测试结果

目前在少量计算图上进行了参数的网格测试。

测试的硬件环境

全部测试在华为云提供的通用计算增强型ECS上运行(c6.8xlarge.2|32vCPUs|64GB)

测试结果

以下为部分测试结果，受限于时间，没有进行进一步的数据分析。

计算图	详细的子图同构检查	命名空间边界	节点数上限	节点数下限	子图模式最小数量	运行时间（秒）	原计算图大小	子图模式个数	子图平均个数	子图模式平均大小	MDL	压缩比
resnet_ms_output.pb	√	√	24	3	2	0.062241554	648	1	2	3	1.001545595	0.00154321
alexnets_output.pb	√	√	24	3	2	0.007840157	68	0	0	0	1	0
finetune_ms_output.pb	√	√	24	3	2	1.728209734	17531	15	42.73333333	4	1.114423749	0.102675261
lenet_custom_ms_output.pb	√	√	24	3	2	0.007078648	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	24	3	2	0.306711912	3161	5	12.2	3.6	1.05436958	0.05156596
bert_pretrain_ms_output_0train.pb	√	√	24	3	2	4.51624465	41694	17	88	4.176470588	1.129612571	0.11474073
resnet_ms_output.pb	√	√	24	3	3	0.060488701	648	0	0	0	1	0
alexnets_output.pb	√	√	24	3	3	0.007082462	68	0	0	0	1	0
finetune_ms_output.pb	√	√	24	3	3	1.728914022	17531	10	63.1	4.2	1.113432836	0.101876676
lenet_custom_ms_output.pb	√	√	24	3	3	0.006036997	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	24	3	3	0.30915308	3161	4	14.75	3.75	1.053666667	0.050933249
bert_pretrain_ms_output_0train.pb	√	√	24	3	3	4.483399153	41694	14	106.5	4	1.129214853	0.114428935
resnet_ms_output.pb	√	√	24	3	4	0.060696602	648	0	0	0	1	0
alexnets_output.pb	√	√	24	3	4	0.006472111	68	0	0	0	1	0
finetune_ms_output.pb	√	√	24	3	4	1.712932348	17531	9	69.77777778	4.111111111	1.11286739	0.101420341
lenet_custom_ms_output.pb	√	√	24	3	4	0.00535655	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	24	3	4	0.290920019	3161	3	18.66666667	3.333333333	1.050166113	0.047769693
bert_pretrain_ms_output_0train.pb	√	√	24	3	4	4.48350215	41694	10	147.9	4.1	1.128389716	0.113781359
resnet_ms_output.pb	√	√	24	4	2	0.062403202	648	0	0	0	1	0
alexnets_output.pb	√	√	24	4	2	0.007834673	68	0	0	0	1	0
finetune_ms_output.pb	√	√	24	4	2	1.727493048	17531	7	28.28571429	5.142857143	1.047001911	0.044891906
lenet_custom_ms_output.pb	√	√	24	4	2	0.007256269	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	24	4	2	0.302264452	3161	2	4	4.5	1.007650622	0.007592534
bert_pretrain_ms_output_0train.pb	√	√	24	4	2	4.526198387	41694	10	62.7	5	1.063595316	0.059792776
resnet_ms_output.pb	√	√	24	4	3	0.06093955	648	0	0	0	1	0
alexnets_output.pb	√	√	24	4	3	0.007035494	68	0	0	0	1	0
finetune_ms_output.pb	√	√	24	4	3	1.722250462	17531	6	32.66666667	5	1.046626866	0.044549655
lenet_custom_ms_output.pb	√	√	24	4	3	0.005998135	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	24	4	3	0.299001455	3161	2	4	4.5	1.007650622	0.007592534
bert_pretrain_ms_output_0train.pb	√	√	24	4	3	4.467306614	41694	8	77.875	4.75	1.063269834	0.059504965
resnet_ms_output.pb	√	√	24	4	4	0.060094595	648	0	0	0	1	0
alexnets_output.pb	√	√	24	4	4	0.006175518	68	0	0	0	1	0
finetune_ms_output.pb	√	√	24	4	4	1.693614006	17531	5	38.6	5	1.046127223	0.04409332
lenet_custom_ms_output.pb	√	√	24	4	4	0.005753279	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	24	4	4	0.296646118	3161	1	5	4	1.004448681	0.004428978
bert_pretrain_ms_output_0train.pb	√	√	24	4	4	4.49622345	41694	6	102.8333333	4.833333333	1.062809075	0.059097232
resnet_ms_output.pb	√	√	24	5	2	0.062052727	648	0	0	0	1	0
alexnets_output.pb	√	√	24	5	2	0.007708549	68	0	0	0	1	0
finetune_ms_output.pb	√	√	24	5	2	1.729995966	17531	6	25	5.333333333	1.035376801	0.034168045
lenet_custom_ms_output.pb	√	√	24	5	2	0.007137299	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	24	5	2	0.301678896	3161	1	3	5	1.003173596	0.003163556
bert_pretrain_ms_output_0train.pb	√	√	24	5	2	4.465279579	41694	7	22	5.428571429	1.01492174	0.014702355
resnet_ms_output.pb	√	√	24	5	3	0.060448408	648	0	0	0	1	0
alexnets_output.pb	√	√	24	5	3	0.00707531	68	0	0	0	1	0
finetune_ms_output.pb	√	√	24	5	3	1.711164713	17531	5	29.6	5.2	1.035010037	0.033825794
lenet_custom_ms_output.pb	√	√	24	5	3	0.006061792	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	24	5	3	0.309580564	3161	1	3	5	1.003173596	0.003163556
bert_pretrain_ms_output_0train.pb	√	√	24	5	3	4.505323172	41694	5	30	5.2	1.014625362	0.014414544
resnet_ms_output.pb	√	√	24	5	4	0.061283827	648	0	0	0	1	0
alexnets_output.pb	√	√	24	5	4	0.006326437	68	0	0	0	1	0
finetune_ms_output.pb	√	√	24	5	4	1.692445755	17531	4	36.25	5.25	1.034521421	0.03336946
lenet_custom_ms_output.pb	√	√	24	5	4	0.005301714	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	24	5	4	0.297692776	3161	0	0	0	1	0
bert_pretrain_ms_output_0train.pb	√	√	24	5	4	4.490851164	41694	4	36.75	5.25	1.014403192	0.014198686
resnet_ms_output.pb	√	√	36	3	2	0.062241793	648	1	2	3	1.001545595	0.00154321
alexnets_output.pb	√	√	36	3	2	0.007721424	68	0	0	0	1	0
finetune_ms_output.pb	√	√	36	3	2	1.767173052	17531	15	42.73333333	4	1.114423749	0.102675261
lenet_custom_ms_output.pb	√	√	36	3	2	0.007422209	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	36	3	2	0.311676025	3161	5	12.2	3.6	1.05436958	0.05156596
bert_pretrain_ms_output_0train.pb	√	√	36	3	2	4.527644157	41694	17	88	4.176470588	1.129612571	0.11474073
resnet_ms_output.pb	√	√	36	3	3	0.061226845	648	0	0	0	1	0
alexnets_output.pb	√	√	36	3	3	0.00723505	68	0	0	0	1	0
finetune_ms_output.pb	√	√	36	3	3	1.763282299	17531	10	63.1	4.2	1.113432836	0.101876676
lenet_custom_ms_output.pb	√	√	36	3	3	0.00603199	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	36	3	3	0.30532074	3161	4	14.75	3.75	1.053666667	0.050933249
bert_pretrain_ms_output_0train.pb	√	√	36	3	3	4.49263525	41694	14	106.5	4	1.129214853	0.114428935
resnet_ms_output.pb	√	√	36	3	4	0.060677052	648	0	0	0	1	0
alexnets_output.pb	√	√	36	3	4	0.006333351	68	0	0	0	1	0
finetune_ms_output.pb	√	√	36	3	4	1.717193365	17531	9	69.77777778	4.111111111	1.11286739	0.101420341
lenet_custom_ms_output.pb	√	√	36	3	4	0.005299091	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	36	3	4	0.294864416	3161	3	18.66666667	3.333333333	1.050166113	0.047769693
bert_pretrain_ms_output_0train.pb	√	√	36	3	4	4.538393021	41694	10	147.9	4.1	1.128389716	0.113781359
resnet_ms_output.pb	√	√	36	4	2	0.063029766	648	0	0	0	1	0
alexnets_output.pb	√	√	36	4	2	0.008029938	68	0	0	0	1	0
finetune_ms_output.pb	√	√	36	4	2	1.718839645	17531	7	28.28571429	5.142857143	1.047001911	0.044891906
lenet_custom_ms_output.pb	√	√	36	4	2	0.007293224	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	36	4	2	0.306421757	3161	2	4	4.5	1.007650622	0.007592534
bert_pretrain_ms_output_0train.pb	√	√	36	4	2	4.491660833	41694	10	62.7	5	1.063595316	0.059792776
resnet_ms_output.pb	√	√	36	4	3	0.061044216	648	0	0	0	1	0
alexnets_output.pb	√	√	36	4	3	0.007163286	68	0	0	0	1	0
finetune_ms_output.pb	√	√	36	4	3	1.717627525	17531	6	32.66666667	5	1.046626866	0.044549655
lenet_custom_ms_output.pb	√	√	36	4	3	0.006210327	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	36	4	3	0.325507164	3161	2	4	4.5	1.007650622	0.007592534
bert_pretrain_ms_output_0train.pb	√	√	36	4	3	4.640462399	41694	8	77.875	4.75	1.063269834	0.059504965
resnet_ms_output.pb	√	√	36	4	4	0.06065321	648	0	0	0	1	0
alexnets_output.pb	√	√	36	4	4	0.006196737	68	0	0	0	1	0
finetune_ms_output.pb	√	√	36	4	4	1.677106857	17531	5	38.6	5	1.046127223	0.04409332
lenet_custom_ms_output.pb	√	√	36	4	4	0.005405188	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	36	4	4	0.299811125	3161	1	5	4	1.004448681	0.004428978
bert_pretrain_ms_output_0train.pb	√	√	36	4	4	4.530343294	41694	6	102.8333333	4.833333333	1.062809075	0.059097232
resnet_ms_output.pb	√	√	36	5	2	0.062182665	648	0	0	0	1	0
alexnets_output.pb	√	√	36	5	2	0.007760525	68	0	0	0	1	0
finetune_ms_output.pb	√	√	36	5	2	1.721920013	17531	6	25	5.333333333	1.035376801	0.034168045
lenet_custom_ms_output.pb	√	√	36	5	2	0.007405996	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	36	5	2	0.307225227	3161	1	3	5	1.003173596	0.003163556
bert_pretrain_ms_output_0train.pb	√	√	36	5	2	4.525276184	41694	7	22	5.428571429	1.01492174	0.014702355
resnet_ms_output.pb	√	√	36	5	3	0.060627222	648	0	0	0	1	0
alexnets_output.pb	√	√	36	5	3	0.007078648	68	0	0	0	1	0
finetune_ms_output.pb	√	√	36	5	3	1.70279026	17531	5	29.6	5.2	1.035010037	0.033825794
lenet_custom_ms_output.pb	√	√	36	5	3	0.00607729	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	36	5	3	0.31686759	3161	1	3	5	1.003173596	0.003163556
bert_pretrain_ms_output_0train.pb	√	√	36	5	3	4.568966866	41694	5	30	5.2	1.014625362	0.014414544
resnet_ms_output.pb	√	√	36	5	4	0.06118679	648	0	0	0	1	0
alexnets_output.pb	√	√	36	5	4	0.006427526	68	0	0	0	1	0
finetune_ms_output.pb	√	√	36	5	4	1.761789322	17531	4	36.25	5.25	1.034521421	0.03336946
lenet_custom_ms_output.pb	√	√	36	5	4	0.005286932	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	36	5	4	0.293881893	3161	0	0	0	1	0
bert_pretrain_ms_output_0train.pb	√	√	36	5	4	4.490770102	41694	4	36.75	5.25	1.014403192	0.014198686
resnet_ms_output.pb	√	√	48	3	2	0.062267303	648	1	2	3	1.001545595	0.00154321
alexnets_output.pb	√	√	48	3	2	0.007829189	68	0	0	0	1	0
finetune_ms_output.pb	√	√	48	3	2	1.734648705	17531	15	42.73333333	4	1.114423749	0.102675261
lenet_custom_ms_output.pb	√	√	48	3	2	0.007406235	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	48	3	2	0.325101614	3161	5	12.2	3.6	1.05436958	0.05156596
bert_pretrain_ms_output_0train.pb	√	√	48	3	2	4.554587364	41694	17	88	4.176470588	1.129612571	0.11474073
resnet_ms_output.pb	√	√	48	3	3	0.061438322	648	0	0	0	1	0
alexnets_output.pb	√	√	48	3	3	0.007080078	68	0	0	0	1	0
finetune_ms_output.pb	√	√	48	3	3	1.686970234	17531	10	63.1	4.2	1.113432836	0.101876676
lenet_custom_ms_output.pb	√	√	48	3	3	0.00618434	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	48	3	3	0.321268797	3161	4	14.75	3.75	1.053666667	0.050933249
bert_pretrain_ms_output_0train.pb	√	√	48	3	3	4.546809673	41694	14	106.5	4	1.129214853	0.114428935
resnet_ms_output.pb	√	√	48	3	4	0.061330557	648	0	0	0	1	0
alexnets_output.pb	√	√	48	3	4	0.006357908	68	0	0	0	1	0
finetune_ms_output.pb	√	√	48	3	4	1.686776638	17531	9	69.77777778	4.111111111	1.11286739	0.101420341
lenet_custom_ms_output.pb	√	√	48	3	4	0.005280972	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	48	3	4	0.299105883	3161	3	18.66666667	3.333333333	1.050166113	0.047769693
bert_pretrain_ms_output_0train.pb	√	√	48	3	4	4.500252247	41694	10	147.9	4.1	1.128389716	0.113781359
resnet_ms_output.pb	√	√	48	4	2	0.061999798	648	0	0	0	1	0
alexnets_output.pb	√	√	48	4	2	0.00780344	68	0	0	0	1	0
finetune_ms_output.pb	√	√	48	4	2	1.711999893	17531	7	28.28571429	5.142857143	1.047001911	0.044891906
lenet_custom_ms_output.pb	√	√	48	4	2	0.007456541	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	48	4	2	0.312227726	3161	2	4	4.5	1.007650622	0.007592534
bert_pretrain_ms_output_0train.pb	√	√	48	4	2	4.475604773	41694	10	62.7	5	1.063595316	0.059792776
resnet_ms_output.pb	√	√	48	4	3	0.0609622	648	0	0	0	1	0
alexnets_output.pb	√	√	48	4	3	0.007066011	68	0	0	0	1	0
finetune_ms_output.pb	√	√	48	4	3	1.715359449	17531	6	32.66666667	5	1.046626866	0.044549655
lenet_custom_ms_output.pb	√	√	48	4	3	0.006078959	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	48	4	3	0.313993692	3161	2	4	4.5	1.007650622	0.007592534
bert_pretrain_ms_output_0train.pb	√	√	48	4	3	4.423980474	41694	8	77.875	4.75	1.063269834	0.059504965
resnet_ms_output.pb	√	√	48	4	4	0.060632229	648	0	0	0	1	0
alexnets_output.pb	√	√	48	4	4	0.006267786	68	0	0	0	1	0
finetune_ms_output.pb	√	√	48	4	4	1.68068099	17531	5	38.6	5	1.046127223	0.04409332
lenet_custom_ms_output.pb	√	√	48	4	4	0.005374432	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	48	4	4	0.298303127	3161	1	5	4	1.004448681	0.004428978
bert_pretrain_ms_output_0train.pb	√	√	48	4	4	4.498241186	41694	6	102.8333333	4.833333333	1.062809075	0.059097232
resnet_ms_output.pb	√	√	48	5	2	0.062193632	648	0	0	0	1	0
alexnets_output.pb	√	√	48	5	2	0.007737875	68	0	0	0	1	0
finetune_ms_output.pb	√	√	48	5	2	1.685749054	17531	6	25	5.333333333	1.035376801	0.034168045
lenet_custom_ms_output.pb	√	√	48	5	2	0.00732398	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	48	5	2	0.319581032	3161	1	3	5	1.003173596	0.003163556
bert_pretrain_ms_output_0train.pb	√	√	48	5	2	4.514687777	41694	7	22	5.428571429	1.01492174	0.014702355
resnet_ms_output.pb	√	√	48	5	3	0.061242819	648	0	0	0	1	0
alexnets_output.pb	√	√	48	5	3	0.007065773	68	0	0	0	1	0
finetune_ms_output.pb	√	√	48	5	3	1.74185586	17531	5	29.6	5.2	1.035010037	0.033825794
lenet_custom_ms_output.pb	√	√	48	5	3	0.006070375	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	48	5	3	0.314492464	3161	1	3	5	1.003173596	0.003163556
bert_pretrain_ms_output_0train.pb	√	√	48	5	3	4.478384495	41694	5	30	5.2	1.014625362	0.014414544
resnet_ms_output.pb	√	√	48	5	4	0.061461926	648	0	0	0	1	0
alexnets_output.pb	√	√	48	5	4	0.006294489	68	0	0	0	1	0
finetune_ms_output.pb	√	√	48	5	4	1.693351507	17531	4	36.25	5.25	1.034521421	0.03336946
lenet_custom_ms_output.pb	√	√	48	5	4	0.005315542	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√	√	48	5	4	0.29899168	3161	0	0	0	1	0
bert_pretrain_ms_output_0train.pb	√	√	48	5	4	4.529016256	41694	4	36.75	5.25	1.014403192	0.014198686
resnet_ms_output.pb	√		24	3	2	0.138673544	648	5	2.2	13.4	1.038461538	0.037037037
alexnets_output.pb	√		24	3	2	0.012963295	68	3	2	6.333333333	1.172413793	0.147058824
finetune_ms_output.pb	√		24	3	2	9.420670271	17531	414	35.65217391	20.00483092	1.314463523	0.239233358
lenet_custom_ms_output.pb	√		24	3	2	0.010452747	57	2	2	5.5	1.163265306	0.140350877
mobilenetv2_ms_output.pb	√		24	3	2	0.981835127	3161	41	6.048780488	10.43902439	1.255361398	0.20341664
bert_pretrain_ms_output_0train.pb	√		24	3	2	22.76951051	41694	555	45.32612613	20.45765766	2.680941358	0.62699669
resnet_ms_output.pb	√		24	3	3	0.141837835	648	5	4	21.4	1.053658537	0.050925926
alexnets_output.pb	√		24	3	3	0.008281469	68	2	3	3	1.214285714	0.176470588
finetune_ms_output.pb	√		24	3	3	9.099755764	17531	396	37.17676768	20.42929293	1.319707919	0.242256574
lenet_custom_ms_output.pb	√		24	3	3	0.006388187	57	1	3	3	1.117647059	0.105263158
mobilenetv2_ms_output.pb	√		24	3	3	0.877743959	3161	35	7.2	12.65714286	1.33940678	0.253400823
bert_pretrain_ms_output_0train.pb	√		24	3	3	22.57780766	41694	529	47.46691871	20.62003781	2.739421813	0.634959467
resnet_ms_output.pb	√		24	3	4	0.124639511	648	5	4.4	21.2	1.2	0.166666667
alexnets_output.pb	√		24	3	4	0.006572247	68	0	0	0	1	0
finetune_ms_output.pb	√		24	3	4	9.186805248	17531	395	37.26582278	20.40759494	1.317823047	0.24117278
lenet_custom_ms_output.pb	√		24	3	4	0.005529404	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√		24	3	4	0.834102392	3161	32	8.9375	14.8125	1.468183929	0.318886428
bert_pretrain_ms_output_0train.pb	√		24	3	4	22.48218322	41694	527	47.63567362	20.68690702	2.737623112	0.634719624
resnet_ms_output.pb	√		24	4	2	0.139594555	648	4	2.25	16	1.0368	0.035493827
alexnets_output.pb	√		24	4	2	0.01286602	68	3	2	6.333333333	1.172413793	0.147058824
finetune_ms_output.pb	√		24	4	2	9.310757399	17531	397	35.94206549	20.73299748	1.297343299	0.229193999
lenet_custom_ms_output.pb	√		24	4	2	0.010356188	57	1	2	8	1.096153846	0.087719298
mobilenetv2_ms_output.pb	√		24	4	2	0.985623598	3161	26	4.692307692	14.73076923	1.192830189	0.161657703
bert_pretrain_ms_output_0train.pb	√		24	4	2	22.81166363	41694	542	43.14944649	20.87638376	2.651446741	0.622847412
resnet_ms_output.pb	√		24	4	3	0.134735107	648	5	4	21.4	1.053658537	0.050925926
alexnets_output.pb	√		24	4	3	0.00809741	68	0	0	0	1	0
finetune_ms_output.pb	√		24	4	3	9.130953789	17531	386	36.90414508	20.88082902	1.303807824	0.233015801
lenet_custom_ms_output.pb	√		24	4	3	0.00630784	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√		24	4	3	0.940241814	3161	26	5.307692308	16	1.271009248	0.213223663
bert_pretrain_ms_output_0train.pb	√		24	4	3	22.37666655	41694	515	45.30679612	21.09902913	2.704592631	0.63025855
resnet_ms_output.pb	√		24	4	4	0.124225616	648	5	4.4	21.2	1.2	0.166666667
alexnets_output.pb	√		24	4	4	0.006614208	68	0	0	0	1	0
finetune_ms_output.pb	√		24	4	4	8.997601032	17531	384	37.08072917	20.90625	1.301388167	0.231589755
lenet_custom_ms_output.pb	√		24	4	4	0.005384922	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√		24	4	4	0.853661537	3161	25	7.12	18.12	1.380952381	0.275862069
bert_pretrain_ms_output_0train.pb	√		24	4	4	22.6061039	41694	515	45.30679612	21.09902913	2.704592631	0.63025855
resnet_ms_output.pb	√		24	5	2	0.13952136	648	4	2.25	16	1.0368	0.035493827
alexnets_output.pb	√		24	5	2	0.012817144	68	3	2	6.333333333	1.172413793	0.147058824
finetune_ms_output.pb	√		24	5	2	9.085716248	17531	394	36.14467005	20.86040609	1.28885458	0.224117278
lenet_custom_ms_output.pb	√		24	5	2	0.010809183	57	1	2	8	1.096153846	0.087719298
mobilenetv2_ms_output.pb	√		24	5	2	1.025201559	3161	22	4.772727273	16.68181818	1.177281192	0.150585258
bert_pretrain_ms_output_0train.pb	√		24	5	2	22.92642212	41694	536	42.78731343	21.06902985	2.635191505	0.620520938
resnet_ms_output.pb	√		24	5	3	0.134394646	648	5	4	21.4	1.053658537	0.050925926
alexnets_output.pb	√		24	5	3	0.00808239	68	0	0	0	1	0
finetune_ms_output.pb	√		24	5	3	9.239853144	17531	384	37.03385417	20.96875	1.295330279	0.227996121
lenet_custom_ms_output.pb	√		24	5	3	0.011878967	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√		24	5	3	0.952285528	3161	22	5.5	18.18181818	1.252377179	0.201518507
bert_pretrain_ms_output_0train.pb	√		24	5	3	22.41912699	41694	512	44.69921875	21.203125	2.689068043	0.628123951
resnet_ms_output.pb	√		24	5	4	0.124053001	648	5	4.4	21.2	1.2	0.166666667
alexnets_output.pb	√		24	5	4	0.006574631	68	0	0	0	1	0
finetune_ms_output.pb	√		24	5	4	8.991466761	17531	382	37.21204188	20.9947644	1.292941957	0.226570076
lenet_custom_ms_output.pb	√		24	5	4	0.005506992	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√		24	5	4	0.841246128	3161	22	7.318181818	20.04545455	1.358401375	0.263840557
bert_pretrain_ms_output_0train.pb	√		24	5	4	22.51612949	41694	512	44.69921875	21.203125	2.689068043	0.628123951
resnet_ms_output.pb	√		36	3	2	0.27671814	648	9	2.111111111	24.44444444	0.971514243	-0.029320988
alexnets_output.pb	√		36	3	2	0.012947321	68	3	2	6.333333333	1.172413793	0.147058824
finetune_ms_output.pb	√		36	3	2	140.3004999	17531	1849	32.90319091	32.97890752	0.274888279	-2.637841538
lenet_custom_ms_output.pb	√		36	3	2	0.011546612	57	2	2	5.5	1.163265306	0.140350877
mobilenetv2_ms_output.pb	√		36	3	2	2.803932905	3161	76	4.618421053	22.85526316	0.926436108	-0.079405252
bert_pretrain_ms_output_0train.pb	√		36	3	2	645.1359303	41694	3160	33.89462025	33.74841772	0.38457778	-1.600254233
resnet_ms_output.pb	√		36	3	3	0.226215124	648	7	3.714285714	29.42857143	1.157142857	0.135802469
alexnets_output.pb	√		36	3	3	0.0082798	68	2	3	3	1.214285714	0.176470588
finetune_ms_output.pb	√		36	3	3	144.8053215	17531	1835	33.13787466	33.17057221	0.274763338	-2.63949575
lenet_custom_ms_output.pb	√		36	3	3	0.007463217	57	1	3	3	1.117647059	0.105263158
mobilenetv2_ms_output.pb	√		36	3	3	2.340086699	3161	79	5.151898734	26.20253165	0.952969551	-0.049351471
bert_pretrain_ms_output_0train.pb	√		36	3	3	596.6075664	41694	3120	34.30544872	33.825	0.388147238	-1.57634192
resnet_ms_output.pb	√		36	3	4	0.184396267	648	7	4.285714286	29.14285714	1.443207127	0.307098765
alexnets_output.pb	√		36	3	4	0.006783247	68	0	0	0	1	0
finetune_ms_output.pb	√		36	3	4	137.8431895	17531	1834	33.15485278	33.17284624	0.274681541	-2.640579545
lenet_custom_ms_output.pb	√		36	3	4	0.005484819	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√		36	3	4	2.073955536	3161	79	6.936708861	26.65822785	1.053315561	0.050616893
bert_pretrain_ms_output_0train.pb	√		36	3	4	636.8523462	41694	3118	34.32552919	33.84477229	0.388111107	-1.576581762
resnet_ms_output.pb	√		36	4	2	0.270419836	648	8	2.125	27.125	0.97005988	-0.030864198
alexnets_output.pb	√		36	4	2	0.01266098	68	3	2	6.333333333	1.172413793	0.147058824
finetune_ms_output.pb	√		36	4	2	141.7555578	17531	1832	32.94050218	33.25709607	0.274131757	-2.647880897
lenet_custom_ms_output.pb	√		36	4	2	0.011081457	57	1	2	8	1.096153846	0.087719298
mobilenetv2_ms_output.pb	√		36	4	2	2.783252239	3161	61	3.68852459	27.73770492	0.891930023	-0.121164189
bert_pretrain_ms_output_0train.pb	√		36	4	2	642.2136412	41694	3147	33.4725135	33.87543692	0.383965079	-1.604403511
resnet_ms_output.pb	√		36	4	3	0.229095221	648	7	3.714285714	29.42857143	1.157142857	0.135802469
alexnets_output.pb	√		36	4	3	0.008169651	68	0	0	0	1	0
finetune_ms_output.pb	√		36	4	3	148.7275791	17531	1825	33.05808219	33.33589041	0.274067473	-2.648736524
lenet_custom_ms_output.pb	√		36	4	3	0.006525755	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√		36	4	3	2.336178541	3161	70	4.185714286	29.18571429	0.917828107	-0.08952863
bert_pretrain_ms_output_0train.pb	√		36	4	3	597.6310258	41694	3106	33.88795879	33.96394076	0.387440296	-1.581042836
resnet_ms_output.pb	√		36	4	4	0.185164452	648	7	4.285714286	29.14285714	1.443207127	0.307098765
alexnets_output.pb	√		36	4	4	0.006605387	68	0	0	0	1	0
finetune_ms_output.pb	√		36	4	4	138.9160748	17531	1823	33.09105869	33.35490949	0.273960401	-2.650162569
lenet_custom_ms_output.pb	√		36	4	4	0.00570488	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√		36	4	4	1.951111078	3161	72	6.111111111	28.95833333	1.007650622	0.007592534
bert_pretrain_ms_output_0train.pb	√		36	4	4	591.9786947	41694	3106	33.88795879	33.96394076	0.387440296	-1.581042836
resnet_ms_output.pb	√		36	5	2	0.270936489	648	8	2.125	27.125	0.97005988	-0.030864198
alexnets_output.pb	√		36	5	2	0.013339043	68	3	2	6.333333333	1.172413793	0.147058824
finetune_ms_output.pb	√		36	5	2	146.7244027	17531	1829	32.97922362	33.30508475	0.273750781	-2.652957618
lenet_custom_ms_output.pb	√		36	5	2	0.010558367	57	1	2	8	1.096153846	0.087719298
mobilenetv2_ms_output.pb	√		36	5	2	2.898170471	3161	57	3.649122807	29.40350877	0.8832076	-0.132236634
bert_pretrain_ms_output_0train.pb	√		36	5	2	632.1272347	41694	3141	33.39223177	33.93314231	0.383622395	-1.606729985
resnet_ms_output.pb	√		36	5	3	0.238684177	648	7	3.714285714	29.42857143	1.157142857	0.135802469
alexnets_output.pb	√		36	5	3	0.008165359	68	0	0	0	1	0
finetune_ms_output.pb	√		36	5	3	146.8073792	17531	1823	33.08118486	33.3680746	0.273690948	-2.653756203
lenet_custom_ms_output.pb	√		36	5	3	0.007680416	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√		36	5	3	2.57319212	3161	66	4.181818182	30.71212121	0.908072393	-0.101233787
bert_pretrain_ms_output_0train.pb	√		36	5	3	603.2939513	41694	3103	33.77666774	33.99355462	0.387120136	-1.583177436
resnet_ms_output.pb	√		36	5	4	0.184869766	648	7	4.285714286	29.14285714	1.443207127	0.307098765
alexnets_output.pb	√		36	5	4	0.00658536	68	0	0	0	1	0
finetune_ms_output.pb	√		36	5	4	143.8490016	17531	1821	33.11422295	33.38714992	0.27358417	-2.655182249
lenet_custom_ms_output.pb	√		36	5	4	0.005692959	57	0	0	0	1	0
mobilenetv2_ms_output.pb	√		36	5	4	2.019337177	3161	69	6.130434783	30.04347826	0.995590551	-0.004428978
bert_pretrain_ms_output_0train.pb	√		36	5	4	607.015888	41694	3103	33.77666774	33.99355462	0.387120136	-1.583177436
resnet_ms_output.pb	√		48	3	2	0.522814989	648	16	2.3125	36.5625	0.737201365	-0.356481481
alexnets_output.pb	√		48	3	2	0.0133183	68	3	2	6.333333333	1.172413793	0.147058824
finetune_ms_output.pb	√		48	3	2	3982.185682	17531	5832	30.14180384	45.22325103	0.065772739	-14.20386743
lenet_custom_ms_output.pb	√		48	3	2	0.014278889	57	2	2	5.5	1.163265306	0.140350877
mobilenetv2_ms_output.pb	√		48	3	2	7.351045847	3161	113	4.088495575	33.68141593	0.617021277	-0.620689655