections = ort_session.run(None, ort_inputs)
detections = detections[0]
for i in range(len(config['network']['task'])):
if config['network']['task'][i].lower() == 'box2d':
detections = detections[0]
# detections = [batch, class, score, x1, y1, x2, y2]
for detection in detections:
if detection[1] > 0.5:
label = int(detection[0])
score = detection[1]
x1 = int(detection[2] * w_output)
y1 = int(detection[3] * h_output)
x2 = int(detection[4] * w_output)
y2 = int(detection[5] * h_output)
txt = f"{config['network']['classes'][label]} : {score:.2f}"
box_color = colormap[label]
cv2.rectangle(img_out, (x1, y1), (x2, y2), box_color, line_thickness)
cv2.putText(img_out, txt, (x1, y1), cv2.FONT_HERSHEY_SIMPLEX, text_scale, box_color, text_thickness)
end = time.time()
cv2.putText(img_out, f"FPS : {1/(end-start):.2f}", (10, 30), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 255), 1)
if config['show_result']:
cv2.imshow('result', img_out)
key = cv2.waitKey(1)
if key == ord('q'):
break
if config['save_video'] is not None:
vc_out.write(img_out)
# img
if config['img_dir'] is not None:
img_list = os.listdir(config['img_dir'])
img_list.sort()
for img_file in img_list:
start = time.time()
try:
img = cv2.imread(f"{config['img_dir']}/{img_file}")
img = cv2.resize(img, (w_input, h_input))
img_out = cv2.resize(img, (w_output, h_output)).astype(np.uint8)
img = np.expand_dims(img.transpose(2, 0, 1).astype(np.float32), axis=0)
except:
continue
ort_inputs = {ort_session.get_inputs()[0].name: img}
detections = ort_session.run(None, ort_inputs)
detections = detections[0]
for i in range(len(config['network']['task'])):
if config['network']['task'][i].lower() == 'box2d':
detections = detections[0]
# detections = [batch, class, score, x1, y1, x2, y2]
for detection in detections:
if detection[1] > 0.5:
label = int(detection[0])
score = detection[1]
x1 = int(detection[2] * w_output)
y1 = int(detection[3] * h_output)
x2 = int(detection[4] * w_output)
y2 = int(detection[5] * h_output)
txt = f"{config['network']['classes'][label]} : {score:.2f}"
box_color = colormap[label]
cv2.rectangle(img_out, (x1, y1), (x2, y2), box_color, line_thickness)
cv2.putText(img_out, txt, (x1, y1), cv2.FONT_HERSHEY_SIMPLEX, text_scale, box_color, text_thickness)
end = time.time()
cv2.putText(img_out, f"FPS : {1/(end-start):.2f}", (10, 30), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 255), 1)
if config['show_result']:
cv2.imshow('result', img_out)
key = cv2.waitKey(1)
if key == ord('q'):
break
if config['save_video'] is not None:
vc_out.write(img_out)
if config['video'] is not None:
vc = cv2.VideoCapture(config['video'])
assert vc.isOpened(), '동영상 파일을 열 수 없습니다.'
while True:
ret, img = vc.read()
if not ret:
vc.release()
break
start = time.time()
img = cv2.resize(img, (w_input, h_input))
img_out = cv2.resize(img, (w_output, h_output)).astype(np.uint8)
img = np.expand_dims(img.transpose(2, 0, 1).astype(np.float32), axis=0)
ort_inputs = {ort_session.get_inputs()[0].name: img}
detections = ort_session.run(None, ort_inputs)
detections = detections[0]
for i in range(len(config['network']['task'])):
if config['network']['task'][i].lower() == 'box2d':
detections = detections[0]
# detections = [batch, class, score, x1, y1, x2, y2]
for detection in detections:
if detection[1] > 0.5:
label = int(detection[0])
score = detection[1]
x1 = int(detection[2] * w_output)
y1 = int(detection[3] * h_output)
x2 = int(detection[4] * w_output)
y2 = int(detection[5] * h_output)
txt = f"{config['network']['classes'][label]} : {score:.2f}"
box_color = colormap[label]
cv2.rectangle(img_out, (x1, y1), (x2, y2), box_color, line_thickness)
cv2.putText(img_out, txt, (x1, y1), cv2.FONT_HERSHEY_SIMPLEX, text_scale, box_color, text_thickness)
end = time.time()
cv2.putText(img_out, f"FPS : {1/(end-start):.2f}", (10, 30), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 255), 1)
if config['show_result']:
cv2.imshow('result', img_out)
key = cv2.waitKey(1)
if key == ord('q'):
vc.release()
break
if config['save_video'] is not None:
vc_out.write(img_out)
if config['save_video'] is not None:
vc_out.release()
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('--model_dir', type=str, default=None)
# Input
parser.add_argument('--img_dir', default=None, type=str)
parser.add_argument('--video', default=None, type=str)
parser.add_argument('--cam', action='store_true')
# CPU or GPU
parser.add_argument('--run_cpu', action='store_true')
# Save Video (저장할 비디오 이름름)
parser.add_argument('--save_video', default=None, type=str)
# Select model weight
parser.add_argument('--best', action='store_true')
parser.add_argument('--last', action='store_true')
# Label Save (추론 결과를 저장할 것인지)
parser.add_argument('--save_label', action='store_true')
parser.add_argument('--label_dir', default=None, type=str) # label save dir
# Show Result (추론 결과를 화면에 보여줄 것인지)
parser.add_argument('--show_result', action='store_true') # GPU Server에서는 False 권장
opt = parser.parse_args()
# Test
opt.model_dir = 'runs/mobilenetv2_ssd_argococo_crop05_1'
opt.best = True
opt.cam = True
# opt.video = 'D:/construct_02.mp4'
# opt.img_dir = 'D:/Argoseye/Test/CH2'
# opt.img_dir = 'sample'
opt.show_result = True
# opt.save_video = 'test_CH2.mp4'
opt.device = 'gpu'
assert opt.model_dir is not None, '모델 경로를 입력해주세요.'
# read txt to dict
with open(f'{opt.model_dir}/config.json', 'r') as f:
config = json.load(f)
config.update(vars(opt))
test(config)