6.9 KiB

Raw Blame History

EfficientNet与ResNeXt101_wsl系列

概述

正在持续更新中......

在预测时，图像的crop_size和resize_short_size如下表所示。

Models	crop_size	resize_short_size
ResNeXt101_32x8d_wsl	224	224
ResNeXt101_32x16d_wsl	224	224
ResNeXt101_32x32d_wsl	224	224
ResNeXt101_32x48d_wsl	224	224
Fix_ResNeXt101_32x48d_wsl	320	320
EfficientNetB0	224	256
EfficientNetB1	240	272
EfficientNetB2	260	292
EfficientNetB3	300	332
EfficientNetB4	380	412
EfficientNetB5	456	488
EfficientNetB6	528	560
EfficientNetB7	600	632
EfficientNetB0_small	224	256

精度、FLOPS和参数量

Models	Top1	Top5	Reference top1	Reference top5	FLOPS (G)	Parameters (M)
ResNeXt101_ 32x8d_wsl	0.826	0.967	0.822	0.964	29.140	78.440
ResNeXt101_ 32x16d_wsl	0.842	0.973	0.842	0.972	57.550	152.660
ResNeXt101_ 32x32d_wsl	0.850	0.976	0.851	0.975	115.170	303.110
ResNeXt101_ 32x48d_wsl	0.854	0.977	0.854	0.976	173.580	456.200
Fix_ResNeXt101_ 32x48d_wsl	0.863	0.980	0.864	0.980	354.230	456.200
EfficientNetB0	0.774	0.933	0.773	0.935	0.720	5.100
EfficientNetB1	0.792	0.944	0.792	0.945	1.270	7.520
EfficientNetB2	0.799	0.947	0.803	0.950	1.850	8.810
EfficientNetB3	0.812	0.954	0.817	0.956	3.430	11.840
EfficientNetB4	0.829	0.962	0.830	0.963	8.290	18.760
EfficientNetB5	0.836	0.967	0.837	0.967	19.510	29.610
EfficientNetB6	0.840	0.969	0.842	0.968	36.270	42.000
EfficientNetB7	0.843	0.969	0.844	0.971	72.350	64.920
EfficientNetB0_ small	0.758	0.926			0.720	4.650

FP16预测速度

Models	batch_size=1 (ms)	batch_size=4 (ms)	batch_size=8 (ms)	batch_size=32 (ms)
ResNeXt101_ 32x8d_wsl	16.063	16.342	24.914	45.035
ResNeXt101_ 32x16d_wsl	16.471	25.235	30.762	67.869
ResNeXt101_ 32x32d_wsl	29.425	37.149	50.834
ResNeXt101_ 32x48d_wsl	40.311	58.414
Fix_ResNeXt101_ 32x48d_wsl	43.960	86.514
EfficientNetB0	1.759	2.748	3.761	10.178
EfficientNetB1	2.592	4.122	5.829	16.262
EfficientNetB2	2.866	4.715	7.064	20.954
EfficientNetB3	3.869	6.815	10.672	34.097
EfficientNetB4	5.626	11.937	19.753	67.436
EfficientNetB5	8.907	21.685	37.248	134.185
EfficientNetB6	13.591	34.093	60.976
EfficientNetB7	20.963	56.397	103.971
EfficientNetB0_ small	1.039	1.665	2.493	7.748

FP32预测速度

Models	batch_size=1 (ms)	batch_size=4 (ms)	batch_size=8 (ms)	batch_size=32 (ms)
ResNeXt101_ 32x8d_wsl	16.325	25.633	37.196	108.535
ResNeXt101_ 32x16d_wsl	25.224	40.929	62.898
ResNeXt101_ 32x32d_wsl	41.047	79.575
ResNeXt101_ 32x48d_wsl	60.610
Fix_ResNeXt101_ 32x48d_wsl	80.280
EfficientNetB0	1.902	3.296	4.361	11.319
EfficientNetB1	2.908	5.093	6.900	18.015
EfficientNetB2	3.324	5.832	8.357	23.371
EfficientNetB3	4.557	8.526	12.485	38.124
EfficientNetB4	6.767	14.742	23.218	77.590
EfficientNetB5	11.097	26.642	43.590
EfficientNetB6	17.582	42.408	74.336
EfficientNetB7	26.529	70.337	126.839
EfficientNetB0_ small	1.171	2.026	2.906	8.506