1name: "VGG_VOC0712_SSD_300x300_deploy"
2layer {
3  name: "input"
4  type: "Input"
5  top: "data"
6  input_param {
7    shape {
8      dim: 1
9      dim: 3
10      dim: 300
11      dim: 300
12    }
13  }
14}
15layer {
16  name: "conv1_1"
17  type: "Convolution"
18  bottom: "data"
19  top: "conv1_1"
20  param {
21    lr_mult: 1
22    decay_mult: 1
23  }
24  param {
25    lr_mult: 2
26    decay_mult: 0
27  }
28  convolution_param {
29    num_output: 64
30    pad: 1
31    kernel_size: 3
32    weight_filler {
33      type: "xavier"
34    }
35    bias_filler {
36      type: "constant"
37      value: 0
38    }
39  }
40}
41layer {
42  name: "relu1_1"
43  type: "ReLU"
44  bottom: "conv1_1"
45  top: "conv1_1"
46}
47layer {
48  name: "conv1_2"
49  type: "Convolution"
50  bottom: "conv1_1"
51  top: "conv1_2"
52  param {
53    lr_mult: 1
54    decay_mult: 1
55  }
56  param {
57    lr_mult: 2
58    decay_mult: 0
59  }
60  convolution_param {
61    num_output: 64
62    pad: 1
63    kernel_size: 3
64    weight_filler {
65      type: "xavier"
66    }
67    bias_filler {
68      type: "constant"
69      value: 0
70    }
71  }
72}
73layer {
74  name: "relu1_2"
75  type: "ReLU"
76  bottom: "conv1_2"
77  top: "conv1_2"
78}
79layer {
80  name: "pool1"
81  type: "Pooling"
82  bottom: "conv1_2"
83  top: "pool1"
84  pooling_param {
85    pool: MAX
86    kernel_size: 2
87    stride: 2
88  }
89}
90layer {
91  name: "conv2_1"
92  type: "Convolution"
93  bottom: "pool1"
94  top: "conv2_1"
95  param {
96    lr_mult: 1
97    decay_mult: 1
98  }
99  param {
100    lr_mult: 2
101    decay_mult: 0
102  }
103  convolution_param {
104    num_output: 128
105    pad: 1
106    kernel_size: 3
107    weight_filler {
108      type: "xavier"
109    }
110    bias_filler {
111      type: "constant"
112      value: 0
113    }
114  }
115}
116layer {
117  name: "relu2_1"
118  type: "ReLU"
119  bottom: "conv2_1"
120  top: "conv2_1"
121}
122layer {
123  name: "conv2_2"
124  type: "Convolution"
125  bottom: "conv2_1"
126  top: "conv2_2"
127  param {
128    lr_mult: 1
129    decay_mult: 1
130  }
131  param {
132    lr_mult: 2
133    decay_mult: 0
134  }
135  convolution_param {
136    num_output: 128
137    pad: 1
138    kernel_size: 3
139    weight_filler {
140      type: "xavier"
141    }
142    bias_filler {
143      type: "constant"
144      value: 0
145    }
146  }
147}
148layer {
149  name: "relu2_2"
150  type: "ReLU"
151  bottom: "conv2_2"
152  top: "conv2_2"
153}
154layer {
155  name: "pool2"
156  type: "Pooling"
157  bottom: "conv2_2"
158  top: "pool2"
159  pooling_param {
160    pool: MAX
161    kernel_size: 2
162    stride: 2
163  }
164}
165layer {
166  name: "conv3_1"
167  type: "Convolution"
168  bottom: "pool2"
169  top: "conv3_1"
170  param {
171    lr_mult: 1
172    decay_mult: 1
173  }
174  param {
175    lr_mult: 2
176    decay_mult: 0
177  }
178  convolution_param {
179    num_output: 256
180    pad: 1
181    kernel_size: 3
182    weight_filler {
183      type: "xavier"
184    }
185    bias_filler {
186      type: "constant"
187      value: 0
188    }
189  }
190}
191layer {
192  name: "relu3_1"
193  type: "ReLU"
194  bottom: "conv3_1"
195  top: "conv3_1"
196}
197layer {
198  name: "conv3_2"
199  type: "Convolution"
200  bottom: "conv3_1"
201  top: "conv3_2"
202  param {
203    lr_mult: 1
204    decay_mult: 1
205  }
206  param {
207    lr_mult: 2
208    decay_mult: 0
209  }
210  convolution_param {
211    num_output: 256
212    pad: 1
213    kernel_size: 3
214    weight_filler {
215      type: "xavier"
216    }
217    bias_filler {
218      type: "constant"
219      value: 0
220    }
221  }
222}
223layer {
224  name: "relu3_2"
225  type: "ReLU"
226  bottom: "conv3_2"
227  top: "conv3_2"
228}
229layer {
230  name: "conv3_3"
231  type: "Convolution"
232  bottom: "conv3_2"
233  top: "conv3_3"
234  param {
235    lr_mult: 1
236    decay_mult: 1
237  }
238  param {
239    lr_mult: 2
240    decay_mult: 0
241  }
242  convolution_param {
243    num_output: 256
244    pad: 1
245    kernel_size: 3
246    weight_filler {
247      type: "xavier"
248    }
249    bias_filler {
250      type: "constant"
251      value: 0
252    }
253  }
254}
255layer {
256  name: "relu3_3"
257  type: "ReLU"
258  bottom: "conv3_3"
259  top: "conv3_3"
260}
261layer {
262  name: "pool3"
263  type: "Pooling"
264  bottom: "conv3_3"
265  top: "pool3"
266  pooling_param {
267    pool: MAX
268    kernel_size: 2
269    stride: 2
270  }
271}
272layer {
273  name: "conv4_1"
274  type: "Convolution"
275  bottom: "pool3"
276  top: "conv4_1"
277  param {
278    lr_mult: 1
279    decay_mult: 1
280  }
281  param {
282    lr_mult: 2
283    decay_mult: 0
284  }
285  convolution_param {
286    num_output: 512
287    pad: 1
288    kernel_size: 3
289    weight_filler {
290      type: "xavier"
291    }
292    bias_filler {
293      type: "constant"
294      value: 0
295    }
296  }
297}
298layer {
299  name: "relu4_1"
300  type: "ReLU"
301  bottom: "conv4_1"
302  top: "conv4_1"
303}
304layer {
305  name: "conv4_2"
306  type: "Convolution"
307  bottom: "conv4_1"
308  top: "conv4_2"
309  param {
310    lr_mult: 1
311    decay_mult: 1
312  }
313  param {
314    lr_mult: 2
315    decay_mult: 0
316  }
317  convolution_param {
318    num_output: 512
319    pad: 1
320    kernel_size: 3
321    weight_filler {
322      type: "xavier"
323    }
324    bias_filler {
325      type: "constant"
326      value: 0
327    }
328  }
329}
330layer {
331  name: "relu4_2"
332  type: "ReLU"
333  bottom: "conv4_2"
334  top: "conv4_2"
335}
336layer {
337  name: "conv4_3"
338  type: "Convolution"
339  bottom: "conv4_2"
340  top: "conv4_3"
341  param {
342    lr_mult: 1
343    decay_mult: 1
344  }
345  param {
346    lr_mult: 2
347    decay_mult: 0
348  }
349  convolution_param {
350    num_output: 512
351    pad: 1
352    kernel_size: 3
353    weight_filler {
354      type: "xavier"
355    }
356    bias_filler {
357      type: "constant"
358      value: 0
359    }
360  }
361}
362layer {
363  name: "relu4_3"
364  type: "ReLU"
365  bottom: "conv4_3"
366  top: "conv4_3"
367}
368layer {
369  name: "pool4"
370  type: "Pooling"
371  bottom: "conv4_3"
372  top: "pool4"
373  pooling_param {
374    pool: MAX
375    kernel_size: 2
376    stride: 2
377  }
378}
379layer {
380  name: "conv5_1"
381  type: "Convolution"
382  bottom: "pool4"
383  top: "conv5_1"
384  param {
385    lr_mult: 1
386    decay_mult: 1
387  }
388  param {
389    lr_mult: 2
390    decay_mult: 0
391  }
392  convolution_param {
393    num_output: 512
394    pad: 1
395    kernel_size: 3
396    weight_filler {
397      type: "xavier"
398    }
399    bias_filler {
400      type: "constant"
401      value: 0
402    }
403    dilation: 1
404  }
405}
406layer {
407  name: "relu5_1"
408  type: "ReLU"
409  bottom: "conv5_1"
410  top: "conv5_1"
411}
412layer {
413  name: "conv5_2"
414  type: "Convolution"
415  bottom: "conv5_1"
416  top: "conv5_2"
417  param {
418    lr_mult: 1
419    decay_mult: 1
420  }
421  param {
422    lr_mult: 2
423    decay_mult: 0
424  }
425  convolution_param {
426    num_output: 512
427    pad: 1
428    kernel_size: 3
429    weight_filler {
430      type: "xavier"
431    }
432    bias_filler {
433      type: "constant"
434      value: 0
435    }
436    dilation: 1
437  }
438}
439layer {
440  name: "relu5_2"
441  type: "ReLU"
442  bottom: "conv5_2"
443  top: "conv5_2"
444}
445layer {
446  name: "conv5_3"
447  type: "Convolution"
448  bottom: "conv5_2"
449  top: "conv5_3"
450  param {
451    lr_mult: 1
452    decay_mult: 1
453  }
454  param {
455    lr_mult: 2
456    decay_mult: 0
457  }
458  convolution_param {
459    num_output: 512
460    pad: 1
461    kernel_size: 3
462    weight_filler {
463      type: "xavier"
464    }
465    bias_filler {
466      type: "constant"
467      value: 0
468    }
469    dilation: 1
470  }
471}
472layer {
473  name: "relu5_3"
474  type: "ReLU"
475  bottom: "conv5_3"
476  top: "conv5_3"
477}
478layer {
479  name: "pool5"
480  type: "Pooling"
481  bottom: "conv5_3"
482  top: "pool5"
483  pooling_param {
484    pool: MAX
485    kernel_size: 3
486    stride: 1
487    pad: 1
488  }
489}
490layer {
491  name: "fc6"
492  type: "Convolution"
493  bottom: "pool5"
494  top: "fc6"
495  param {
496    lr_mult: 1
497    decay_mult: 1
498  }
499  param {
500    lr_mult: 2
501    decay_mult: 0
502  }
503  convolution_param {
504    num_output: 1024
505    pad: 6
506    kernel_size: 3
507    weight_filler {
508      type: "xavier"
509    }
510    bias_filler {
511      type: "constant"
512      value: 0
513    }
514    dilation: 6
515  }
516}
517layer {
518  name: "relu6"
519  type: "ReLU"
520  bottom: "fc6"
521  top: "fc6"
522}
523layer {
524  name: "fc7"
525  type: "Convolution"
526  bottom: "fc6"
527  top: "fc7"
528  param {
529    lr_mult: 1
530    decay_mult: 1
531  }
532  param {
533    lr_mult: 2
534    decay_mult: 0
535  }
536  convolution_param {
537    num_output: 1024
538    kernel_size: 1
539    weight_filler {
540      type: "xavier"
541    }
542    bias_filler {
543      type: "constant"
544      value: 0
545    }
546  }
547}
548layer {
549  name: "relu7"
550  type: "ReLU"
551  bottom: "fc7"
552  top: "fc7"
553}
554layer {
555  name: "conv6_1"
556  type: "Convolution"
557  bottom: "fc7"
558  top: "conv6_1"
559  param {
560    lr_mult: 1
561    decay_mult: 1
562  }
563  param {
564    lr_mult: 2
565    decay_mult: 0
566  }
567  convolution_param {
568    num_output: 256
569    pad: 0
570    kernel_size: 1
571    stride: 1
572    weight_filler {
573      type: "xavier"
574    }
575    bias_filler {
576      type: "constant"
577      value: 0
578    }
579  }
580}
581layer {
582  name: "conv6_1_relu"
583  type: "ReLU"
584  bottom: "conv6_1"
585  top: "conv6_1"
586}
587layer {
588  name: "conv6_2"
589  type: "Convolution"
590  bottom: "conv6_1"
591  top: "conv6_2"
592  param {
593    lr_mult: 1
594    decay_mult: 1
595  }
596  param {
597    lr_mult: 2
598    decay_mult: 0
599  }
600  convolution_param {
601    num_output: 512
602    pad: 1
603    kernel_size: 3
604    stride: 2
605    weight_filler {
606      type: "xavier"
607    }
608    bias_filler {
609      type: "constant"
610      value: 0
611    }
612  }
613}
614layer {
615  name: "conv6_2_relu"
616  type: "ReLU"
617  bottom: "conv6_2"
618  top: "conv6_2"
619}
620layer {
621  name: "conv7_1"
622  type: "Convolution"
623  bottom: "conv6_2"
624  top: "conv7_1"
625  param {
626    lr_mult: 1
627    decay_mult: 1
628  }
629  param {
630    lr_mult: 2
631    decay_mult: 0
632  }
633  convolution_param {
634    num_output: 128
635    pad: 0
636    kernel_size: 1
637    stride: 1
638    weight_filler {
639      type: "xavier"
640    }
641    bias_filler {
642      type: "constant"
643      value: 0
644    }
645  }
646}
647layer {
648  name: "conv7_1_relu"
649  type: "ReLU"
650  bottom: "conv7_1"
651  top: "conv7_1"
652}
653layer {
654  name: "conv7_2"
655  type: "Convolution"
656  bottom: "conv7_1"
657  top: "conv7_2"
658  param {
659    lr_mult: 1
660    decay_mult: 1
661  }
662  param {
663    lr_mult: 2
664    decay_mult: 0
665  }
666  convolution_param {
667    num_output: 256
668    pad: 1
669    kernel_size: 3
670    stride: 2
671    weight_filler {
672      type: "xavier"
673    }
674    bias_filler {
675      type: "constant"
676      value: 0
677    }
678  }
679}
680layer {
681  name: "conv7_2_relu"
682  type: "ReLU"
683  bottom: "conv7_2"
684  top: "conv7_2"
685}
686layer {
687  name: "conv8_1"
688  type: "Convolution"
689  bottom: "conv7_2"
690  top: "conv8_1"
691  param {
692    lr_mult: 1
693    decay_mult: 1
694  }
695  param {
696    lr_mult: 2
697    decay_mult: 0
698  }
699  convolution_param {
700    num_output: 128
701    pad: 0
702    kernel_size: 1
703    stride: 1
704    weight_filler {
705      type: "xavier"
706    }
707    bias_filler {
708      type: "constant"
709      value: 0
710    }
711  }
712}
713layer {
714  name: "conv8_1_relu"
715  type: "ReLU"
716  bottom: "conv8_1"
717  top: "conv8_1"
718}
719layer {
720  name: "conv8_2"
721  type: "Convolution"
722  bottom: "conv8_1"
723  top: "conv8_2"
724  param {
725    lr_mult: 1
726    decay_mult: 1
727  }
728  param {
729    lr_mult: 2
730    decay_mult: 0
731  }
732  convolution_param {
733    num_output: 256
734    pad: 0
735    kernel_size: 3
736    stride: 1
737    weight_filler {
738      type: "xavier"
739    }
740    bias_filler {
741      type: "constant"
742      value: 0
743    }
744  }
745}
746layer {
747  name: "conv8_2_relu"
748  type: "ReLU"
749  bottom: "conv8_2"
750  top: "conv8_2"
751}
752layer {
753  name: "conv9_1"
754  type: "Convolution"
755  bottom: "conv8_2"
756  top: "conv9_1"
757  param {
758    lr_mult: 1
759    decay_mult: 1
760  }
761  param {
762    lr_mult: 2
763    decay_mult: 0
764  }
765  convolution_param {
766    num_output: 128
767    pad: 0
768    kernel_size: 1
769    stride: 1
770    weight_filler {
771      type: "xavier"
772    }
773    bias_filler {
774      type: "constant"
775      value: 0
776    }
777  }
778}
779layer {
780  name: "conv9_1_relu"
781  type: "ReLU"
782  bottom: "conv9_1"
783  top: "conv9_1"
784}
785layer {
786  name: "conv9_2"
787  type: "Convolution"
788  bottom: "conv9_1"
789  top: "conv9_2"
790  param {
791    lr_mult: 1
792    decay_mult: 1
793  }
794  param {
795    lr_mult: 2
796    decay_mult: 0
797  }
798  convolution_param {
799    num_output: 256
800    pad: 0
801    kernel_size: 3
802    stride: 1
803    weight_filler {
804      type: "xavier"
805    }
806    bias_filler {
807      type: "constant"
808      value: 0
809    }
810  }
811}
812layer {
813  name: "conv9_2_relu"
814  type: "ReLU"
815  bottom: "conv9_2"
816  top: "conv9_2"
817}
818layer {
819  name: "conv4_3_norm"
820  type: "Normalize"
821  bottom: "conv4_3"
822  top: "conv4_3_norm"
823  norm_param {
824    across_spatial: false
825    scale_filler {
826      type: "constant"
827      value: 20
828    }
829    channel_shared: false
830  }
831}
832layer {
833  name: "conv4_3_norm_mbox_loc"
834  type: "Convolution"
835  bottom: "conv4_3_norm"
836  top: "conv4_3_norm_mbox_loc"
837  param {
838    lr_mult: 1
839    decay_mult: 1
840  }
841  param {
842    lr_mult: 2
843    decay_mult: 0
844  }
845  convolution_param {
846    num_output: 16
847    pad: 1
848    kernel_size: 3
849    stride: 1
850    weight_filler {
851      type: "xavier"
852    }
853    bias_filler {
854      type: "constant"
855      value: 0
856    }
857  }
858}
859layer {
860  name: "conv4_3_norm_mbox_loc_perm"
861  type: "Permute"
862  bottom: "conv4_3_norm_mbox_loc"
863  top: "conv4_3_norm_mbox_loc_perm"
864  permute_param {
865    order: 0
866    order: 2
867    order: 3
868    order: 1
869  }
870}
871layer {
872  name: "conv4_3_norm_mbox_loc_flat"
873  type: "Flatten"
874  bottom: "conv4_3_norm_mbox_loc_perm"
875  top: "conv4_3_norm_mbox_loc_flat"
876  flatten_param {
877    axis: 1
878  }
879}
880layer {
881  name: "conv4_3_norm_mbox_conf"
882  type: "Convolution"
883  bottom: "conv4_3_norm"
884  top: "conv4_3_norm_mbox_conf"
885  param {
886    lr_mult: 1
887    decay_mult: 1
888  }
889  param {
890    lr_mult: 2
891    decay_mult: 0
892  }
893  convolution_param {
894    num_output: 84
895    pad: 1
896    kernel_size: 3
897    stride: 1
898    weight_filler {
899      type: "xavier"
900    }
901    bias_filler {
902      type: "constant"
903      value: 0
904    }
905  }
906}
907layer {
908  name: "conv4_3_norm_mbox_conf_perm"
909  type: "Permute"
910  bottom: "conv4_3_norm_mbox_conf"
911  top: "conv4_3_norm_mbox_conf_perm"
912  permute_param {
913    order: 0
914    order: 2
915    order: 3
916    order: 1
917  }
918}
919layer {
920  name: "conv4_3_norm_mbox_conf_flat"
921  type: "Flatten"
922  bottom: "conv4_3_norm_mbox_conf_perm"
923  top: "conv4_3_norm_mbox_conf_flat"
924  flatten_param {
925    axis: 1
926  }
927}
928layer {
929  name: "conv4_3_norm_mbox_priorbox"
930  type: "PriorBox"
931  bottom: "conv4_3_norm"
932  bottom: "data"
933  top: "conv4_3_norm_mbox_priorbox"
934  prior_box_param {
935    min_size: 30.0
936    max_size: 60.0
937    aspect_ratio: 2
938    flip: true
939    clip: false
940    variance: 0.1
941    variance: 0.1
942    variance: 0.2
943    variance: 0.2
944    step: 8
945    offset: 0.5
946  }
947}
948layer {
949  name: "fc7_mbox_loc"
950  type: "Convolution"
951  bottom: "fc7"
952  top: "fc7_mbox_loc"
953  param {
954    lr_mult: 1
955    decay_mult: 1
956  }
957  param {
958    lr_mult: 2
959    decay_mult: 0
960  }
961  convolution_param {
962    num_output: 24
963    pad: 1
964    kernel_size: 3
965    stride: 1
966    weight_filler {
967      type: "xavier"
968    }
969    bias_filler {
970      type: "constant"
971      value: 0
972    }
973  }
974}
975layer {
976  name: "fc7_mbox_loc_perm"
977  type: "Permute"
978  bottom: "fc7_mbox_loc"
979  top: "fc7_mbox_loc_perm"
980  permute_param {
981    order: 0
982    order: 2
983    order: 3
984    order: 1
985  }
986}
987layer {
988  name: "fc7_mbox_loc_flat"
989  type: "Flatten"
990  bottom: "fc7_mbox_loc_perm"
991  top: "fc7_mbox_loc_flat"
992  flatten_param {
993    axis: 1
994  }
995}
996layer {
997  name: "fc7_mbox_conf"
998  type: "Convolution"
999  bottom: "fc7"
1000  top: "fc7_mbox_conf"
1001  param {
1002    lr_mult: 1
1003    decay_mult: 1
1004  }
1005  param {
1006    lr_mult: 2
1007    decay_mult: 0
1008  }
1009  convolution_param {
1010    num_output: 126
1011    pad: 1
1012    kernel_size: 3
1013    stride: 1
1014    weight_filler {
1015      type: "xavier"
1016    }
1017    bias_filler {
1018      type: "constant"
1019      value: 0
1020    }
1021  }
1022}
1023layer {
1024  name: "fc7_mbox_conf_perm"
1025  type: "Permute"
1026  bottom: "fc7_mbox_conf"
1027  top: "fc7_mbox_conf_perm"
1028  permute_param {
1029    order: 0
1030    order: 2
1031    order: 3
1032    order: 1
1033  }
1034}
1035layer {
1036  name: "fc7_mbox_conf_flat"
1037  type: "Flatten"
1038  bottom: "fc7_mbox_conf_perm"
1039  top: "fc7_mbox_conf_flat"
1040  flatten_param {
1041    axis: 1
1042  }
1043}
1044layer {
1045  name: "fc7_mbox_priorbox"
1046  type: "PriorBox"
1047  bottom: "fc7"
1048  bottom: "data"
1049  top: "fc7_mbox_priorbox"
1050  prior_box_param {
1051    min_size: 60.0
1052    max_size: 111.0
1053    aspect_ratio: 2
1054    aspect_ratio: 3
1055    flip: true
1056    clip: false
1057    variance: 0.1
1058    variance: 0.1
1059    variance: 0.2
1060    variance: 0.2
1061    step: 16
1062    offset: 0.5
1063  }
1064}
1065layer {
1066  name: "conv6_2_mbox_loc"
1067  type: "Convolution"
1068  bottom: "conv6_2"
1069  top: "conv6_2_mbox_loc"
1070  param {
1071    lr_mult: 1
1072    decay_mult: 1
1073  }
1074  param {
1075    lr_mult: 2
1076    decay_mult: 0
1077  }
1078  convolution_param {
1079    num_output: 24
1080    pad: 1
1081    kernel_size: 3
1082    stride: 1
1083    weight_filler {
1084      type: "xavier"
1085    }
1086    bias_filler {
1087      type: "constant"
1088      value: 0
1089    }
1090  }
1091}
1092layer {
1093  name: "conv6_2_mbox_loc_perm"
1094  type: "Permute"
1095  bottom: "conv6_2_mbox_loc"
1096  top: "conv6_2_mbox_loc_perm"
1097  permute_param {
1098    order: 0
1099    order: 2
1100    order: 3
1101    order: 1
1102  }
1103}
1104layer {
1105  name: "conv6_2_mbox_loc_flat"
1106  type: "Flatten"
1107  bottom: "conv6_2_mbox_loc_perm"
1108  top: "conv6_2_mbox_loc_flat"
1109  flatten_param {
1110    axis: 1
1111  }
1112}
1113layer {
1114  name: "conv6_2_mbox_conf"
1115  type: "Convolution"
1116  bottom: "conv6_2"
1117  top: "conv6_2_mbox_conf"
1118  param {
1119    lr_mult: 1
1120    decay_mult: 1
1121  }
1122  param {
1123    lr_mult: 2
1124    decay_mult: 0
1125  }
1126  convolution_param {
1127    num_output: 126
1128    pad: 1
1129    kernel_size: 3
1130    stride: 1
1131    weight_filler {
1132      type: "xavier"
1133    }
1134    bias_filler {
1135      type: "constant"
1136      value: 0
1137    }
1138  }
1139}
1140layer {
1141  name: "conv6_2_mbox_conf_perm"
1142  type: "Permute"
1143  bottom: "conv6_2_mbox_conf"
1144  top: "conv6_2_mbox_conf_perm"
1145  permute_param {
1146    order: 0
1147    order: 2
1148    order: 3
1149    order: 1
1150  }
1151}
1152layer {
1153  name: "conv6_2_mbox_conf_flat"
1154  type: "Flatten"
1155  bottom: "conv6_2_mbox_conf_perm"
1156  top: "conv6_2_mbox_conf_flat"
1157  flatten_param {
1158    axis: 1
1159  }
1160}
1161layer {
1162  name: "conv6_2_mbox_priorbox"
1163  type: "PriorBox"
1164  bottom: "conv6_2"
1165  bottom: "data"
1166  top: "conv6_2_mbox_priorbox"
1167  prior_box_param {
1168    min_size: 111.0
1169    max_size: 162.0
1170    aspect_ratio: 2
1171    aspect_ratio: 3
1172    flip: true
1173    clip: false
1174    variance: 0.1
1175    variance: 0.1
1176    variance: 0.2
1177    variance: 0.2
1178    step: 32
1179    offset: 0.5
1180  }
1181}
1182layer {
1183  name: "conv7_2_mbox_loc"
1184  type: "Convolution"
1185  bottom: "conv7_2"
1186  top: "conv7_2_mbox_loc"
1187  param {
1188    lr_mult: 1
1189    decay_mult: 1
1190  }
1191  param {
1192    lr_mult: 2
1193    decay_mult: 0
1194  }
1195  convolution_param {
1196    num_output: 24
1197    pad: 1
1198    kernel_size: 3
1199    stride: 1
1200    weight_filler {
1201      type: "xavier"
1202    }
1203    bias_filler {
1204      type: "constant"
1205      value: 0
1206    }
1207  }
1208}
1209layer {
1210  name: "conv7_2_mbox_loc_perm"
1211  type: "Permute"
1212  bottom: "conv7_2_mbox_loc"
1213  top: "conv7_2_mbox_loc_perm"
1214  permute_param {
1215    order: 0
1216    order: 2
1217    order: 3
1218    order: 1
1219  }
1220}
1221layer {
1222  name: "conv7_2_mbox_loc_flat"
1223  type: "Flatten"
1224  bottom: "conv7_2_mbox_loc_perm"
1225  top: "conv7_2_mbox_loc_flat"
1226  flatten_param {
1227    axis: 1
1228  }
1229}
1230layer {
1231  name: "conv7_2_mbox_conf"
1232  type: "Convolution"
1233  bottom: "conv7_2"
1234  top: "conv7_2_mbox_conf"
1235  param {
1236    lr_mult: 1
1237    decay_mult: 1
1238  }
1239  param {
1240    lr_mult: 2
1241    decay_mult: 0
1242  }
1243  convolution_param {
1244    num_output: 126
1245    pad: 1
1246    kernel_size: 3
1247    stride: 1
1248    weight_filler {
1249      type: "xavier"
1250    }
1251    bias_filler {
1252      type: "constant"
1253      value: 0
1254    }
1255  }
1256}
1257layer {
1258  name: "conv7_2_mbox_conf_perm"
1259  type: "Permute"
1260  bottom: "conv7_2_mbox_conf"
1261  top: "conv7_2_mbox_conf_perm"
1262  permute_param {
1263    order: 0
1264    order: 2
1265    order: 3
1266    order: 1
1267  }
1268}
1269layer {
1270  name: "conv7_2_mbox_conf_flat"
1271  type: "Flatten"
1272  bottom: "conv7_2_mbox_conf_perm"
1273  top: "conv7_2_mbox_conf_flat"
1274  flatten_param {
1275    axis: 1
1276  }
1277}
1278layer {
1279  name: "conv7_2_mbox_priorbox"
1280  type: "PriorBox"
1281  bottom: "conv7_2"
1282  bottom: "data"
1283  top: "conv7_2_mbox_priorbox"
1284  prior_box_param {
1285    min_size: 162.0
1286    max_size: 213.0
1287    aspect_ratio: 2
1288    aspect_ratio: 3
1289    flip: true
1290    clip: false
1291    variance: 0.1
1292    variance: 0.1
1293    variance: 0.2
1294    variance: 0.2
1295    step: 64
1296    offset: 0.5
1297  }
1298}
1299layer {
1300  name: "conv8_2_mbox_loc"
1301  type: "Convolution"
1302  bottom: "conv8_2"
1303  top: "conv8_2_mbox_loc"
1304  param {
1305    lr_mult: 1
1306    decay_mult: 1
1307  }
1308  param {
1309    lr_mult: 2
1310    decay_mult: 0
1311  }
1312  convolution_param {
1313    num_output: 16
1314    pad: 1
1315    kernel_size: 3
1316    stride: 1
1317    weight_filler {
1318      type: "xavier"
1319    }
1320    bias_filler {
1321      type: "constant"
1322      value: 0
1323    }
1324  }
1325}
1326layer {
1327  name: "conv8_2_mbox_loc_perm"
1328  type: "Permute"
1329  bottom: "conv8_2_mbox_loc"
1330  top: "conv8_2_mbox_loc_perm"
1331  permute_param {
1332    order: 0
1333    order: 2
1334    order: 3
1335    order: 1
1336  }
1337}
1338layer {
1339  name: "conv8_2_mbox_loc_flat"
1340  type: "Flatten"
1341  bottom: "conv8_2_mbox_loc_perm"
1342  top: "conv8_2_mbox_loc_flat"
1343  flatten_param {
1344    axis: 1
1345  }
1346}
1347layer {
1348  name: "conv8_2_mbox_conf"
1349  type: "Convolution"
1350  bottom: "conv8_2"
1351  top: "conv8_2_mbox_conf"
1352  param {
1353    lr_mult: 1
1354    decay_mult: 1
1355  }
1356  param {
1357    lr_mult: 2
1358    decay_mult: 0
1359  }
1360  convolution_param {
1361    num_output: 84
1362    pad: 1
1363    kernel_size: 3
1364    stride: 1
1365    weight_filler {
1366      type: "xavier"
1367    }
1368    bias_filler {
1369      type: "constant"
1370      value: 0
1371    }
1372  }
1373}
1374layer {
1375  name: "conv8_2_mbox_conf_perm"
1376  type: "Permute"
1377  bottom: "conv8_2_mbox_conf"
1378  top: "conv8_2_mbox_conf_perm"
1379  permute_param {
1380    order: 0
1381    order: 2
1382    order: 3
1383    order: 1
1384  }
1385}
1386layer {
1387  name: "conv8_2_mbox_conf_flat"
1388  type: "Flatten"
1389  bottom: "conv8_2_mbox_conf_perm"
1390  top: "conv8_2_mbox_conf_flat"
1391  flatten_param {
1392    axis: 1
1393  }
1394}
1395layer {
1396  name: "conv8_2_mbox_priorbox"
1397  type: "PriorBox"
1398  bottom: "conv8_2"
1399  bottom: "data"
1400  top: "conv8_2_mbox_priorbox"
1401  prior_box_param {
1402    min_size: 213.0
1403    max_size: 264.0
1404    aspect_ratio: 2
1405    flip: true
1406    clip: false
1407    variance: 0.1
1408    variance: 0.1
1409    variance: 0.2
1410    variance: 0.2
1411    step: 100
1412    offset: 0.5
1413  }
1414}
1415layer {
1416  name: "conv9_2_mbox_loc"
1417  type: "Convolution"
1418  bottom: "conv9_2"
1419  top: "conv9_2_mbox_loc"
1420  param {
1421    lr_mult: 1
1422    decay_mult: 1
1423  }
1424  param {
1425    lr_mult: 2
1426    decay_mult: 0
1427  }
1428  convolution_param {
1429    num_output: 16
1430    pad: 1
1431    kernel_size: 3
1432    stride: 1
1433    weight_filler {
1434      type: "xavier"
1435    }
1436    bias_filler {
1437      type: "constant"
1438      value: 0
1439    }
1440  }
1441}
1442layer {
1443  name: "conv9_2_mbox_loc_perm"
1444  type: "Permute"
1445  bottom: "conv9_2_mbox_loc"
1446  top: "conv9_2_mbox_loc_perm"
1447  permute_param {
1448    order: 0
1449    order: 2
1450    order: 3
1451    order: 1
1452  }
1453}
1454layer {
1455  name: "conv9_2_mbox_loc_flat"
1456  type: "Flatten"
1457  bottom: "conv9_2_mbox_loc_perm"
1458  top: "conv9_2_mbox_loc_flat"
1459  flatten_param {
1460    axis: 1
1461  }
1462}
1463layer {
1464  name: "conv9_2_mbox_conf"
1465  type: "Convolution"
1466  bottom: "conv9_2"
1467  top: "conv9_2_mbox_conf"
1468  param {
1469    lr_mult: 1
1470    decay_mult: 1
1471  }
1472  param {
1473    lr_mult: 2
1474    decay_mult: 0
1475  }
1476  convolution_param {
1477    num_output: 84
1478    pad: 1
1479    kernel_size: 3
1480    stride: 1
1481    weight_filler {
1482      type: "xavier"
1483    }
1484    bias_filler {
1485      type: "constant"
1486      value: 0
1487    }
1488  }
1489}
1490layer {
1491  name: "conv9_2_mbox_conf_perm"
1492  type: "Permute"
1493  bottom: "conv9_2_mbox_conf"
1494  top: "conv9_2_mbox_conf_perm"
1495  permute_param {
1496    order: 0
1497    order: 2
1498    order: 3
1499    order: 1
1500  }
1501}
1502layer {
1503  name: "conv9_2_mbox_conf_flat"
1504  type: "Flatten"
1505  bottom: "conv9_2_mbox_conf_perm"
1506  top: "conv9_2_mbox_conf_flat"
1507  flatten_param {
1508    axis: 1
1509  }
1510}
1511layer {
1512  name: "conv9_2_mbox_priorbox"
1513  type: "PriorBox"
1514  bottom: "conv9_2"
1515  bottom: "data"
1516  top: "conv9_2_mbox_priorbox"
1517  prior_box_param {
1518    min_size: 264.0
1519    max_size: 315.0
1520    aspect_ratio: 2
1521    flip: true
1522    clip: false
1523    variance: 0.1
1524    variance: 0.1
1525    variance: 0.2
1526    variance: 0.2
1527    step: 300
1528    offset: 0.5
1529  }
1530}
1531layer {
1532  name: "mbox_loc"
1533  type: "Concat"
1534  bottom: "conv4_3_norm_mbox_loc_flat"
1535  bottom: "fc7_mbox_loc_flat"
1536  bottom: "conv6_2_mbox_loc_flat"
1537  bottom: "conv7_2_mbox_loc_flat"
1538  bottom: "conv8_2_mbox_loc_flat"
1539  bottom: "conv9_2_mbox_loc_flat"
1540  top: "mbox_loc"
1541  concat_param {
1542    axis: 1
1543  }
1544}
1545layer {
1546  name: "mbox_conf"
1547  type: "Concat"
1548  bottom: "conv4_3_norm_mbox_conf_flat"
1549  bottom: "fc7_mbox_conf_flat"
1550  bottom: "conv6_2_mbox_conf_flat"
1551  bottom: "conv7_2_mbox_conf_flat"
1552  bottom: "conv8_2_mbox_conf_flat"
1553  bottom: "conv9_2_mbox_conf_flat"
1554  top: "mbox_conf"
1555  concat_param {
1556    axis: 1
1557  }
1558}
1559layer {
1560  name: "mbox_priorbox"
1561  type: "Concat"
1562  bottom: "conv4_3_norm_mbox_priorbox"
1563  bottom: "fc7_mbox_priorbox"
1564  bottom: "conv6_2_mbox_priorbox"
1565  bottom: "conv7_2_mbox_priorbox"
1566  bottom: "conv8_2_mbox_priorbox"
1567  bottom: "conv9_2_mbox_priorbox"
1568  top: "mbox_priorbox"
1569  concat_param {
1570    axis: 2
1571  }
1572}
1573layer {
1574  name: "mbox_conf_reshape"
1575  type: "Reshape"
1576  bottom: "mbox_conf"
1577  top: "mbox_conf_reshape"
1578  reshape_param {
1579    shape {
1580      dim: 0
1581      dim: -1
1582      dim: 21
1583    }
1584  }
1585}
1586layer {
1587  name: "mbox_conf_softmax"
1588  type: "Softmax"
1589  bottom: "mbox_conf_reshape"
1590  top: "mbox_conf_softmax"
1591  softmax_param {
1592    axis: 2
1593  }
1594}
1595layer {
1596  name: "mbox_conf_flatten"
1597  type: "Flatten"
1598  bottom: "mbox_conf_softmax"
1599  top: "mbox_conf_flatten"
1600  flatten_param {
1601    axis: 1
1602  }
1603}
1604
1605
1606