1name: "VGG_text_longer_conv_300x300_deploy"
2input: "data"
3input_shape {
4  dim: 1
5  dim: 3
6  dim: 700
7  dim: 700
8}
9layer {
10  name: "conv1_1"
11  type: "Convolution"
12  bottom: "data"
13  top: "conv1_1"
14  param {
15    lr_mult: 0
16    decay_mult: 0
17  }
18  param {
19    lr_mult: 0
20    decay_mult: 0
21  }
22  convolution_param {
23    num_output: 64
24    pad: 1
25    kernel_size: 3
26    weight_filler {
27      type: "xavier"
28    }
29    bias_filler {
30      type: "constant"
31      value: 0
32    }
33  }
34}
35layer {
36  name: "relu1_1"
37  type: "ReLU"
38  bottom: "conv1_1"
39  top: "conv1_1"
40}
41layer {
42  name: "conv1_2"
43  type: "Convolution"
44  bottom: "conv1_1"
45  top: "conv1_2"
46  param {
47    lr_mult: 0
48    decay_mult: 0
49  }
50  param {
51    lr_mult: 0
52    decay_mult: 0
53  }
54  convolution_param {
55    num_output: 64
56    pad: 1
57    kernel_size: 3
58    weight_filler {
59      type: "xavier"
60    }
61    bias_filler {
62      type: "constant"
63      value: 0
64    }
65  }
66}
67layer {
68  name: "relu1_2"
69  type: "ReLU"
70  bottom: "conv1_2"
71  top: "conv1_2"
72}
73layer {
74  name: "pool1"
75  type: "Pooling"
76  bottom: "conv1_2"
77  top: "pool1"
78  pooling_param {
79    pool: MAX
80    kernel_size: 2
81    stride: 2
82  }
83}
84layer {
85  name: "conv2_1"
86  type: "Convolution"
87  bottom: "pool1"
88  top: "conv2_1"
89  param {
90    lr_mult: 0
91    decay_mult: 0
92  }
93  param {
94    lr_mult: 0
95    decay_mult: 0
96  }
97  convolution_param {
98    num_output: 128
99    pad: 1
100    kernel_size: 3
101    weight_filler {
102      type: "xavier"
103    }
104    bias_filler {
105      type: "constant"
106      value: 0
107    }
108  }
109}
110layer {
111  name: "relu2_1"
112  type: "ReLU"
113  bottom: "conv2_1"
114  top: "conv2_1"
115}
116layer {
117  name: "conv2_2"
118  type: "Convolution"
119  bottom: "conv2_1"
120  top: "conv2_2"
121  param {
122    lr_mult: 0
123    decay_mult: 0
124  }
125  param {
126    lr_mult: 0
127    decay_mult: 0
128  }
129  convolution_param {
130    num_output: 128
131    pad: 1
132    kernel_size: 3
133    weight_filler {
134      type: "xavier"
135    }
136    bias_filler {
137      type: "constant"
138      value: 0
139    }
140  }
141}
142layer {
143  name: "relu2_2"
144  type: "ReLU"
145  bottom: "conv2_2"
146  top: "conv2_2"
147}
148layer {
149  name: "pool2"
150  type: "Pooling"
151  bottom: "conv2_2"
152  top: "pool2"
153  pooling_param {
154    pool: MAX
155    kernel_size: 2
156    stride: 2
157  }
158}
159layer {
160  name: "conv3_1"
161  type: "Convolution"
162  bottom: "pool2"
163  top: "conv3_1"
164  param {
165    lr_mult: 1
166    decay_mult: 1
167  }
168  param {
169    lr_mult: 2
170    decay_mult: 0
171  }
172  convolution_param {
173    num_output: 256
174    pad: 1
175    kernel_size: 3
176    weight_filler {
177      type: "xavier"
178    }
179    bias_filler {
180      type: "constant"
181      value: 0
182    }
183  }
184}
185layer {
186  name: "relu3_1"
187  type: "ReLU"
188  bottom: "conv3_1"
189  top: "conv3_1"
190}
191layer {
192  name: "conv3_2"
193  type: "Convolution"
194  bottom: "conv3_1"
195  top: "conv3_2"
196  param {
197    lr_mult: 1
198    decay_mult: 1
199  }
200  param {
201    lr_mult: 2
202    decay_mult: 0
203  }
204  convolution_param {
205    num_output: 256
206    pad: 1
207    kernel_size: 3
208    weight_filler {
209      type: "xavier"
210    }
211    bias_filler {
212      type: "constant"
213      value: 0
214    }
215  }
216}
217layer {
218  name: "relu3_2"
219  type: "ReLU"
220  bottom: "conv3_2"
221  top: "conv3_2"
222}
223layer {
224  name: "conv3_3"
225  type: "Convolution"
226  bottom: "conv3_2"
227  top: "conv3_3"
228  param {
229    lr_mult: 1
230    decay_mult: 1
231  }
232  param {
233    lr_mult: 2
234    decay_mult: 0
235  }
236  convolution_param {
237    num_output: 256
238    pad: 1
239    kernel_size: 3
240    weight_filler {
241      type: "xavier"
242    }
243    bias_filler {
244      type: "constant"
245      value: 0
246    }
247  }
248}
249layer {
250  name: "relu3_3"
251  type: "ReLU"
252  bottom: "conv3_3"
253  top: "conv3_3"
254}
255layer {
256  name: "pool3"
257  type: "Pooling"
258  bottom: "conv3_3"
259  top: "pool3"
260  pooling_param {
261    pool: MAX
262    kernel_size: 2
263    stride: 2
264  }
265}
266layer {
267  name: "conv4_1"
268  type: "Convolution"
269  bottom: "pool3"
270  top: "conv4_1"
271  param {
272    lr_mult: 1
273    decay_mult: 1
274  }
275  param {
276    lr_mult: 2
277    decay_mult: 0
278  }
279  convolution_param {
280    num_output: 512
281    pad: 1
282    kernel_size: 3
283    weight_filler {
284      type: "xavier"
285    }
286    bias_filler {
287      type: "constant"
288      value: 0
289    }
290  }
291}
292layer {
293  name: "relu4_1"
294  type: "ReLU"
295  bottom: "conv4_1"
296  top: "conv4_1"
297}
298layer {
299  name: "conv4_2"
300  type: "Convolution"
301  bottom: "conv4_1"
302  top: "conv4_2"
303  param {
304    lr_mult: 1
305    decay_mult: 1
306  }
307  param {
308    lr_mult: 2
309    decay_mult: 0
310  }
311  convolution_param {
312    num_output: 512
313    pad: 1
314    kernel_size: 3
315    weight_filler {
316      type: "xavier"
317    }
318    bias_filler {
319      type: "constant"
320      value: 0
321    }
322  }
323}
324layer {
325  name: "relu4_2"
326  type: "ReLU"
327  bottom: "conv4_2"
328  top: "conv4_2"
329}
330layer {
331  name: "conv4_3"
332  type: "Convolution"
333  bottom: "conv4_2"
334  top: "conv4_3"
335  param {
336    lr_mult: 1
337    decay_mult: 1
338  }
339  param {
340    lr_mult: 2
341    decay_mult: 0
342  }
343  convolution_param {
344    num_output: 512
345    pad: 1
346    kernel_size: 3
347    weight_filler {
348      type: "xavier"
349    }
350    bias_filler {
351      type: "constant"
352      value: 0
353    }
354  }
355}
356layer {
357  name: "relu4_3"
358  type: "ReLU"
359  bottom: "conv4_3"
360  top: "conv4_3"
361}
362layer {
363  name: "pool4"
364  type: "Pooling"
365  bottom: "conv4_3"
366  top: "pool4"
367  pooling_param {
368    pool: MAX
369    kernel_size: 2
370    stride: 2
371  }
372}
373layer {
374  name: "conv5_1"
375  type: "Convolution"
376  bottom: "pool4"
377  top: "conv5_1"
378  param {
379    lr_mult: 1
380    decay_mult: 1
381  }
382  param {
383    lr_mult: 2
384    decay_mult: 0
385  }
386  convolution_param {
387    num_output: 512
388    pad: 1
389    kernel_size: 3
390    weight_filler {
391      type: "xavier"
392    }
393    bias_filler {
394      type: "constant"
395      value: 0
396    }
397  }
398}
399layer {
400  name: "relu5_1"
401  type: "ReLU"
402  bottom: "conv5_1"
403  top: "conv5_1"
404}
405layer {
406  name: "conv5_2"
407  type: "Convolution"
408  bottom: "conv5_1"
409  top: "conv5_2"
410  param {
411    lr_mult: 1
412    decay_mult: 1
413  }
414  param {
415    lr_mult: 2
416    decay_mult: 0
417  }
418  convolution_param {
419    num_output: 512
420    pad: 1
421    kernel_size: 3
422    weight_filler {
423      type: "xavier"
424    }
425    bias_filler {
426      type: "constant"
427      value: 0
428    }
429  }
430}
431layer {
432  name: "relu5_2"
433  type: "ReLU"
434  bottom: "conv5_2"
435  top: "conv5_2"
436}
437layer {
438  name: "conv5_3"
439  type: "Convolution"
440  bottom: "conv5_2"
441  top: "conv5_3"
442  param {
443    lr_mult: 1
444    decay_mult: 1
445  }
446  param {
447    lr_mult: 2
448    decay_mult: 0
449  }
450  convolution_param {
451    num_output: 512
452    pad: 1
453    kernel_size: 3
454    weight_filler {
455      type: "xavier"
456    }
457    bias_filler {
458      type: "constant"
459      value: 0
460    }
461  }
462}
463layer {
464  name: "relu5_3"
465  type: "ReLU"
466  bottom: "conv5_3"
467  top: "conv5_3"
468}
469layer {
470  name: "pool5"
471  type: "Pooling"
472  bottom: "conv5_3"
473  top: "pool5"
474  pooling_param {
475    pool: MAX
476    kernel_size: 3
477    stride: 1
478    pad: 1
479  }
480}
481layer {
482  name: "fc6"
483  type: "Convolution"
484  bottom: "pool5"
485  top: "fc6"
486  param {
487    lr_mult: 1
488    decay_mult: 1
489  }
490  param {
491    lr_mult: 2
492    decay_mult: 0
493  }
494  convolution_param {
495    num_output: 1024
496    pad: 6
497    kernel_size: 3
498    weight_filler {
499      type: "xavier"
500    }
501    bias_filler {
502      type: "constant"
503      value: 0
504    }
505    dilation: 6
506  }
507}
508layer {
509  name: "relu6"
510  type: "ReLU"
511  bottom: "fc6"
512  top: "fc6"
513}
514layer {
515  name: "fc7"
516  type: "Convolution"
517  bottom: "fc6"
518  top: "fc7"
519  param {
520    lr_mult: 1
521    decay_mult: 1
522  }
523  param {
524    lr_mult: 2
525    decay_mult: 0
526  }
527  convolution_param {
528    num_output: 1024
529    kernel_size: 1
530    weight_filler {
531      type: "xavier"
532    }
533    bias_filler {
534      type: "constant"
535      value: 0
536    }
537  }
538}
539layer {
540  name: "relu7"
541  type: "ReLU"
542  bottom: "fc7"
543  top: "fc7"
544}
545layer {
546  name: "conv6_1"
547  type: "Convolution"
548  bottom: "fc7"
549  top: "conv6_1"
550  param {
551    lr_mult: 1
552    decay_mult: 1
553  }
554  param {
555    lr_mult: 2
556    decay_mult: 0
557  }
558  convolution_param {
559    num_output: 256
560    pad: 0
561    kernel_size: 1
562    stride: 1
563    weight_filler {
564      type: "xavier"
565    }
566    bias_filler {
567      type: "constant"
568      value: 0
569    }
570  }
571}
572layer {
573  name: "conv6_1_relu"
574  type: "ReLU"
575  bottom: "conv6_1"
576  top: "conv6_1"
577}
578layer {
579  name: "conv6_2"
580  type: "Convolution"
581  bottom: "conv6_1"
582  top: "conv6_2"
583  param {
584    lr_mult: 1
585    decay_mult: 1
586  }
587  param {
588    lr_mult: 2
589    decay_mult: 0
590  }
591  convolution_param {
592    num_output: 512
593    pad: 1
594    kernel_size: 3
595    stride: 2
596    weight_filler {
597      type: "xavier"
598    }
599    bias_filler {
600      type: "constant"
601      value: 0
602    }
603  }
604}
605layer {
606  name: "conv6_2_relu"
607  type: "ReLU"
608  bottom: "conv6_2"
609  top: "conv6_2"
610}
611layer {
612  name: "conv7_1"
613  type: "Convolution"
614  bottom: "conv6_2"
615  top: "conv7_1"
616  param {
617    lr_mult: 1
618    decay_mult: 1
619  }
620  param {
621    lr_mult: 2
622    decay_mult: 0
623  }
624  convolution_param {
625    num_output: 128
626    pad: 0
627    kernel_size: 1
628    stride: 1
629    weight_filler {
630      type: "xavier"
631    }
632    bias_filler {
633      type: "constant"
634      value: 0
635    }
636  }
637}
638layer {
639  name: "conv7_1_relu"
640  type: "ReLU"
641  bottom: "conv7_1"
642  top: "conv7_1"
643}
644layer {
645  name: "conv7_2"
646  type: "Convolution"
647  bottom: "conv7_1"
648  top: "conv7_2"
649  param {
650    lr_mult: 1
651    decay_mult: 1
652  }
653  param {
654    lr_mult: 2
655    decay_mult: 0
656  }
657  convolution_param {
658    num_output: 256
659    pad: 1
660    kernel_size: 3
661    stride: 2
662    weight_filler {
663      type: "xavier"
664    }
665    bias_filler {
666      type: "constant"
667      value: 0
668    }
669  }
670}
671layer {
672  name: "conv7_2_relu"
673  type: "ReLU"
674  bottom: "conv7_2"
675  top: "conv7_2"
676}
677layer {
678  name: "conv8_1"
679  type: "Convolution"
680  bottom: "conv7_2"
681  top: "conv8_1"
682  param {
683    lr_mult: 1
684    decay_mult: 1
685  }
686  param {
687    lr_mult: 2
688    decay_mult: 0
689  }
690  convolution_param {
691    num_output: 128
692    pad: 0
693    kernel_size: 1
694    stride: 1
695    weight_filler {
696      type: "xavier"
697    }
698    bias_filler {
699      type: "constant"
700      value: 0
701    }
702  }
703}
704layer {
705  name: "conv8_1_relu"
706  type: "ReLU"
707  bottom: "conv8_1"
708  top: "conv8_1"
709}
710layer {
711  name: "conv8_2"
712  type: "Convolution"
713  bottom: "conv8_1"
714  top: "conv8_2"
715  param {
716    lr_mult: 1
717    decay_mult: 1
718  }
719  param {
720    lr_mult: 2
721    decay_mult: 0
722  }
723  convolution_param {
724    num_output: 256
725    pad: 1
726    kernel_size: 3
727    stride: 2
728    weight_filler {
729      type: "xavier"
730    }
731    bias_filler {
732      type: "constant"
733      value: 0
734    }
735  }
736}
737layer {
738  name: "conv8_2_relu"
739  type: "ReLU"
740  bottom: "conv8_2"
741  top: "conv8_2"
742}
743layer {
744  name: "pool6"
745  type: "Pooling"
746  bottom: "conv8_2"
747  top: "pool6"
748  pooling_param {
749    pool: AVE
750    global_pooling: true
751  }
752}
753layer {
754  name: "conv4_3_norm"
755  type: "Normalize"
756  bottom: "conv4_3"
757  top: "conv4_3_norm"
758  norm_param {
759    across_spatial: false
760    scale_filler {
761      type: "constant"
762      value: 20
763    }
764    channel_shared: false
765  }
766}
767layer {
768  name: "conv4_3_norm_mbox_loc"
769  type: "Convolution"
770  bottom: "conv4_3_norm"
771  top: "conv4_3_norm_mbox_loc"
772  param {
773    lr_mult: 1
774    decay_mult: 1
775  }
776  param {
777    lr_mult: 2
778    decay_mult: 0
779  }
780  convolution_param {
781    num_output: 48
782    weight_filler {
783      type: "xavier"
784    }
785    bias_filler {
786      type: "constant"
787      value: 0
788    }
789    pad_h: 0
790    pad_w: 2
791    kernel_h: 1
792    kernel_w: 5
793    stride_h: 1
794    stride_w: 1
795  }
796}
797layer {
798  name: "conv4_3_norm_mbox_loc_perm"
799  type: "Permute"
800  bottom: "conv4_3_norm_mbox_loc"
801  top: "conv4_3_norm_mbox_loc_perm"
802  permute_param {
803    order: 0
804    order: 2
805    order: 3
806    order: 1
807  }
808}
809layer {
810  name: "conv4_3_norm_mbox_loc_flat"
811  type: "Flatten"
812  bottom: "conv4_3_norm_mbox_loc_perm"
813  top: "conv4_3_norm_mbox_loc_flat"
814  flatten_param {
815    axis: 1
816  }
817}
818layer {
819  name: "conv4_3_norm_mbox_conf"
820  type: "Convolution"
821  bottom: "conv4_3_norm"
822  top: "conv4_3_norm_mbox_conf"
823  param {
824    lr_mult: 1
825    decay_mult: 1
826  }
827  param {
828    lr_mult: 2
829    decay_mult: 0
830  }
831  convolution_param {
832    num_output: 24
833    weight_filler {
834      type: "xavier"
835    }
836    bias_filler {
837      type: "constant"
838      value: 0
839    }
840    pad_h: 0
841    pad_w: 2
842    kernel_h: 1
843    kernel_w: 5
844    stride_h: 1
845    stride_w: 1
846  }
847}
848layer {
849  name: "conv4_3_norm_mbox_conf_perm"
850  type: "Permute"
851  bottom: "conv4_3_norm_mbox_conf"
852  top: "conv4_3_norm_mbox_conf_perm"
853  permute_param {
854    order: 0
855    order: 2
856    order: 3
857    order: 1
858  }
859}
860layer {
861  name: "conv4_3_norm_mbox_conf_flat"
862  type: "Flatten"
863  bottom: "conv4_3_norm_mbox_conf_perm"
864  top: "conv4_3_norm_mbox_conf_flat"
865  flatten_param {
866    axis: 1
867  }
868}
869layer {
870  name: "conv4_3_norm_mbox_priorbox"
871  type: "PriorBox"
872  bottom: "conv4_3_norm"
873  bottom: "data"
874  top: "conv4_3_norm_mbox_priorbox"
875  prior_box_param {
876    min_size: 30.0
877    aspect_ratio: 2
878    aspect_ratio: 3
879    aspect_ratio: 5
880    aspect_ratio: 7
881    aspect_ratio: 10
882    flip: false
883    clip: true
884    variance: 0.1
885    variance: 0.1
886    variance: 0.2
887    variance: 0.2
888    offset_h: 0.5
889    offset_w: 0.5
890    offset_h: 1.0
891    offset_w: 0.5
892  }
893}
894layer {
895  name: "fc7_mbox_loc"
896  type: "Convolution"
897  bottom: "fc7"
898  top: "fc7_mbox_loc"
899  param {
900    lr_mult: 1
901    decay_mult: 1
902  }
903  param {
904    lr_mult: 2
905    decay_mult: 0
906  }
907  convolution_param {
908    num_output: 56
909    weight_filler {
910      type: "xavier"
911    }
912    bias_filler {
913      type: "constant"
914      value: 0
915    }
916    pad_h: 0
917    pad_w: 2
918    kernel_h: 1
919    kernel_w: 5
920    stride_h: 1
921    stride_w: 1
922  }
923}
924layer {
925  name: "fc7_mbox_loc_perm"
926  type: "Permute"
927  bottom: "fc7_mbox_loc"
928  top: "fc7_mbox_loc_perm"
929  permute_param {
930    order: 0
931    order: 2
932    order: 3
933    order: 1
934  }
935}
936layer {
937  name: "fc7_mbox_loc_flat"
938  type: "Flatten"
939  bottom: "fc7_mbox_loc_perm"
940  top: "fc7_mbox_loc_flat"
941  flatten_param {
942    axis: 1
943  }
944}
945layer {
946  name: "fc7_mbox_conf"
947  type: "Convolution"
948  bottom: "fc7"
949  top: "fc7_mbox_conf"
950  param {
951    lr_mult: 1
952    decay_mult: 1
953  }
954  param {
955    lr_mult: 2
956    decay_mult: 0
957  }
958  convolution_param {
959    num_output: 28
960    weight_filler {
961      type: "xavier"
962    }
963    bias_filler {
964      type: "constant"
965      value: 0
966    }
967    pad_h: 0
968    pad_w: 2
969    kernel_h: 1
970    kernel_w: 5
971    stride_h: 1
972    stride_w: 1
973  }
974}
975layer {
976  name: "fc7_mbox_conf_perm"
977  type: "Permute"
978  bottom: "fc7_mbox_conf"
979  top: "fc7_mbox_conf_perm"
980  permute_param {
981    order: 0
982    order: 2
983    order: 3
984    order: 1
985  }
986}
987layer {
988  name: "fc7_mbox_conf_flat"
989  type: "Flatten"
990  bottom: "fc7_mbox_conf_perm"
991  top: "fc7_mbox_conf_flat"
992  flatten_param {
993    axis: 1
994  }
995}
996layer {
997  name: "fc7_mbox_priorbox"
998  type: "PriorBox"
999  bottom: "fc7"
1000  bottom: "data"
1001  top: "fc7_mbox_priorbox"
1002  prior_box_param {
1003    min_size: 60.0
1004    max_size: 114.0
1005    aspect_ratio: 2
1006    aspect_ratio: 3
1007    aspect_ratio: 5
1008    aspect_ratio: 7
1009    aspect_ratio: 10
1010    flip: false
1011    clip: true
1012    variance: 0.1
1013    variance: 0.1
1014    variance: 0.2
1015    variance: 0.2
1016    offset_h: 0.5
1017    offset_w: 0.5
1018    offset_h: 1.0
1019    offset_w: 0.5
1020  }
1021}
1022layer {
1023  name: "conv6_2_mbox_loc"
1024  type: "Convolution"
1025  bottom: "conv6_2"
1026  top: "conv6_2_mbox_loc"
1027  param {
1028    lr_mult: 1
1029    decay_mult: 1
1030  }
1031  param {
1032    lr_mult: 2
1033    decay_mult: 0
1034  }
1035  convolution_param {
1036    num_output: 56
1037    weight_filler {
1038      type: "xavier"
1039    }
1040    bias_filler {
1041      type: "constant"
1042      value: 0
1043    }
1044    pad_h: 0
1045    pad_w: 2
1046    kernel_h: 1
1047    kernel_w: 5
1048    stride_h: 1
1049    stride_w: 1
1050  }
1051}
1052layer {
1053  name: "conv6_2_mbox_loc_perm"
1054  type: "Permute"
1055  bottom: "conv6_2_mbox_loc"
1056  top: "conv6_2_mbox_loc_perm"
1057  permute_param {
1058    order: 0
1059    order: 2
1060    order: 3
1061    order: 1
1062  }
1063}
1064layer {
1065  name: "conv6_2_mbox_loc_flat"
1066  type: "Flatten"
1067  bottom: "conv6_2_mbox_loc_perm"
1068  top: "conv6_2_mbox_loc_flat"
1069  flatten_param {
1070    axis: 1
1071  }
1072}
1073layer {
1074  name: "conv6_2_mbox_conf"
1075  type: "Convolution"
1076  bottom: "conv6_2"
1077  top: "conv6_2_mbox_conf"
1078  param {
1079    lr_mult: 1
1080    decay_mult: 1
1081  }
1082  param {
1083    lr_mult: 2
1084    decay_mult: 0
1085  }
1086  convolution_param {
1087    num_output: 28
1088    weight_filler {
1089      type: "xavier"
1090    }
1091    bias_filler {
1092      type: "constant"
1093      value: 0
1094    }
1095    pad_h: 0
1096    pad_w: 2
1097    kernel_h: 1
1098    kernel_w: 5
1099    stride_h: 1
1100    stride_w: 1
1101  }
1102}
1103layer {
1104  name: "conv6_2_mbox_conf_perm"
1105  type: "Permute"
1106  bottom: "conv6_2_mbox_conf"
1107  top: "conv6_2_mbox_conf_perm"
1108  permute_param {
1109    order: 0
1110    order: 2
1111    order: 3
1112    order: 1
1113  }
1114}
1115layer {
1116  name: "conv6_2_mbox_conf_flat"
1117  type: "Flatten"
1118  bottom: "conv6_2_mbox_conf_perm"
1119  top: "conv6_2_mbox_conf_flat"
1120  flatten_param {
1121    axis: 1
1122  }
1123}
1124layer {
1125  name: "conv6_2_mbox_priorbox"
1126  type: "PriorBox"
1127  bottom: "conv6_2"
1128  bottom: "data"
1129  top: "conv6_2_mbox_priorbox"
1130  prior_box_param {
1131    min_size: 114.0
1132    max_size: 168.0
1133    aspect_ratio: 2
1134    aspect_ratio: 3
1135    aspect_ratio: 5
1136    aspect_ratio: 7
1137    aspect_ratio: 10
1138    flip: false
1139    clip: true
1140    variance: 0.1
1141    variance: 0.1
1142    variance: 0.2
1143    variance: 0.2
1144    offset_h: 0.5
1145    offset_w: 0.5
1146    offset_h: 1.0
1147    offset_w: 0.5
1148  }
1149}
1150layer {
1151  name: "conv7_2_mbox_loc"
1152  type: "Convolution"
1153  bottom: "conv7_2"
1154  top: "conv7_2_mbox_loc"
1155  param {
1156    lr_mult: 1
1157    decay_mult: 1
1158  }
1159  param {
1160    lr_mult: 2
1161    decay_mult: 0
1162  }
1163  convolution_param {
1164    num_output: 56
1165    weight_filler {
1166      type: "xavier"
1167    }
1168    bias_filler {
1169      type: "constant"
1170      value: 0
1171    }
1172    pad_h: 0
1173    pad_w: 2
1174    kernel_h: 1
1175    kernel_w: 5
1176    stride_h: 1
1177    stride_w: 1
1178  }
1179}
1180layer {
1181  name: "conv7_2_mbox_loc_perm"
1182  type: "Permute"
1183  bottom: "conv7_2_mbox_loc"
1184  top: "conv7_2_mbox_loc_perm"
1185  permute_param {
1186    order: 0
1187    order: 2
1188    order: 3
1189    order: 1
1190  }
1191}
1192layer {
1193  name: "conv7_2_mbox_loc_flat"
1194  type: "Flatten"
1195  bottom: "conv7_2_mbox_loc_perm"
1196  top: "conv7_2_mbox_loc_flat"
1197  flatten_param {
1198    axis: 1
1199  }
1200}
1201layer {
1202  name: "conv7_2_mbox_conf"
1203  type: "Convolution"
1204  bottom: "conv7_2"
1205  top: "conv7_2_mbox_conf"
1206  param {
1207    lr_mult: 1
1208    decay_mult: 1
1209  }
1210  param {
1211    lr_mult: 2
1212    decay_mult: 0
1213  }
1214  convolution_param {
1215    num_output: 28
1216    weight_filler {
1217      type: "xavier"
1218    }
1219    bias_filler {
1220      type: "constant"
1221      value: 0
1222    }
1223    pad_h: 0
1224    pad_w: 2
1225    kernel_h: 1
1226    kernel_w: 5
1227    stride_h: 1
1228    stride_w: 1
1229  }
1230}
1231layer {
1232  name: "conv7_2_mbox_conf_perm"
1233  type: "Permute"
1234  bottom: "conv7_2_mbox_conf"
1235  top: "conv7_2_mbox_conf_perm"
1236  permute_param {
1237    order: 0
1238    order: 2
1239    order: 3
1240    order: 1
1241  }
1242}
1243layer {
1244  name: "conv7_2_mbox_conf_flat"
1245  type: "Flatten"
1246  bottom: "conv7_2_mbox_conf_perm"
1247  top: "conv7_2_mbox_conf_flat"
1248  flatten_param {
1249    axis: 1
1250  }
1251}
1252layer {
1253  name: "conv7_2_mbox_priorbox"
1254  type: "PriorBox"
1255  bottom: "conv7_2"
1256  bottom: "data"
1257  top: "conv7_2_mbox_priorbox"
1258  prior_box_param {
1259    min_size: 168.0
1260    max_size: 222.0
1261    aspect_ratio: 2
1262    aspect_ratio: 3
1263    aspect_ratio: 5
1264    aspect_ratio: 7
1265    aspect_ratio: 10
1266    flip: false
1267    clip: true
1268    variance: 0.1
1269    variance: 0.1
1270    variance: 0.2
1271    variance: 0.2
1272    offset_h: 0.5
1273    offset_w: 0.5
1274    offset_h: 1.0
1275    offset_w: 0.5
1276  }
1277}
1278layer {
1279  name: "conv8_2_mbox_loc"
1280  type: "Convolution"
1281  bottom: "conv8_2"
1282  top: "conv8_2_mbox_loc"
1283  param {
1284    lr_mult: 1
1285    decay_mult: 1
1286  }
1287  param {
1288    lr_mult: 2
1289    decay_mult: 0
1290  }
1291  convolution_param {
1292    num_output: 56
1293    weight_filler {
1294      type: "xavier"
1295    }
1296    bias_filler {
1297      type: "constant"
1298      value: 0
1299    }
1300    pad_h: 0
1301    pad_w: 2
1302    kernel_h: 1
1303    kernel_w: 5
1304    stride_h: 1
1305    stride_w: 1
1306  }
1307}
1308layer {
1309  name: "conv8_2_mbox_loc_perm"
1310  type: "Permute"
1311  bottom: "conv8_2_mbox_loc"
1312  top: "conv8_2_mbox_loc_perm"
1313  permute_param {
1314    order: 0
1315    order: 2
1316    order: 3
1317    order: 1
1318  }
1319}
1320layer {
1321  name: "conv8_2_mbox_loc_flat"
1322  type: "Flatten"
1323  bottom: "conv8_2_mbox_loc_perm"
1324  top: "conv8_2_mbox_loc_flat"
1325  flatten_param {
1326    axis: 1
1327  }
1328}
1329layer {
1330  name: "conv8_2_mbox_conf"
1331  type: "Convolution"
1332  bottom: "conv8_2"
1333  top: "conv8_2_mbox_conf"
1334  param {
1335    lr_mult: 1
1336    decay_mult: 1
1337  }
1338  param {
1339    lr_mult: 2
1340    decay_mult: 0
1341  }
1342  convolution_param {
1343    num_output: 28
1344    weight_filler {
1345      type: "xavier"
1346    }
1347    bias_filler {
1348      type: "constant"
1349      value: 0
1350    }
1351    pad_h: 0
1352    pad_w: 2
1353    kernel_h: 1
1354    kernel_w: 5
1355    stride_h: 1
1356    stride_w: 1
1357  }
1358}
1359layer {
1360  name: "conv8_2_mbox_conf_perm"
1361  type: "Permute"
1362  bottom: "conv8_2_mbox_conf"
1363  top: "conv8_2_mbox_conf_perm"
1364  permute_param {
1365    order: 0
1366    order: 2
1367    order: 3
1368    order: 1
1369  }
1370}
1371layer {
1372  name: "conv8_2_mbox_conf_flat"
1373  type: "Flatten"
1374  bottom: "conv8_2_mbox_conf_perm"
1375  top: "conv8_2_mbox_conf_flat"
1376  flatten_param {
1377    axis: 1
1378  }
1379}
1380layer {
1381  name: "conv8_2_mbox_priorbox"
1382  type: "PriorBox"
1383  bottom: "conv8_2"
1384  bottom: "data"
1385  top: "conv8_2_mbox_priorbox"
1386  prior_box_param {
1387    min_size: 222.0
1388    max_size: 276.0
1389    aspect_ratio: 2
1390    aspect_ratio: 3
1391    aspect_ratio: 5
1392    aspect_ratio: 7
1393    aspect_ratio: 10
1394    flip: false
1395    clip: true
1396    variance: 0.1
1397    variance: 0.1
1398    variance: 0.2
1399    variance: 0.2
1400    offset_h: 0.5
1401    offset_w: 0.5
1402    offset_h: 1.0
1403    offset_w: 0.5
1404  }
1405}
1406layer {
1407  name: "pool6_mbox_loc"
1408  type: "Convolution"
1409  bottom: "pool6"
1410  top: "pool6_mbox_loc"
1411  param {
1412    lr_mult: 1
1413    decay_mult: 1
1414  }
1415  param {
1416    lr_mult: 2
1417    decay_mult: 0
1418  }
1419  convolution_param {
1420    num_output: 56
1421    weight_filler {
1422      type: "xavier"
1423    }
1424    bias_filler {
1425      type: "constant"
1426      value: 0
1427    }
1428    pad_h: 0
1429    pad_w: 2
1430    kernel_h: 1
1431    kernel_w: 5
1432    stride_h: 1
1433    stride_w: 1
1434  }
1435}
1436layer {
1437  name: "pool6_mbox_loc_perm"
1438  type: "Permute"
1439  bottom: "pool6_mbox_loc"
1440  top: "pool6_mbox_loc_perm"
1441  permute_param {
1442    order: 0
1443    order: 2
1444    order: 3
1445    order: 1
1446  }
1447}
1448layer {
1449  name: "pool6_mbox_loc_flat"
1450  type: "Flatten"
1451  bottom: "pool6_mbox_loc_perm"
1452  top: "pool6_mbox_loc_flat"
1453  flatten_param {
1454    axis: 1
1455  }
1456}
1457layer {
1458  name: "pool6_mbox_conf"
1459  type: "Convolution"
1460  bottom: "pool6"
1461  top: "pool6_mbox_conf"
1462  param {
1463    lr_mult: 1
1464    decay_mult: 1
1465  }
1466  param {
1467    lr_mult: 2
1468    decay_mult: 0
1469  }
1470  convolution_param {
1471    num_output: 28
1472    weight_filler {
1473      type: "xavier"
1474    }
1475    bias_filler {
1476      type: "constant"
1477      value: 0
1478    }
1479    pad_h: 0
1480    pad_w: 2
1481    kernel_h: 1
1482    kernel_w: 5
1483    stride_h: 1
1484    stride_w: 1
1485  }
1486}
1487layer {
1488  name: "pool6_mbox_conf_perm"
1489  type: "Permute"
1490  bottom: "pool6_mbox_conf"
1491  top: "pool6_mbox_conf_perm"
1492  permute_param {
1493    order: 0
1494    order: 2
1495    order: 3
1496    order: 1
1497  }
1498}
1499layer {
1500  name: "pool6_mbox_conf_flat"
1501  type: "Flatten"
1502  bottom: "pool6_mbox_conf_perm"
1503  top: "pool6_mbox_conf_flat"
1504  flatten_param {
1505    axis: 1
1506  }
1507}
1508layer {
1509  name: "pool6_mbox_priorbox"
1510  type: "PriorBox"
1511  bottom: "pool6"
1512  bottom: "data"
1513  top: "pool6_mbox_priorbox"
1514  prior_box_param {
1515    min_size: 276.0
1516    max_size: 330.0
1517    aspect_ratio: 2
1518    aspect_ratio: 3
1519    aspect_ratio: 5
1520    aspect_ratio: 7
1521    aspect_ratio: 10
1522    flip: false
1523    clip: true
1524    variance: 0.1
1525    variance: 0.1
1526    variance: 0.2
1527    variance: 0.2
1528    offset_h: 0.5
1529    offset_w: 0.5
1530    offset_h: 1.0
1531    offset_w: 0.5
1532  }
1533}
1534layer {
1535  name: "mbox_loc"
1536  type: "Concat"
1537  bottom: "conv4_3_norm_mbox_loc_flat"
1538  bottom: "fc7_mbox_loc_flat"
1539  bottom: "conv6_2_mbox_loc_flat"
1540  bottom: "conv7_2_mbox_loc_flat"
1541  bottom: "conv8_2_mbox_loc_flat"
1542  bottom: "pool6_mbox_loc_flat"
1543  top: "mbox_loc"
1544  concat_param {
1545    axis: 1
1546  }
1547}
1548layer {
1549  name: "mbox_conf"
1550  type: "Concat"
1551  bottom: "conv4_3_norm_mbox_conf_flat"
1552  bottom: "fc7_mbox_conf_flat"
1553  bottom: "conv6_2_mbox_conf_flat"
1554  bottom: "conv7_2_mbox_conf_flat"
1555  bottom: "conv8_2_mbox_conf_flat"
1556  bottom: "pool6_mbox_conf_flat"
1557  top: "mbox_conf"
1558  concat_param {
1559    axis: 1
1560  }
1561}
1562layer {
1563  name: "mbox_priorbox"
1564  type: "Concat"
1565  bottom: "conv4_3_norm_mbox_priorbox"
1566  bottom: "fc7_mbox_priorbox"
1567  bottom: "conv6_2_mbox_priorbox"
1568  bottom: "conv7_2_mbox_priorbox"
1569  bottom: "conv8_2_mbox_priorbox"
1570  bottom: "pool6_mbox_priorbox"
1571  top: "mbox_priorbox"
1572  concat_param {
1573    axis: 2
1574  }
1575}
1576layer {
1577  name: "mbox_conf_reshape"
1578  type: "Reshape"
1579  bottom: "mbox_conf"
1580  top: "mbox_conf_reshape"
1581  reshape_param {
1582    shape {
1583      dim: 0
1584      dim: -1
1585      dim: 2
1586    }
1587  }
1588}
1589layer {
1590  name: "mbox_conf_softmax"
1591  type: "Softmax"
1592  bottom: "mbox_conf_reshape"
1593  top: "mbox_conf_softmax"
1594  softmax_param {
1595    axis: 2
1596  }
1597}
1598layer {
1599  name: "mbox_conf_flatten"
1600  type: "Flatten"
1601  bottom: "mbox_conf_softmax"
1602  top: "mbox_conf_flatten"
1603  flatten_param {
1604    axis: 1
1605  }
1606}
1607layer {
1608  name: "detection_out"
1609  type: "DetectionOutput"
1610  bottom: "mbox_loc"
1611  bottom: "mbox_conf_flatten"
1612  bottom: "mbox_priorbox"
1613  top: "detection_out"
1614  include {
1615    phase: TEST
1616  }
1617  detection_output_param {
1618    num_classes: 2
1619    share_location: true
1620    background_label_id: 0
1621    nms_param {
1622      nms_threshold: 0.45
1623      top_k: 400
1624    }
1625    code_type: CENTER_SIZE
1626    keep_top_k: 200
1627    confidence_threshold: 0.01
1628  }
1629}
1630