Pruned-MobileNet_v2

Apply the pruning strategy of Network Slimming for MobileNet_v2.

The Caffe implementation of the algorithm is available in link.

Results

The size of the input image is 224x224.

Step	Prune Ratio	L1 value	Parameters	Top1 Accuracy	Speed
0	0	0	9.8MB	93.24%	152.0ms
1	0	0.001	9.8MB	92.68%	152.0ms
2	0.3	0.001	7.2MB	91.84%	-
3	0.3 + 0.7 * 0.3 = 0.51	0.001	5.4MB	91.26%	-
4	0.51 + 0.49 * 0.2 = 0.608	0	4.6MB	92.13%	79.3ms
5	0.608 (merging BN)	-	4.5MB	92.13%	58.0ms

Model	Speed on PC	Speed on iPhone7p (using NCNN)
ResNet50	about 1000ms	277.8ms
MobileNet v2	152.0ms	41.4ms
Pruned MobileNet v2 (with BN)	79.3ms	18.1ms
Pruned MobileNet v2 (merging BN)	58.0ms	13.6ms (about 5ms when the input is resized into `96x96` )

If you have any problems, please feel free to contact me via eezywu@163.com.

用户评价

全部评价

还没有评论，说两句吧！

热门资源