forked from awslabs/deeplearning-benchmark
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtask_config_template.cfg
executable file
·294 lines (252 loc) · 27.9 KB
/
task_config_template.cfg
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
[metrics_parameters_inference]
patterns = ['Prediction-Time: (\d+\.\d+|\d+) milliseconds']
metrics = ['prediction_time']
compute_method = ['average']
[metrics_parameters_imperative_hybrid_top_1]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'training: accuracy=(\d+\.\d+|\d+)', 'validation: accuracy=(\d+\.\d+|\d+)', 'time cost: (\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'validation_acc', 'total_training_time']
compute_method = ['average', 'last', 'last', 'total']
[metrics_parameters_images_top_1]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)' ]
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc']
compute_method = ['average', 'last', 'total', 'last']
[metrics_parameters_images_top_5]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)', 'Train-top_k_accuracy_5=(\d+\.\d+|\d+)','Validation-top_k_accuracy_5=(\d+\.\d+|\d+)' ]
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc','Train-top_k_accuracy_5','Validation-top_k_accuracy_5']
compute_method = ['average', 'last', 'total', 'last','last','last']
[resnet50_cifar10_symbolic]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)' ]
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc']
compute_method = ['average', 'last', 'total', 'last']
command_to_execute = python image_classification/image_classification.py --model resnet50_v1 --dataset cifar10 --mode symbolic --gpus 8 --epochs 20 --log-interval 50 --kvstore device
num_gpus = 8
[resnet50_cifar10_symbolic_fp16_batch_size64]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc']
compute_method = ['average', 'last', 'total','last']
command_to_execute = python image_classification/image_classification.py --model resnet50_v1 --dataset cifar10 --mode symbolic --gpus 8 --epochs 20 --log-interval 50 --dtype float16 --batch-size 64 --kvstore device
num_gpus = 8
[resnet50_cifar10_symbolic_fp32_batch_size32]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc']
compute_method = ['average', 'last', 'total','last']
command_to_execute = python image_classification/image_classification.py --model resnet50_v1 --dataset cifar10 --mode symbolic --gpus 8 --epochs 20 --log-interval 50 --batch-size 32 --kvstore device
num_gpus = 8
[resnet50_cifar10_symbolic_fp16_batch_size32]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc']
compute_method = ['average', 'last', 'total','last']
command_to_execute = python image_classification/image_classification.py --model resnet50_v1 --dataset cifar10 --mode symbolic --gpus 8 --epochs 20 --log-interval 50 --dtype float16 --batch-size 32 --kvstore device
num_gpus = 8
[resnet50_cifar10_symbolic_fp32_batch_size16]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc']
compute_method = ['average', 'last', 'total','last']
command_to_execute = python image_classification/image_classification.py --model resnet50_v1 --dataset cifar10 --mode symbolic --gpus 8 --epochs 20 --log-interval 50 --batch-size 16 --kvstore device
num_gpus = 8
[resnet50_cifar10_symbolic_fp32_batch_size64]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc']
compute_method = ['average', 'last', 'total','last']
command_to_execute = python image_classification/image_classification.py --model resnet50_v1 --dataset cifar10 --mode symbolic --gpus 8 --epochs 20 --log-interval 50 --batch-size 64 --kvstore device
num_gpus = 8
[resnet50_cifar10_hybrid]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'training: accuracy=(\d+\.\d+|\d+)', 'validation: accuracy=(\d+\.\d+|\d+)', 'time cost: (\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'validation_acc', 'total_training_time']
compute_method = ['average', 'last', 'last', 'total']
command_to_execute = python image_classification/image_classification.py --model resnet50_v1 --dataset cifar10 --mode hybrid --gpus 8 --epochs 20 --log-interval 50 --kvstore device
num_gpus = 8
[resnet50_cifar10_imperative]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'training: accuracy=(\d+\.\d+|\d+)', 'validation: accuracy=(\d+\.\d+|\d+)', 'time cost: (\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'validation_acc', 'total_training_time']
compute_method = ['average', 'last', 'last', 'total']
command_to_execute = python image_classification/image_classification.py --model resnet50_v1 --dataset cifar10 --gpus 8 --epochs 20 --log-interval 50 --kvstore device
num_gpus = 8
[dawnbench_cifar10_symbolic]
patterns = ['Epoch \d+, Batch \d+, Speed=(\d+\.\d+|\d+)', 'Epoch \d+, Training accuracy=(\d+\.\d+|\d+)', 'Epoch \d+, Validation accuracy=(\d+\.\d+|\d+)', 'Epoch \d+, Duration=(\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'validation_acc', 'total_training_time']
compute_method = ['average', 'last', 'last', 'total']
command_to_execute = python dawnbench/cifar10.py --gpus 4 --early-stopping-acc 0.86 --epochs 400 --lr 0.05 --total-batch-size 256
num_gpus = 4
[lstm_ptb_imperative]
patterns = ['time cost (\d+\.\d+|\d+)', 'valid loss (\d+\.\d+|\d+)', 'valid ppl (\d+\.\d+|\d+)', 'test loss (\d+\.\d+|\d+)', 'test ppl (\d+\.\d+|\d+)']
metrics = ['total_training_time', 'validation_loss', 'validation_perplexity', 'test_loss', 'test_perplexity']
compute_method = ['total', 'last', 'last', 'last', 'last']
command_to_execute = python word_language_model/word_language_model.py --gpus 8 --nhid 650 --emsize 650 --dropout 0.5 --epochs 40 --data word_language_model/data/ptb. --mode imperative --kvstore device
num_gpus = 8
[lstm_ptb_hybrid]
patterns = ['time cost (\d+\.\d+|\d+)', 'valid loss (\d+\.\d+|\d+)', 'valid ppl (\d+\.\d+|\d+)', 'test loss (\d+\.\d+|\d+)', 'test ppl (\d+\.\d+|\d+)']
metrics = ['total_training_time', 'validation_loss', 'validation_perplexity', 'test_loss', 'test_perplexity']
compute_method = ['total', 'last', 'last', 'last', 'last']
command_to_execute = python word_language_model/word_language_model.py --gpus 8 --nhid 650 --emsize 650 --dropout 0.5 --epochs 40 --data word_language_model/data/ptb. --mode hybrid --kvstore device
num_gpus = 8
[lstm_ptb_symbolic]
patterns = ['Time cost=(\d+\.\d+|\d+)', 'Train-perplexity=(\d+\.\d+|\d+)', 'Validation-perplexity=(\d+\.\d+|\d+)', 'Speed: (\d+\.\d+|\d+) samples/sec']
metrics = ['total_training_time', 'train_perplexity', 'validation_perplexity', 'speed']
compute_method = ['total', 'last', 'last', 'average']
command_to_execute = python word_language_model/lstm_bucketing.py --num-hidden 650 --num-embed 650 --gpus 8 --epochs 25 --kv-store device
num_gpus = 8
[resnet50_imagenet_symbolic_fp16_batch_size32_p3_16]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc']
compute_method = ['average', 'last', 'total','last']
command_to_execute = python /home/ubuntu/mxnet/example/image-classification/train_imagenet.py --data-train /home/ubuntu/imagenet/imagenet1k-train.rec --data-val /home/ubuntu/imagenet/imagenet1k-val.rec --gpus 1,0,2,3,4,5,6,7 --batch-size 256 --data-nthreads 15 --num-epochs 80 --dtype float16
num_gpus = 1
[resnet50_imagenet_symbolic_fp16_batch_size64_p3_16]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc']
compute_method = ['average', 'last', 'total','last']
command_to_execute = python /home/ubuntu/mxnet/example/image-classification/train_imagenet.py --data-train /home/ubuntu/imagenet/imagenet1k-train.rec --data-val /home/ubuntu/imagenet/imagenet1k-val.rec --gpus 1,0,2,3,4,5,6,7 --batch-size 512 --data-nthreads 15 --num-epochs 80 --dtype float16
num_gpus = 1
[resnet50_imagenet_symbolic_fp16_batch_size128_p3_16]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc']
compute_method = ['average', 'last', 'total','last']
command_to_execute = python /home/ubuntu/mxnet/example/image-classification/train_imagenet.py --data-train /home/ubuntu/imagenet/imagenet1k-train.rec --data-val /home/ubuntu/imagenet/imagenet1k-val.rec --gpus 1,0,2,3,4,5,6,7 --batch-size 1024 --data-nthreads 32 --num-epochs 80 --dtype float16
num_gpus = 1
[resnet50_imagenet_symbolic_fp32_batch_size32_p3_16]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc']
compute_method = ['average', 'last', 'total','last']
command_to_execute = python /home/ubuntu/mxnet/example/image-classification/train_imagenet.py --data-train /home/ubuntu/imagenet/imagenet1k-train.rec --data-val /home/ubuntu/imagenet/imagenet1k-val.rec --gpus 1,0,2,3,4,5,6,7 --batch-size 256 --data-nthreads 15 --num-epochs 80
num_gpus = 1
[resnet50_imagenet_symbolic_fp32_batch_size64_p3_16]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc']
compute_method = ['average', 'last', 'total','last']
command_to_execute = python /home/ubuntu/mxnet/example/image-classification/train_imagenet.py --data-train /home/ubuntu/imagenet/imagenet1k-train.rec --data-val /home/ubuntu/imagenet/imagenet1k-val.rec --gpus 1,0,2,3,4,5,6,7 --batch-size 512 --data-nthreads 15 --num-epochs 80
num_gpus = 1
[resnet50_imagenet_symbolic_fp16_batch_size32_p3_8]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc']
compute_method = ['average', 'last', 'total','last']
command_to_execute = python /home/ubuntu/mxnet/example/image-classification/train_imagenet.py --data-train /home/ubuntu/imagenet/imagenet1k-train.rec --data-val /home/ubuntu/imagenet/imagenet1k-val.rec --gpus 1,0,2,3 --batch-size 256 --data-nthreads 15 --num-epochs 80 --dtype float16
num_gpus = 1
[resnet50_imagenet_symbolic_fp16_batch_size64_p3_8]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc']
compute_method = ['average', 'last', 'total','last']
command_to_execute = python /home/ubuntu/mxnet/example/image-classification/train_imagenet.py --data-train /home/ubuntu/imagenet/imagenet1k-train.rec --data-val /home/ubuntu/imagenet/imagenet1k-val.rec --gpus 1,0,2,3 --batch-size 512 --data-nthreads 15 --num-epochs 80 --dtype float16
num_gpus = 1
[resnet50_imagenet_symbolic_fp32_batch_size32_p3_8]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc']
compute_method = ['average', 'last', 'total','last']
command_to_execute = python /home/ubuntu/mxnet/example/image-classification/train_imagenet.py --data-train /home/ubuntu/imagenet/imagenet1k-train.rec --data-val /home/ubuntu/imagenet/imagenet1k-val.rec --gpus 1,0,2,3 --batch-size 256 --data-nthreads 15 --num-epochs 80
num_gpus = 1
[resnet50_imagenet_symbolic_fp32_batch_size64_p3_8]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc']
compute_method = ['average', 'last', 'total','last']
command_to_execute = python /home/ubuntu/mxnet/example/image-classification/train_imagenet.py --data-train /home/ubuntu/imagenet/imagenet1k-train.rec --data-val /home/ubuntu/imagenet/imagenet1k-val.rec --gpus 1,0,2,3 --batch-size 512 --data-nthreads 15 --num-epochs 80
num_gpus = 1
[metrics_parameters_distributed_top_k]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Train-top_k_accuracy_\d=(\d+\.\d+|\d+)','Validation-top_k_accuracy_\d=(\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'validation_acc', 'total_training_time','training_acc_top5','validation_acc_top5']
compute_method = ['average_aggregate', 'last', 'last', 'total','last','last']
[metrics_parameters_distributed]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)']
metrics = ['speed', 'training_acc', 'validation_acc', 'total_training_time']
compute_method = ['average_aggregate', 'last', 'last', 'total']
[tensorflow_resnet50_p3_2xlg]
patterns = ['images/sec: (\d+\.\d+)', 'time: (\d+\.\d+)']
metrics = ['Images per sec', 'Time']
compute_method = ['average', 'last']
command_to_execute = python tensorflow_benchmark/tf_cnn_benchmarks/tf_cnn_benchmarks.py --data_name=imagenet --num_gpus=1 --batch_size=32 --model=resnet50 --variable_update=parameter_server --print_training_accuracy=True --num_batches=100
num_gpus = 1
[tensorflow_resnet50_p3_8xlg]
patterns = ['images/sec: (\d+\.\d+)', 'time: (\d+\.\d+)']
metrics = ['Images per sec', 'Time']
compute_method = ['average', 'last']
command_to_execute = python tensorflow_benchmark/tf_cnn_benchmarks/tf_cnn_benchmarks.py --data_name=imagenet --num_gpus=4 --batch_size=32 --model=resnet50 --variable_update=parameter_server --print_training_accuracy=True --num_batches=100
num_gpus = 4
[tensorflow_resnet50_p3_16xlg]
patterns = ['images/sec: (\d+\.\d+)', 'time: (\d+\.\d+)']
metrics = ['Images per sec', 'Time']
compute_method = ['average', 'last']
command_to_execute = python tensorflow_benchmark/tf_cnn_benchmarks/tf_cnn_benchmarks.py --data_name=imagenet --num_gpus=8 --batch_size=32 --model=resnet50 --variable_update=parameter_server --print_training_accuracy=True --num_batches=100
num_gpus = 8
[tensorflow_resnet152_p3_2xlg]
patterns = ['images/sec: (\d+\.\d+)', 'time: (\d+\.\d+)']
metrics = ['Images per sec', 'Training']
compute_method = ['average', 'last']
command_to_execute = python tensorflow_benchmark/tf_cnn_benchmarks/tf_cnn_benchmarks.py --data_name=imagenet --num_gpus=1 --batch_size=32 --model=resnet152 --variable_update=parameter_server --print_training_accuracy=True --num_batches=100
num_gpus = 1
[tensorflow_resnet152_p3_8xlg]
patterns = ['images/sec: (\d+\.\d+)', 'time: (\d+\.\d+)']
metrics = ['Images per sec', 'Training']
compute_method = ['average', 'last']
command_to_execute = python tensorflow_benchmark/tf_cnn_benchmarks/tf_cnn_benchmarks.py --data_name=imagenet --num_gpus=4 --batch_size=32 --model=resnet152 --variable_update=parameter_server --print_training_accuracy=True --num_batches=100
num_gpus = 4
[tensorflow_resnet152_p3_16xlg]
patterns = ['images/sec: (\d+\.\d+)', 'time: (\d+\.\d+)']
metrics = ['Images per sec', 'Training time']
compute_method = ['average', 'last']
command_to_execute = python tensorflow_benchmark/tf_cnn_benchmarks/tf_cnn_benchmarks.py --data_name=imagenet --num_gpus=8 --batch_size=32 --model=resnet152 --variable_update=parameter_server --print_training_accuracy=True --num_batches=100
num_gpus = 8
[tensorflow_resnet56_p3_2xlg_fp16]
patterns = ['images/sec: (\d+\.\d+)', 'time: (\d+\.\d+)']
metrics = ['Images per sec', 'Training time']
compute_method = ['average', 'last']
command_to_execute = python tensorflow_benchmark/tf_cnn_benchmarks/tf_cnn_benchmarks.py --data_name=cifar10 --data_dir=cifar-10-batches-py --num_gpus=1 --batch_size=32 --model=resnet56 --variable_update=replicated --print_training_accuracy=True --use_fp16=True --use_tf_layers=False --target_accuracy=0.94
num_gpus = 1
[tensorflow_resnet56_p3_8xlg_fp16]
patterns = ['images/sec: (\d+\.\d+)', 'time: (\d+\.\d+)']
metrics = ['Images per sec', 'Training time']
compute_method = ['average', 'last']
command_to_execute = python tensorflow_benchmark/tf_cnn_benchmarks/tf_cnn_benchmarks.py --data_name=cifar10 --data_dir=cifar-10-batches-py --num_gpus=4 --batch_size=32 --model=resnet56 --variable_update=replicated --print_training_accuracy=True --use_fp16=True --use_tf_layers=False --target_accuracy=0.94
num_gpus = 4
[tensorflow_resnet56_p3_16xlg_fp16]
patterns = ['images/sec: (\d+\.\d+)', 'time: (\d+\.\d+)']
metrics = ['Images per sec', 'Training time']
compute_method = ['average', 'last']
command_to_execute = python tensorflow_benchmark/tf_cnn_benchmarks/tf_cnn_benchmarks.py --data_name=cifar10 --data_dir=cifar-10-batches-py --num_gpus=8 --batch_size=32 --model=resnet56 --variable_update=replicated --print_training_accuracy=True --use_fp16=True --use_tf_layers=False --target_accuracy=0.94
num_gpus = 8
[tensorflow_resnet50_p3_2xlg_fp16]
patterns = ['images/sec: (\d+\.\d+)', 'time: (\d+\.\d+)']
metrics = ['Images per sec', 'Time']
compute_method = ['average', 'last']
command_to_execute = python tensorflow_benchmark/tf_cnn_benchmarks/tf_cnn_benchmarks.py --data_name=imagenet --num_gpus=1 --batch_size=32 --model=resnet50 --variable_update=replicated --print_training_accuracy=True --num_batches=100 --use_fp16=True --use_tf_layers=False
num_gpus = 1
[tensorflow_resnet50_p3_8xlg_fp16]
patterns = ['images/sec: (\d+\.\d+)', 'time: (\d+\.\d+)']
metrics = ['Images per sec', 'Time']
compute_method = ['average', 'last']
command_to_execute = python tensorflow_benchmark/tf_cnn_benchmarks/tf_cnn_benchmarks.py --data_name=imagenet --num_gpus=4 --batch_size=32 --model=resnet50 --variable_update=replicated --print_training_accuracy=True --num_batches=100 --use_fp16=True --use_tf_layers=False
num_gpus = 4
[tensorflow_resnet50_p3_16xlg_fp16]
patterns = ['images/sec: (\d+\.\d+)', 'time: (\d+\.\d+)']
metrics = ['Images per sec', 'Time']
compute_method = ['average', 'last']
command_to_execute = python tensorflow_benchmark/tf_cnn_benchmarks/tf_cnn_benchmarks.py --data_name=imagenet --num_gpus=8 --batch_size=32 --model=resnet50 --variable_update=replicated --print_training_accuracy=True --num_batches=100 --use_fp16=True --use_tf_layers=False
num_gpus = 8
[mms_resnet18_cpu]
patterns = ['Throughput_concurrency_50_req_25 :(\d+\.\d+|\d+)','Average_latency_concurrency_50_req_25 :(\d+\.\d+|\d+)','Median_latency_concurrency_50_req_25 :(\d+\.\d+|\d+)','P90_latency_concurrency_50_req_25 :(\d+\.\d+|\d+)','Error_rate_concurrency_50_req_25 :(\d+\.\d+|\d+)','Throughput_concurrency_50_req_50 :(\d+\.\d+|\d+)','Average_latency_concurrency_50_req_50 :(\d+\.\d+|\d+)','Median_latency_concurrency_50_req_50 :(\d+\.\d+|\d+)','P90_latency_concurrency_50_req_50 :(\d+\.\d+|\d+)','Error_rate_concurrency_50_req_50 :(\d+\.\d+|\d+)','Throughput_concurrency_50_req_100 :(\d+\.\d+|\d+)','Average_latency_concurrency_50_req_100 :(\d+\.\d+|\d+)','Median_latency_concurrency_50_req_100 :(\d+\.\d+|\d+)','P90_latency_concurrency_50_req_100 :(\d+\.\d+|\d+)','Error_rate_concurrency_50_req_100 :(\d+\.\d+|\d+)','Throughput_concurrency_100_req_25 :(\d+\.\d+|\d+)','Average_latency_concurrency_100_req_25 :(\d+\.\d+|\d+)','Median_latency_concurrency_100_req_25 :(\d+\.\d+|\d+)','P90_latency_concurrency_100_req_25 :(\d+\.\d+|\d+)','Error_rate_concurrency_100_req_25 :(\d+\.\d+|\d+)', 'Throughput_concurrency_100_req_50 :(\d+\.\d+|\d+)','Average_latency_concurrency_100_req_50 :(\d+\.\d+|\d+)','Median_latency_concurrency_100_req_50 :(\d+\.\d+|\d+)','P90_latency_concurrency_100_req_50 :(\d+\.\d+|\d+)','Error_rate_concurrency_100_req_50 :(\d+\.\d+|\d+)', 'Throughput_concurrency_100_req_100 :(\d+\.\d+|\d+)','Average_latency_concurrency_100_req_100 :(\d+\.\d+|\d+)','Median_latency_concurrency_100_req_100 :(\d+\.\d+|\d+)','P90_latency_concurrency_100_req_100 :(\d+\.\d+|\d+)','Error_rate_concurrency_100_req_100 :(\d+\.\d+|\d+)' ]
metrics = ['Throughput_concurrency_50_req_25','Average_latency_concurrency_50_req_25','Median_latency_concurrency_50_req_25','P90_latency_concurrency_50_req_25','Error_rate_concurrency_50_req_25','Throughput_concurrency_50_req_50','Average_latency_concurrency_50_req_50','Median_latency_concurrency_50_req_50','P90_latency_concurrency_50_req_50','Error_rate_concurrency_50_req_50','Throughput_concurrency_50_req_100','Average_latency_concurrency_50_req_100','Median_latency_concurrency_50_req_100','P90_latency_concurrency_50_req_100','Error_rate_concurrency_50_req_100','Throughput_concurrency_100_req_25','Average_latency_concurrency_100_req_25','Median_latency_concurrency_100_req_25','P90_latency_concurrency_100_req_25','Error_rate_concurrency_100_req_25','Throughput_concurrency_100_req_50','Average_latency_concurrency_100_req_50','Median_latency_concurrency_100_req_50','P90_latency_concurrency_100_req_50','Error_rate_concurrency_100_req_50','Throughput_concurrency_100_req_100','Average_latency_concurrency_100_req_100','Median_latency_concurrency_100_req_100','P90_latency_concurrency_100_req_100','Error_rate_concurrency_100_req_100']
compute_method = ['last', 'last', 'last', 'last', 'last','last', 'last', 'last', 'last', 'last','last', 'last', 'last', 'last', 'last', 'last', 'last', 'last', 'last', 'last', 'last', 'last', 'last', 'last', 'last','last', 'last', 'last', 'last', 'last']
command_to_execute =sudo bash /mxnet-model-server/ci/benchmark_ai.sh -g 0
num_gpus = 0
[mms_resnet18_gpu]
patterns = ['Throughput_concurrency_50_req_25 :(\d+\.\d+|\d+)', 'Average_latency_concurrency_50_req_25 :(\d+\.\d+|\d+)','Median_latency_concurrency_50_req_25 :(\d+\.\d+|\d+)','P90_latency_concurrency_50_req_25 :(\d+\.\d+|\d+)','Error_rate_concurrency_50_req_25 :(\d+\.\d+|\d+)','Throughput_concurrency_50_req_50 :(\d+\.\d+|\d+)','Average_latency_concurrency_50_req_50 :(\d+\.\d+|\d+)','Median_latency_concurrency_50_req_50 :(\d+\.\d+|\d+)','P90_latency_concurrency_50_req_50 :(\d+\.\d+|\d+)','Error_rate_concurrency_50_req_50 :(\d+\.\d+|\d+)','Throughput_concurrency_50_req_100 :(\d+\.\d+|\d+)','Average_latency_concurrency_50_req_100 :(\d+\.\d+|\d+)','Median_latency_concurrency_50_req_100 :(\d+\.\d+|\d+)','P90_latency_concurrency_50_req_100 :(\d+\.\d+|\d+)','Error_rate_concurrency_50_req_100 :(\d+\.\d+|\d+)','Throughput_concurrency_100_req_25 :(\d+\.\d+|\d+)','Average_latency_concurrency_100_req_25 :(\d+\.\d+|\d+)','Median_latency_concurrency_100_req_25 :(\d+\.\d+|\d+)','P90_latency_concurrency_100_req_25 :(\d+\.\d+|\d+)','Error_rate_concurrency_100_req_25 :(\d+\.\d+|\d+)','Throughput_concurrency_100_req_50 :(\d+\.\d+|\d+)','Average_latency_concurrency_100_req_50 :(\d+\.\d+|\d+)','Median_latency_concurrency_100_req_50 :(\d+\.\d+|\d+)','P90_latency_concurrency_100_req_50 :(\d+\.\d+|\d+)','Error_rate_concurrency_100_req_50 :(\d+\.\d+|\d+)','Throughput_concurrency_100_req_100 :(\d+\.\d+|\d+)','Average_latency_concurrency_100_req_100 :(\d+\.\d+|\d+)','Median_latency_concurrency_100_req_100 :(\d+\.\d+|\d+)','P90_latency_concurrency_100_req_100 :(\d+\.\d+|\d+)', 'Error_rate_concurrency_100_req_100 :(\d+\.\d+|\d+)' ]
metrics = ['Throughput_concurrency_50_req_25','Average_latency_concurrency_50_req_25','Median_latency_concurrency_50_req_25','P90_latency_concurrency_50_req_25','Error_rate_concurrency_50_req_25','Throughput_concurrency_50_req_50','Average_latency_concurrency_50_req_50','Median_latency_concurrency_50_req_50','P90_latency_concurrency_50_req_50','Error_rate_concurrency_50_req_50','Throughput_concurrency_50_req_100','Average_latency_concurrency_50_req_100','Median_latency_concurrency_50_req_100','P90_latency_concurrency_50_req_100','Error_rate_concurrency_50_req_100','Throughput_concurrency_100_req_25','Average_latency_concurrency_100_req_25','Median_latency_concurrency_100_req_25','P90_latency_concurrency_100_req_25','Error_rate_concurrency_100_req_25','Throughput_concurrency_100_req_50','Average_latency_concurrency_100_req_50','Median_latency_concurrency_100_req_50','P90_latency_concurrency_100_req_50','Error_rate_concurrency_100_req_50','Throughput_concurrency_100_req_100','Average_latency_concurrency_100_req_100','Median_latency_concurrency_100_req_100','P90_latency_concurrency_100_req_100','Error_rate_concurrency_100_req_100']
compute_method = ['last', 'last', 'last', 'last', 'last','last', 'last', 'last', 'last', 'last','last', 'last', 'last', 'last', 'last','last', 'last', 'last', 'last', 'last', 'last', 'last', 'last', 'last', 'last','last', 'last', 'last', 'last', 'last']
command_to_execute =sudo bash /mxnet-model-server/ci/benchmark_ai.sh -g 1
num_gpus = 4
[onnx_mxnet_import_model_inference_test_cpu]
patterns = ['Average_inference_time_bvlc_alexnet_cpu: (\d+\.\d+|\d+)', 'Average_inference_time_bvlc_googlenet_cpu: (\d+\.\d+|\d+)', 'Average_inference_time_bvlc_reference_caffenet_cpu: (\d+\.\d+|\d+)', 'Average_inference_time_bvlc_reference_rcnn_ilsvrc13_cpu: (\d+\.\d+|\d+)', 'Average_inference_time_densenet121_cpu: (\d+\.\d+|\d+)', 'Average_inference_time_resnet50_cpu: (\d+\.\d+|\d+)', 'Average_inference_time_shufflenet_cpu: (\d+\.\d+|\d+)', 'Average_inference_time_squeezenet_cpu: (\d+\.\d+|\d+)', 'Average_inference_time_vgg19_cpu: (\d+\.\d+|\d+)']
metrics = ['Average_inference_time_bvlc_alexnet_cpu','Average_inference_time_bvlc_googlenet_cpu','Average_inference_time_bvlc_reference_caffenet_cpu','Average_inference_time_bvlc_reference_rcnn_ilsvrc13_cpu','Average_inference_time_densenet121_cpu','Average_inference_time_resnet50_cpu','Average_inference_time_shufflenet_cpu','Average_inference_time_squeezenet_cpu','Average_inference_time_vgg19_cpu']
compute_method = ['last', 'last', 'last', 'last', 'last', 'last', 'last', 'last', 'last']
command_to_execute = sudo bash ./onnx_benchmark/setup.sh "cpu" && python3 ./onnx_benchmark/import_benchmarkscript.py "cpu"
num_gpus = 1
[onnx_mxnet_import_model_inference_test_gpu]
patterns = ['Average_inference_time_bvlc_alexnet_gpu: (\d+\.\d+|\d+)', 'Average_inference_time_bvlc_googlenet_gpu: (\d+\.\d+|\d+)', 'Average_inference_time_bvlc_reference_caffenet_gpu: (\d+\.\d+|\d+)', 'Average_inference_time_bvlc_reference_rcnn_ilsvrc13_gpu: (\d+\.\d+|\d+)', 'Average_inference_time_densenet121_gpu: (\d+\.\d+|\d+)', 'Average_inference_time_resnet50_gpu: (\d+\.\d+|\d+)', 'Average_inference_time_shufflenet_gpu: (\d+\.\d+|\d+)', 'Average_inference_time_squeezenet_gpu: (\d+\.\d+|\d+)', 'Average_inference_time_vgg19_gpu: (\d+\.\d+|\d+)']
metrics =['Average_inference_time_bvlc_alexnet_gpu','Average_inference_time_bvlc_googlenet_gpu','Average_inference_time_bvlc_reference_caffenet_gpu','Average_inference_time_bvlc_reference_rcnn_ilsvrc13_gpu','Average_inference_time_densenet121_gpu','Average_inference_time_resnet50_gpu','Average_inference_time_shufflenet_gpu','Average_inference_time_squeezenet_gpu','Average_inference_time_vgg19_gpu']
compute_method = ['last', 'last', 'last', 'last', 'last', 'last', 'last', 'last', 'last']
command_to_execute = sudo bash ./onnx_benchmark/setup.sh "gpu" && python3 ./onnx_benchmark/import_benchmarkscript.py "gpu"
num_gpus = 1
[test_resnet50_imagenet-480px-256px-q95_p3.16x_fp16_docker]
patterns = ['Speed: (\d+\.\d+|\d+) samples/sec', 'Train-accuracy=(\d+\.\d+|\d+)', 'Time cost=(\d+\.\d+|\d+)', 'Validation-accuracy=(\d+\.\d+|\d+)' ]
metrics = ['speed', 'training_acc', 'total_training_time', 'validation_acc']
compute_method = ['average', 'last', 'total', 'last']
command_to_execute = bash image_classification/scripts/test.sh
num_gpus = 8