-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmeasured_times.txt
305 lines (259 loc) · 13.7 KB
/
measured_times.txt
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
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
===> Using 'cuda' for computation.
####################################################
# Procesing ResNetModel #
####################################################
Converting ResNetModel to TensorRT FP32...
Converting ResNetModel to TensorRT FP16...
Converting ResNetModel to TensorRT INT8...
Testing inference times over PyTorch FP32 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP32] Inference time mean (713.6965589735244) and std (214.11265622551164) in ms
-Model Throughput (FPS)->1.4094799522292842
-Quality PSNR (24.876483917236328)
Testing inference times over PyTorch FP16 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (316.87693888346354) and std (95.06439351246223) in ms
-Model Throughput (FPS)->3.165328409886266
-Quality PSNR (24.880208333333332)
Testing inference times over TensorRT FP32 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP32] Inference time mean (654.9191507975261) and std (196.47746079143502) in ms
-Model Throughput (FPS)->1.5293166510130296
-Quality PSNR (24.876483917236328)
Testing inference times over TensorRT FP16 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (199.56831648084852) and std (59.87106907273105) in ms
-Model Throughput (FPS)->5.032866652979478
-Quality PSNR (24.87505594889323)
Testing inference times over TensorRT INT8 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (131.79612053765192) and std (39.53926034010163) in ms
-Model Throughput (FPS)->7.635942883937202
-Quality PSNR (11.173668225606283)
####################################################
# Procesing ResNetModel_LateFusion #
####################################################
Converting ResNetModel_LateFusion to TensorRT FP32...
Converting ResNetModel_LateFusion to TensorRT FP16...
Converting ResNetModel_LateFusion to TensorRT INT8...
Testing inference times over PyTorch FP32 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP32] Inference time mean (709.6418721516927) and std (212.89891094626057) in ms
-Model Throughput (FPS)->1.4161138961395785
-Quality PSNR (29.06027348836263)
Testing inference times over PyTorch FP16 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (309.73040534125437) and std (92.92092223219956) in ms
-Model Throughput (FPS)->3.237385289359356
-Quality PSNR (29.0625)
Testing inference times over TensorRT FP32 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP32] Inference time mean (654.3450202094184) and std (196.30445041629125) in ms
-Model Throughput (FPS)->1.532311480109916
-Quality PSNR (29.06027348836263)
Testing inference times over TensorRT FP16 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (198.53113928900825) and std (59.55962444158243) in ms
-Model Throughput (FPS)->5.057825665454412
-Quality PSNR (29.064523696899414)
Testing inference times over TensorRT INT8 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (132.06692911783855) and std (39.6204786973482) in ms
-Model Throughput (FPS)->7.618972329258744
-Quality PSNR (11.676202456156412)
####################################################
# Procesing InceptionModel #
####################################################
Converting InceptionModel to TensorRT FP32...
Converting InceptionModel to TensorRT FP16...
Converting InceptionModel to TensorRT INT8...
Testing inference times over PyTorch FP32 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP32] Inference time mean (206.09207712809246) and std (61.8279550041952) in ms
-Model Throughput (FPS)->4.852217646331363
-Quality PSNR (26.970356623331707)
Testing inference times over PyTorch FP16 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (127.51789923773872) and std (38.25595796476287) in ms
-Model Throughput (FPS)->7.842415447443938
-Quality PSNR (26.973958333333332)
Testing inference times over TensorRT FP32 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP32] Inference time mean (170.04439798990884) and std (51.01354605450997) in ms
-Model Throughput (FPS)->5.913935103256817
-Quality PSNR (26.970356623331707)
Testing inference times over TensorRT FP16 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (67.36133202446831) and std (20.208965646084728) in ms
-Model Throughput (FPS)->15.038388494953578
-Quality PSNR (26.98090934753418)
Testing inference times over TensorRT INT8 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (47.43168178134494) and std (14.230125132923376) in ms
-Model Throughput (FPS)->21.423570185046593
-Quality PSNR (11.28984022140503)
####################################################
# Procesing InceptionModel_LateFusion #
####################################################
Converting InceptionModel_LateFusion to TensorRT FP32...
Converting InceptionModel_LateFusion to TensorRT FP16...
Converting InceptionModel_LateFusion to TensorRT INT8...
Testing inference times over PyTorch FP32 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP32] Inference time mean (205.77322540283203) and std (61.732344449212604) in ms
-Model Throughput (FPS)->4.86435611575846
-Quality PSNR (28.551069895426433)
Testing inference times over PyTorch FP16 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (126.81714664035373) and std (38.04553973950078) in ms
-Model Throughput (FPS)->7.888571638574177
-Quality PSNR (28.557291666666668)
Testing inference times over TensorRT FP32 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP32] Inference time mean (174.0003409491645) and std (52.20025377334641) in ms
-Model Throughput (FPS)->5.774612847430889
-Quality PSNR (28.551069895426433)
Testing inference times over TensorRT FP16 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (67.97459250556098) and std (20.392743201650227) in ms
-Model Throughput (FPS)->14.883968111865956
-Quality PSNR (28.55426279703776)
Testing inference times over TensorRT INT8 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (47.775210783216686) and std (14.333873628704184) in ms
-Model Throughput (FPS)->21.26019302693908
-Quality PSNR (11.3268035252889)
####################################################
# Procesing CNNModel #
####################################################
Converting CNNModel to TensorRT FP32...
Converting CNNModel to TensorRT FP16...
Converting CNNModel to TensorRT INT8...
Testing inference times over PyTorch FP32 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP32] Inference time mean (676.9898613823784) and std (203.09934810863805) in ms
-Model Throughput (FPS)->1.4818924420080957
-Quality PSNR (28.138348897298176)
Testing inference times over PyTorch FP16 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (297.5791741265191) and std (89.27524089772663) in ms
-Model Throughput (FPS)->3.3695695678036075
-Quality PSNR (28.140625)
Testing inference times over TensorRT FP32 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP32] Inference time mean (624.1758117675781) and std (187.25496369845808) in ms
-Model Throughput (FPS)->1.605219756491499
-Quality PSNR (28.1383482615153)
Testing inference times over TensorRT FP16 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (191.39212714301215) and std (57.41810999376537) in ms
-Model Throughput (FPS)->5.246914877540829
-Quality PSNR (28.132497151692707)
Testing inference times over TensorRT INT8 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (0.8212302241060468) and std (0.25563751328355055) in ms
-Model Throughput (FPS)->2301.679215049279
-Quality PSNR (11.548112233479818)
####################################################
# Procesing CNNModel_LateFusion #
####################################################
Converting CNNModel_LateFusion to TensorRT FP32...
Converting CNNModel_LateFusion to TensorRT FP16...
Converting CNNModel_LateFusion to TensorRT INT8...
Testing inference times over PyTorch FP32 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP32] Inference time mean (674.4915452745225) and std (202.3500568856989) in ms
-Model Throughput (FPS)->1.490537630650067
-Quality PSNR (28.88698450724284)
Testing inference times over PyTorch FP16 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (290.5791920979818) and std (87.17528791609847) in ms
-Model Throughput (FPS)->3.4503814673880977
-Quality PSNR (28.875)
Testing inference times over TensorRT FP32 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP32] Inference time mean (623.3835984971788) and std (187.0174458618178) in ms
-Model Throughput (FPS)->1.6072911805672179
-Quality PSNR (28.88698387145996)
Testing inference times over TensorRT FP16 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (190.20902065700955) and std (57.06319201706028) in ms
-Model Throughput (FPS)->5.27953754502481
-Quality PSNR (28.87971369425456)
Testing inference times over TensorRT INT8 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (127.67194951375326) and std (38.302549378732735) in ms
-Model Throughput (FPS)->7.8834847514307285
-Quality PSNR (11.66981569925944)
####################################################
# Procesing AttentionModel #
####################################################
Converting AttentionModel to TensorRT FP32...
[TensorRT] ERROR: 3: [network.cpp::addPoolingNd::884] Error Code 3: Internal Error (Parameter check failed at: optimizer/api/network.cpp::addPoolingNd::884, condition: allDimsGtEq(windowSize, 1) && volume(windowSize) < MAX_KERNEL_DIMS_PRODUCT(nbSpatialDims)
)
######################## EXCEPTION OCURRED ##################################
'NoneType' object has no attribute 'stride_nd'
################################### END EXCEPTION ###############################
Testing inference times over PyTorch FP32 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP32] Inference time mean (557.9050306532118) and std (167.37596809272432) in ms
-Model Throughput (FPS)->1.800494063719683
-Quality PSNR (28.10034116109212)
Testing inference times over PyTorch FP16 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (276.3602077907986) and std (82.90840814791437) in ms
-Model Throughput (FPS)->3.628099272064351
-Quality PSNR (28.104166666666668)
####################################################
# Procesing AttentionModel_LateFusion #
####################################################
Converting AttentionModel_LateFusion to TensorRT FP32...
[TensorRT] ERROR: 3: [network.cpp::addPoolingNd::884] Error Code 3: Internal Error (Parameter check failed at: optimizer/api/network.cpp::addPoolingNd::884, condition: allDimsGtEq(windowSize, 1) && volume(windowSize) < MAX_KERNEL_DIMS_PRODUCT(nbSpatialDims)
)
######################## EXCEPTION OCURRED ##################################
'NoneType' object has no attribute 'stride_nd'
################################### END EXCEPTION ###############################
Testing inference times over PyTorch FP32 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP32] Inference time mean (559.3840549045138) and std (167.8175970904957) in ms
-Model Throughput (FPS)->1.7971972148495867
-Quality PSNR (29.166518529256184)
Testing inference times over PyTorch FP16 ...
-Warming up GPU...
-Measuring times over 100 iterations...
-[MODEL IN FP16] Inference time mean (275.3964569091797) and std (82.61926734019029) in ms
-Model Throughput (FPS)->3.642987085009465
-Quality PSNR (29.171875)